Conference PaperPDF Available

Should GPs refer directly for colonoscopy for suspected colorectal cancer rather than to urgent outpatient clinic?

Authors:

Abstract

Introduction: Urgent referrals are made by GPs to outpatient clinic for ‘red flag’ symptoms of colorectal cancer (CRC). Colonoscopy and radiological investigations may then be arranged by clinic specialists according to patient suitability. We investigated whether any patient factors predicted a final diagnosis of CRC, in order to determine which patients might be suitable for direct investigation by GPs, rather than being first referred to clinic. Methods: A prospective observational study investigated patients with suspected CRC at a UK NHS Trust (May 2017–August 2018). Characteristics included gender, age, and 6 pre-defined referral categories within our regional network: (i) >40 years old, weight loss (WL) and abdominal pain (AP); (ii) <50 years old, rectal bleeding (RB) and AB, iron deficiency anaemia (IDA), change in bowel habit (CBH), or unintended WL; (iii) >50 years old, RB; (iv) >60 years old, IDA or CBH >6 weeks; (v) rectal/abdominal mass; and (vi) anal mass/ulcer. Binomial multivariate logistic regression was used to investigate characteristics associated with final diagnosis of CRC. Results: There were 605 patients (52% female; mean age 65 years), of which 3.1% had a final diagnosis of CRC. None of the 6 pre-defined referral categories, age or gender were significant predictors of CRC. Conclusions: Overall cancer rate was low, and no patient characteristics on GP referral predicted CRC. These data may support a pathway in which GPs may refer directly for colonoscopy or radiological investigations based on any of the predefined ‘red-flag’ criteria rather than to a colorectal clinic first.
2019 ACPGBI Poster Abstracts
P001
Use of FIT test to triage patients with suspected colo-rectal
cancers at a primary healthcare level
N. Joshi
1
, F. Reid
1
, S. Rai
1
& K. McEwan
2
1
Stepping Hill Hospital, Stockport, United Kingdom,
2
Park View group Practice,
Stockport, United Kingdom
Quantitative faecal immunochemical tests (FITs) use antibodies that specifically
recognise the globin of human haemoglobin to determine the amount of Hb in a
faecal sample. The primary indication for this assessment is the use of tests for the
presence of occult blood in faeces as a triage step in the investigation for possible
Colo-rectal cancer (CRC) according to the NICE guideline recommendation for
CRC screening (Point 1.3.4).
We tested 207 patients with a FIT test in a period of 1 month at a GP practice.
Of these, 36 were tested Positive FIT (>10 score), while 171 were negative.
Of the 36 positives, 32 were referred as 2WW. All 32 underwent further inves-
tigations (CT/Colonoscopy). 3 (9.375%) came back as malignancy, 7 (21.8%) came
as adenomas while 15 (47%) had diverticulosis.
Of the 171 FIT negatives, 17 were referred as 2WW and 5 were referred as
routine referrals. All the 2WW patients had Normal further investigations. Of the 5
routine referrals, 3 had normal further investigations while 2 patients were yet to be
seen.
FIT does pick up higher % of cancers than 2WW referrals, which is about 5
6% on average. The relatively high proportion of FIT false positives observed when
the target condition is CRC may be mitigated by the potential to diagnose other
bowel pathologies in these patients.
90% of negative FIT were not referred, but none of them have presented to us
as an emergency. And all those with negative FIT referred to us were true nega-
tives.
P002
Keeping FIT: early clinical outcomes of a novel two week wait
pathway for colorectal cancer using faecal immunochemical testing
R. McKinney
1
, C. Chapman
2
, J. Morling
2,3,4
, J. Weller
2
, A. Tangri
5
,
J. A. Simpson
1
, D. H. Humes
1,3,4
& A. Banerjea
1
1
Nottingham Colorectal Service, Nottingham University Hospitals NHS Trust,
Nottingham, United Kingdom,
2
Eastern Hub, Bowel Cancer Screening Programme,
Nottingham, United Kingdom,
3
National Institute for Health Research Nottingham
Digestive Diseases Biomedical Research Centre, Nottingham University Hospitals
NHS Trust, Nottingham, United Kingdom,
4
Division of Epidemiology and Public
Health, School of Medicine, University of Nottingham, Nottingham, United
Kingdom,
5
Nottingham City Clinical Commissioning Group, Nottingham, United
Kingdom
Purpose: We introduced direct GP access to faecal immunochemical testing (FIT)
in November 2017. We aimed to evaluate the initial 6-month outcomes for patients
undertaking FIT.
Methods: FIT (OC-Sensor
TM
) was recommended for all symptoms other than rectal
bleeding and rectal mass. No age restrictions were applied. GPs were advised to con-
sider alternatives to a two week wait (2WW) colorectal cancer (CRC) referral after
negative FIT results. All FIT requests were automatically logged prospectively. Elec-
tronic records of every patient with a FIT result between 7/11/17 and 10/5/18
were searched retrospectively.
Results: 25 (30.9%) of 81 patients with results 150 lgHb/gFaeces were diagnosed
with CRC a significantly higher diagnostic rate than lower FIT strata (Chi square
<0.001).
22.2% of patients with FIT 100149.9 lgHb/gFaeces had CRC. Overall CRC
detection between 4 and 149.9 lgHb/gFaeces was 2.5% and below NICE’s 3%
threshold. Three of four CRC patients with FIT results <80 lgHb/gFaeces had 1
marker of risk: anaemia, thrombocytosis or palpable mass on digital rectal examina-
tion.
Detection of non-colorectal cancers was significantly higher in patients with FIT
<4lgHb/gFaeces (P=0.032). The cost per CRC diagnosis for initial investigations
was £1,456 with FIT 150 lgHb/gFaeces (compared to £8,636 per CRC with FIT
10149.9 lgHb/gFaeces and £203,752 per CRC with FIT <4lgHb/gFaeces). Only
1 CRC was detected with FIT <4lgHb/gFaeces and the diagnosis of significant
bowel pathology was 5.3% in this group.
Discussion: Introduction of FIT appears safe and cost-effective in the short term.
Longer follow-up is needed to establish the miss rate after negative FIT results.
P003
Keeping FIT: faecal haemoglobin measurement with FIT has
stratification value in the diagnosis of colorectal cancer in all
symptom and age groups
H. A. H. Ibrahim, C. Chapman, J. Morling, J. Weller, A. Tangri, J. A. Simpson,
D. H. Humes & A. Banerjea
Queens Medical Centre, Nottingham, United Kingdom
Introduction: We incorporated FIT into our pathways for urgent investigation in
symptomatic patients in November 2017. We aimed to evaluate the value of FIT in
different symptom and age groups.
Methods: FIT (OC-Sensor
TM
) was recommended for all symptoms other than rectal
bleeding and rectal mass. No age restrictions were applied. All FIT requests were
automatically logged prospectively. Electronic records of every patient who had a
FIT result between 7/11/17 and 10/5/18 were searched retrospectively to ascertain
outcomes. Data were statistically analysed on GraphPad Prism.
Results: There were 1,947 FIT requests between 7th November 2017 and 10th
May 2018, representing 1,934 patients with mean age 66.3 (0.3 SEM) years
44.4% were male. 22 kits were unsuitable for analysis.
FIT 150 lgHb/gFaeces had CRC detection rate of 30.9%, significantly higher
than other FIT strata for all age groups, including those under the age of 60 years
(Chi square 22.38, P<0.0001).
FIT had similar discriminatory value in all symptom categories. There were no
CRC diagnoses in 305 patients with negative FIT (<4lgHb/gFaeces) investigated
for iron deficiency anaemia. CRC detection in 273 patients with negative FIT
investigated for altered bowel habit was 0.4%. In 531 FIT negative patients who
were investigated, diagnosis of other cancer was more likely than CRC (5.1% vs
0.2%, Chi square 22.93, P<0.0001).
Conclusion: FIT has diagnostic value in all symptom groups not just “low risk”
symptoms. It also has value in younger patients. NICE referral criteria for urgent
investigation shall need revision if these data are replicated widely.
P004
Keeping FIT: a 12 month impact assessment of the introduction of
FIT into 2WW practice in Nottingham
A. H. Khawaja, C. Chapman, J. Morling, J. Weller, A. Tangri, J. A. Simpson,
D. H. Humes & A. Banerjea
Queens Medical Centre, Nottingham, United Kingdom
Introduction: We introduced direct GP access to FIT testing in November 2017,
to improve “rule out” and “rule in” for two week wait (2WW) referrals and guide
“first test selection” in referred patients. We evaluated the impact on our 2WW
pathway.
Methods: FIT (OC-Sensor
TM
) was recommended for all symptoms except rectal
bleeding and rectal mass. It was required in the Nottingham University Hospitals
(NUH) 2WW pathway but not in the neighbouring Treatment Centre (TC). FIT
requests from 6/11/17 were prospectively logged. Prospectively collected data on
2WW referrals and first tests to 31/10/18, and cancer diagnoses between 1/11/17
31/12/18 were collated and analysed on GraphPad Prism.
Results: 5,437 FIT requests were recorded (rising through the year) and 4,680 FIT
results were evaluated. 522 CRC’s were recorded to 31/12/18 and 111 CRC diag-
noses (21.3%) were associated with a FIT result. 2WW referrals across Nottingham
rose by 33%.
3,338 FIT results (71.3%) were “negative” (<4lgHb/gFaeces). 1 confirmed and
1 likely CRC diagnosis have been recorded in patients with FIT <4lgHb/gFaeces,
yielding a CRC rate of <0.1% to date.
We used significantly less colonoscopy (P=0.016, Paired ttest), more CT
colonography (P<0.01, Paired ttest) and more flexible sigmoidoscopy (P<0.01,
Paired ttest) in our NUH 2WW pathway. There was no drop in CRC diagnosis
numbers and proportion of CRC diagnosed on 2WW pathways increased.
Conclusion: These early data suggest FIT at this threshold appears safe and effec-
tive in improving selection of patients for 2WW referral and identifying those most
likely to benefit from colonoscopy.
ª2019 The Authors
Colorectal Disease ª2019 The Association of Coloproctology of Great Britain and Ireland. 21 (Suppl. 2), 10–59
10
P005
Digital rectal examination for collecting stool sample for faecal
immunochemical test does not affect the diagnostic accuracy of
the test
M. Klimovskij, A. Khan & R. Harshen
East Sussex Healthcare Trust, Hastings, United Kingdom
Aim: FIT is a very accurate test to detect blood in the stool. At present, the collec-
tion of sample is by patients at home. We have investigated the accuracy of stool
samples obtained for FIT at the time of rectal examination as a point of care investi-
gation in our patients referred on 2-week wait CRC pathway.
Methods: Patients referred on the urgent 2-week wait colorectal cancer pathway
based on NG 12 symptoms were seen in the clinic and referred for investigations as
deemed suitable. A FIT was done at the time of examination in clinic where a rectal
examination was performed routinely. The results of the test were matched with the
FIT (10 µg/g) to ascertain the accuracy of the test to diagnose colorectal cancer.
More importantly, the negative predictive value of FIT was assessed to corroborate
the existing evidence of its accuracy as an effective rule out test.
Results: A total of 757 patients (83.2%; M:F 1:2; median age 71 years, IQR 63
80) were included. The accuracy analyses for CRC were sensitivity of 87.1%, speci-
ficity of 84.2%, positive predictive value of 23.13%, negative predictive value of
99.18%, and AUC of 87.4%.
Conclusion: DRE is a perfectly acceptable way of obtaining stool samples for per-
forming F IT test without impacting on its accuracy.
P006
Proposal of management strategies for symptomatic patients
referred on 2-week wait suspected colorectal cancer pathway
based on faecal immunochemical test results
A. Khan, M. Klimovskij & R. Harshen
East Sussex Healthcare Trust, Hastings, United Kingdom
Aim: FIT is being recognized as a very accurate test for ruling out colorectal cancer
(CRC) with a negative predictive rate of over 99%. We have incorporated FIT in
our colorectal referral protocol and are assessing its impact on the 62-day pathway.
Methods: All colorectal cancer referrals have a FIT in the primary care at the time
of referral. FIT-positive patients (10 µg/g) are referred directly for colonoscopy.
FIT-negative patients are seen in the colorectal clinic within 2 weeks. A new refer-
ral proforma has been devised for this after consultation with the local clinical care
groups. Awareness and education has been promoted through series of meetings and
FAQs. We are auditing the entire pathway very closely and the preliminary results
will be presented at the ACPGBI conference.
Results: This pathway was devised following a local study in our trust. A total of
757 patients (83.2%; M:F 1:2; median age 71 years, IQR 6380) were included.
The accuracy analyses for CRC were sensitivity of 87.1%, specificity of 84.2%, posi-
tive predictive value of 23.13%, negative predictive value of 99.18%, and AUC of
87.4%.
Conclusion: The detection rate of CRC from 2-week wait referral criteria based
on NG 12 guidelines is only 4 to 6%. By incorporating this very accurate and easily
performed test we are focussing our investigative resources to prioritize the patients
with significantly higher risk of CRC. We have ensured that we have built in
appropriate safety netting to avoid missing any CRC in the lower risk group.
P007
The role of bi-directional endoscopy during the same session in
patients with iron deficiency anaemia
N. Randhawa
1
, C. K. Khoo
2
, A. Varma
2
& M. Saedon
2
1
Nottingham University Hospital NHS Trust, Nottingham, United Kingdom,
2
Grantham and District Hospital, Grantham, United Kingdom
Introduction: The latest British Society of Gastroenterology (BSG) recommends
urgent upper gastrointestinal (UGI) and lower gastrointestinal (LGI) endoscopies for
patients with iron deficiency anaemia (IDA). This is the commonest indication for
simultaneous bi-directional endoscopy.
Aim: The aim of this study was to determine the utility of bi-directional endoscopy
in patients with IDA.
Methods: All unselected consecutive patients who had simultaneous bidirectional
endoscopies between July 2015 and December 2016 at a UK Trust (4 endoscopy
units) were recruited into the study. Demographic and clinical data which included
endoscopic findings and histology were analysed using SPSS.
Results: During the study period, 1,655 (male n=785, female n=870, age 63
(1695)) patients underwent bi-directional endoscopy. 972 patients were anaemic
(male =538, female =434) and 798 had IDA (male =402, female =396).
Overall, a significantly higher proportion of neoplastic lesion was detected in
LGI cohort than UGI cohort (n=61 vs. n=15, Pvalue <0.05).
A similar trend was seen in all anaemic patients (LGI n=49 vs. UGI n=9,
P<0.05).
This difference was also observed in the iron deficiency anaemia cohort (LGI
n=44 vs UGI n=5, Pvalue <0.05).
Conclusion: Our data suggest a more pragmatic approach to investigating patie nts
presenting with iron deficiency anaemia rather than subjecting all to bi-directional
endoscopy. It seems prudent that the recommendation for urgent endoscopic inves-
tigation for anaemia be revisited.
P008
Screening and diagnosis of colorectal cancer: a surgeon’s
perspective
H. Kumar
1
& K. Heer
2
1
Cairns Hospital, Cairns, Qld, Australia,
2
HK Surgical, Cairns, Qld, Australia
Purpose: To review the results of current CRC screening and diagnostic pro-
grammes in the UK and compare these to those used for other cancers and other
countries’ CRC programmes, to look for directions for the future.
Methods: Review of 2WW clinic outcomes (Public Health UK); JAG endoscopy
data; NBOCA summaries, Scottish CRC screening data (Information Service Divi-
sion, NHS Scotland) and some international CRC screening data.
Observations and comments:
1Given the incidence of CRC in the UK has remained stable over 25 years
placing emphasis on effective screening seems logical.
2NBOCA data demonstrates improved operative results year on year, with
90-day elective mortality down to 3.2%. Thus, further improvement in survival
would require detecting T0N0 disease instead of early detection of cancer. FOB/
FIT-based programmes are not effective in achieving this aim.
3Despite increased consultant numbers annual colonoscopies performed in the
UK remain stable (900,000/year) and less than 10% are performed for screening.
42WW clinics, 31- and 62-day targets have not produced results that justify
the massive resource diversion they mandate.
5Data from Germany and USA have demonstrated the benefit of colono-
scopy-based screening programmes in reducing colorectal cancer incidence and
hence mortality.
Conclusions: Replacing arbitrary non-clinical targets and faecal testing with colo-
noscopy-based screening will reduce CRC incidence and mortality. Such screening
is achievable in the UK through public-private-partnerships. The time has come to
“make every colonoscope count and count every colonoscopy!
P009
Emergency presentations of colorectal cancer in a district general
hospital
N. Anscomb, E. Chidambarasamy & S. Vasudevan
Colchester General Hospital, Colchester, United Kingdom
Background: Colchester General Hospital sees patients from a range of socioeco-
nomic backgrounds, including from Tendring District, a region encompassing some
of the poorest areas of the country. We conducted an audit to investigate route of
referral to diagnosis of colorectal cancer with hypothesis that a higher proportion of
emergency presentations are from lower socioeconomic areas.
Methods: Retrospective audit of all cases of colorectal cancer diagnosed at Colch-
ester General Hospital, in 12 months March 2017March 2018 (n=280).
Cases were analysed by referral route; National Screening Programme (NSP)
(referrals n=1,114, cancer diagnosis n=72/280; 25.7%), GP referral (n=160/280;
57.1% (2 week-wait: total referrals n=3,875 cancer diagnosis n=151, routine:
n=9)) Emergency presentation (n=48/280, 17.1%) figures compared to national
averages and analysis of stage and management intent for each referral category per-
formed. Postcode population used and percentag e of cases from each district diag-
nosed as emergency compared.
Results: Previous screening/investigation in only 4.17% emergency cases. Analysis
by postcode showed statistically significant difference between percentage emergency
presentations from Tendring District vs more affluent districts (22/154,164 vs 26/
381,886, P=0.0155 (Chi-squared)). Stage of referral was higher in emergency pre-
sentations than NSP (66.67% vs 22.22%, P=<0.0001 (Chi-squared)) but similar for
GP referrals (66.67% vs 57.50%, P=0.3002 (Chi-squared)).
Conclusions: 17.1% of cancer cases diagnosed as emergency, a figure higher than
national targets. Only 4.17% of these had previous investigation. Statistically signifi-
cant higher percentage of emergency presentations were from lower socioeconomic
areas highlighting that screening strategies targeted to specific population groups may
be of benefit in increasing uptake and reducing number of cases presenting as emer-
gencies.
2019 ACPGBI Poster Abstracts
ª2019 The Authors
Colorectal Disease ª2019 The Association of Coloproctology of Great Britain and Ireland. 21 (Suppl. 2), 10–59 11
P010
PRE-therapeutic MRI assessment of early stage rectal cancer and
significant rectal polyps to avoid major resectional surgery: a new
approach to the management of early stage rectal cancer the
PRESERVE trial
A. Shaw & On behalf of The PRESERVE Group
Royal Marsden Hospital, London, United Kingdom, London, United Kingdom
Background: Annually, 2000 UK patients are diagnosed with significant polyps
and early colorectal cancer (SPECC), defined as cancer without spread beyond the
bowel wall, yet only 15% undergo local excision (LE) with rectal preservation. This
trial will use out new MRI reporting system to significantly improve identification
of SPECC, the pathway for treatment and follow-up.
Objectives and research questions: To demonstrate an improvement in rectal
preservation rates in patients with pT1 and pT2 tumours.
Methods: Design:
Multicentre, prospective randomised trial
Intervention: MRI staging protocol as specified for use in this trial. Specialist
reporting of the tumour depth as seen on MRI will be used as a guide by the
colorectal MDT to identify patients suitable for organ preservation by local exci-
sion.
Patients will be randomised following local excision if they are found to have
moderate-risk features on post-operative pathology assessment.
Inclusion criteria:
Rectal lesion mrT3b
Exclusion criteria:
Neoadjuvant therapy
Metastatic disease
MRI contraindications
Outcome parameters:
Primary:
To demonstrate an improvement in rectal preservation rates in patients with
pT1 and pT2 rectal tumours
Secondary:
Accuracy of MRI
Stoma-free rates at 1 year
Readmission rates
Quality of life at 1 year
Group size calculation:
146
Time frame and funding
Feasibility grant funded.
Recruitment start April 2019.
P011
Use of CT colonography as first line investigation for suspected
colorectal cancer
E. Howie, S. Griffiths, T. Edwards & L. Hunt
Musgrove Park Hospital, Taunton, United Kingdom
Aims: To evaluate the use of CT Colonography (CTC) as a first-line investigation
for suspected colorectal cancer.
Methods: Data including indication, subsequent need for endoscopy, colonic and
extra-colonic findings were collected and analysed from 100 consecutive CTC scans.
Results: The main referrals for investigation were: change in bowel habit, anaemia
and bleeding. 36% of scans had significant pathology with 15 patients requiring fur-
ther endoscopic evaluation after CTC for polypectomy or biopsy. 10 patients had a
likely malignancy on CTC 3 progressed straight to surgery and 4 had a further
colonoscopy. 21 patients had significant extra-colonic disease requiring further inves-
tigation or management.
Conclusion: CTC is a safe and effective test. CTC and colonoscopy have similar
sensitivity and specificity for detecting colorectal neoplasms. CTCs detect clinically
significant polyps whereas colonoscopy may lead to unnecessary removal of small
polyps. An additional benefit of CTC is the ability to diagnose extra-colonic pathol-
ogy. Complication rates are comparable in both tests. The overall cost of colono-
scopy (£518) versus CTC (£434), taking into consideration further endoscopy rate,
is similar. Our results are similar to the national SIGGAR trial, demonstrating that
CTC is a safe alternative to colonoscopy. This also shows that conclusions from
large trials can be applied to smaller populations in district general hospitals. CTC is
especially useful in frail patients with co-morbidities who are at risk of complication
from bowel preparation. Individual patient choice is also important and suitable
patients should be educated and involved in the decision process.
P012
The old man and the CT: a retrospective case review of
investigations and outcomes from two week wait target referrals
for colorectal cancer in 80100 year olds
Poster not displayed.
P013
Implementation of a significant polyp and early colorectal cancer
multi-disciplinary team meeting: optimizing decision-making for
clinicians and patients
L. Longstaff, O. Aly, M. Dattani, S. Arnold, A. Venkatasubramaniam, B. Moran
& F. Di Fabio
Colorectal Surgery and Peritoneal Malignancy Unit, Basingstoke North Hampshire
Hospital, Basingstoke, United Kingdom
Introduction: The concept of Significant Polyps and Early Colorectal Cancer
(SPECC) encompasses a spectrum of colorectal lesions where definitive diagnosis
may be problematic. Optimal assessment and planning help to minimise under- or
over-treatment. We report on the introduction of a dedicated SPECC Multi-Disci-
plinary Team Meeting (MDT) as part of a Colorectal Cancer MDT.
Methods: We prospectively analysed the management and outcomes of all signifi-
cant colorectal polyps, defined as non-pedunculated lesions of >20 mm size, diag-
nosed and discussed in a SPECC-MDT which commenced in June 2018.
Results: There were 21 dedicated SPECC-MDTs in 7 months encompassing 42
SPECCs discussions in 35 patients (M/F:19/16; median age: 70 years). Three
SPECCs were discussed on multiple occasions due to lesions and patient-related
complexity. SPECCs were colonic in 22 (63%) and rectal in 13 (37%) patients.
Biopsies were taken from 23 (66%) lesions and all were benign. Colonic SPECCs
were managed by endoscopic resection in 17 (77%), while 4 (18%) required surgery.
One patient (5%) was not deemed fit for any treatment. Rectal SPECCs were man-
aged by endoscopic resection in 5 (38%), TEMS/TAMIS in 4 (31%) and surgery in
4 (31%). At final histology, adenocarcinoma was present in 8/35 (23%) SPECCs. Of
these, 3 (37.5%) had endoscopic resection as definitive treatment and 5 (62.5%) had
resectional surgery.
Conclusion: The implementation of a SPECC-MDT has enhanced the assessment
and definitive management for clinicians and patients. A SPECC-MDT aids optimal
management and is an excellent platform for recruiting patients in forthcoming clini-
cal trials focusing on rectal SPECC.
P014
Variability in height of rectal tumours; radiology versus endoscopy
who is more accurate?
P. A. Blake, J. L. Waterman, A. Kadhim & P. N. Haray
Prince Charles Hospital, Merthyr Tydfil, United Kingdom
Background: The accurate measurement and staging of rectal cancer, in particular
the distal margin of low rectal tumours, is of paramount importance to optimise
oncological surgical resection whilst preserving function. The lower the tumour the
greater the technical challenges, the operative time and the possibility of a temporary
or even a permanent stoma. Accurate localisation is also essential when considering
neo-adjuvant chemoradiotherapy. The objective was to compare tumour height on
Magnetic Resonance Imaging (MRI) with endoscopic measurement.
Design and methods: A retrospective analysis of rectal tumour heights on pre-
operative endoscopy and MRI in patients undergoing radical surgery with curative
intent. Any tumour recorded as being within 15 cm of the anal verge (AV) on
either MRI or endoscopy was included. MRI measurements were reported by two
specialist gastrointestinal radiologists whereas there were multiple endoscopists.
Results: Records of 81 patients with histologically confirmed rectal adenocarci-
noma were reviewed. On MRI median tumour height from the AV was 10.75 cm
(3.518 cm) compared with 23 cm (145 cm) on endoscopy. On comparing endo-
scopy with MRI, median difference was 12 cm (024 cm). Only rectal surgeons
documented tumour height in relation to the rectal folds. On no occasion was it
documented whether the tumour had been measured during insertion or withdrawal
of the endoscope.
Conclusion: Precise localisation of rectal tumours is imperative to plan surgery and
give informed counsel to patients. This study demonstrates the urgent need for a
standardised protocol for all endoscopists to use while recording the distal extent of
rectal tumours.
2019 ACPGBI Poster Abstracts
ª2019 The Authors
Colorectal Disease ª2019 The Association of Coloproctology of Great Britain and Ireland. 21 (Suppl. 2), 10–59
12
P015
Endoscopic tattooing of colorectal lesions are essential and impacts
quality of care and patient outcome
A. K. Nahid, P. Faraj, M. Imtiaz, G. Harinath, B. Aravind, A. K. Shrestha &
P. Basnyat
William Harvey Hospital, Kent, United Kingdom
Introduction: National Bowel Cancer Screening Programme (BCSP) published
guidelines advocates use of endoscopic tattooing for suspected malignant lesions to
assists identification and facilitates laparoscopic resections. However, endoscopic tat-
tooing practices are variable in endoscopic units resulting in re-endoscopy and at
times delaying of patient management.
Objectives: To assess the adherence to tattoo protocol for significant colonic
lesions at an endoscopy unit at a large DGH.
Method: Retrospective analysis of 143 consecutive patients with the colonic lesions
between January 2017 and December 2018 were performed. Data were collected
through reviewing patient’s notes, histopathology and endoscopy reports. Data on
lesions, complications, number and site of tattoo placed and any repeat endoscopy
for a tattoo were collected.
Results: 121 (85%) had malignant lesions and 12 (15%) were benign lesions. Only
in 78 (55%) colonoscopy report mentioned that tattoo had been placed. Out of these
78 cases in only 36 (46%) patients, the report stated the site of tattoo in relation to
the lesion. Unfortunately, 17% of all the cases required re-endoscopy to tattoo the
lesion prior to surgery.
Conclusions: Our study highlights the need for uniformity of tattoo practise
among endoscopist despite the BCSP guidelines. It remains that a significant propor-
tion of colorectal lesions are not tattooed during their first endoscopy. Some patients
had to have repeat test just for the purpose of tattoo which also had impact on
patient pathway. Active involvement and participation of all endoscopists in the
Colorectal MDT and the Complex Polyp MDT may help to improve Tattoo ser-
vice.
P016
Polyp progression in paediatric patients with familial adenomatous
polyposis syndrome a single center experience
C. Anele
1,2,3
, J. Xiang
4
, M. Hawkins
1
, S. K. Clark
1,2,3
, O. Faiz
1,2,3
,
A. Latchford
1,2,3
& W. Hyer
1,2
1
The Polyposis Registry, St Mark’s Hospital, London, United Kingdom,
2
St Mark’s
Hospital and Academic Institute, London, United Kingdom,
3
Department of Surgery
and Cancer, Imperial College London, London, United Kingdom,
4
Imperial College
London, London, United Kingdom
Introduction: Colectomy at a premalignant stage is the cornerstone of manage-
ment of familial adenomatous polyposis (FAP). Prior to surgery, annual colonoscopy
surveillance is recommended in children with FAP. This study aims to evaluate ade-
noma progression in children and factors influencing timing of colectomy.
Method: Patients with FAP under <18 years old at first colonoscopy and had under-
gone >1 surveillance colonoscopy were identified. Demographic, endoscopic, genetic
and surgical data were retrieved. Cumulative polyp counts were categorized as: 0100,
101200, 201500, 5011,000 and >1,000 whilst accounting for any polypectomies.
Follow-up time was split into: 2 years, 25 years, >5 years. Polyp progression was
defined as an increase in polyp-count category compared to the previous year.
Results: Eighty-four patients (50% male; mean age at first colonoscopy 13 years
1.97 were identified of which 83 (98.8%) were called up for screening due to fam-
ily history of FAP. At first colonoscopy, 67 (79.8%) had <100 adenomas. 29/83
(34.5%) had colonic polyps despite rectal sparing. The median increase in polyp
count/year was 26% (95% CI: 20% to 32%) P<0.001. At each colonoscopy,
approximately 80% of polyp counts remained in the same category compared to the
previous years across all three different time periods. Of the 45 (54%) patients who
had undergone surgery, 41 (91.1) underwent total colectomy and IRA.
Conclusion: Our results suggest that adenoma numbers remain relatively stable in
the majority of children under surveillance. Annual surveillance may not be required
and tailoring surveillance interval to phenotype may be a more appropriate strategy.
P017
Patients with normal previous colonic imaging have a low risk of
cancer and should not be referred to 2 week clinics
K. G. Flashman, A. Senapati, B. G. Ellis & M. R. Thompson
Queen Alexandra Hospital, Portsmouth, United Kingdom
Aim: The 2015 NICE GP Guidelines suggested that patients with >3% risk of can-
cer should be referred to the Two-week clinic (TWC)
1
but took no account of nor-
mal previous colonic imaging (PCI) which identifies patients at low life-time risk of
having cancer.
2
This study determined the risk of cancer in patients referred to the
TWC according to whether they had PCI and whether symptoms had resolved.
Method: Data were collected on 866 consecutive TWC patients (June 2017Oct
2018) including whether they had PCI. Significant polyps were >10 mm or three
smaller polyps.
Results: All 866 patients had flexible sigmoidoscopy (FS), 15% (134) had CTC, 4%
(34) had colonoscopy. 22 (81%) cancers were diagnosed by FS and 5 (19%) by
CTC. 239 (28%) patients had PCI and 92 (11%) had resolution of symptoms. Bowel
cancer was diagnosed in 4.3% (27/627) of patients not having PCI, 0% (0/239) in
those having PCI and 1.3% (1/92) in those with symptom resolution. Diagnosis of
significant polyps or proctocolitis was made in 2.5% (6/239) in PCI. 7.3% (18/587)
in no PCI and 4.3% (4/92) in those with symptom resolution.
Conclusion: Patients who have had PCI or whose symptoms have resolved are
unlikely to have cancer and do not need referral to the TWC. Those with persistent
symptoms could safely be referred to routine clinics for advice. This could dramati-
cally reduce the number of patients who suffer anxiety and morbidity from over
investigation as well as inappropriate referrals to TWC and surgical clinics.
1. BMJ. 2015;350:h2418.
2. Lancet. 2017;389:12991311.
P018
The two week rule colorectal cancer pathway: an update on recent
practice, the unsustainable burden on endoscopy and a call for
change
W. Maclean
1,2
& I. Jourdan
1,2
1
Royal Surrey County Hospital, Guildford, United Kingdom,
2
Minimal Access and
Therapy Training Unit (MATTU), Guildford, United Kingdom
Introduction: Survival for colorectal cancer is improved by earlier detection . The
need for rapid assessment and diagnosis has created a surge in the two-week rule
(TWR) referral rate and has subsequently placed a greater burden on endoscopy ser-
vices.
Between 2009 and 2014, a mean of 709 patients annually were referred to
Royal Surrey County Hospital with a detection rate of 53 cancers per year giving a
positive predictive value (PPV) for these patients of 7.5%.
Aims: We aimed to assess what impact the recent changes in NICE referral criteria
had on local cancer detection rate and endoscopy services.
Method: A prospectively maintained database of patients referred under the TWR
pathway for the financial year April 20172018 was analysed and the data cross-
referenced with endoscopy reports and other investigations.
Results: There were 1,414 referrals which is double that of previous years. 80.6%
underwent endoscopy as primary investigation. 62 cancers were identified, 51 being
of colorectal and anal origin (PPV 3.7%). 88 patients were diagnosed with other sig-
nificant colorectal disease defined as high-risk adenomas, colitis and benign ulcers. A
total of 10.6% of all patients referred therefore had a significant finding of any sort.
Conclusions: Since implementation of the updated NICE referral criteria, a dra-
matic rise in TWR referral rate has not increased cancer detection. It has however
placed significant pressure on diagnostic services. This signifies and paves the way
for a less invasive, cheaper yet specific ‘rule out’ test that can enable clinicians to
triage and reduce referral to endoscopy.
P019
Weight loss, one of the 2015 NICE GP referral criteria for the 2-
week clinic (TWC), is of little additional clinical value in identifying
patients with >3% risk of having bowel cancer
K. G. Flashman, A. Senapati, B. G. Ellis & M. R. Thompson
Queen Alexandra Hospital, Portsmouth, United Kingdom
Aim: Weight loss on multivariate analysis has independent predictive value (PV) for
bowel cancer with Odds Ratio of 1.66 (1.391.95).
1
To be of additional clinical
value it should significantly increase PV of the common symptom combinations of
bowel cancer.
1
Methods: 12,192 patients referred to a colorectal clinic were studied.
1
The effect
of presence/absence of weight loss was determined on the PV of symptom combina-
tions of change in bowel habit (C), rectal bleeding (B), abdominal pain (A) and peri-
anal symptoms (P) without an abdominal/rectal mass or IDA.
Results: 49% of patients with cancer had weight loss varying from 23% in patients
with pile symptoms +B-C-A+P to 59% in patients with +C-B+A. It also varied in
patients without cancer from 12% in +B-C-A+P to 49% in +C-B+A. There was lit-
tle effect of presence or absence of weight loss on PVs of symptom combinations
without abdominal or rectal mass or IDA; +C+B-A, 14.2% v 14.2%; +C-B-A, 5.0%
v 3.7%; +B-C-P, 8.8% v 5.8%; +B-C+P, 1.9% v 1.0%; +B-C+A+P, 1.6% v 1.3%
apart from +A-C-B, 6.6% v 1.9%.
Conclusion: As weight loss is common in patients with +A/+C but without can-
cer, it has little additional effect on PVs of symptom combinations, especially with-
out mass/IDA. It therefore cannot identify patients with >3% risk eligible for TWC
referral. The NICE guidelines need to clarify that weight loss is only of value in
those presenting with abdominal pain alone (+A-B-C) without mass/IDA to avoid
inappropriate referrals.
1. BrJSurg; 2017;104:1393404.
2019 ACPGBI Poster Abstracts
ª2019 The Authors
Colorectal Disease ª2019 The Association of Coloproctology of Great Britain and Ireland. 21 (Suppl. 2), 10–59 13
P020
Short term outcomes of a protocol of ESD/Hybrid ESD as the
primary resection strategy for all large (20 mm) rectal adenomas
A. Emmanuel, C. Lapa, S. Williams, S. Gulati, M. Burt, B. Hayee & A. Haji
King’s College Hospital, London, United Kingdom
Introduction: ESD is rarely practiced in western centres. Given its technic al diffi-
culty, many western experts believe indications for ESD are limited. However,
histopathologic diagnostic and treatment uncertainty resulting in over- or under-
treatment can have grave consequences in the rectum. As a result, our unit recently
opted for the exclusive use of ESD/Hybrid ESD to resect all large rectal adenomas.
We report short term outcomes using this protocol for 12 months.
Methods: Endoscopic resection (ER) of large (20 mm) colorectal adenomas were
analysed and outcomes compared after adoption of an exclusive ESD resection strategy
for 12 months (Period 2) compared to earlier resections (Period 1) when resection
strategy was based on lesion morphology, surface characteristics and ER experience.
Results: ER was performed for 185 rectal adenomas (period 1 n=154, Period 2
n=31), mean size 63 mm (range 20160 mm). ESD/Hybrid ESD was used for
97% of ER in Period 2 versus 61% in Period 1 (P<0.001). A trainee performed part
or all of 52% of ERs in Period 2 versus 7% in Period 1 (P<0.001). There were no
differences between time periods in complications (P=0.2), significant perforations
(P=1), post procedure bleeding (P=0.35) or risk of stenosis (P=0.66). Rates of
submucosal invasive cancer were similar (12.9% versus 7.8%, P=0.36).
Conclusions: ESD/Hybrid ESD for all large rectal adenomas, even when incorpo-
rating ESD training, is feasible and safe with sufficient expertise and experience. Data
for long term outcomes are desirable to evaluate potential benefits in oncological
results, reduced recurrence and potential fewer additional procedures.
P021
How does a recorded diagnosis of dementia affect the presentation
and management of patients over 65 years of age diagnosed with
colorectal cancer in England between April 2014 and March 2017?
Poster not displayed.
P022
Recurrence after advanced colorectal endoscopic resection results
in a substantial cost burden
A. Emmanuel, S. Williams, C. Lapa, S. Gulati, M. Burt, B. Hayee & A. Haji
King’s College Hospital, London, United Kingdom
Introduction: Little is known about the cost of treating recurrence after colorectal
endoscopic resection (ER) and few, if any, studies examine the cost associated with
ER taking into account treatment for recurrence. We evaluated the total cost of ER
at a tertiary centre.
Methods: ER (EMR and ESD) of large (20 mm) colorectal tumours with at least
the first surveillance colonoscopy were included. Procedure costs were based on
NHS national tariffs. Total costs included the cost of all endoscopi c or surgical treat-
ments and surveillance to 12 months.
Results: 473 patients had undergone at least the first surveillance colonoscopy after
ER. Overall, recurrence occurred in 69 (14.7%). Patients with recurrence required a
median of 2 ERs (range 26). 8 patients required surgery for recurrence. Mean cost
of treating those with recurrence was £3,976 versus £1,539 for those without
(P<0.001). 51% of lesions had been subjected to prior failed attempts at resection
or extensive biopsy samples and these were significantly more likely to recur (OR
3.8, 95% CI 2.16.9, P<0.001). The cost of treating these patients was significantly
more than those with minimal or no manipulation (P=0.001). Recurrence was less
likely after ESD (OR 0.25, P=0.01).
Conclusion: There is a substantial cost associated with the treatment of recurrence
after ER which has not been considered in studies evaluating the efficacy of EMR.
Injudicious prior manipulation contributes to the heavy cost burden. ESD should be
evaluated more closely as it results in fewer subsequent treatments for recurrence
and the possibility of longer surveillance intervals.
P023
The value of thrombocytosis in risk stratification of two week wait
referrals for colorectal cancer
N. Hanbali, K. Premji, J. Bunce, S. Mashlab, J. Voll, J. A. Simpson, D. J. Humes
& A. Banerjea
Nottingham Colorectal Service, Nottingham University Hospitals NHS Trust,
Nottingham, United Kingdom
Introduction: Primary care studies suggest that thrombocytosis (platelet counts
>400 910
9
/l) are associated with colorectal cancer (CRC). We aimed to establish
whether this marker has significant value in secondary care.
Methods: 2,978 patients referred to our colorectal two week wait (2WW) pathway
between August 2014 and August 2017 were reviewed. Patient demographics were
recorded prospectively and local electronic records systems were reviewed to retrieve
full blood counts (FBC’s) and cancer diagnoses. Patients with no recent platelet
count at time of referral or incomplete records were excluded. Statistical analyses
were carried out on GraphPad Prism.
Results: 2,220 patients were included in this evaluation. There was no significant
difference in the age distribution of those with thrombocytosis and those without.
There were significantly more females in the thrombocytosis group (71.8% v 53.9 %,
Chi squared 22.90, P<0.0001). 133 CRC were detected (6.0%) and CRC detec-
tion rate was significantly higher in patients with thrombocytosis (12.4% vs 5.4%,
Chi squared 14.86, P=0.0001). The CRC diagnosis rate was significantly higher in
females with thrombocytosis (11.0% v 3.1%, Chi square 18.62, P<0.0001) but did
not reach significance in males with thrombocytosis (15.8% v 8.0%, Chi square
3.31, P=0.069).
24 (of 13517.8%) patients with anaemia and thrombocytosis had CRC but
only 1 (of 671.5%) patient with thrombocytosis and no anaemia had CRC.
Conclusions: Thrombocytosis appears to have stratification value in the 2WW
population and particularly in females in this dataset. Further evaluation of its value
alone, or in combination with other stratification tests, is required.
P024
Sensitivity and specificity of anaemia blood tests in detecting
malignancy amongst patients referred via a 2-week wait suspected
lower gastrointestinal cancer referral pathway
K. Wanigasooriya
1,2
, S. Vayalapra
2
, N. Konda
2
, R. S. Karri
2
, P. Nightingale
1
,
A. Beggs
1,2
& T. Ismail
1,2
1
University Hospitals Birmingham NHS Foundation Trust, Birmingham, United
Kingdom,
2
College of Medical and Dental Sciences, University of Birmingham,
Birmingham, United Kingdom
Introduction: Unexplained anaemia is an indication for 2-week wait lower gas-
trointestinal (LGI) referral. We evaluated the significance of blood tests used for
anaemia classification into microcytic or iron deficient (IDA); in predicting a malig-
nant diagnosis amongst patients referred within the two-week-wait (2WW) LGI
cancer referral pathway.
Methods: A retrospective patient sample spanning a 5-year period identified
through clinical coding at a United Kingdom tertiary centre. Electronic patient
records interrogated for relevant data. Local laboratory reference ranges used.
Results: 462 patients (male =276; median age =76; benign =231; malig-
nant =231) included in the analysis. Most malignancies were colorectal adenocarci-
noma (95%). 239 (52%) were anaemic at presentation of whom 126 (53%) had
malignant pathology. 64% of patients with a microcytic anaemia had malignant
pathology compared to 52% with a low ferritin (confirmed IDA). Confirmed IDA
has a sensitivity of 31% (95% [CI] =19.1444.81%) and specificity of 79% (95%
[CI] =66.9887.89%) for a malignant diagnosis. Microcytic anaemia demonstrated a
sensitivity of 46% (95% [CI] =36.6754.75%) specificity of 72% (95%
[CI] =62.4379.76%). Unclassified anaemia has sensitivity, specificity as well as pos-
itive predictive values 54%, 51% and 53% respectively. ROC area under the curve
for unclassified anaem ia, microcytic anaemia, IDA were 0.528 (95% [CI] =0.476
0.581), 0.556 (95% [CI] =0.5040.609) and 0.548 (95% [CI] =0.4450.652).
Conclusion: Anaemia remains a frequent finding amongst patients referred as
2WW. Malignant pathology may present with non-IDA. In isolation, blood tests for
anaemia, microcytic anaemia or IDA are not useful diagnostic tests in detecting
malignancy amongst patients referred on this pathway.
P025
Variability in comfort scores between patients and healthcare
professionals for screening colonoscopy
D. Naumann
1
, S. Potter-Concannon
2
& S. Karandikar
2
1
University of Birmingham, Birmingham, United Kingdom,
2
University Hospitals
Birmingham NHS Foundation Trust, Birmingham, United Kingdom
Background: Patient comfort is key during colonoscopy to improve patient well-
being and ensure high quality. Patient, colonoscopist and Specialist Screening Practi-
tioner (SSPs) record their perceptions of patient comfort on a validated 1 (best) 5
2019 ACPGBI Poster Abstracts
ª2019 The Authors
Colorectal Disease ª2019 The Association of Coloproctology of Great Britain and Ireland. 21 (Suppl. 2), 10–59
14
(worst) ordinal scale at our UK screening centre. We investigated agreement
between these scores and comfort levels recalled by patients, and which factors influ-
ence them.
Methods: A prospective observational study included 498 bowel cancer screening
colonoscopies (April 2017March 2018). Inter-rater agreement of discomfort scores
between endoscopist, patient, and SSP was investigated using Cohen’s Kappa statis-
tic. Multivariate ordinal logistic regression was used to investigate the effects of
patient and colonoscopy factors on comfort scores.
Results: The median age was 68 (IQR 6471), and 320/498 (64.3%) were male.
SSPs had superior comfort score agreement with patients (0.638; “moderate agree-
ment”) than endoscopists had with the same patients (0.526; “weak agreement”).
Male patients reported lower scores than female patients (OR 0.483 (OR 0.499
(95% CI 0.3440.723); P<0.001). Endoscopists reported lower scores when there
was better bowel prep (OR 0.512 (95% CI 0.2790.938); P=0.030). No other fac-
tor was significantly associated with comfort scores.
Conclusion: There is considerable variability in perceived comfort levels between
healthcare providers and patients during screening colonoscopy. Female gender and
poor quality of bowel prep were factors associated with worse comfort. Endoscopist
that undertake screening colonoscopies may wish to consider both patient and
healthcare provider comfort scores in order to improve patient experience whilst
ensuring optimal quality assurance.
P026
Anaemia in patients referred within the two-week wait suspected
lower gastrointestinal cancer referral pathway is seldom due to
malignancy
K. Wanigasooriya
1,2
, N. Konda
2
, S. Vayalapra
2
, R. S. Karri
2
, P. Nightingale
1
,
A. Beggs
1,2
& T. Ismail
1,2
1
University Hospitals Birmingham NHS Foundation Trust, Birmingham, United
Kingdom,
2
College of Medical and Dental Sciences, University of Birmingham,
Birmingham, United Kingdom
Introduction: Anaemia is frequently detected in patients referred via the two-week
wait (2WW) suspected lower GI (LGI) cancer referral pathway. We explored the
factors associated with anaemia within a sample from this patient cohort.
Methods: A retrospective sample of patients referred over a 5-year period analysed.
Equal numbers of malignant and benign (or normal) cases randomly selected . Rele-
vant data obtained from electronic patient records. Statistical analysis using SPSS
(IBM, New York, USA).
Results: 239 of this 462 patient sample (male =276; median age =76) were anae-
mic at presentation. Binary logistic regression identified age, gender, chronic kidney
disease (CKD), aspirin usage, clopidogrel usage as significant predictors of anaemia
[Chi-Square =86.911, df =13 and P<0.001]. These factors were significant at the
5% level. Presence of malignancy, rectal bleeding, anticoagulant use, Ischaemic heart
disease, asthma/COPD, diabetes mellitus, recent surgery were not significant predic-
tors of anaemia. The odds ratio (OR) for CKD is 3.014 (95% CI: 1.2374.192), for
aspirin use 2.277 (95% CI: 1.4526.256) and clopidogrel use 4.209 (95% CI: 1.090
16.250). This model predicted 70% of anaemic patients correctly and 65% of patients
who were not anaemic correctly.
Conclusion: Anaemia is a common finding and indication for referral of patients
on the 2WW LGI pathway. However, malignancy is often not the cause of anaemia
in these patients. Rectal bleeding was not associated with anaemia. Alongside
advancing population age; prevalent antiplatelet use as well as comorbidities such as
CKD likely contributes to the volume of anaemia referrals as well as the investiga-
tion burden on patients referred with anaemia on this pathway.
P027
Accuracy of preoperative location of distal colorectal cancer: when
is sigmoid cancer actually rectal cancer? A unicentric retrospective
analysis
Y. Abou El Ella, Z. Ahmed, S. Azam, H. Blege, P. Yang, D. Zabrzycki,
O. Oluwajobi & P. Waterland
Russell’s Hall Hospital, Dudley, United Kingdom
Background: Accurate diagnosis of colorectal cancer is important as tumour loca-
tion determines management strategy. Early lesions may be missed on imaging and
precise location may be solely inferred from endoscopy with potentia l error.
This study aims to assess accuracy of initial distal sigmoid tumour localisation.
Methods: An electronic database of distal colorectal cancer from January 2014 to
November 2018 was created. Demographics, endoscopy reports and staging investi-
gations were recorded. Outcomes were assessed to determine inaccuracy between
initial endoscopic and operative localisation.
Results: A total of 113 patients were localised to distal sigmoid cancer via endo-
scopy.
Regarding CT staging: 22.1% (25/113) were non-visible. Visible cancers were
54% (61/113) sigmoid cancers, 15.9% (18) recto-sigmoid and 4.4% (5/113) rectal
cancers.
Pre-operative MRI was performed in 19/113 patients and showed: 21% (4/19)
sigmoid, 10.5% (2/19) recto-sigmoid and 63% (12/19) rectal cancers (2 mid, 10
upper rectal).
20% (5/25) of patients with non-visible lesions had MRI and 60% (3/5)
revealed upper rectal cancers. Of the remaining 20 patients, 70% (14/20) had sig-
moid, 20% (4) rectal and 10% (2) recto-sigmoid cancers at operation.
Intra-operatively, 18.6% (21/113) were identified as rectal cancer. Looking at
these: 38% (8/21) were non-visible on CT, 50% (4/8) of these had an MRI, with
75% (3/4) rectal cancers. Overall 9/21 (42.8%) had an MRI, 88.8% (8/9) showed a
rectal cancer and one non-visible lesion.
Conclusion: Endoscopic localisation of distal sigmoid cancer which is non-visible
on CT is unreliable. In these selected patients a pre-operative MRI may be consid-
ered to improve staging accuracy.
P028
Outcomes of large colonic polypectomy using OVESCO clips
B. Edgar, R. McBride, P. Allen & K. McCallion
Ulster Hospital, Dundonald, United Kingdom
Background: The Over-the-Scope-Clip system (OVESCO) is a novel endoscopic
method for mechanical soft tissue compression within the gastrointestinal tract. It
produces full-thickness closure of a perforation, in contrast to through-the-scope
clips whose closure is confined to the mucosal and submucosal layers. There is little
literature available documenting the use of OVESCO clips following endoscopic
removal of large colonic polyps.
Methods: 18 patients had OVESCO clips used to facilitate endoscopic resection of
large colonic polyps over a 16-month period. Endoscopy and histology reports and
electronic care records were reviewed for patient characteristics, histologic al diagno-
sis, complications and recurrence rates.
Results: 10 male and 8 female patients were included aged between 5185 years
(mean 71). Median polyp size was 30 mm and 4 of the 18 polyps contained adeno-
carcinoma. 14 patients had comorbidities which might preclude surgical resection. 2
patients had post-procedural bleeding. 1 underwent laparotomy for colonic perfora-
tion.
Average time to endoscopic follow up was 14 weeks. 3 patients had evidence of
recurrent disease at surveillance endoscopy and 1 required further surgical resection.
8 patients still had clips in situ at surveillance colonoscopy.
Conclusion: OVESCO clips can be used safely to facilitate endoscopic resection of
large colonic polyps as an alternative to surgery in patients with multiple comorbidi-
ties.
P029
Outcome of patients
M. Abdelrehman
1
, T. Munro
2
, S. Krishan
2
& J. Shabbir
1
1
University Hospital Bristol, Bristol, United Kingdom,
2
Bristol University, Bristol,
United Kingdom
Background: Colorectal Cancer (CRC) is the third most common cancer in the
UK. The incidence of CRC is increasing in younger population, whilst the overall
incidence of CRC is decreasing. It is important to evaluate this patient set as current
evidence supporting the referral, treatment and outcome of this group is limited.
Aim: To investigate the presentation, surgical treatment and outcome of cons ecu-
tive patients presenting with Colorectal Cancer (CRC) under 50 years of age
between Oct 2012 to June 2018.
Patients and methods: One hundred and twenty-three consecutive patients with
CRC diagnosed under the age of 50 were initially included. Eigh ty seven under-
went surgery and were included in the final analysis. The data was collected from
prospectively collected Enhanced Recovery Database and where needed case notes
were reviewed.
Results: Fifty two patients presented with locally advanced disease (60.0%) and dis-
tant metastasis were seen in 21 (24%). Twenty-eight patients developed post op
complications, with (n=7) requiring a re-operation. Average length of hospital stay
was 9.5 days. Mean overall survival from operation was 34 months. Recurrence
occurred in 14 (16%) cases with mean disease-free survival of 24.5 months.
Conclusion: A high proportion of younger CRC patients presented with advanced
disease. It is suggested that clinicians should be wary of CRC as a diagnosis in this
age group.
P030
Rectal cancer in patients under the age of 50 years: an advanced
disease?
Poster not displayed.
2019 ACPGBI Poster Abstracts
ª2019 The Authors
Colorectal Disease ª2019 The Association of Coloproctology of Great Britain and Ireland. 21 (Suppl. 2), 10–59 15
P031
Detection of malignancy in patients who have received follow-up
after a presentation of CT-confirmed acute diverticulitis
G. M. Monnery, A. C. James & C. Hoh
University Hospitals of Leicester, Leicester, United Kingdom
Aims: We reviewed the records of patients in whom a CT scan had confirmed a
diagnosis of acute diverticulitis. We then looked at the follow-up that these patients
received, with the aim of discerning how frequently further investigations such as
endoscopy went on to identify malignancy.
Methodology: Patients who had CT-confirmed diverticulitis with an acute presen-
tation between December 2016 and December 2018 were identified. Information
regarding the follow-up of these patients was collated from discharge letters, outpa-
tient clinic letters and requests/reports from further imaging/endoscopy. Histologica l
reports were also reviewed for patients in whom samples had been taken during
endoscopy or emergency surgery.
Results: Trustwide, 281 patients were identified as having a CT-confirmed diagno-
sis of acute diverticulitis; 136 of these were discharged with no follow-up. Of the
remaining patients, 103 underwent endoscopy and 16 had emergency surgery during
that admission. From histological samples taken during endoscopy, 3 patients were
identified as having malignancy- these were found to be colorectal, a recurrence of
Lymphoma and a metastasis from a suspected gynaecological primary. In all 16
patients who underwent emergency surgery, no malignancy was identified and his-
tology was consistent with diverticular perforation.
Conclusions: Malignancy was identified in 3 patients with CT-confirmed acute
diverticulitis who subsequently underwent an endoscopy. The gynaecological malig-
nancy was discernible on CT, therefore 2 incidences of occult malignancy were
identified in patients with CT-confirmed acute diverticulitis (1.94% of patients who
underwent endoscopy). Nevertheless, the inherent risks of endoscopy/ other investi -
gations require careful consideration when planning follow-up for patients.
P032
Does colorectal cancer location change with age? An epidemiologic
study of resectable colorectal cancer in the United States
D. E. Kearney, C. E. Cauley, A. Aiello, M. F. Kalady, J. Church, S. R. Steele &
M. A. Valente
Cleveland Clinic Foundation, Cleveland, OH, USA
Introduction: Recent single-institution studies have shown that colorectal cancer
(CRC) in patients <50 years is more common on the left side. This finding, how-
ever, has not been demonstrated on a population-wide basis. It is also unknown as
to whether this finding is equal across all racial groups, as previous studies have
shown a higher percentage of right-sided tumors in African-Americans.
Methods: We used the US Nationwide Inpatient Sample to identify all patients
with a diagnosis of colon or rectal cancer who underwent a colon or rectal resection
from 20002014 by ICD-9 procedure and diagnosis codes. Logistic regression mod-
els were used to determine the odds of a patient having left-sided CRC based on
age and race.
Results: Using population estimates, 1,650,542 patients underwent a CRC resec-
tion, with a mean age of 68.5. 61% of patients <50 years old had a left-sided tumor
compared to 45% of patients 50 (OR =1.9). The difference increased with age as
only 38% of patients 70 had a left-sided CRC (<50 vs 70 OR =2.5). The esti-
mated incidence of CRC in the <50 cohort increased over the study period (15.6%)
due to an increase in left-sided tumors. The distribution of CRC varied with race,
with African-Americans having a higher-odds of right-sided tumors (OR =1.13)
and Asians/Pacific Islanders having a lower-odds of right-sided tumors (OR =0.59)
compared to whites.
Conclusion: The incidence of CRC <50 is increasing. The majority of these
tumors were left sided, and the absolute number and proportion of left-sided CRC
has increased over time.
P033
Should GPs refer directly for colonoscopy for suspected colorectal
cancer rather than to urgent outpatient clinic?
Q. Ain, D. Naumann, T. Rahim, P. Jadeja, K. Hitchcock, C. Evans & A. Bajwa
University Hospitals Coventry and Warwickshire NHS Trust, Coventry, United
Kingdom
Introduction: Urgent referrals are made by GPs to outpatient clinic for ‘red flag’
symptoms of colorectal cancer (CRC). Colonoscopy and radiological investigations
may then be arranged by clinic specialists according to patient suitability. We inves-
tigated whether any patient factors predicted a final diagnosis of CRC, in order to
determine which patients might be suitable for direct investigation by GPs, rather
than being first referred to clinic.
Methods: A prospective observational study investigated patients with suspected
CRC at a UK NHS Trust (May 2017August 2018). Characteristics included gen-
der, age, and 6 pre-defined referral categories within our regional network: (i)
>40 years old, weight loss (WL) and abdominal pain (AP); (ii) <50 years old, rectal
bleeding (RB) and AB, iron deficiency anaemia (IDA), change in bowel habit
(CBH), or unintended WL; (iii) >50 years old, RB; (iv) >60 years old, IDA or
CBH >6 weeks; (v) rectal/abdominal mass; and (vi) anal mass/ulcer. Binomial multi-
variate logistic regression was used to investigate characteristics associated with final
diagnosis of CRC.
Results: There were 605 patients (52% female; mean age 65 years), of which 3.1%
had a final diagnosis of CRC. None of the 6 pre-defined referral categories, age or
gender were significant predictors of CRC.
Conclusions: Overall cancer rate was low, and no patient characteristics on GP
referral predicted CRC. These data may support a pathway in which GPs may refer
directly for colonoscopy or radiological investigations based on any of the pre-
defined ‘red-flag’ criteria rather than to a colorectal clinic first.
P034
Is it the colorectal team’s duty to diagnose non-colorectal cancers?
A review of 4,037 patients referred via the two weeks wait system
L. R. Wingfield
1
, J. Merchant
2
& A. Goede
2
1
University of Oxford, Oxford, United Kingdom,
2
Stoke Mandeville Hospital,
Aylesbury, United Kingdom
Aims: Bowel cancer is the fourth most common cancer in the UK (12% of all new
cancer cases in 2015). Early diagnosis is important, however, symptoms can be non-
specific with a low detection rate. We examined the two week-wait (2ww) colorec-
tal cancer pathway in Buckinghamshire NHS Trust during an 18-month period to
assess the waiting time to clinic, colorectal cancer detection rate, and other cancers
picked up during this screening process.
Methods: We retrospectively reviewed all 2ww patient records for the Bucking-
hamshire NHS Trust during January 2017 to August 2018. Medical record s, labora-
tory tests, and imaging results were reviewed.
Results: A total of 4,037 patients (average age =70.08, range: 23.399.04) were
referred over the time period. The average waiting time from referral to clinic was
12.29 days. Average time from referral to first treatment of colorectal patients
59.93 days (range: 7183) and in non-colorectal cancers was 48.03 days (range: 18
183). 134 colorectal cancers were detected (3.32%), in line with national data. Thirty
(0.74%) other malignancies were diagnosed including pancreas (n=7), renal (n=7),
gynaecological (n=3), lymphoma (n=2), lung (n=2), and others (n=5). 18.3%
of cancers diagnosed were not colorectal in origin.
Conclusions: Colorectal 2ww clinics can still safety achieve national targets for
bowel cancer screening, whilst serving a potentially unmet need in diagnosing and
referring patients for other types of cancers. Around 1 in 5 cancers diagnosed were
non-colorectal cancer in nature, and would not be diagnosed with Faecal Immuno-
chemical Testing and colonoscopy alone.
P035
The two week wait telephone conversation: is the Department of
Health missing something?
L. R. Wingfield
1
, N. Londsdale
2
, J. Merchant
2
& A. Goede
2
1
University of Oxford, Oxford, United Kingdom,
2
Stoke Mandeville Hospital,
Aylesbury, United Kingdom
Aims: Currently, NHS England is implementing a new standard to streamline the
colorectal cancer pathway. Tactics employed include a telephone triage system to
determine the initial investigations patients should have. We examined the 2-week-
wait patient pathway (2ww) in Buckinghamshire Healthcare NHS Trust (BHT) to
assess feasibility and safety of telephone triage.
Methods: We retrospectively audited patients referred to BHT via 2ww in June
2018 (our busiest month to date) by examining medical records, laboratory tests, and
imaging results.
Results: A total of 232 patients were referred and 209 attended clinic. Following
clinic, 107 patients underwent colonoscopy, 54 had CT colon, 18 had flexible sig-
moidoscopy, 6 had CTAP, 2 gastroscopy, 1 MRI, and 1 US as their primary inves-
tigation. Average time for first referral to patient being reviewed in clinic was
10 days. The average time to first investigation was 23.7 days. 14 colorectal cancers
were detected (6.69%). Other malignancies diagnosed include hepatic cancer
(n=1), melanoma (n=1) and prostate cancer (n=1). Primary benign conditions
diagnosed include inflammatory bowel disease (n=2), diverticular disease (n=53;
25.35%), haemorrhoids (n=24, 11.48%), and colorectal polyps (n=14; 6.69%).
Conclusions: Implementing a telephone triage system may reduce time to initial
investigation, however, face-to-face appointments aids selection of appropriate tests
and allows assessment of patient fitness. Many patients have other diagnoses the
initial assessment is invaluable in determining the need for further review. We would
argue that relentless focus on cancer targets ignores the value of good clinical assess-
ment, which aids assessment of patients, including the 93% without cancer.
2019 ACPGBI Poster Abstracts
ª2019 The Authors
Colorectal Disease ª2019 The Association of Coloproctology of Great Britain and Ireland. 21 (Suppl. 2), 10–59
16
P036
Two week suspected colorectal cancer referrals: how can we
improve our service?
B. Y. Khor, B. Das & A. Prabhudesai
General Surgery Department, The Hillingdon Hospital, London, United Kingdom
Background: The 2-week rule is useful pathway for investigating patients with
suspected colorectal cancer (CRC). In our unit, it was noted that the number of
these referrals had increased significantly but the diagnosis of CRC had not.
Aim: To examine the appropriateness of GP referrals to the Rapid Access Clinic
(RAC) for suspected CRC.
Methods: Data was collected prospectively for all patients referred to our unit as a
2-week wait by GPs with suspected CRC from 1st March 2018 to 30th May 2018.
Appropriateness of referrals was evaluated by Consultant surgeons running the RAC
and reasons for referrals deemed inappropriate were given. Patient demographics and
investigation results were recorded.
Results: 229 2-week wait referrals were included, of which 70 (30.6%) were
deemed ‘inappropriate’. Reasons for inappropriate referrals included: no red flag
symptoms (n=22), symptoms with a known cause (n=15), chronic symptoms
(n=17), recent or ongoing diagnostic investigations (n=12), or significant co-mor-
bidities (n=4). From 70 inappropriate referrals, 64 went on to have further investi-
gations, including CT colonography (n=27), endoscopy (n=29) and CT abdomen
and pelvis (n=8). None of these patients were diagnosed with CRC or any other
serious pathology.
Conclusion: Approximately every 1 in 3 referrals to the RAC for suspected CRC
were deemed inappropriate and there was a low incidence of serious patholo gy in
these patients. Inappropriate referrals have important implications in terms of
resource utilisation. Following this audit, we plan to provide education and feedback
to GPs with regards to current referral practice and review the referral criteria.
P037
Leicicarbon A suppositories an acceptable bowel preparation for
flexible sigmoidoscopy?
C. Brennan
1,2
, M. Douglas
2
, L. Robertson
2
& G. McFarlane
2
1
St Johns Hospital, Livingston, United Kingdom,
2
Gilbert Bain Hospital, Lerwick,
United Kingdom
Background: Poor bowel preparation has been associated with lower quality indi-
cators of colonoscopy performance, increased patient discomfort and lower adenoma
detection.
Frail elderly patients struggle to self administer enemas. Enemas must also be
used with caution in certain patient groups. Suppositories are easier to insert inde-
pendently. Leicicarbon A Suppository functions by a physical induction of reflex
bowel evacuation caused by carbon dioxide release as the suppository contacts mois-
ture and offers a lower side effect profile.
Aims: Primary aim was to assess whether Leicicarbon A suppositories were an
acceptable bowel preparation to the patient which offered good quality bowel prepa-
ration and allowed adequate examination of the bowel compared to current practice.
Methods: Randomised single (operator) blind prospective study. Patients were
assigned either Leicicarcbon A suppository or Phosphate Enema according to com-
puter randomisation. Outcome measures were 1) Length scope reached, 2) Adequate
examination length reached for indication, 3) Quality of bowel preparation and 4)
Patient experience
Results: 20 patients used Lecicarbon A suppository (LC) and 19 patients used phos-
phate enema (PE). Mean length of scope reached (LC 50 cm, PE 50 cm), adequate
examination length reached for procedure indication (LC 45%, PE 63%), inadequate
bowel prep (LC 35%, PE 21%), Patient experienced some side effects (LC9, PE5).
Conclusions: Bowel preparation with Leicicarbon A suppositories was less accept-
able to patients with more reported side effects. There was a higher rate of inade-
quate bowel preparation and lower rate of adequate examination for procedure with
the Leicicarbon A Suppositories.
P038
Tissue expression of extrinsic coagulation factors in colorectal
cancer
P. A. Rees
1,2
, H. W. Clouston
1
, H. Shaker
1
, J. Castle
1
, S. E. Duff
2
&C.
C. Kirwan
1,2
1
University of Manchester, Manchester, United Kingdom,
2
Manchester University
NHS Foundation Trust, Manchester, United Kingdom
Introduction: Tissue factor (TF), the initiator of the extrinsic clotting pathway, is
over expressed in colorectal cancer (CRC). The relevance of this to tumour biology
is poorly understood.
Aim: To investigate the cellular distribution and clinical significance of extrinsic
coagulation factors in CRC.
Materials and methods: The CHAMPion study (UKCRN no 8685), is a
prospective cohort study of patients undergoing curative resection for CRC.
Tumour and normal tissue was immunohistochemically stained for TF, thrombin,
PAR-1 and PAR-2. The relationship between cellular expression (dichotomised into
present/absent), adverse histological features and clinical outcomes was explored.
Results: Of the 159 patients recruited (100 male, median age 69), 147 had satisfac-
tory tissue samples for analysis.
TF expression was more common in tumour epithelium (tumour 93/146
[63.7%] vs. normal 13/133 [9.8%], P<0.001) and stroma (tumour 128/146 [87.7%]
vs. normal 70/132 [53.0%], P<0.001) compared to normal tissue.
In the stroma, TF expression inversely correlated with stage (T1 11/13 [85%]
vs. T2 37/39 [95%] vs. T3 61/66 [92%] vs. T4 19/28 [68%], P=0.006), however
the TF cellular receptor, PAR-2, was more frequent in poorly differentiated
tumours (poor 6/7 [86%] vs well/mod 41/140 [29%], P=0.004).
During follow-up (median 23 months [range 284]) 13/159 (8.0%) patients
died. On multivariable analysis, correcting for stage and EMVI status, epithelial TF
expression was associated with increased mortality (hazard ratio 7.37 [95% CI 1.33
40.80, P=0.022).
Conclusions: TF is more commonly expressed in the tumour epithelium and
stroma compared to normal tissue. Epithelial TF expression is an independent pre-
dictor of reduced survival.
P039
An in vitro model of HIPEC: facilitating pharmacological
assessment in colorectal peritoneal metastases
N. Un Nahar, U. Arshad, C. Goldring & P. A. Sutton
University of Liverpool, Liverpool, United Kingdom
Purpose: Hyperthermic intraperitoneal chemotherapy is a promising adjunct to
cytoreductive surgery in the management of colorectal peritoneal metastases. A dee-
per understanding of the intrinsic effects of HIPEC and its mechanisms of action,
efficacy and resistance are still needed. We aimed to establish an in vitro model of
HIPEC to facilitate pharmacological assessment.
Methods: Three human colon cancer cell lines (HCT116, DLD-1, LoVo) and a
murine colon cancer cell line (CT26) were studied. In addition to their growth in
monolayers, HCT116 and CT26 cells were grown in a three dimensional setting as
spheroids. The anti-tumour efficacy of oxaliplatin and mitomycin C as single agents
or in combination was assessed, following incubation for 90 min at either 37°Cor
43°C. Cell viability was analysed using the CellTiter-Glo
chemiluminescence assay
(Promega).
Results: LoVo cells were the most sensitive to the effects of hyperthermia. Admin-
istered as a single agent, oxaliplatin (dose range 120 lmol/l) showed enhanced
cytotoxicity in LoVo cells at 43°C compared to 37°C(P<0.05). No difference was
seen in CT26 cells. The results observed for HCT116 and CT26 cultured in a
monolayer were consistent with those seen in their respective 3D models. The addi-
tion of mitomycin C to hyperthermia and oxaliplatin showed no synergistic tumori-
cidal activity in any of the cell lines.
Conclusions: The combination of hyperthermia and oxaliplatin resulted in an
increased tumoricidal effect in the LoVo cell line monolayer. This in vitro model
presents an opportunity to further investigate the mechanism, efficacy and biological
basis of HIPEC.
P040
Biomarkers of the future in colorectal cancer
F. Kamel
1,2
, P. Nisar
1
& M. Soloviev
2
1
Ashford & St Peters NHS Trust, Chertsey, United Kingdom,
2
Royal Holloway
University of London, Egham, United Kingdom
Aims: To identify a comprehensive list of predictive biomarkers that may have an
association with CRC.
Methods: A thorough PubMed search was performed to identify 20 initial
biomarkers that are linked to CRC. For transcription factors (TFs) these were
entered into the TRRUST online database to identify upstream TFs associated with
them. These upstream TF’s were themselves inputted into the database to reveal
their downstream targets. For biological pathways (BPs) the initial markers were
inputted into the KEGG online database. After the pathways were identified a search
on the NCBI online database was performed to reveal the proteins involved in each
pathway. The results of the two mechanisms were compared to reveal a number of
predictive biomarkers.
Results: An initial PubMed search revealed 20 initial genes known to be involved
in CRC. 268 upstream transcription factors were identified from the initial 20 genes.
A total of, 2,069 downstream target genes were found from these. 141 biological
pathways were found from the original 20 genes. 265,111 proteins were found to be
involved in these pathways. Proteins involved in 70 or more different pathways were
then chosen leaving a total of 323 coding genes for these proteins. Combination of
these genes revealed 150 common future predictive biomarkers.
Conclusion: The 150 predictive biomarkers provide a baseline. Further investiga-
tion into whether each one is directly involved in CRC will now be carried out.
This will be in the form of qPCR to identify genes that are overexpressed in CRC
tissue.
2019 ACPGBI Poster Abstracts
ª2019 The Authors
Colorectal Disease ª2019 The Association of Coloproctology of Great Britain and Ireland. 21 (Suppl. 2), 10–59 17
P041
Biomarker status predicts outcome following cytoreductive surgery
and hyperthermic intraperitoneal chemotherapy for colorectal
cancer peritoneal metastases
D. Bhullar
1,2
, S. T. O’Dwyer
1,2
, M. S. Wilson
1
, C. R. Selvasekar
1
,A.
G. Renehan
1,2
, M. P. Saunders
1,2
, J. Barriuso
2,3
& O. Aziz
1,2
1
The Christie Colorectal & Peritoneal Oncology Centre, Manchester, United
Kingdom,
2
University of Manchester, Manchester, United Kingdom,
3
The Christie
NHSFT, Manchester, United Kingdom
Introduction: Patients with colorectal cancer peritoneal metastases (CRPM) have
lower median overall survival (mOS) of 16.3 compared to 24.6 and 19.1 months for
lung and liver metastases respectively on systemic chemotherapy. RAS gene muta-
tion status selects patients for anti-EGFR treatment, yet its role in predicting
response from cytoreductive surgery with hyperthermic intraperitoneal chemother-
apy (CRS/HIPEC) isn’t known.
Methods: A prospective database was used to collect information on patients
undergoing CRS/HIPEC for CRPM at national peritoneal tumour centre. Patients
were discussed at a specialist multi-disciplinary team meeting and underwent surgery
with curative intent. Data on tumour subtype, biomarker status, peritoneal cancer
index, and completeness of cytoreduction were collected. KaplanMeier curves were
used to calculate mOS from date of CRS/HIPEC and Cox regression was used for
multivariable analysis (MVA).
Results: Between 2004 and 2017, 195 patients underwent CRS/HIPEC for
CRPM, with a median follow-up of 19.8 months. Mutational status was available
for KRAS (n=108), NRAS (n=79) and BRAF (n=42). Mutation rates were:
KRAS =43.5%, NRAS =3.8%, and BRAF =14.3%. KRAS wild-type (WT)
tumours had a higher mOS (35.9 months) compared to mutants (29.8 months).
NRAS WT tumours had a higher mOS (35.9 months) compared to mutants
(14.3 months). Finally, BRAF WT tumours had a higher mOS (31.1 months) com-
pared to mutants (17.2 months). MVA showed KRAS mutation as an independent
negative prognostic factor for mOS (HR: 2.362, 95% CI 1.2334.525, P=0.010).
Discussion: KRAS mutation status predicts outcome following CRS/HIPEC for
CRPM, and could be used to personalize intraperitoneal treatments.
P042
Vitamin D influences gene expression in normal rectal mucosa and
impacts on biological mechanisms relevant to cancer prevention
P. G. Vaughan-Shaw, G. Grimes, M. Timofeeva, V. Svinti, S. M. Farrington &
M. G. Dunlop
Colon Cancer Genetics Group, MRC IGMM, University of Edinburgh, Edinburgh,
United Kingdom
Background: Controversy surrounds the role of vitamin D deficiency in the aetiol-
ogy of colorectal cancer. We investigated the effect of vitamin D on gene expression
in human rectal mucosa.
Methods: We sampled normal rectal mucosa and plasma in 178 volunteers,
extracted mucosal RNA and performed gene expression profiling (IlluminaHT-12
microarray). Plasma vitamin D (25OHD) was measured (liquid chromatography
mass spectrometry) and correlations sought between 25OHD and gene expression
using linear regression.
50 subjects were given vitamin D supplements (3,200 IU), then re-sam pled after
12-weeks to assess effects on expression. We performed mean-rank gene-set testing
for enrichment between correlative and supplementation datasets and gene-ontology
pathway analysis to investigate biological relevance.
Results: In the correlative study, expression of 533 genes was associated with circulat-
ing plasma 25OHD (P<0.01). Pathway analysis revealed 412 significantly enriched
biological processes, consistent with pleiotropic effects of vitamin D. Top-ranked
processes were highly relevant to carcinogenesis (eg. cell motility (P=8.8e-11); cell
migration (P=2.4e-10); biological adhesion (P=2.2e-10)). The 533 candidate
effector genes were tested in the supplementation study. These genes were significantly
enriched in differential expression analysis after vitamin D (up-regulation P=9.46e-
05, down-regulation P=0.003), consistent with beneficial effects of supplementa-
tion. Biological processes identified in the correlative dataset and relevant to carcino-
genesis were also enriched after vitamin D supplementation (P<0.01).
Conclusions: Plasma vitamin D level is associated with differential gene expression
in the rectum. Supplementation beneficially induces gene expression and biological
processes relevant to carcinogenesis. These findings underscore the rationale for
chemoprevention trials of vitamin D in colorectal cancer with mucosal gene expres-
sion as an intermediate biomarker.
P043
Lipid changes in colorectal tumours observed by direct mass
spectrometry analysis of tissue samples
Y. Shanneik
1
, E. Jones
2
, D. Forster
1
, K. Williams
3
, S. Pringle
4
,
A. W. McMahon
1
& O. Aziz
5
1
Wolfson Molecular Imaging Centre, University of Manchester, Manchester, United
Kingdom,
2
Waters Corporation, Wilmslow, United Kingdom,
3
Division of Pharmacy
& Optometry, University of Manchester, Manchester, United Kingdom,
4
Waters
Corporation, Manchester, United Kingdom,
5
The Christie Colorectal and Peritoneal
Oncology Centre, Manchester, United Kingdom
Introduction: Membrane lipids are linked to cancer progression and due to the
uncontrollable cell proliferation in cancers, there is an increased demand for these
lipids. Phosphatidylinositols (PI) are an important group of phospholipids known to
be upregulated in colorectal cancer (CRC).
Desorption electrospray ionisation mass spectrometry (DESI-MS) and rapid
evaporative ionisation mass spectrometry (REIMS) can generate spectra directly from
tissues, offering the prospect of real-time lipid ‘fingerprint’. This study aims to
demonstrate the use of DESI-MS and REIMS to determine lipid changes in CRC
xenografts before and after radiotherapy.
Methods: Fresh frozen irradiated and non-irradiated xenografts (HCT116) were
cut into serial sections. DESI-MS imaging in positive ion mode was undertaken on
Xevo G2-XS quadrupole Time of Flight (Q-ToF) (Waters, UK) followed by tan-
dem mass spectrometry (MS/MS). Serial sections were used for immunofluorescence
and Haematoxylin & Eosin (H&E) staining. Remaining sample blocks were used for
REIMS analysis (Q-ToF (Waters, UK)).
Results: Immunofluorescence and H&E staining allowed identification of hypoxic
and necrotic areas on the xenografts. Multivariate analysis of DESI-MS imaging
enabled identification of lipids PI (36:2) and PI (34:1) where they were upregulated
and PI (38:4) were downregulated in irradiated groups. Spectra from DESI-MS and
REIMS exhibited many of the same differentiators between the groups. This sug-
gests DESI-MS and REIMS can be used to determine lipid composition in CRC
Xenografts before and after radiotherapy.
Conclusion: This study suggests DESI-MS and REIMS may be used to provide a
‘lipid fingerprint’ in CRC that may be used to assess treatment response.
P044
Characteristics of cancer development in Lynch syndrome: a single
centre experience
R. Mak, Y. Maeda, F. V. N. Din, M. G. Dunlop & D. Speake
Academic Coloproctology, Western General Hospital, Edinburgh, United Kingdom
Purpose: Lynch syndrome (LS) is an autosomal dominant disorder associated with
mutations in DNA mismatch repair (MMR) genes. We describe the characteristics
of cancers developed in our LS patient cohort.
Methods: Data of patients diagnosed with LS following cancer and those identified
as mutation carriers (MLH1, MSH2, MHS6, PMS2) after cascade screening were
included.
Results: Some 124 patients (63 females), 66 of known pedigrees and 58 with can-
cer presentation, have been under surveillance for total of 881 years
(11,078 months).
In total, 93 cancers were diagnosed in 63 patients: the commonest cancer was
colorectal (CRC) (n=50, 54%), followed by endometrial (n=15, 16%), ovarian
(n=10, 11%), urothelial (n=5, 5%), and small bowel (n=4, 4%). Median age at
the time of first cancer diagnosis was 46 years (2474).
Mutation frequency was as follows: MSH2 (n=47, 38%), MLH1 (n=42,
34%), MSH6 (n=32, 26%) and PMS2 (n=3, 2%).
The incidence of CRC according to mutation was as follows: MLH1 (n=18,
45%), MSH2 (n=13, 35%), and MSH6 (n=8, 20%). Female and MSH6 carrier
had reduced risk of developing CRC (odd ratio (OR) 0.37, 95% confidence interval
(CI): 0.170.81, P=0.01, OR 0.35, 95% CI: 0.120.93, P=0.04, respectively).
There were 7 metachronous CRC (18%), 3 found under surveillance.
Conclusion: One in five patients with CRC developed metachronous CRC
despite surveillance. Up to one in 20 patients develop small bowel cancer as a conse-
quence of LS either as index presentation of cancer or after colonic resection for
cancer, the role of small bowel surveillance in LS is debated.
2019 ACPGBI Poster Abstracts
ª2019 The Authors
Colorectal Disease ª2019 The Association of Coloproctology of Great Britain and Ireland. 21 (Suppl. 2), 10–59
18
P045
Differentially methylated epigenetic profiles in blood and tumour
DNA from patients with rectal cancer
F. D. McDermott
1
, R. Dbeis
1
, C. Jones
1
, G. Shireby
2
, B. Knight
1
, J. Burrage
2
,
J. Mill
2
& N. J. Smart
1
1
Royal Devon & Exeter Foundation Trust, Exeter, United Kingdom,
2
Epigenetics
Department, University of Exeter Medical School, Exeter, United Kingdom
Background: Epigenetic research studies are adding to our understanding of cancer
pathways and helping to identify potential new biomarkers. The Risk Stratification for
Rectal Cancer Treatment (RIST) epigenetic study identified differentially methylated
probes (DMPs) and regions (DMRs) in matched rectal tumour and adjacent normal
mucosal samples. Some DMPs/DMRs were novel while others associated with col-
orectal cancer. The aim was to compare methylation profiles of matched RIST bloods
samples against RIST tissue and a healthy control population ‘EXTEND’.
Methods: DNA extracted from blood, underwent bisulfite modification and
methylation profiling (Infinium Methylation 450k/EPIC BeadChip). Methylation
(b) values presented as mean differences or absolute value with S.E.M. QC and sta-
tistical analyses with ‘R’. Unpaired t-tests to compare differential methylation
between RIST blood and EXTEND samples (significance P<0.0003 (Bonferroni
correction 0.05/161).
Results: Following Quality Control (QC): RIST tissues n=15, RIST blood
n=13 and EXTEND blood n=898 and 161/ 176 DMPS were analysed. 29 signif-
icant DMPs were identified in RIST blood, 18 annotated to UCSC genes, 17
hypermethylated and 1 hypomethylated. Genes associated with colorectal cancer
included ZNF625, ZMYND8, SLIT3, PRDM14, OPLAH, C9orf70, C9orf50 (hy-
permethylated) and KSR1 hypomethylated. All differential methylation patterns sim-
ilar in tissue and blood excluding ZMYND8 which was hypermethylated in blood
and hypomethylated in tissue.
Conclusion: Our study identified novel and previously described DMPs in both
rectal cancer tissue and associated blood samples. Although small, the sample size
provides useful data to target further research pending the results of a larger epigen-
ome wide association study (EWAS).
P046
Morphological and molecular risk markers for coexistent
adenocarcinoma in low-grade dysplastic areas of high-grade
colorectal adenomas
A. Emmanuel, S. Diaz-Cano, S. Gulati, S. Papagrigoriadis, B. Hayee & A. Haji
King’s College Hospital, London, United Kingdom
Introduction: Successful endoscopic resection (ER) relies on endoscopic diagnosis
to predict the risk of invasive cancer. However, a detailed evaluation of histopatho-
logical features and the molecular profile of the dysplastic mucosa to predict coexis-
tent invasive cancer is not available.
Methods: ER of large colorectal adenomas were analysed. A subset containing
high-grade dysplasia, intramucosal cancer or invasive cancer was identified and sub-
jected to detailed histopathological analysis: ulceration, distribution of high-grade
dysplasia, dysplastic nuclear grade, presence/distribution of necrosis, and distribution
of tumour-infiltrating lymphocytes (TIL). Microdissection, DNA extraction and
next-generation sequencing using a clinically relevant tumour panel of 24 genes
were performed separately for two areas with the highest morphological grade from
each lesion.
Results: ER was performed for 418 large (20 mm) adenomas. Histopathological
genetic evaluation was available in 70 high grade cases. Coexistent adenocarcinoma
significantly correlated with adenomatous mucosa featuring ulceration, mixed inter-
face/interstitial TIL, multifocal high nuclear grade, infiltrative edges, and multifocal
intraluminal necrosis. Multifocal intraluminal necrosis and high nuclear grade in the
adjacent low-grade dysplastic mucosa were driven by cooperative genetic abnormali-
ties of high-impact (FLT4), moderate impact (KRAS/NRAS for infiltrative edges,
FLT4, TP53, ERBB2), and low impact (FGFR3, PDGFA).
Conclusions: The dysplastic stage of high-grade adenomas is characterized by mul-
tiple cooperative genetic mutations. A subset of these identify a risk of coexistent
adenocarcinoma with a close correlation between genetic markers of angiogenesis
(FLT4), receptor activation (RAS/ERBB2), genome maintenance (TP53) and stro-
mal reaction (FGFR3, PDGFRA) with morphological features defined by high
nuclear grade, intraluminal necrosis, and inflammatory stromal reaction.
P047
Inflammatory reaction patterns and molecular genetics in high-
grade colorectal adenomas
A. Emmanuel, S. Diaz-Cano, S. Gulati, S. Papagrigoriadis, B. Hayee & A. Haji
King’s College Hospital, London, United Kingdom
Introduction: Inflammatory stromal changes are frequent findings in large colorec-
tal adenomas that can contribute to neoplastic progression, but their role is not fully
established in high-grade adenomas as predictors of coexistent invasi ve malignancy.
Methods: ER of large colorectal adenomas were analysed. A subset containing
high-grade dysplasia, intramucosal cancer or invasive cancer was identified and sub-
jected to detailed histopathological analysis: ulceration, distribution of high-grade
dysplasia, dysplastic nuclear grade, presence/distribution of necrosis, and distribution
of tumour-infiltrating lymphocytes (TIL). Microdissection, DNA extraction and
next-generation sequencing using a human clinically relevant tumour panel of 24
genes were performed separately for two areas with the highest morphological grade
from each lesion.
Results: ER was performed for 418 large (20 mm) adenomas. Histopathological
genetic evaluation was available in 70 high-grade cases. The extension of TIL posi-
tively correlated with the presence of coexistent adenocarcinoma, along with multi-
focal intraluminal necrosis and high nuclear grade. The presence of multifocal high-
grade dysplasia was driven by different cooperative sets of genetic abnormalities
regarding the TIL pattern: high-impact (TP53), and moderate impact (FLT4,
ERBB2, RET/RAS/RAF/ERK) for an interface pattern, and high-impact (TP53,
PDFGRA, FGFR3), and moderate impact (TP53, ERBB2, KRAS, RET/FGFR3)
for an interstitial pattern.
Conclusions: A more extensive lymphocytic response contributes to the progres-
sion of adenomas and is associated with morphological changes of increased risk of
adenocarcinoma (multifocal necrosis and high nuclear grade and abnormal TP53).
An interface pattern expresses proliferative (receptor kinase) and vascular (FLT4)
profile and the interstitial pattern outlines predominantly stromal (PDGFRA,
FGFR3) features.
P048
Pharmacological inhibition of acid ceramidase; a novel
radiosensitizer in a 3D colorectal cancer model
R. Clifford
1
, N. Govindarajah
1
, D. Bowden
1
, P. Sutton
1
, J. Parsons
1
&
D. Vimalachandran
2
1
Institute of Cancer Medicine, The University of Liverpool, United Kingdom,
2
The
Countess of Chester Hospital, Chester, United Kingdom
Introduction: We have previously utilized proteomic and immuno-histochemical
data to validate that high acid ceramidase (AC) expression confers poorer responses
to neoadjuvant chemoradiotherapy in colorectal cancer. Biological (siRNA) knock-
down of AC improved radiosensitivity in-vitro, as did pharmacological inhibitio n
with oral 5-FU derivative Carmofur.
Aim: To assess the radiosensitizing potential of LCL521, a novel small molecular
inhibitor (SMI), for locally advanced rectal cancer, in a 3D model.
Method: An ELISA activity assay was performed to establish the optimal dose of
LCL521 in three colorectal cancer cell lines with differential AC expression (HCT
116, LIM 1215, HT 29). Standard clonogenic assays were performed, alongside
spheroid volume measurement, across cell lines to assess cell survival following
LCL521 and increasing x-ray radiation.
Results: Activity assays revealed reduction in expression of AC to 18% with 10 lm
concentration of LCL521 in HCT115, 11.5% HT29, and 21% HT 29. Clonogenic
assays demonstrated reduced colony formation efficiency (colonies/number of cells
plated CFE) and improved radiosensitivity with LCL521 use across all cell lines.
LCL521 dosing 2 h pre-irradiation of 3D spheroid models improved radiosensitivity
across cell lines. HCT116 spheroid volume at day 15 post-LCL521
2.36 910
5
mm v control 4.15 910
5
mm.
Conclusions: Initial work has shown that pharmacological inhibition of AC with
LCL521 produces comparative radiosensitizing effects to biological (siRNA) inhibi-
tion in-vitro using these cell lines. Further work is needed to recapitulate these find-
ings in complex organoid models and ultimately in-vivo to establish a clinical role of
this novel SMI in locally invasive rectal cancer.
P049
Routine mismatch repair testing (MMR) in colorectal cancer (CRC)
are we failing patients?
T. Subramaniam, E. Richards, A. Krishnamoorthy, A. Aikoye, J. Francombe &
S. Ramcharan
South Warwickshire NHS Foundation Trust, Warwick, United Kingdom
The introduction of Mismatch Repair (MMR) Testing of all colorectal cancers
(CRC), as recommended by NICE (UK), aims at earlier identification of Lynch
Syndrome (LS) through pathways of cascade genetic testing and ultimately to pre-
vent familial CRC and associated tumours. We examined MMR results and their
use in care pathways of patients and their relatives.
Aim: To Assess compliance with NICE recommendations, and appropriate commu-
nication and counselling to patients and their relatives affected by CRC.
Methodology: Review of the prospectively accrued database identified all CRC
patients at our institution since commencing testing in July 2017 to June 2018.
Immunohistochemistry (IHC) detected the expression of MLH1, MSH2, MSH6 and
PMS2. For tumours with absent or reduced staining of 1 or more proteins, compre-
hensive screening for constitutional mutations in MMR genes was performe d at
the regional genetic laboratory. MMR results, follow-up and counselling data were
collected and analysed.
2019 ACPGBI Poster Abstracts
ª2019 The Authors
Colorectal Disease ª2019 The Association of Coloproctology of Great Britain and Ireland. 21 (Suppl. 2), 10–59 19
Results: Of 191 CRCs only 1 received pre-screening counselling and 106 were
tested by IHC histology. Those not tested were younger (67 versus 72 years,
P<0.05) and less likely to have undergone surgical excision (20.1% versus 30.2%,
P=0.128). More MMR positive tumours were from the right (57.8%) and trans-
verse colons (45%) and negative ones were more likely from the left colon/rectum
(67.8%, P<0.05). Most (89.5%) expressed MLH1 and PMS2 genes. Only 8 (42%)
were referred to the Family History Clinic.
Conclusion: Better national guidance and training of CRC teams is urgently
needed on managing and counselling patients and their relatives on MMR testing.
P050
Knockout of mucin 1 glycoprotein using CRISPR/CAS9 sensitises
rectal cancer cells to chemotherapy
I. S. Reynolds
1,2
, E. W. Kay
3,4
, D. A. McNamara
1,5
, J. H. M. Prehn
2
&
J. P. Burke
1
1
Department of Colorectal Surgery, Beaumont Hospital, Dublin 9, Ireland,
2
Department of Physiology & Medical Physics, Royal College of Surgeons in Ireland,
Dublin, Ireland,
3
Department of Pathology, Beaumont Hospital, Dublin 9, Ireland,
4
Department of Pathology, Royal College of Surgeons in Ireland, Dublin, Ireland,
5
Department of Surgery, Royal College of Surgeons in Ireland, Dublin, Ireland
Introduction: MUC1 is a heavily glycosylated O-glycoprotein that functions to
protect and lubricate epithelial cells. It has been implicated in chemotherapy resis-
tance and inhibition of apoptosis in several cancer types. The aim of this study was
to determine the association between MUC1 and response to chemotherapy in rec-
tal cancer cells.
Materials and methods: Using CRISPR/CA S9, the MUC1 gene was knocked
out from the CCL-235 rectal cancer cell line. The expression of MUC1 was assessed
using western blot. The IC50 for 5-FU and oxaliplatin was calculated using the
MTT assay in KO and wildtype (WT) controls. A dose response relationship for 5-
FU and oxaliplatin was determined in both the KO and WT cell lines and compar-
isons made using ANOVA.
Results: KO cells expressed no MUC1 relative to WT cells. The IC50 for 5-FU
in KO1 and KO2 was 64.52 lmol/l and 41.06 lmol/l respectively, versus
129.9 lmol/l in the WT. The IC50 for oxaliplatin in KO1 and KO2 was
5.27 lmol/l and 7.36 lmol/l versus 5.56 lmol/l in the WT. Following treatment
with 5-FU at 10, 25, 50, 250 and 500 lmol/l a reduced percentage of viable cells
were found in KO1 and KO2 when compared to WT cells at all doses (P<0.05).
Following treatment with oxaliplatin at 2, 5, 10, 50 and 100 lmol/l a reduced per-
centage of viable cells were found in KO1 and KO2 when compared to the WT at
50 and 100 lmol/l only (P<0.01).
Conclusions: MUC1 KO in rectal cancer cells increases their sensitivity to
chemotherapy and is a potential therapeutic target in rectal cancer.
P051
Utilisation of mismatch repair immunohistochemistry in clinical
practice correlation of endoscopic biopsy and resected specimens
C. Anele
1,2
, I. Al-Bakir
1
, D. Georgiou
3
, M. Moorghen
1
, H. Thomas
3
,
S. K. Clark
1,2
, O. Faiz
1,2
& A. Latchford
1,2
1
St Mark’s Hospital and Academic Institute, London, United Kingdom,
2
Department
of Surgery and Cancer, Imperial College London, London, United Kingdom,
3
Family
Cancer Clinic, St Mark’s Hospital, London, United Kingdom
Aim: Mismatch repair Immunohistochemistry (MMR IHC) is predominantly per-
formed on resected CRC specimens (SR) as molecular screening for Lynch syn-
drome (LS). There are scant data evaluating the reliability of performing MMR
IHC on pre-operative endoscopic biopsies (PB) or SR that have undergone neo-
adjuvant treatment. We aimed to evaluate concordance of MMR IHC between PB
and matched SR specimen.
Methods: Paired CRC PB and SR specimens were analysed for MMR IHC
(MLH1, MSH2, PMS2 and MSH6). Abnormal expression was defined as complete
loss of MMR protein. Concordance between PB and SR results was analysed.
Results: A total of 97 paired cases were analysed; 40/97 (41.2%) had undergone
neo-adjuvant chemoradiotherapy for rectal cancer.
In the 57 cases without preoperative treatment, 20 had abnormal MMR on SR
(13 loss of MLH1 and PMS2, 3 loss of MSH6 and PMS2, one loss of MSH6, one
loss of PMS2 and 2 loss of all 4 MMR proteins). Concordant staining patterns
between PB and SR were observed in 56/57 (98%).
In the cases who had neo-adjuvant chemoradiotherapy, nine had complete
pathological tumour response and were excluded. All SR showed proficient MMR
expression. There was 100% concordance in MMR IHC status between PB and
post-neoadjuvant therapy SR.
Conclusion: MMR IHC on endoscopy colorectal cancer biopsies appears to be as
reliable as that on resected specimens. Neoadjuvant chemoradiotherapy does not
appear to induce MMR protein loss, but it may result in complete tumour regres-
sion, therefore in the setting of neo-adjuvant therapy, MMR IHC on endoscopic
biopsies is recommended.
P052
Curtailing pro-tumourigenic and invasive signals via an orphan
nuclear receptor in colorectal cancer
M. Ismaiel
1,2
, B. Murphy
1,2
, H. Giffney
2
, S. Aldhafiri
2
, S. Fattah
2
, A. Baird
2
,
D. Crean
2
& D. Winter
1,2
1
Department of Surgery, St. Vincent’s University Hospital, Dublin, Ireland,
2
Schools
of Medicine, Veterinary Medicine and UCD Conway Institute, University College
Dublin, Dublin, Ireland
Introduction: Inflammatory processes are pivotal pathogenic factors in colorectal
cancer. Orphan nuclear receptors type 4A1 (NR4A1) are emerging as regulators,
concurrently repressing pro-inflammatory processes while activating resolution path-
ways. Whether this could improve cancer-related immune dysregulation is
unknown.
Aim: Curtailing pro-tumourigenic and invasive signals via an orphan nuclear recep-
tor in colorectal cancer.
Methods: Tumour and normal control tissue (n=20) obtained from patients
undergoing colorectal resection were exposed to a NR4A1 agonist (Cytosporone B
(CsnB) 4100 lmol/l) ex-vivo. The supernatant was collected and RNA was
extracted from tissues at 8 h. A cytokine/chemokine array was used to examine 104
secreted proteins associated with tumour inflammation, angiogenesis, fibrosis and
growth factors. Quantitative enzyme-linked immunosorbent assay (ELISA) and
qRT-PCR were used. Viability studies were performed on colorectal cancer cell
lines for toxicity experiments.
Results: Cytokine/chemokine array analysis revealed 50/104 were increased in
tumours including inflammatory (IL-8, TNF-a), angiogenic factors (angiopoietin 1,
vascular endothelial growth factor), and growth factors (fibroblast growth factor 7,
leukemia inhibitory factor). Of those, 30/50 were repressed by 50% by the
NR4A1 agonist. Multiple targets identified from the array were confirmed using
quantitative ELISA and/or qRT-PCR including cytokines (e.g. IL-8, TNF-a, IL-
23, IL-6), and chemokines (e.g. CCL3, CCL4, CCL20). Viability studies confirmed
that CsnB is non toxic.
Conclusion: Activation of an orphan nuclear receptor (NR4A1) represses pro-
tumourigenic mediators such as cytokines, chemokines, growth factors, and angio-
genic factors.
P053
The detection of intestinal anastomotic leaks with the serum C-
reactive protein concentration and computerised tomography
Y. M. Ho
1,2
, A. Kirubakaran
1
, H. Labib
1
& J. Laycock
1
1
Brighton and Sussex University Hospital, Brighton, United Kingdom,
2
Griffith
University, Gold Coast, Qld, Australia
Introduction: Serum C-reactive protein concentration (CRP) is a useful tool to
detect an anastomotic leak after colorectal surgery. There was a continuous debate
about the cut-off of CRP and the clinical applications of such cut-offs.
Methods: Patients who underwent an abdominal operation which involved an
intestine anastomosis between December 2016 and October 2018 were included in
the study. Post-operative dates (POD) were verified with the receipt times of the
corresponding histopathology specimens on the electronic pathology system, from
which the POD CRP results were also obtained.
Result: 327 patients were recruited during the study period. There were 122 left-
sided colonic resections, 117 ileocolic anastomoses, and 59 small intestine anasto-
moses. Indications of the procedures were illustrated in the following table.
Among left-sided colonic anastomoses, the Receiver operating characteristic
(ROC) curve was drawn with the anastomotic leakage as the state variable. The area
under the curve (AUC) was 0.52, 0.67, 0.79, 0.83 and 0.87, respectively. The calcu-
lated CRP ratio cut-offs of all the rest of PODs (2, 3, 4 and 5) were highly signifi-
cant in the laparoscopic group and the overall group (P<0.001). 30% of CT
performed would have been spared under the protocol.
Conclusion: Our cohort confirmed the incidence of anastomotic leaks in colorectal
surgery. POD 3 CRP below 234 mg/l, or CPR ratio (POD2/POD1) below 1.9
had a negative predictive value of 1 in open surgery (P=0.001 and <0.0001 respec-
tively). CRP ratio enabled the early exclusion of an anastomotic leak and spared
additional CRP and CT scans.
2019 ACPGBI Poster Abstracts
ª2019 The Authors
Colorectal Disease ª2019 The Association of Coloproctology of Great Britain and Ireland. 21 (Suppl. 2), 10–59
20
P054
Vitamin D deficiency is an easily modifiable risk factor which
predicts survival after colorectal cancer surgery independent of
systemic inflammatory response
P. G. Vaughan-Shaw
1
, L. Zgaga
2
, L. Y. Ooi
1
, E. Theodoratou
3
, M. Timofeeva
1
,
F. O’Sullivan
2
, F. V. N. Din
1
& M. G. Dunlop
1
1
MRC Human Genetics Unit, Institute of Genetics and Molecular Medicine,
Edinburgh, United Kingdom,
2
Department of Public Health and Primary Care,
Trinity College Dublin, Dublin, Ireland,
3
Centre for Global Health Research, Usher
Institute for Population Health Sciences and Informatics, University of Edinburgh,
Edinburgh, United Kingdom
Objective: To elucidate the relationship between circulating vitamin D (25OHD),
systemic inflammatory response and survival after colorectal cancer (CRC) surgery,
we assayed perioperative levels of plasma 25OHD and C-Reactive Protein (CRP) in
CRC patients.
Design: We assayed 25OHD and CRP in perioperative plasma samples from 92
patients undergoing CRC resection. We assayed 25OHD and CRP in two cohorts
of CRC patients (n=1,994, n=2,100) and conducted survival analysis. 22 CRC
patients received oral vitamin D supplementation (3,200 IU OD 12/52).
Results: Serial sampling revealed postoperative falls in 25OHD (day 12;
17.3 nmol/l; P=3.64e-9). CRP increased postoperatively yet the surgically-
induced fall in 25OHD was independent of CRP. In cohort analyses, 25OHD was
lower in the 12-months following operation (mean =48.82 nmol/l) than preopera-
tively (54.76 nmol/l; P=3.85e-3) with recovery after 24 months (52.22 nmol/l;
P=2.78e-3). Survival analyses demonstrated associations between 25OHD tertile
and CRC-mortality (HR =0.69, 95%CI: 0.520.91), and all-cause mortality
(HR =0.68, 95%CI: 0.550.84). Survival associations were strengthened after
adjustment for CRP and replicated in an independent cohort. We confirm ed a sur-
vival interaction between 25OHD and a genetic variant in vitamin D receptor
(rs11568820; P=0.01) supporting a causal relationship between vitamin D and sur-
vival. A survival model was developed with improvement in predictive model per-
formance with 25OHD (AUC 0.77 vs. 0.74). Finally, we demonstrated significant
response to vitamin D supplementation in CRC patients (25OHD 34.589 nmol/l;
P=6.40e-05).
Conclusions: Plasma 25OHD is a clinically-relevant survival biomarker indepen-
dent of CRP. This study establishes that vitamin D deficiency is an easily modifiable
risk factor associated with CRC survival and provides rationale for a trial of vitamin
D supplementation after CRC surgery.
P055
The use of the electronic frailty index to predict adverse outcomes
in elective colorectal cancer surgery
T. R. W. Ward
1,2
, P. A. Sutton
1
& D. Vimalachandran
1
1
Countess of Chester Hospital, Chester, United Kingdom,
2
University of Liverpool,
Liverpool, United Kingdom
Background: The National Bowel Cancer Audit 2017 demonstrated lower 2-year
survival in patients who did not undergo resection of their primary tumour
(1)
. These
patients are often very frail with multiple comorbidities. Our aim was to assess
whether frailty, as scored using the electronic frailty index, correlates with 5-year
mortality in colorectal cancer surgery
(2)
.
Methods: The degree of frailty for patients aged 70 and over who underwent a
resection for colorectal cancer between 1st January31st December 2010 was calcu-
lated. Data for each of the 36 parameters required to calculate the electronic frailty
index score was obtained retrospectively using electronic patient records. Kaplan
Meir survival analysis was performed with differences between groups examined for
using the log rank test and statistical significance defined at a level of 0.05.
Results: 116 patients were studied with a median age of 77 (7481); 53% of the
cohort was male. 44 patients (38%) were classed as fit, 43 (37%) had mild frailty, 25
(22%) had moderate frailty, and 4 (3%) had severe frailty. 5-year survival for the
cohort was 66.4%. For patients classified as fit, the 5-year survival was 75.0%,
65.1%, for the mildly frail, 64.0%, for the moderately frail and 0.0% for the severely
frail (P=0.023).
Conclusions: Frailty is a common geriatric syndrome and is increasingly more
common with an ageing population. Severe frailty has shown to be a significant pre-
dictor of adverse outcomes following colorectal cancer resections. A formal frailty
scoring system should be used to assist the surgeon in risk stratifying their patients.
P056
Pre-operative psoas major muscle size is negatively effected by
neoadjuvant chemoradiotherapy in patients with rectal cancer
G. Simpson, T. Marks, S. Blacker, J. Wilson & C. Magee
Wirral University Teaching Hospital, Wirral, United Kingdom
Introduction: Pre-operative skeletal muscle mass and sarcopenia represented by
psoas dimensions can predict outcomes following surgery. We aim to assess the
association between psoas major size and outcome in patients undergoing resection
for rectal cancer, and the effect neoadjuvant treatment has on pre-operative muscle
bulk.
Methods: A retrospective analysis of a prospectively maintained database of all
patients who underwent rectal cancer resection at a single centre between 2014
2017. Psoas Major and vertebral body cross sectional area was calculated at the L3
level and psoas major to L3 ratio calculated (PM:L3). Outcome measures included
30 and 90-day mortality, complications, inpatient stay and anastomotic leak.
Results: 121 patients were included. Median Age was 72 years (IQR:6478 years).
Male:Female ratio was 82:39. 30-day mortality was 0%. 90-day mortality was 0.83%.
Sixty-one patients underwent neoadjuvant chemoradiotherapy (50.4%). Thirty one
patients underwent APER (25.6%), 1 underwent proctocolectomy (0.83%), 1 under-
went completion proctectomy (0.83%) and 88 patients underwent anterior resection
(72.7%). Significant muscle loss during neoadjuvant therapy occurred (median
loss:25.9%, IQR:12.636.8%)(P<0.0001). No correlation was observed between
PM:L3 and inpatient stay. Patients with PM:L3 in the lowest quartile had a chest
infection rate of 11.1% and a complication rate 37.1% compared with 6.2% and
26.8% for those in the upper quartiles. Anastomotic leak rate in the PM:L3 lowest
quartile was 11.4% compared to 23.5% in patients in the upper quartiles.
Conclusion: Patients who received neoadjuvant chemoradiotherapy experienced a
significant reduction in muscle mass during treatment. This may be targeted with a
prehabilitation programme to reduce muscle loss and improve outcomes.
P057
Predicting hospital readmission following colorectal resection for
colorectal cancer
Poster not displayed.
P058
Colorectal cancer outcomes and survival in young vs elderly
patients: a population-based study
C. Anele
1,2
, C. Arhi
2
, G. Worley
1,2
, P. Ayling
3
, S. Clark
1,2
, A. Latchford
1,2
&
O. Faiz
1,2
1
St Mark’s Hospital and Academic Institute, London, United Kingdom,
2
Department
of Surgery and Cancer, Imperial College London, London, United Kingdom,
3
Faculty
of Medicine, School of Public Health, Imperial College London, London, United
Kingdom
Aim: Studies have reported an increase in the incidence of colorectal cancer (CRC)
in young adults. This group of patients are also thought to have poor histological
features and prognosis. We aimed to evaluate differences in clinicopathological fea-
tures and survival in young versus older patients.
Method: All patients over the age of 18 diagnosed with CRC 19972012 were
identified from the National Cancer Intelligence Network (NCIN) linked with
Hospital episode statistics (HES) database. Patients were stratified into three age
groups: (1) 1840 years, (2) 4160 years and (3) >60 years. Clinicopathological fea-
tures and overall survival were evaluated and compared between the groups.
Results: A total of 391,976 CRC patients were included: 5,307 (1.4%) in group 1,
65,538 (16.7%) in group 2, and 321,086 (81.9%) in group 3. Young CRC patients
presented with a higher incidence of poorly differentiated tumours (Group 1
24.7%, Group 16%, Group 3 15.9%, P=0.0001) and more advanced (UICC
stage 3&4) disease (Group 1 60.5%, Group 2 55%, Group 3 49%, P=0.001).
The 5-year OS were 62%, 60.1%, and 41% in groups 1, 2, and 3 respectively
(P<0.001). Multivariate analysis revealed older age (>40 years) was an independent
predictor of poor prognosis (HR, 1.2; 95% CI, 1.131.27; P<0.001) in group 2
and (HR, 2.3; 95% CI, 2.172.44; P<0.001) in the group 3.
Conclusion: Our data suggest that although young patients aged 1840 years with
CRC present with poorer pathological features and more advanced disease, they
have better prognosis.
P059
The impact of age on the clinicopathological characteristics and
prognosis of colorectal: UK single center retrospective study
C. Anele
1,2
, A. Askari
1,2
, L. Navaratne
1
, K. Patel
1,2
, I. Jenkins
1,2
, O. Faiz
1,2
&
A. Latchford
1,2
1
St Mark’s Hospital and Academic Institute, London, United Kingdom,
2
Department
of Surgery and Cancer, Imperial College London, London, United Kingdom
Introduction: Colorectal cancer (CRC) is uncommon in patients <40 years old.
Studies have worse histological features and survival in this age group. We aim to
evaluate age-related differences in clinicopathological features and prognosis in
patients diagnosed with CRC.
Method: We performed a single center retrospective review of patients diagnosed
with CRC between 2004 and 2013. Patients were stratified into three age groups:
(1) 1840 years, (2) 4160 years and (3) >60 years. Clinicopathological
2019 ACPGBI Poster Abstracts
ª2019 The Authors
Colorectal Disease ª2019 The Association of Coloproctology of Great Britain and Ireland. 21 (Suppl. 2), 10–59 21
characteristics, survival outcomes were compared between the three groups. We
excluded patients with familial predisposition syndromes, IBD and recurrent CRC.
Results: Overall, 1,328 patients were included (57.2% male). There were 28 (2.1%)
patients in group 1, 287 (21.6%) in group 2 and 1,013 (76.3%) in group 3. Group 1
had the highest proportion of rectal tumours (57.1% in group 1, 50.2% in group 2
and 31.9% in group 3; P<0.001). There were no statistically significant differences
between the groups in terms of histology and disease stage. After a median follow-
up of 71 months (IQR 3097), the 5-year over all survival (OS) in group 1, group 2
and group 3 were 69%, 77%, 60% respectively (P<0.001). The 5-year CRC dis-
ease-free survival was 44%, 78%, 77% respectively (P=0.022). Multivariate analysis
revealed that age was not an independent predictor of OS and DFS.
Conclusion: Young patients under the age of 40 years are more likely to present
with rectal cancer. Although histological features and disease stage are comparable,
they have worse OS and DFS compared to older patients.
P060
An evaluation to quantify oligometastatic disease in colorectal
cancer patients: a prospective single institute audit
C. Khanna
1
, A. Ansell
2
& S. T. O’Dwyer
1,3
1
The Christie NHSFT, Manchester, United Kingdom,
2
The Christie NHSFT,
Manchester, NH, USA,
3
The University of Manchester, Manchester, United
Kingdom
Background: Colorectal cancer (CRC) is the second most common cause of can-
cer-related deaths in the UK, with 40,000 new cases diagnosed every year
1
. At diag-
nosis, 25% of patients will have metastases involving lymph nodes, single organs
and/or peritoneum
2
. The term oligometastasis (OM) has been ascribed when multi-
ple secondary tumours occur in extracolonic sites
3
. Cancer guidelines recommend
case discussion at multidisciplinary team (MDT) meetings however, most MDTs are
organ-specific hence patients with OM may engage multiple MDTs. We evaluated
cases of OM through specialized MDTs.
Methods: CRC patients with OM discussed in colorectal, pelvic and peritoneal
MDT meetings from July-October 2018 were included. The primary and metastatic
sites were recorded, time from diagnosis to MDT and definitive treatment recom-
mendations reviewed.
Results: From 309 CRC patients 60, median age 57 years, 31 females, had OM.
The primary site was rectum/rectosigmoid 42%, sigmoid 23%, right colon 23%, left
colon 12%. The number of metastatic sites ranged from 18: 20% involved a single
site/organ, 37% two sites, 20% three sites. The predominant sites were peritoneum/
omentum 58%, liver 40% and lung 35%. MDT recommendations included systemic
chemotherapy in 42%, surgery 23% and combined treatments in 17% of patients.
Conclusion: In this series, OM was present in 20% of patients. Active treatments
were recommended in over 80% of patients the outcome of which is yet to be eval-
uated.
1Bowel cancer statistics. https://www.cancerresearchuk.org/health-
professional/cancer-statistics/statistics-by-cancer-type/bowel-cancer
2Nboca.org.uk. https://www.nboca.org.uk/content/uploads/2017/
06/nati-clin-audi-bowe-canc-2016-rep-v2
3NCI Dictionary of Cancer Terms. https://www.cancer.gov/publi
cations/dictionaries/cancer-terms/def/oligometastasis
P061
Identifying factors that might confound trials evaluating
cytoreductive surgery and hyperthermic intraperitoneal
chemotherapy (HIPEC) for colorectal peritoneal metastases: a UK
two-centre pre-trial IDEAL framework scoping collaborative
A. G. Renehan
1,2
, L. A. Malcomson
2
, S. Crane
3
, T. Cecil
3
, B. Moran
3
,M.
S. Wilson
1,2
, S. T. O’Dwyer
1,2
& F. Mohamed
3
1
Colorectal and Peritoneal Oncology Centre, The Christie NHS Foundation Trust,
Manchester, United Kingdom,
2
Division of Cancer Sciences, Faculty of Biology,
Medicine and Health, School of Medical Sciences, Manchester Academic Health
Science Centre, University of Manchester, Manchester, United Kingdom,
3
Department of Surgery, Hampshire Hospitals Foundation Trust, The Pelican
Cancer Foundation, The Ark, Basingstoke, United Kingdom
Background: Interpretation of trials of complex interventions, such as cytoreduc-
tive surgery (CRS) and HIPEC for colorectal peritoneal metastases (CRPM), may
be confounded (despite randomisation) because key factors in the intervention are
unbalanced. We used the IDEAL framework to evaluate the effects of variations in
patient factors and intervention on morbidity and mortality, as a requisite to trial
development (ClinicalTrials.gov NCT03733184).
Methods: We prospectively collected data identifying eleven ‘core’ variables from
200 patients with CRPM undergoing CRS and HIPEC at two UK centres (100
each centre: 201316). Primary endpoint was median overall survival (OS); sec-
ondary outcomes were Clavien-Dindo (C-D) complication grades 3 to 4; and 30-
day mortality.
Results: There were between-centre differences for the following: gender
(P=0.010); performance status (P=0.003); histological nodal positivity at syn-
chronous surgery (P<0.0001); HIPEC regimen selection (P<0.0001); blood trans-
fusion usage (P<0.0001); and stoma formation (P<0.0001). The median
peritoneal cancer index was 7 (IQR: 414). Median OS was 31.4 months, with no
between-centre difference. The combined C-D grade 3 and 4 complication rate was
12%, with no between-centre difference. There were no 30-day mortalities.
Interpretation: For patients with CRPM undergoing CRS and HIPEC, there
were differences in six of eleven putative confounders, but no differences in survival
and complications between centres. These results support the principle that, at least
in high-volume centres, favourable and equivalent outcomes are achievable for a
complex intervention like CRS and HIPEC, and support a platform for national tri-
als.
P062
Highly selective use of defunctioning stomas in anterior resection
in the era of laparoscopic management is safe and results in low
overall stoma rates
Z. Hussain
1
, A. Ghosh
1
, A. Emmanuel
2
& J. Ellul
1
1
Princess Royal University Hospital, London, United Kingdom,
2
King’s College
Hospital, London, United Kingdom
Introduction: The majority of patients undergoing anterior resection of the rectum
receive a defunctioning ileostomy in international practice. However, defunctioning
stomas are associated with a considerable increase risk of post-operative complica-
tions. Highly selective use of defunctioning stomas may be desirab le in the era of
laparoscopic management of both the primary operation and subsequent re-opera-
tions. We report outcomes of more selective use of defunctioning stomas in anterior
resection.
Methods: Analysis of prospectively collected data from a single laparoscopic col-
orectal surgeon. Low anterior resections (LAR) were included. Patients initially
managed without a defunctioning stoma were compared to those with a defunction-
ing stoma.
Results: 101 LAR were performed, 96% laparoscopic. 55% were managed without
an initial defunctioning stoma. Anastomotic leak occurred in 8 (14%) patien ts with-
out initial defunctioning stoma vs 4 (9%) with initial stoma (P=0.40). There were
no differences between groups in overall reoperations (P=0.98), readmissions
(P=0.75), or initial length of stay (P=0.43).
8 patients initially managed without a stoma developed an anastomotic leak and
were managed with conservative management (n=1), laparoscopic lavage and
defunctioning stoma (n=6), laparotomy and stoma (n=1). Of patients initially
managed with a stoma, 55% remained with a stoma 1 year later.
There were no mortalities in either group.
Conclusion: Highly selective use of defunctioning stomas in anterior resection is
safe and results in far fewer patients ultimately requiring a stoma. Furthermore,
almost all instances of anastomotic leak can be managed with laparoscopic surgery
without recourse to laparotomy and anastomotic take down.
P063
Comparison of outcomes for operative and non-operative
management of colorectal cancer in octogenarians
M. Abdelrahman, T. McCabe, H. Y. LIM, A. Mohammed, M. Hamed &
J. Shabbir
University Hospitals Bristol, Bristol, United Kingdom
Background: Colorectal cancer is extremely common in the elderly with over
43% of cases in the UK occurring in patients over 75 years old. Our aim was to
carry out a pragmatic study to review outcomes for patients with this diagnosis who
are 80 years or older.
Method: Data was collected retrospectively at the Bristol Royal Infirmary (Septem-
ber 2012August 2017). Data were collected from electronic hospital records.
Patients who had primary anal squamous cell cancer and those lost to follow-up
(moved area) were excluded.
Results: 217 patients’ records in total were reviewed (M: F, 50:50). Ninety-four
patients had operative intervention and 123 were treated conservatively. In the oper-
ative group, the median age was 83 years (8092). The median survival was
30 months (082) in the operative group. In the non-operative group, the median
age was 86 years (8097). One hundred and seven patients’ records were analysed.
Sixteen patients were excluded (anal cancer n=9, lost to follow-up n=7). Median
survival was 10 months (067) in the non-operative group.
Conclusion: This study shows favourable outcomes of resectional surgery among
octogenarians. The operative decision making should consider individual patient’s
comorbidities.
2019 ACPGBI Poster Abstracts
ª2019 The Authors
Colorectal Disease ª2019 The Association of Coloproctology of Great Britain and Ireland. 21 (Suppl. 2), 10–59
22
P064
Peri-operative anaemia in colorectal cancer patients: an audit of
current practice and development of a local pathway
R. Saunders, D. J. Smith, C. J. Walsh & J. I. Wilson
Wirral University Teaching Hospital, Liverpool, United Kingdom
Aims: The incidence of iron deficiency anaemia (IDA) in colorectal cancer patients
is high, potentially causing increased mortality and length of stay. Current guidance
recommends a target pre-operative haemoglobin (Hb) of >130 for male and female
patients. Our aim was to evaluate local current practice and develop a pathway for
use in this cohort.
Methods: Electronic records of all patients undergoing open, laparoscopi c and
robotic elective colorectal cancer resections from April to August 2018 in a single
centre were reviewed. The multi-disciplinary team was interviewed to establish local
current practice.
Results: 40 patients were included in the analysis. 21 (52%) were male with a med-
ian age of 71 (IQR 6378). Pre-o perative anaemia