Available via license: CC BY 4.0
Content may be subject to copyright.
Silvia de Sanjosé et al., 2023 eLife. https://doi.org/10.7554/eLife.91469.1 1 of 26
Epidemiology and Global Health
Design of the HPV-Automated Visual
Evaluation (PAVE) Study: Validating a
Novel Cervical Screening Strategy
Silvia de Sanjosé , Rebecca B. Perkins, Nicole G. Campos, Federica Inturrisi, Didem Egemen,
Brian Befano, Ana Cecilia Rodriguez, Jose Jerónimo, Li C. Cheung, Kanan Desai, Paul Han, Akiva P Novetsky,
Abigail Ukwuani, Jenna Marcus, Syed Rakin Ahmed, Nicolas Wentzensen, Jayashree Kalpathy-Cramer,
Mark Schiffman, for the PAVE Study Group
Division of Cancer Epidemiology and Genetics, National Cancer Institute, National Institutes of Health, Rockville,
MD, USA • ISGlobal, Barcelona, Spain • University Chobanian and Avedisian School of Medicine/Boston Medical
Center, Boston, MA, USA • Center for Health Decision Science, Harvard T.H. Chan School of Public Health, Boston,
MA, USA • Information Management Services Inc, 3901 Calverton Blvd Suite 200, Calverton, MD, USA • Department
of Epidemiology, University of Washington School of Public Health, Seattle, Washington, USA • Division of Cancer
Control and Population Sciences, National Cancer Institute, National Institutes of Health, Rockville, MD, USA •
Westchester Medical Center/New York Medical College, Valhalla, NY, USA • Feinberg School of Medicine at
Northwestern University, Chicago, IL, USA • Athinoula A. Martinos Center for Biomedical Imaging, Department of
Radiology, Massachusetts General Hospital, Boston, MA, USA • Harvard Graduate Program in Biophysics, Harvard
Medical School, Harvard University, Cambridge, MA, USA • Massachusetts Institute of Technology, Cambridge, MA,
USA • Geisel School of Medicine at Dartmouth, Dartmouth College, Hanover, NH, USA • University of Colorado
Anschutz Medical Campus, Aurora, CO, USA
https://en.wikipedia.org/wiki/Open_access
Copyright information
Abstract
Objective
To describe the HPV-Automated Visual Evaluation (PAVE) Study, an international, multi-
centric study designed to evaluate a novel cervical screen-triage-treat strategy for resource-
limited settings as part of a global strategy to reduce cervical cancer burden. The PAVE
strategy involves: 1) screening with self-sampled HPV testing; 2) triage of HPV-positive
participants with a combination of extended genotyping and visual evaluation of the cervix
assisted by deep-learning-based automated visual evaluation (AVE); and 3) treatment with
thermal ablation or excision (Large Loop Excision of the Transformation Zone). The PAVE
study has two phases: efficacy (2023-2024) and effectiveness (planned to begin in 2024-2025).
The efficacy phase aims to refine and validate the screen-triage portion of the protocol. The
effectiveness phase will examine implementation of the PAVE strategy into clinical practice,
cost-effectiveness, and health communication.
Reviewed Preprint
Published from the original
preprint after peer review
and assessment by eLife.
About eLife's process
Reviewed preprint posted
October 16, 2023 (this version)
Posted to medRxiv
October 9, 2023
Sent for peer review
August 17, 2023
Silvia de Sanjosé et al., 2023 eLife. https://doi.org/10.7554/eLife.91469.1 2 of 26
Study design
Phase 1 Efficacy: Nonpregnant women, aged 25-49 years, without prior hysterectomy, are
being screened at nine study sites in resource-limited settings. Eligible and consenting
participants perform self-collection of vaginal specimens for HPV testing using a FLOQSwab
(Copan). Swabs are transported dry and undergo testing for HPV using a newly-redesigned
isothermal DNA amplification HPV test (ScreenFire), which has been designed to provide HPV
genotyping by hierarchical risk groups: HPV16, else HPV18/45, else HPV31/33/35/52/58, else
HPV39/51/56/59/68. HPV-negative individuals are considered negative for precancer/cancer
and do not undergo further testing. HPV-positive individuals undergo pelvic examination
with collection of cervical images and targeted biopsies of all acetowhite areas or
endocervical sampling in the absence of visible lesions. Cervical images are used to refine a
deep learning AVE algorithm that classifies images as normal, indeterminate, or precancer+.
AVE classifications are validated against the histologic endpoint of high-grade precancer
determined by biopsy. The combination of HPV genotype and AVE classification is used to
generate a risk score that corresponds to the risk of precancer (lower, medium, high, highest).
During the efficacy phase, clinicians and patients will receive HPV testing results but not AVE
results or risk scores. Treatment during the efficacy phase will be performed per local
standard of care: positive Visual Inspection with Acetic Acid impression, high-grade
colposcopic impression or CIN2+ on colposcopic biopsy, HPV positivity, or HPV 16,18/45
positivity. The sensitivity of the PAVE strategy for detection of precancer will be compared to
current SOC at a given level of specificity.
Phase 2 Effectiveness: The AVE software will be downloaded to the new dedicated image
analysis and thermal ablation devices (Liger Iris) into which the HPV genotype information
can be entered to provide risk HPV-AVE risk scores for precancer to clinicians in real time.
The effectiveness phase will examine clinician use of the PAVE strategy in practice, including
feasibility and acceptability for clinicians and patients, cost-effectiveness, and health
communication.
Conclusion
The goal of the PAVE study is to validate a screen-triage-treat protocol using novel biomarkers
to provide an accurate, feasible, cost-effective strategy for cervical cancer prevention in
resource-limited settings.
PAVE Study Group
Brazil
Ana Ribeiro - ana-ribeiro.dantas@fiocruz.br
Tainá Raiol - taina.raiol@fiocruz.br
Center for Women’s Integrated Health, Oswaldo Cruz Foundation (Fiocruz), Brasília, DF,
Brazil.
MARCO Clinical and Molecular Research Center, University Hospital of Brasília/EBSERH,
Federal District, Brazil
Silvia de Sanjosé et al., 2023 eLife. https://doi.org/10.7554/eLife.91469.1 3 of 26
Cambodia
Te Vantha, MD, Director of Takeo Provincial Hospital,Cambodia
Thay Sovannara, MD, Medical Practitioner, Raffles Medical Group, Cambodia
Judith Norman, MD, Director of Women’s Health, Mercy Medical Center, Cambodia
judynorman@gmail.com
Dr. Andrew T. Goldstein, Director, Gynecologic Cancers Research Foundation.
drg.cvvd@gmail.com
Dominican Republic
Margaret M. Madeleine, MPH, PhD
Program in Epidemiology, Fred Hutchinson Cancer Center
mmadelei@fredhutch.org
Yeycy Donastorg, MD
Instituto Dermatológico y Cirugía de la Piel “Dr. Huberto Bogaert Díaz”, HIV Vaccine Trials
Research Unit, Santo Domingo, Dominican Republic. ydonastorg@gmail.com
El Salvador
Miriam Cremer MD; Basic Health International, Pittsburgh, PA 15205, USA. Ob/Gyn and
Women’s Health Institute, Cleveland Clinic, Cleveland, OH 44195, USA.
miriam.cremer@gmail.com
Karla Alfaro, MD Basic Health International, El Salvador, kalfaro@basichealth.org
Honduras
Miriam Cremer MD; Basic Health International, Pittsburgh, PA 15205, USA. Ob/Gyn and
Women’s Health Institute, Cleveland Clinic, Cleveland, OH 44195, USA.
miriam.cremer@gmail.com
Karla Alfaro, MD Basic Health International, El Salvador, kalfaro@basichealth.org.
Jaqueline Figueroa, MD, Programa Nacional contra el Cáncer, Tegucigalpa, Honduras.
jacqueline_figueroan@yahoo.com
Eswatini
Eyrun F. Kjetland, MD, PhD, Professor, Departments of Global Health and Infectious Diseases
Ullevaal, Centre for imported and Tropical Diseases, Oslo University Hospital Ullevaal, Oslo,
Norway; College of Health Sciences, Discipline of Public Health, Nelson Mandela School of
Medicine, University of KwaZulu-Natal, Durban, South Africa;Centre for Bilharzia and
Tropical Health Research (non-profit), BRIGHT Academy, Durban, South Africa
e.f.kjetland@medisin.uio.no
Teresa Norris, Founder and President, HPV Global Action, tnorris@hpvglobalaction.org
Silvia de Sanjosé et al., 2023 eLife. https://doi.org/10.7554/eLife.91469.1 4 of 26
Zeev Rosberger, PhD, Department of Oncology, Psychology and Psychiatry, McGill University,
Montreal, Canada, zeev.rosberger@mcgill.ca
Amelie McFadyen, MA, Chief Executive Officer, HPV Global Action,
ameliemcfadyen@hpvglobalaction.org
Marc Steben, MD, Ecole de Sante Publique, Université de Montréal; International society for
STD research, marc@marcsteben.com
Malawi
Amna Haider, MD, Epidemiologist, Department of Epidemiology and Training, Epicentre,
Dubai, UAE, amna.haider@epicentre.msf.org
George Kassim Chilinda, MD, Médecins Sans Frontières, Operational Centre Paris, Blantyre,
Malawi, gchilinda@gmail.com
Henry B.K.Phiri, MD-Sexual and reproductive health department, Ministry of Health, Malawi,
henryphiri06@gmail.com
Nigeria
Ajenifuja Kayode Olusegun, MD, Obafemi Awolowo University Teaching Hospital, Ile-Ife,
Osun state Nigeria, ajenifujako@yahoo.com
Adepiti Clement Akinfolarin, MD, Obafemi Awolowo University Teaching Hospital, Ile-Ife,
Osun state Nigeria, akinfolarindepiti@yahoo.co.uk
Adekunbiola Banjo, MD, College of Medicine University of Lagos, Lagos,
aafbanjo@cmul.edu.ng
Moharson-Bello Imran, MD, College of Medicine, University of Ibadan, Oyo state, Nigeria,
imranmorhasonbello@gmail.com
Oyinloye Temitope,MD, Obafemi Awolowo University Teaching Hospitals Complex, Ile-Ife,
Osun state, Nigeria, projectcoordinator.itoju@gmail.com
Bola-Oyebamiji Sekinat, MD, College of Medicine, Osun state University, Osogbo, Osun state.
Adeyemo Marydiya, MD, College of Medicine, Osun state University, Osogbo, Osun state
Tanzania
Karen Yeates-MD, MPH, Department of Medicine, Queen’s University, Kingston, Ontario,
Canada, yeatesk@queensu.ca
Safina Yuma, MD, Cervical Cancer Focal Person, Ministry of Health, Tanzania,
sychande@yahoo.com
Bariki Mchome, MD, Head, Reproductive Health Centre, Kilimanjaro Christian Medical
Centre, Kilimanjaro, Tanzania, barikimchome@gmail.com
Alex Mremi, MD, Head, Department of Pathology, Kilimanjaro Christian Medical Centre,
Kilimanjaro, Tanzania, alexmremi@gmail.com
Silvia de Sanjosé et al., 2023 eLife. https://doi.org/10.7554/eLife.91469.1 5 of 26
eLife assessment
This important study will provide evidence about a novel screen-triage-treat strategy
for cervical cancer prevention. The strategy would contribute to improving access to
cervical cancer prevention to vulnerable women with low access to health care, and,
therefore, at the highest risk of cervical cancer. However, the current protocol
description is currently incomplete and missing key information for clarity and
reproducibility.
Introduction
Global burden of cervical cancer
Cervical cancer causes substantial morbidity and mortality worldwide, with approximately
600,000 incident cases and 340,000 deaths each year.1 Globally, cervical cancer is caused by
persistent infection with one of ∼13 carcinogenic human papillomavirus (HPV) types.2 Cervical
cancer rates vary greatly worldwide due to uneven access to effective preventive measures;
nearly 85% of cervical cancer cases and almost 90% of cervical cancer deaths occur in low- and
middle-income countries (LMIC).3 The World Health Organization (WHO) has called for the
global elimination of cervical cancer, based on an advanced understanding of the natural history
of the causal carcinogenic types of cervical HPV infection and existence of effective preventive
technologies, including prophylactic HPV vaccination and cervical screening.2 ,4 ,5 However,
translation of the HPV-based prevention methods has not yet occurred in many LMIC.
While prophylactic vaccination will eventually decrease cervical cancer rates6 if high uptake
can be achieved in LMIC, maximum potential health benefits of vaccinating adolescents today will
not be achieved for 40 years. However, rapid implementation of a broad, effective cervical
screening campaign for adult women in the highest burden areas will advance cancer control by
20 years (Figure 1 ). The US Cancer Moonshot initiative for Accelerated Control of Cervical
Cancer has supported development of the new screening methods.7
Screening using HPV testing
The detection of carcinogenic cervical/vaginal HPV DNA is currently the most sensitive screening
method to distinguish those with an appreciable risk of precancer or cancer from those at low
risk.9 ,10 The WHO currently recommends using either screen-treat or screen-triage-treat
strategies. HPV testing is preferred to using visual inspection with acetic acid (VIA) as the primary
screening method where resources permit.4 There is a growing consensus that to achieve broad
screening coverage, HPV testing of self-collected cervicovaginal specimens would be optimal for
many populations.3 ,4 ,7 ,11 The results from meta-analyses comparing the performance of
self-collected to clinician-collected samples, using PCR-based HPV detection, showed similar
sensitivity and specificity for the detection of cervical precancer. 11 As of 2022, seven LMIC were
recommending HPV self-collection.12
There is broad consensus that HPV testing is the preferred screening method due to its high
negative predictive value and reproducibility. Treatment of all HPV-positive women with thermal
ablation (i.e., screen-treat strategy) is an option under current WHO guidelines4 ; however, this
may substantially overtreat infections, only the minority of which would progress to cancer.13
To make the best use of limited resources and concentrate on those at highest risk, triage strategies
are critical to determine which HPV-positive women are at higher risk of cervical cancer. Triage
with cytology or dual stain, as used in high resource settings, is unlikely to be a feasible solution in
Silvia de Sanjosé et al., 2023 eLife. https://doi.org/10.7554/eLife.91469.1 6 of 26
Figure 1.
Timing and deaths averted with one-time prevention campaigns: vaccination only, screening only,
or both
Footnote Figure 1: Projection of the relative timing of health benefits, measured as deaths averted, accrued by vaccination
and/or screening applied through one-time campaigns. Three scenarios were examined: 1) a one-time screening campaign
providing effective management for approximately 25% of 30-to 49-year-old women in 2027 (i.e., 20 birth cohorts) (green
line); 2) vaccinating 90% of 9-to 14-year-old girls in 2027 (i.e., 6 birth cohorts) with a bivalent HPV16/18 vaccination (orange
line); and 3) both a screening campaign and HPV vaccination for respective birth cohorts in 2027 (blue line). We considered
cervical cancer deaths averted over the lifetime of cohorts subject to the intervention, and conservatively assumed that
deaths averted due to screening would only occur after age 50, to account for prevalent cancers. Projections were developed
for the ∼65 LMIC with age-standardized cervical cancer incidence greater than 10 per 100,000 women.8 Even a relatively
short-term intervention that combines screening and vaccination could avert over 1.2 million deaths over the lifetime of
intervention cohorts, and implementation of effective screening campaigns could lead to reductions in cancer mortality
almost immediately.
Silvia de Sanjosé et al., 2023 eLife. https://doi.org/10.7554/eLife.91469.1 7 of 26
the majority of low-resource settings. Cervical visual examination using visual techniques,
including VIA, are often used as triage methods. However, these techniques are subject to human
error, have low accuracy for precancer, and require continuous training and quality control
measures.14 ,15 HPV genotyping is a newer, more accurate method of triage, as genotype
carcinogenicity varies predictably across populations.16 ,17 HPV16 is the most carcinogenic,
followed by HPV18/45, followed by HPV31/33/35/52/58, followed by HPV39/51/56/59/68.18
Automated Visual Evaluation (AVE) using an Artificial Intelligence (AI) algorithm shows promise
as a relatively simple and fast triage method that could be used in conjunction with HPV
genotyping to generate a highly accurate composite triage test.19 ,20
HPV-AVE (PAVE) strategy
The US National Cancer Institute (NCI) is currently undertaking a multi-centric study designed to
evaluate a novel cervical screening and triage strategy for resource-limited settings, including
settings with high HIV prevalence, as part of a global strategy to reduce cervical cancer burden.
The PAVE strategy aims to target cervical precancer accurately and affordably by 1) self-sampled
HPV screening; 2) triage among HPV-positive participants by combination of extended genotyping
and visual evaluation assisted by deep-learning-based AVE; and 3) treatment using thermal
ablation or excision (Large Loop Excision of the Transformation Zone (LLETZ). PAVE utilizes the
concept of risk-based management, defined as patient management determined by their risk of
precancer/cancer to minimize overtreatment in low-risk patients and concentrate treatment
resources on high-risk patients (Figure 2 ). This manuscript describes the study protocol,
structure, and logic of the PAVE strategy.
Methods
The PAVE study has two phases: efficacy (2023-2024) and effectiveness (planned to begin in 2024).
The efficacy phase aims to refine and validate the screen-triage portion of the protocol. The
effectiveness phase will research the introduction of the PAVE strategy into clinical practice.
Phase 1: Efficacy
Setting: Study design and locations
The study aims to recruit up tens of thousands women in nine countries: Brazil, Cambodia,
Dominican Republic, El Salvador, Eswatini, Honduras, Malawi, Nigeria and Tanzania (Figure 3 ).
Criteria for study site selection included: a) existing screening programs, b) willingness to research
self-sampled HPV for screening, c) capacity to run the HPV test d) availability of pathology services
to process biopsies and e) access to treatment services including ablation, excision, and cancer
treatment. Outreach and recruitment activities under the protocol include awareness campaigns
to inform the eligible female population in the catchment area. Research protocol details including
recruitment strategies, number of visits, and institutional review board approval are under the
control of individual study sites. The PAVE project is integrated into the screening activities at all
study sites, and at select sites is also integrated into other ongoing research studies.
Ethical and regulatory aspects
This multi-centric study is designed to function as a consortium. All ethical oversight of
recruitment and clinical data collection will be done by the local sites under guidance of their own
Institutional Review Boards and will follow local guidelines. All participants will have an informed
consent for participation in the study and can drop their participation at any time during the
Silvia de Sanjosé et al., 2023 eLife. https://doi.org/10.7554/eLife.91469.1 8 of 26
Figure 2.
Risk-based PAVE screen-triage-treat strategy provides risk stratification to assist in the
management of screening participants.
Note: hrHPV: refers to those HPV types considered as having a high potential capacity to induce cervical cancer when the
infection is persistent over time. It includes HPV 16,18,45,31,33,45,52,58,39,51,56,59 and 68
Figure 3.
Map of PAVE study sites
Silvia de Sanjosé et al., 2023 eLife. https://doi.org/10.7554/eLife.91469.1 9 of 26
Figure 4.
Schematic of PAVE protocol elements
Figure 5.
Theoretical approach to compare the HPV AVE strategy and the standard of care (SOC) screening
and triage outcome
Footnote: Figure 5 is a hypothetical example showing how PAVE and SOC will be compared. Under a specific specificity
value (which will be determined by the SOC), we will compare the difference between the sensitivities of PAVE and SOC. In
this example, we had three categories for HPV genotype groups and three categories for AVE (normal-indeterminate-
precancer/cancer), and in total nine PAVE categories.
Silvia de Sanjosé et al., 2023 eLife. https://doi.org/10.7554/eLife.91469.1 10 of 26
process. All study documents written in languages other than English are officially translated into
English for study records. Compiled analysis of de-identified data by NCI research staff for study
purposes is deemed non-human subjects research by NIH.
Protocol overview
In-country elements: Patient enrollment and data collection
The steps of the PAVE protocol include 1) determination of study eligibility, 2) informed consent, 3)
self-sampled HPV testing, 4) cervical image collection and biopsy collection for those testing HPV
positive, and 5) treatment as indicated per local protocols.
Enrollment
Eligibility criteria are: individuals with a cervix aged 30-49 years (general population) or 25-49
years if living with HIV (WLHIV), not currently pregnant, and able to understand study risks,
benefits, and alternatives and provide informed consent in their native language. Those eligible
for and interested in study participation undergo informed consent per local protocols. Those who
choose to enroll provide basic demographic information (age, parity, HIV status if known).
HPV self-collection
Participants self-collect a vaginal sample for HPV testing using a FLOQSwab (Copan) following
instruction by study personnel. Self-samples (FLOQSwab) are delivered dry for testing. All sites
intend to use the ScreenFire HPV risk-stratification (RS) assay (ScreenFire) (US patent 11091799,
Atila Biosystems Inc, Sunnyvale, CA, US).21 ,22 However, other HPV tests may be acceptable
alternatives if they can provide genotyping information in following groups: HPV16, HPV18/45,
HPV31/33/35/52/58, and HPV39/51/56/59/68.
HPV tests are run onsite or in local laboratories in few days, with results returned to women
quickly per local protocols. Women screening HPV-negative are informed of their results,
reassured about their low subsequent risk of cervical cancer, and their participation in the study
ends at most sites. The exception is El Salvador, at which 5% of those screening HPV negative
undergo colpscopic examination. On average, approximately 80% of participants will screen HPV-
negative, but this varies by study population.
Triage of HPV-positive results: image collection and biopsy
Women with HPV positive results undergo speculum exam with application of 5% acetic acid.
Cervical images will be collected by a trained study provider using a dedicated device (Iris, Liger
Medical LLC, Lehi, UT, US). Local clinicians also record their VIA assessment (negative, positive,
suspicion of cancer) or colposcopy impression (normal, low-grade, high-grade or more severe).
Following image capture, pathology specimens are collected. Biopsies will be collected from up to
four acetowhite areas for each participant. If no acetowhite areas are seen, then an endocervical
sample (using curette or brush) or cytology will be collected. Sites using colposcopy will collect
punch biopsies per standard practice. Sites using VIA will use Softbiopsy/SoftECC brush biopsy ™
(Histologic LLC, Anaheim, CA, US), a device that is simpler to learn and perform and is associated
with lower bleeding risk. All women that have an HPV positive test are expected to have a
histologic diagnosis (biopsy, endocervical sample (ECC), and/or excisional tissue diagnosis). HPV-
positive women with a negative triage evaluation initially, but who are later identified by PAVE
activities to have a CIN2+, will be flagged and the clinical sites will be notified to permit “safety
net” recall for adequate management.
Silvia de Sanjosé et al., 2023 eLife. https://doi.org/10.7554/eLife.91469.1 11 of 26
Treatment
Treatment is provided for women meeting criteria per local protocols: VIA-positive or suspicion of
cancer in sites using VIA, CIN2+ on biopsy and/or high-grade colposcopy impression in sites using
colposcopy/biopsy, or HPV-positive and acetowhite changes or HPV 16,18/45 positive or HPV-
positive for sites using screen-treat protocols. For sites using VIA, treatment decisions for those
screening VIA-positive will be based on the standard of care, most commonly an adaptation of the
WHO visual assessment for treatment (VAT) criteria for use of ablation. For sites using LLETZ,
treatment decisions will follow local protocols. Table 1 describes screening, triage, biopsy, and
treatment protocols for each site.
Central elements: data management, HPV testing,
AVE algorithm development, quality assurance,
statistical analysis
Data management
Study data including demographic survey information, HPV test results (negative/positive,
genotype for positive results), VIA or colposcopy impression, and local pathology results (cytology,
biopsy and/or excisional specimen) are collected using DHIS2, Redcap or WEMA platforms. The
collected data are associated with the corresponding images obtained during the triage visits. To
ensure confidentiality, all personal identification information is removed from datasets before
transferring outside the country of origin. De-identified datasets are securely transferred to a
common server handled by the NGO specialized in country adaptation of DHIS2 information
systems Enlace Hispano Americano de Salud (EHAS) and from there compiled data are transferred
to Information Management Services (IMS), the NCI data management contractor, for data storage
and analytic support during the course of the study. The noteworthy element in this study design
element is that data rights (and residual biospecimens) will remain with the study site partners
within their countries. The data on loan for PAVE analyses will be stored securely and data can be
withdrawn and destroyed at any time by study site partners. This arrangement is important to
many aspects of international data and biospecimen sharing.
HPV testing/typing
The ScreenFire HPV test is a new assay designed to detect the 13 high-risk HPV (hrHPV) genotypes
grouped into the four risk groups described above, and specifically engineered to provide risk
stratification by HPV genotype based on carcinogenicity. ScreenFire includes an internal control
for sample quality guidance. The ScreenFire HPV assay is an isothermal, multiplex nucleic acid
amplification method that uses 3NT technology to reduce false positivity and increase assay
performance. ScreenFire can be run on 1 to 96 samples per batch, requires only basic pipetting
skills, and takes approximately 2.5 hours in total including sample preparation, pipetting, and
DNA amplification with readout.
Validation
ScreenFire was compared against reference research HPV DNA assays in 2078 stored specimens.
Overall concordance for both viral types was >90%, and sensitivity for CIN3+ was 94.2%, similar to
Linear Array (93.1%) and TypeSeq (95.9%), indicating excellent performance.22 Regulatory
approval will require additional comparisons in screening settings. Additional clinical studies will
be nested within the PAVE protocol. In El Salvador, 5% of HPV-negative women will undergo
colposcopic evaluation. Comparison of ScreenFire to other WHO pre-qualified HPV tests (CareHPV
and Abbott) may also be performed on subsets of patients.
Silvia de Sanjosé et al., 2023 eLife. https://doi.org/10.7554/eLife.91469.1 12 of 26
Table 1.
Site-specific primary triage and treatment protocols Biopsy and Treatment Protocols
Note: Prior to the PAVE study, El Salvador screened with primary HPV screening (clinician-collected CareHPV), Brazil and DR
screened with cytology, and Cambodia, Eswatini, Honduras, Malawi, Nigeria, and Tanzania screened with VIA. All sites are
introducing self-sampled HPV testing with ScreenFire as part of the PAVE protocol. In El Salvador, women are continuing to
screen initially with both ScreenFire and CareHPV. Triage testing is performed in all women with HPV-positive results. In El
Salvador, triage testing is also performed on 5% of those testing HPV negative.
Silvia de Sanjosé et al., 2023 eLife. https://doi.org/10.7554/eLife.91469.1 13 of 26
Development of AVE algorithm using clinical images
Cervical images will be transferred via the digital camera to a secure server using a specially
designed script. Images and non-PHI data will be shared and downloaded on a loan-basis to the
PAVE AI team at NCI and the AI collaborators to train the AVE algorithm. Because images from the
Iris device have not been used previously with the AI algorithm, a pilot phase will take place to
retrain the AI-based algorithm. Due to the similarities of the Iris device to previously tested digital
cameras, we expect to be able to retrain the algorithm successfully as done in our prior work.23
During the PAVE study, four AI algorithms are being developed and evaluated: (1) cervix detector,
(2) image quality classifier, (3) disease classifier to identify precancer, and (4) treatability/SCJ
classifier.
(1) Cervix detector. This algorithm is designed to display a bounding box on the screen,
aiding healthcare providers ensure that the cervix is centralized within the image. This
feature simplifies the process of locating the cervix within the digital image, enhancing
efficiency and accuracy of image collection.
(2) Image quality classifier. This algorithm aims to identify images that may be unsuitable
for accurate disease assessment due to factors like obstruction or inadequate visual
sharpness. By flagging such images, it helps ensure that only good quality images are used
when training the algorithm. If shown to be useful, the image quality classifier could
provide feedback in real time to clinicians when taking images to ensure adequate image
quality.
(3) Disease classifier. This algorithm aims to visually distinguish precancerous changes
from lesser abnormalities, and classifies cervical images as normal, indeterminate, or
precancer+. AVE results are assessed on repeatability (correct classification of replicate
images of the same patient) and accuracy (correct classification of AVE based on
histopathology, as well as minimization of extreme misclassification of normal as
precancer or vice versa). Accuracy is defined as correct classification of participants to <
precancer or precancer+. Precancer+ is rigorously defined to include HPV-positive
histologic CIN3, AIS, cancer, and CIN2 diagnoses confirmed by expert pathologic review
and positive for the 8 most carcinogenic HPV genotypes. CIN2 is qualified because although
CIN2 is the threshold for treatment in most clinical practice worldwide, it is a less
reproducible pathologic diagnosis and may regress without treatment, especially when
associated with lower carcinogenicity HPV genotypes.24 ,25
The current prototype algorithm is the result of several years of development. Earlier algorithms
were limited by poor repeatability, misclassification including grave errors (i.e., cases with
precancer called normal or vice versa), overfitting, and lack of portability (defined as the ability of
the algorithm to accurately classify images from different image capture devices and study
settings other than the dataset on which it was trained).20 Additional techniques have been
applied to develop the prototype version of the AVE algorithm that will be refined and tested in
the PAVE study,26 resulting in improved reliability and consistency of model predictions across
repeat images from the same woman27 . The disease classifier algorithm is trained using
histology as the truth standard for defining the presence or absence of precancer. The outcome
definition for the purpose of training a three-class ordinal classification algorithm includes
normal (HPV-positive with histology normal, HPV negative, HPV negative cervicitis),
indeterminate (low-grade HPV-related abnormalities, CIN1), and precancer+, as determined by
histology among HPV positives (defined above). To ensure portability, the algorithm will undergo
external validation using datasets distinct from the training set of images. Early data indicate that
while our algorithm may function among patients across diverse geographies28 a dedicated
device may be needed because AVE fails to transfer without retraining.23 The images collected
Silvia de Sanjosé et al., 2023 eLife. https://doi.org/10.7554/eLife.91469.1 14 of 26
in PAVE will be used to refine and externally validate the prototype AVE algorithm.27 We will
test repeatability, accuracy, and calibration of the model before the algorithm is tested in clinical
settings during the effectiveness phase.29
Treatability/SCJ classifier. This algorithm aims to classify the SCJ is as fully visible, partially visible,
or not visible, using expert colposcopic impression as the truth standard. The goal is to assist
providers in determining treatment eligibility. SCJ visibility is critical (i.e., necessary although not
sufficient) for eligibility for thermal ablation procedures; ablation should not be performed when
the SCJ is not fully visible.
Pathology quality assurance
Pathology readings are performed locally with centralized quality assurance on a subset of cases.
Histotechnology adequacy via slide review from all participating laboratories includes assessment
of specimen preparation, staining adequacy, and clarity/readability of scanned images by the
assigned referent NCI study pathologist in collaboration with pathologists at each study site. Local
pathologists involved in the PAVE project are asked to complete a performance competency review
which include providing diagnoses on 20 standardized cases. Issues with either slide preparation
or interpretation were addressed via videoconference between the NCI expert pathologist and
local laboratories.
To assure histopathology reading standardization the following cases receive review by an expert
gynecologic pathologist, making use of a Motic whole slide scan review collection and transfer: 1)
all cases with histology CIN2+ or high-grade squamous intraepithelial lesion (HSIL); 2) all HPV16+
with <CIN2 pathology; 3) all cases read as precancer+ on AVE; and 4) 5% of biopsies read as
normal. Results will be classified for the study purposes as normal, low-grade (CIN1), high-grade
(CIN2, CIN3), adenocarcinoma in situ (AIS), or invasive cervical cancer (squamous,
adenosquamous, adenocarcinoma, or other).
Statistical analysis
The primary objective of the PAVE protocol is to compare the sensitivity at a given specificity level
of the PAVE approach for triaging HPV-positive women to current SOC (HPV testing without
genotyping triaged by VIA or colposcopic impression). The PAVE screen-triage-treat protocol
combines the three-class classification label (normal, indeterminate, precancer+) provided by the
AVE algorithm with the four hrHPV risk group strata to create 12 strata of risk of precancer (Table
2 ).
The strata from Table 2 will be plotted as an ROC-like curve of sensitivity versus (1-specificity).
The ROC curve for each study site will yield an area under the curve (AUC) for both the PAVE and
the SOC strategies.
The 12-stratum AVE risk score will be compared to visual assessment of SOC: VIA (negative or
positive) or colposcopic impression (less than high grade vs. high grade+).29 At each site, we will
compare the sensitivity of the two approaches, at the (1-specificity) value produced by SOC visual
evaluation. We will compile these values for the consortium and calculate the average (weighted
by study size). We hypothesize that the difference of 2 sensitivities (PAVE minus SOC) conducted on
the overall consortium data will be significantly greater than zero (the null hypothesis of no
difference in sensitivity), indicating that PAVE detects more precancer than SOC triage at the same
level of specificity. Where available, the analysis will be stratified by HIV status. Throughout the
study we will be checking for the consistency of results at the different steps across sites (e.g.,
performance of HPV results, quality of the images, histopathology reports). We will assess the
reproducibility of the PAVE strategy across different sites by measuring the variability of the AUC
values.
Silvia de Sanjosé et al., 2023 eLife. https://doi.org/10.7554/eLife.91469.1 15 of 26
Table 2.
HPV-AVE risk strata
The PAVE protocol will be compared to SOC VIA or colposcopy screen-triage protocols using the visual impressions recorded
during the study. In SOC scenarios, participants are classified as positive or negative in the HPV test and as normal or
abnormal in visual evaluations (e.g., VIA negative or positive, colposcopic impression less than high-grade or high-grade+)
(Table 3 ).
Table 3.
Risk strata for participants with an HPV positive women and visual Standard of Care (SOC) as the
triage (i.e. VIA, colposcopy) test
Note: Participants with a negative test for HPV will not have a VIA nor colposcopy assessment. see Table 1 for SOC at
individual PAVE sites.
Silvia de Sanjosé et al., 2023 eLife. https://doi.org/10.7554/eLife.91469.1 16 of 26
Additional analyses that may be performed during the efficacy phase include:
● Development of treatability algorithm: In addition to defining SCJ visibility, an AI
algorithm to assess whether a lesion fulfills WHO ablation criteria is in development. If AI
algorithm development is successful, output will be compared to the VIA or colposcopy
assessment of eligibility for ablation or referral for surgical management. Accuracy of the
three-class classification label provided by the AVE treatability algorithm (treatable,
uncertain, not treatable) will be compared to the against a truth label based on the
evaluation of three experts.
● Impact of HIV status on PAVE algorithm: We will evaluate the impact of HIV in the
PAVE performance for the accuracy in detection precancer+ by comparing the AUC in
women with and without HIV, with some consideration of role of effective ART.23
Phase 2: Effectiveness
During the efficacy phase, the PAVE algorithm is undergoing evaluation and development, and
clinicians will not be provided with HPV genotyping, AI algorithm outputs, or risk strata. When the
efficacy phase is complete, if the PAVE algorithm outperforms the SOC, we will begin the
effectiveness phase. Sites that chose to participate in the effectiveness phase will test the following
screen-treat-triage protocol. Screening: self-sampled HPV testing with genotyping (ScreenFire or
equivalent test). Triage of HPV positive individuals: image collection using the Iris device, upon
which the AI algorithms have been installed. The AI algorithms will guide the clinician in taking
high-quality images (cervix identifier, Image quality classifier) and then provide a disease
classification score of normal, indeterminate, or precancer+ and an SCJ visibility assessment of
fully visible, partially visible, or not visible. The clinician will then enter the HPV genotyping test
result and the device will output a risk category using the strata in Table 2 (lower, medium,
high, highest). The risk category and SCJ visibility assessment are designed as clinical management
tools to aid clinicians in determining which patients are most likely to benefit from treatment, and
among those needing treatment, whether ablation can be considered. This phase will assess the
feasibility and acceptability of the PAVE strategy in clinical practice.
Cost-effectiveness analysis
To inform decisionmakers designing cervical cancer prevention programs in resource-limited
settings, we will analyze the cost-effectiveness (i.e., cost per precancer detected) and affordability
(the impact on a payer’s budget) of the PAVE strategy at several sites. Micro-costing efforts to
estimate the cost of all resource ingredients used for a screening episode are underway with
technical assistance provided to study sites. A microsimulation model of genotype-specific HPV
natural history and cervical carcinogenesis is being developed specifically for evaluation of novel
biomarker triage tests, including AVE.30 By adapting this natural history model to setting-
specific HPV prevalence patterns (by genotype and age); overlaying screening, triage, and
treatment strategies; and using setting-specific healthcare delivery data inputs for uptake,
adherence to management, and costs, we will evaluate the cost per precancer detected by the
PAVE strategy relative to SOC. We will explore the health and economic implications of applying
different management approaches (e.g., triage, treatment, and follow up) to different risk strata,
depending on health system capacity.
The development of the microsimulation model and the micro-costing tools for the PAVE
consortium will serve as the basis for estimating the real-world costs and health benefits of
implementing novel screening and management strategies. These tools can be adapted to different
settings, with refinement of management algorithms, health care delivery variables, and cost
estimates as implementation and scale-up occur. An early exercise to approximate the potential
costs and benefits of a highly effective screening campaign delivered to women aged 30-49 years
in the ∼65 highest burden LMIC (Figure 1 ; Supplementary Materials) and an HPV vaccination
Silvia de Sanjosé et al., 2023 eLife. https://doi.org/10.7554/eLife.91469.1 17 of 26
program delivered to girls ages 9-14 years found that the number of screening or adolescent HPV
vaccinations needed to avert one cervical cancer death was similar for each intervention (i.e., 293
for screening; 278 for vaccination). Assuming a bundled cost of US$15 per woman screened and
managed appropriately, a onetime screening campaign that achieves 40% uptake and ∼60%
adherence to recommended treatment for screen-positive women yielded a financial cost of
∼US$2.5 billion to avert ∼570,000 deaths, or US$4,400 per death averted. On a similar order of
magnitude, a onetime single-dose bivalent HPV vaccination campaign achieving 90% coverage of
girls aged 9-14 years in the same countries (US$4.50 vaccine cost; US$7 delivery cost) would cost
∼US$2.0 billion and avert ∼640,000 deaths, or US$3,200 per death averted. Of note, these ballpark
estimates are undiscounted and do not account for cancer treatment cost offsets. While data are
not yet available on the costs of implementing novel highly effective screening strategies for adult
women and single-dose HPV vaccination for female adolescents, these data are forthcoming from
the PAVE consortium and single-dose vaccination studies. Refining these cost and effectiveness
estimates, and obtaining country-specific data, is a high priority and a critical component of the
PAVE consortium objectives.
Stakeholder knowledge and attitudes regarding
cervical cancer prevention and screening
interventions in the PAVE Study
Effective dissemination and implementation of the PAVE strategy in the future will require clear
and consistent strategies for communicating information to healthcare providers and patients.
The field of HPV and screening is rapidly evolving and constantly being enriched with new
scientific information. However, this influx of information can sometimes lead to unclear or
conflicting messages, which in turn may diminish the effectiveness of interventions aimed at
improving screening rates and optimizing management strategies. Furthermore, both healthcare
providers and patients can differ in their knowledge and perceptions of cervical cancer risk,
tolerance of risk, and their personal values and attitudes regarding cervical cancer prevention and
screening, all of which can influence uptake of prevention strategies such as the PAVE strategy. To
address these challenges, it will be necessary to provide healthcare providers with accurate,
consistent information about the latest scientific advances and guidelines, and as well as training
and tools that can help them effectively inform and engage patients in cervical cancer prevention
and screening.
To address these needs and prepare for broader dissemination and implementation of the PAVE
strategy, the Communication and Retention Workgroup is conducting a mixed-methods pilot study
utilizing both qualitative interviews and survey questionnaires administered to scientific experts
as well as key stakeholders—patients and healthcare providers—at four participating sites (Brazil,
El Salvador, Nigeria, Tanzania). The specific objectives of this pilot study are to explore
stakeholders’ knowledge, perceptions, and attitudes regarding cervical cancer prevention and
screening, and their preferences for information and participation in decision making. The study
will generate evidence to enable the future development of effective, ethical strategies for
engaging eligible members of LMIC communities in cervical cancer prevention and screening,
including using the PAVE strategy.
Silvia de Sanjosé et al., 2023 eLife. https://doi.org/10.7554/eLife.91469.1 18 of 26
Discussion
Timeline and Next Steps
At the time of this writing, the study has been initiated in five countries. It is expected that a
preliminary interim analysis on the efficacy of the PAVE strategy will be completed by the end of
2023, and field recruitment will be completed by July 2024. The efficacy phase is designed to assess
the validity of the PAVE protocol by refining the AVE protocol using histopathology specimens and
demonstrate the superiority of PAVE for detecting precancer and minimizing unnecessary
referrals compared to the SOC. Following the initial efficacy phase, the effectiveness phase is
planned to examine additional factors including program feasibility, acceptability, and cost-
effectiveness.
Regulatory considerations
The rapid advancement of AI in healthcare has prompted significant ethical and regulatory
discussions. Global ethical principles specific to AI in healthcare are in development and are
rapidly evolving, regulatory authorities such as the WHO and U.S. Food and Drug Administration
(FDA) are responsible for developing guidelines that ensure the safe, effective, and appropriate
use of AI technologies in healthcare and therapeutic development. In addition to complying with
existing ethical principles in medicine, AI solutions must demonstrate scientific validity, ensuring
their effectiveness and reliability. It is imperative that AI technologies do not perpetuate
discrimination or bias, or exclude specific population segments. One major concern is the
potential creation of a permanent digital identity, linked to individuals’ health and personal data,
without obtaining proper consent.31 ,32
The Iris device, used in PAVE, is an example of an AI-based system in healthcare. It will
incorporate AI algorithms that aim to generate scores evaluating image quality, presence of
precancerous lesions, and visibility of the SCJ based on digital cervix images. These scores provide
additional information to assist healthcare providers in making informed decisions. To arrive at a
comprehensive decision, the provider will need to integrate the HPV information, gynecological
examination findings, scoring system outputs generated from the AI algorithms, and the patient’s
medical history. Considering its functionality, the AI algorithm used in the PAVE system, based on
FDA specifications, can be considered an assistant to medical management offering the joint
evaluation of the HPV data and the AVE outcome. As WHO is actively evaluating screening and
triage approaches, it is expected that clear regulatory guidance on the use of the algorithms will be
available before their implementation .31
Competition and commercial aspects
The PAVE strategy utilizes certain products of significant commercial importance. The selection of
each device or assay is based on factors such as accuracy to achieve its purpose (i.e. high
sensitivity of the ScreenFire HPV test to detect precancer, high-quality image capture by the Iris
device) and affordability. It is important to note that the PAVE project does not preclude future
commercial developments. The presence of competition with similar assays is desirable as long as
they can provide similar qualities. If the PAVE strategy is proven to be more accurate than the SOC
for screening and triage, there will likely be an increased demand for devices and assays.
Procurement of these resources may present challenges and require effective communication
with the respective Ministries of Health to ensure a smooth introduction in countries where the
need is evident.
Silvia de Sanjosé et al., 2023 eLife. https://doi.org/10.7554/eLife.91469.1 19 of 26
Dissemination
After the demonstration of efficacy and effectiveness, we expect to have a screen-treat-triage
protocol fulfilling WHO principles that can be used at the discretion of local health authorities in
LMIC for cervical cancer prevention programs. If the PAVE strategy is proven to outperform the
local SOC, and be feasible, acceptable, and affordable for resource-limited settings, then countries
may switch their current SOC to the PAVE strategy. De-implementation of existing strategies, such
as cytology, colposcopy, and VIA, as well as implementation of self-sampled HPV with AVE triage
will require buy-in from stakeholders and policymakers as well as substantial investment in
educating and retraining the laboratory and clinical workforces. Cytology/colposcopy programs,
though effective, are limited in scope and are costly to maintain, therefore switching may be
attractive to health ministers. However, laboratories will need funding for the purchase of
equipment for running HPV testing and materials for self-sampling, and the cytotechnologist and
pathologist workforce will be reduced. Countries with existing VIA programs will require
significant introduction costs such as laboratory machinery and training of healthcare personnel,
and recurrent costs including reagents and self-sampling materials. In all settings, program
continuation beyond the initial study will require local governments and programs to address
issues of procurement and implementation, as well as de-implementation of existing strategies.
In conclusion, the PAVE project will develop and validate a strategy using self-sampled HPV with
genotyping and AVE to identify precancer in a large group of women from many different settings.
The PAVE objective is to create an accurate, feasible, cost-effective screening and triage protocol
for cervical cancer prevention in resource-limited settings. If proven effective, cost-effective,
feasible, and acceptable, the strategy can have a major impact in reducing cervical cancer among
non-vaccinated adult women.
Contributors: All authors affirm that this commentary is an honest, accurate, and transparent
account of the study being reported; that no important aspects of the study have been omitted,
and that any discrepancies from the study as planned (and, if relevant, registered) have been
explained. This research was conducted with the appr approval of the institutional review board
of the National Cancer Institute.
Declaration of interests: No authors have a conflict of interest to declare related to this work
Data Availability
All data produced in the present study are available upon reasonable request to the authors
Funding
National Cancer Institute Cancer Cures Moonshot Initiative. No commercial support was obtained.
Brian Befano was supported by NCI/NIH under Grant T32CA09168.
Silvia de Sanjosé et al., 2023 eLife. https://doi.org/10.7554/eLife.91469.1 20 of 26
References
Singh D, Vignat J, Lorenzoni V, et al. (2023) Global estimates of incidence and mortality of
cervical cancer in 2020: a baseline analysis of the WHO Global Cervical Cancer Elimination
Initiative Lancet Glob Health 11:e197–e206 https://doi.org/10.1016/S2214-109X(22)00501-0
Schiffman M, Castle PE, Jeronimo J, Rodriguez AC, Wacholder S (2007) Human papillomavirus
and cervical cancer The Lancet 370:890–907 https://doi.org/10.1016/S0140-6736(07)61416-0
IARC (2022) Cervical Cancer Screening IARC Handbooks of Cancer Prevention Volume 18
World Health Organization (2021) WHO guideline for screening and treatment of cervical
pre-cancer lesions for cervical cancer prevention, second edition
WHO (2020) Global strategy to accelerate the elimination of cervical cancer as a public
health problem
Lei J, Ploner A, Elfström KM, et al. (2020) HPV Vaccination and the Risk of Invasive Cervical
Cancer N Engl J Med 383:1340–1348 https://doi.org/10.1056/NEJMoa1917338
Perkins RB, Smith DL, Jeronimo J, et al. (2023) Use of risk-based cervical screening programs
in resource-limited settings Cancer Epidemiol 84 https://doi.org/10.1016/j.canep.2023.102369
Ferlay J, Ervik M, Lam F, Colombet M, Mery L, Piñeros M, Znaor A, Soerjomataram I, Bray F
(2020) Global Cancer Observatory: Cancer Today
Demarco M, Egemen D, Hyun N, et al. (2022) Contribution of Etiologic Cofactors to CIN3+
Risk Among Women With Human Papillomavirus-Positive Screening Test Results J Low
Genit Tract Dis 26:127–134 https://doi.org/10.1097/LGT.0000000000000667
Castle PE, Glass AG, Rush BB, et al. (2012) Clinical Human Papillomavirus Detection
Forecasts Cervical Cancer Risk in Women Over 18 Years of Follow-Up Journal of Clinical
Oncology 30:3044–3050 https://doi.org/10.1200/JCO.2011.38.8389
Arbyn M, Smith SB, Temin S, Sultana F, Castle P, Self-Sampling Collaboration on, Testing HPV
(2018) Detecting cervical precancer and reaching underscreened women by using HPV
testing on self samples: updated meta-analyses BMJ 363 https://doi.org/10.1136/bmj.k4823
Serrano B, Ibáñez R, Robles C, Peremiquel-Trillas P, de Sanjosé S, Bruni L (2022) Worldwide use
of HPV self-sampling for cervical cancer screening Preventive Medicine
154 https://doi.org/10.1016/j.ypmed.2021.106900
Demarco M, Hyun N, Carter-Pokras O, et al. (2020) A study of type-specific HPV natural
history and implications for contemporary cervical cancer screening programs
EClinicalMedicine 22 https://doi.org/10.1016/j.eclinm.2020.100293
Sankaranarayanan R, Nene BM, Dinshaw K, et al. (2003) Early detection of cervical cancer
with visual inspection methods: a summary of completed and on-going studies in India
Salud Publica Mex 45:S399–407
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
Silvia de Sanjosé et al., 2023 eLife. https://doi.org/10.7554/eLife.91469.1 21 of 26
Catarino R, Schäfer S, Vassilakos P, Petignat P, Arbyn M (2018) Accuracy of combinations of
visual inspection using acetic acid or lugol iodine to detect cervical precancer: a meta-
analysis BJOG 125:545–553 https://doi.org/10.1111/1471-0528.14783
Sankaranarayanan R, Nene BM, Shastri SS, et al. (2009) HPV Screening for Cervical Cancer in
Rural India N Engl J Med 360:1385–1394 https://doi.org/10.1056/NEJMoa0808516
Qiao YL, Sellors JW, Eder PS, et al. (2008) A new HPV-DNA test for cervical-cancer screening
in developing regions: a cross-sectional study of clinical accuracy in rural China Lancet
Oncol 9:929–936 https://doi.org/10.1016/S1470-2045(08)70210-9
Guan P, Howell-Jones R, Li N, et al. (2012) Human papillomavirus types in 115,789 HPV-
positive women: a meta-analysis from cervical infection to cancer Int J Cancer 131:2349–
2359 https://doi.org/10.1002/ijc.27485
Desai KT, Ajenifuja KO, Banjo A, et al. (2020) Design and feasibility of a novel program of
cervical screening in Nigeria: self-sampled HPV testing paired with visual triage Infect
Agent Cancer 15 https://doi.org/10.1186/s13027-020-00324-5
Desai KT, Befano B, Xue Z, et al. (2022) The development of “automated visual evaluation”
for cervical cancer screening: The promise and challenges in adapting deep-learning for
clinical testing: Interdisciplinary principles of automated visual evaluation in cervical
screening Int J Cancer 150:741–752 https://doi.org/10.1002/ijc.33879
Desai KT, Adepiti CA, Schiffman M, et al. (2022) Redesign of a rapid, low-cost HPV typing
assay to support risk-based cervical screening and management Int J Cancer 151:1142–
1149 https://doi.org/10.1002/ijc.34151
Inturrisi F, de Sanjose S, Desai KT, Dagnal C, Egemen D, Befano B, Rodrigue AC, Jerónimo J,
Zuna RE, Hoffman A, Nozzari SF, Walker JL, Perkins RB, Wentzensen N, Palefsky JM, Schiffman M
(1970) A rapid HPV typing assay to support cervical cancer screening and risk-based
management: a cross-sectional validation study
Parham G, Egemen D, Befano B, Mwanahamuntu M, Rodriguez AC, Antani S, Chisele S,
Munalula MK, Kaunga F, de Sanjose S, Schiffman M, Sahasrabuddhe V (1970) Validation in
Zambia of a cervical screening strategy including HPV genotyping and artificial
intelligence (AI)-based automated visual evaluation
Lee MH, Finlayson SJ, Gukova K, Hanley G, Miller D, Sadownik LA (2018) Outcomes of
Conservative Management of High Grade Squamous Intraepithelial Lesions in Young
Women Journal of Lower Genital Tract Disease 22:212–
218 https://doi.org/10.1097/LGT.0000000000000399
Kylebäck K, Ekeryd-Andalen A, Greppe C, Björkenfeldt Havel C, Zhang C, Strander B (2022)
Active expectancy as alternative to treatment for cervical intraepithelial neoplasia grade
2 in women aged 25 to 30 years: ExCIN2-a prospective clinical multicenter cohort study
Am J Obstet Gynecol. Published online June 29 https://doi.org/10.1016/j.ajog.2022.06.051
Ahmed SR, Befano B, Lemay A, et al. (2023) Reproducible and Clinically Translatable Deep
Neural Networks for Cancer Screening Res Sq. Published online March
3https://doi.org/10.21203/rs.3.rs-2526701/v1
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
Silvia de Sanjosé et al., 2023 eLife. https://doi.org/10.7554/eLife.91469.1 22 of 26
Lemay A, Hoebel K, Bridge CP, et al. (2022) Improving the repeatability of deep learning
models with Monte Carlo dropout NPJ Digit Med 5https://doi.org/10.1038/s41746-022-00709-
3
Ahmed SR, Egemen D, Befano B, Rodriguez AC, Jeronimo J, de Sanjose S, Kalpathy-Cramer J,
Schiffman M (1970) Assessing generalizability of an ai-based visual test for cervical cancer
screening
Egemen D, Perkins R, Cheung L, Befano B, Rodriguez AC, Desai K, Lemay A, Ahmed SR, Antani S,
Jeronimo J, Wentzensen N, Kalpathy-Cramer J, de Sanjose S, Schiffman M (1970) AI-based
image analysis in clinical testing: lessons from cervical cancer screening
Campos NG, Demarco M, Bruni L, et al. (2021) A proposed new generation of evidence-based
microsimulation models to inform global control of cervical cancer Prev Med
144 https://doi.org/10.1016/j.ypmed.2021.106438
WHO (2021) Ethics and governance of artificial intelligence for health
FDA (1970) Artificial Intelligence and Machine Learning in Software as a Medical Device
Article and author information
Silvia de Sanjosé
Division of Cancer Epidemiology and Genetics, National Cancer Institute, National Institutes of
Health, Rockville, MD, USA, ISGlobal, Barcelona, Spain
For correspondence: desanjose.silvia@gmail.com
ORCID iD: 0000-0002-5909-676X
Rebecca B. Perkins
University Chobanian and Avedisian School of Medicine/Boston Medical Center, Boston, MA,
USA
Nicole G. Campos
Center for Health Decision Science, Harvard T.H. Chan School of Public Health, Boston, MA, USA
Federica Inturrisi
Division of Cancer Epidemiology and Genetics, National Cancer Institute, National Institutes of
Health, Rockville, MD, USA
ORCID iD: 0000-0002-3661-0842
Didem Egemen
Division of Cancer Epidemiology and Genetics, National Cancer Institute, National Institutes of
Health, Rockville, MD, USA
ORCID iD: 0000-0001-5617-4585
Brian Befano
Information Management Services Inc, 3901 Calverton Blvd Suite 200, Calverton, MD, USA,
Department of Epidemiology, University of Washington School of Public Health, Seattle,
Washington, USA
ORCID iD: 0000-0003-3614-210X
27.
28.
29.
30.
31.
32.
Silvia de Sanjosé et al., 2023 eLife. https://doi.org/10.7554/eLife.91469.1 23 of 26
Ana Cecilia Rodriguez
Division of Cancer Epidemiology and Genetics, National Cancer Institute, National Institutes of
Health, Rockville, MD, USA
Jose Jerónimo
Division of Cancer Epidemiology and Genetics, National Cancer Institute, National Institutes of
Health, Rockville, MD, USA
ORCID iD: 0000-0002-7734-1837
Li C. Cheung
Division of Cancer Epidemiology and Genetics, National Cancer Institute, National Institutes of
Health, Rockville, MD, USA
ORCID iD: 0000-0003-1625-4331
Kanan Desai
Division of Cancer Epidemiology and Genetics, National Cancer Institute, National Institutes of
Health, Rockville, MD, USA
ORCID iD: 0000-0002-8992-5944
Paul Han
Division of Cancer Control and Population Sciences, National Cancer Institute, National
Institutes of Health, Rockville, MD, USA
ORCID iD: 0000-0003-0165-1940
Akiva P Novetsky
Westchester Medical Center/New York Medical College, Valhalla, NY, USA
ORCID iD: 0000-0002-2990-8499
Abigail Ukwuani
Division of Cancer Epidemiology and Genetics, National Cancer Institute, National Institutes of
Health, Rockville, MD, USA
Jenna Marcus
Feinberg School of Medicine at Northwestern University, Chicago, IL, USA
ORCID iD: 0000-0002-7363-0199
Syed Rakin Ahmed
Athinoula A. Martinos Center for Biomedical Imaging, Department of Radiology, Massachusetts
General Hospital, Boston, MA, USA, Harvard Graduate Program in Biophysics, Harvard Medical
School, Harvard University, Cambridge, MA, USA, Massachusetts Institute of Technology,
Cambridge, MA, USA, Geisel School of Medicine at Dartmouth, Dartmouth College, Hanover,
NH, USA
ORCID iD: 0000-0002-1615-8633
Nicolas Wentzensen
Division of Cancer Epidemiology and Genetics, National Cancer Institute, National Institutes of
Health, Rockville, MD, USA
ORCID iD: 0000-0003-1251-0836
Silvia de Sanjosé et al., 2023 eLife. https://doi.org/10.7554/eLife.91469.1 24 of 26
Jayashree Kalpathy-Cramer
Athinoula A. Martinos Center for Biomedical Imaging, Department of Radiology, Massachusetts
General Hospital, Boston, MA, USA, University of Colorado Anschutz Medical Campus, Aurora,
CO, USA
ORCID iD: 0000-0001-8906-9618
Mark Schiffman
Division of Cancer Epidemiology and Genetics, National Cancer Institute, National Institutes of
Health, Rockville, MD, USA
ORCID iD: 0000-0002-4625-2508
for the PAVE Study Group
Copyright
© 2023, de Sanjosé et al.
This article is distributed under the terms of the Creative Commons Attribution License
(https://creativecommons.org/licenses/by/4.0/) , which permits unrestricted use and
redistribution provided that the original author and source are credited.
Editors
Reviewing Editor
Talía Malagón
McGill University, Canada
Senior Editor
Eduardo Franco
McGill University, Canada
Reviewer #1 (Public Review):
Summary:
A description of a modern protocol for cervical screening that likely could be used in any
country of the world, based on self-sampling, extended HPV genotyping and AI-assisted visual
inspection - which is probably the best available combination today.
Strengths:
Modern, optimised protocol, designed for global use. Innovative.
Weaknesses:
The protocol is not clear. I could not even find how many women were going to be enrolled,
the timelines of the study, the statistical methods ("comparing" is not statistics) or the power
calculations.
Tables 2 and 3 are too schematic - surely the authors must have an approximate idea of what
the actual numbers are behind the green, red and yellow colors.
Silvia de Sanjosé et al., 2023 eLife. https://doi.org/10.7554/eLife.91469.1 25 of 26
Figure 1 comparing screening and vaccination is somewhat misleading. They screen 20 birth
cohorts but vaccinate only 5 birth cohorts. Furthermore, the theoretical gains of screening
has not really been attained in any country in practice. Modelling can be a difficult task and
the commentary does not provide any detail on how to evaluate what was done. It just seems
unnecessary to attack vaccination as a motivation on why screening needs to be modernised.
Reviewer #2 (Public Review):
Summary:
This manuscript describes the study protocol, structure and logic of the PAVE strategy. The
PAVE study is a multicentric study to evaluate a novel cervical screen-triage-treat strategy for
resource-limited settings as part of a global strategy to reduce cervical cancer burden. The
PAVE strategy involves: 1) screening with self-sampled HPV testing; 2) triage of HPV-positive
participants with a combination of extended genotyping and visual evaluation of the cervix
assisted by deep-learning-based automated visual evaluation (AVE); and 3) treatment with
thermal ablation or excision (Large Loop Excision of the Transformation Zone). The PAVE
study has two phases: efficacy (2023-2024) and effectiveness (planned to begin in 2024-2025).
The efficacy phase aims to refine and validate the screen-triage portion of the protocol. The
effectiveness phase will examine implementation of the PAVE strategy into clinical practice.
Strengths and weaknesses:
The Pave Study develops and evaluates a novel strategy that combines HPV self-collection,
that has been proven effective to increase screening coverage in different settings, with
genotyping and Automated Visual Evaluation as triage. The proposed strategy combined
three key innovations to improve an important step in the cervical cancer care continuum. If
the strategy is effective it will contribute to enhancing cervical cancer prevention in low
resource settings.
As the authors mentioned, despite the existence of effective preventive technologies (e.g.,
HPV vaccine and HPV test) translation of the HPV prevention methods has not yet occurred in
many Low-Middle-Income Countries. So, in this context, new screen-triage-treat strategies are
needed and if PAVE strategy were effective, it could be a landmark for cervical cancer
prevention.
The PAVE Study is a solid and important study that is aimed to be carried out in nine
countries and recruit tens of thousands of women. It is a study with a large and diverse
sample that can provide useful information for the development of this new screen-triage-
treat strategy. Another strength is the fact that the PAVE project is integrated into the
screening activities placed in the selected countries that will allow to evaluate efficacy and
effectiveness in real-word context.
The manuscript does not present results because its aim is to describe the study protocol,
structure and logic of the PAVE strategy.
Phase 1 aims to evaluate the efficacy of the strategy. Methods are well described and are
consistent with the study aims.
Phase 2 aims to evaluate the implementation of the PAVE strategy in clinical practice. The
inclusion of implementation evaluation in this type of studies is an important milestone in
the field of cervical cancer prevention. It has been shown that many strategies that have
proven to be effective in controlled studies face barriers when they are implemented in real
life. In that sense, the results of phase 2 are key to ensure the future implementation of the
strategy.
Silvia de Sanjosé et al., 2023 eLife. https://doi.org/10.7554/eLife.91469.1 26 of 26
However, some aspects of Phase 2 need to be clarified and extended. Although authors
mentioned that implementation outcomes, such as acceptability and feasibility will be
evaluated, more information is needed about method (i.e. qualitative/quantitative), data
collection tools (i.e., survey, semi-structure interviews, focus groups, etc.) and frameworks
that will be used to evaluate these implementation outcomes.
Reviewer #3 (Public Review):
Summary:
Despite being preventable and treatable, cervical cancer remains the second most common
cause of cancer death globally. This cancer, and associated deaths, occur overwhelmingly in
low- and middle-income countries (LMIC), reflecting a lack of access to vaccination, screening
and treatment services. Cervical screening is the second pillar in the WHO strategy to
eliminate cervical cancer as a public health problem and will be critical in delivering early
gains in cervical cancer prevention as the impact of vaccination will not be realized for
several decades. However, screening strategies implemented in high income countries are
not feasible or affordable in LMICs. This ambitious multi-center study aims to address these
issues by developing and systematically evaluating a novel approach to cervical screening.
The approach, based on primary screening with self-collected specimens for HPV testing, is
focused on optimizing triage of people in whom HPV is detected, so that sensitivity for the
detection of pre-cancer and cancer is maximized while treatment of people without pre-
cancer or cancer is minimized.
Strengths:
The triage proposed for this study builds on the authors' previously published work in
designing the ScreenFire test to appropriately group the 13 detected genotypes into four
channels and to develop automated visual evaluation (AVE) of images of the cervix, taken by
health workers.
The move from mobile telephone devices to a dedicated device to acquire and evaluate
images overcomes challenges previously encountered whereby updates of mobile phone
models required retraining of the AVE algorithm.
The separation of the study into two phases, an efficacy phase in which screen positive
people will be triaged and treated according to local standard of care and the performance of
AVE will be evaluated against biopsy outcomes will be followed by the second phase in which
the effectiveness, cost-effectiveness, feasibility and acceptability will be evaluated.
The setting in a range of low resource settings which are geographically well spread and
reflective of where the global cancer burden is highest.