Prevalence of noncardiac findings on clinical cardiovascular MRI.
ABSTRACT The purpose of our study was to determine the prevalence and significance of noncardiac findings on clinical cardiovascular MRI and to identify the cardiovascular MRI sequences that most frequently depict noncardiac findings.
Images from 495 clinical cardiovascular MRI studies performed during 2006 were reviewed specifically for noncardiac findings by a cardiovascular imaging fellowship-trained radiologist without knowledge of the prior study interpretation. Noncardiac findings were classified as benign (e.g., gynecomastia), indeterminate (e.g., pleural effusion), or worrisome (e.g., lung nodule). The cardiovascular MRI sequences depicting the noncardiac finding were recorded.
On image review, 295 noncardiac findings were identified in 212 (43%) of 495 studies, including 148 benign, 133 indeterminate, and 14 worrisome noncardiac findings. Of these, 47% of indeterminate and 57% of worrisome noncardiac findings were not previously known. Cardiovascular MRI sequences that most frequently showed noncardiac findings included the single-shot fast steady-state free precession (SSFP) scout images (63% of all noncardiac findings) and axial T1-weighted fast spin-echo thoracic images (60% of all noncardiac findings), with 99% of management-changing noncardiac findings visualized on one of these two sequences.
Noncardiac findings on clinical cardiovascular MRI are common. Although only a small minority of studies contain management-changing noncardiac findings, the vast majority of management-changing noncardiac findings are seen on thoracic SSFP scout and axial T1-weighted thoracic fast spin-echo images.
- [Show abstract] [Hide abstract]
ABSTRACT: To determine the frequency and clinical impact of noncardiovascular incidental findings (IFs) detected on preoperative computed tomographic angiography (CTA) of the chest/abdomen/pelvis performed in elderly patients with severe aortic stenosis being considered for transcatheter aortic valve replacement (TAVR).Mayo Clinic Proceedings 06/2014; 89(6):747-53. · 5.81 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: Objective This study sought to determine the prevalence of significant and non-significant non-cardiac findings in patients undergoing pre-procedural dual source computed tomography (DSCT) prior to transcatheter aortic valve implantation (TAVI). Methods 204 patients (80.5 ± 5.1 years, 106 male) underwent pre-procedural DSCT of the thoraco-abdominal aorta and the pelvic arterial vessels. Non-cardiac findings were recorded and categorized as non-significant (group A), incidental findings requiring follow-up examinations (group B) and significant findings with a demand for clinical treatment (group C). Results In 60/204 DSCT examinations (29.4 %) no non-cardiac findings were observed. Out of the remaining 144 examinations (70.6%), 260 revealed non-cardiac findings. 35 out of 204 patients (17.1%) demonstrated a total of 37 clinically significant non-cardiac findings. Eight malignancies were detected; five of them were incidentally diagnosed on DSCT and changed patient management. A total of 223 non-significant findings were observed in 116/204 patients (56.9 %) (group A), the most frequent findings were pleural effusions or colorectal diverticulosis. The prevalence of incidental and significant findings on DSCT prior TAVI increased with patient age (r2= 0.69; p = 0.01). Conclusion Significant non-cardiac findings are common in patients referred to routine pre-procedural DSCT for planning TAVI (17.1%).Journal of Cardiovascular Computed Tomography 05/2014; · 4.51 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: A broad spectrum of congenital coronary anomalies may be discovered on imaging and sometimes in the emergency setting on computed tomography (CT). Most coronary artery anomalies are of academic interest; however, a minority can cause morbidity and mortality and symptoms such as angina, myocardial infarction, or arrhythmias. These anomalies are usually discovered as an incidental finding on CT examinations as part of the diagnostic workup for other pathology or on dedicated coronary computed tomography angiography (CCTA) as part of the evaluation for a coronary cause of chest pain. The purpose of this pictorial review is to demonstrate the types of coronary anomalies and to enhance the clinicians' understanding of the imaging classifications and clinical implications.Emergency Radiology 05/2014;
W380 AJR:196, April 2011
has been well described; however, there
are few data  on the prevalence of non-
cardiac findings on clinical cardiovascular
MRI, their ultimate clinical importance, or
the cardiovascular MRI sequences most
likely to show the noncardiac finding.
Current cardiovascular MRI training guide-
lines do not specifically address the issue of
training in noncardiac pathology . This
has led to different institutional approaches
for interpreting clinical cardiovascular MRI
studies. At some institutions, either radiolo-
gists or cardiologists perform the interpreta-
tion. Other sites use the combined expertise of
both specialties, either by independent serial
reading sessions or a combined reading ses-
sion attended by both specialists. Our institu-
tion has a combined strategy, with a radiolo-
gist and cardiologist collaborating at a daily
scheduled cardiovascular MRI readout.
The purpose of this study was to assess the
prevalence and significance of noncardiac find-
Prevalence of Noncardiac Findings
on Clinical Cardiovascular MRI
Benjamin P. Romney2,3
Daniel N. Costa1
Neil M. Rofsky1
Warren J. Manning1,2
Khosa F, Romney BP, Costa DN, Rofsky NM,
1Department of Radiology, Beth Israel Deaconess
Medical Center, Harvard Medical School, 330 Brookline
Ave, Boston, MA 02215. Address correspondence to
W. J. Manning (email@example.com).
2Department of Medicine, Cardiovascular Division,
Beth Israel Deaconess Medical Center, Harvard Medical
School, Boston, MA.
3Tufts University School of Medicine, Boston, MA.
Cardiopulmonary Imaging • Original Research
This is a Web exclusive article.
AJR 2011; 196:W380–W386
© American Roentgen Ray Society
structural and functional assessment of the
heart [1–3]. Because cardiovascular MRI
is a cross-sectional modality, it provides
complementary information on structures
adjacent to the heart, including the medi-
astinum, lung fields, chest wall, and upper
abdomen. The acquisition of imaging field
data outside the heart offers the opportu-
nity for detection of noncardiac abnormal-
ities. These noncardiac findings may repre-
sent unsuspected important disease or
benign “incidentalomas,” and carry medi-
colegal, ethical, and financial implications
[4–6]. The prevalence of incidental find-
ings on such imaging modalities as CT
colonography [6–8], renal MR angiogra-
phy (MRA) , and epicardial coronary
calcium/coronary CT angiography [10–16]
ardiovascular MRI has been
increasingly incorporated into
clinical practice as a noninva-
sive method that offers superior
Keywords: cardiac MRI, cardiologists, radiologists
Received May 28, 2009; accepted after revision
November 24, 2010.
F. Khosa and B. P. Romney contributed equally to this
OBJECTIVE. The purpose of our study was to determine the prevalence and significance
of noncardiac findings on clinical cardiovascular MRI and to identify the cardiovascular MRI
sequences that most frequently depict noncardiac findings.
MATERIALS AND METHODS. Images from 495 clinical cardiovascular MRI stud-
ies performed during 2006 were reviewed specifically for noncardiac findings by a cardio-
vascular imaging fellowship-trained radiologist without knowledge of the prior study inter-
pretation. Noncardiac findings were classified as benign (e.g., gynecomastia), indeterminate
(e.g., pleural effusion), or worrisome (e.g., lung nodule). The cardiovascular MRI sequences
depicting the noncardiac finding were recorded.
RESULTS. On image review, 295 noncardiac findings were identified in 212 (43%) of 495
studies, including 148 benign, 133 indeterminate, and 14 worrisome noncardiac findings. Of
these, 47% of indeterminate and 57% of worrisome noncardiac findings were not previously
known. Cardiovascular MRI sequences that most frequently showed noncardiac findings in-
cluded the single-shot fast steady-state free precession (SSFP) scout images (63% of all non-
cardiac findings) and axial T1-weighted fast spin-echo thoracic images (60% of all noncar-
diac findings), with 99% of management-changing noncardiac findings visualized on one of
these two sequences.
CONCLUSION. Noncardiac findings on clinical cardiovascular MRI are common. Al-
though only a small minority of studies contain management-changing noncardiac findings,
the vast majority of management-changing noncardiac findings are seen on thoracic SSFP
scout and axial T1-weighted thoracic fast spin-echo images.
Khosa et al.
Noncardiac Findings on Cardiovascular MRI
AJR:196, April 2011 W381
Noncardiac Findings on Cardiovascular MRI
ings and the cardiovascular MRI sequences
that most frequently depict them by examining
a consecutive series of clinical cardiovascular
MRI studies performed at a single academic
institution. We also examined the impact of
different reading session structures on the re-
porting of noncardiac findings.
Materials and Methods
This HIPAA-compliant study was approved by
our institution’s committee on clinical investiga-
tions. Informed consent was waived for this ret-
Cardiovascular MRI Subjects
The images from consecutive clinical cardio-
vascular MRI studies performed during the cal-
endar year 2006 at our institution were selected
for review. These included 495 studies on 443 sub-
jects (63% males), aged 51 ± 16 years (range, 12–
84 years). The clinical indications for cardiovas-
cular MRI referral are summarized in Table 1.
Cardiovascular MRI Sequences
Clinical cardiovascular MRI studies were carried
out using a standardized protocol based on indica-
tion. All imaging was performed using a 1.5-T MR
system (Achieva, Philips Healthcare) with a com-
mercial five-channel cardiac coil. All studies includ-
ed initial survey single-shot steady-state free preces-
sion (SSFP) thoracic (scout) images in the axial (n =
9), coronal (n = 9), and sagittal (n = 9) planes, field of
view, 450 × 450; assessment of left ventricular and
right ventricular systolic function using SSFP long-
axis, contiguous short-axis, 4-chamber, and left
ventricular outflow tract views; axial T1-weighted
fast spin-echo thoracic images; and velocity-encod-
ed flow imaging of the proximal aorta and the prox-
imal main pulmonary artery.
Additional sequences were performed as dictat-
ed by the clinical indication and included T1-weight-
ed fast spin-echo axial images with fat suppression,
T1-weighted fast spin-echo sagittal images, T1-
weighted axial images after gadolinium contrast ad-
ministration, T2-weighted fast spin-echo axial im-
ages (with or without fat suppression), T2-weighted
STIR images, myocardial late gadolinium en-
hancement images, myocardial perfusion imag-
ing, tagged CSPAMM imaging, pulmonary vein
MRA, aortic MRA, and coronary artery imaging.
Original report—At our center, cardiovascu-
lar MRI studies are scheduled to be reviewed at a
daily combined readout session attended by both
a body MR fellowship-trained radiologist (with-
out specific fellowship training in thoracic or car-
diac radiology) and a cardiovascular MRI level 2
or level 3 trained cardiologist  (as well as ra-
diology and cardiology trainees). The digital im-
ages are reviewed including all full field of view
localizer views. The report syntax provided infor-
mation as to whether a staff radiologist was pres-
ent for the combined readout. When not present at
the initial interpretation, the body MR fellowship-
trained radiologist reviewed the images at a later
time (serial evaluation) and before electronic sign-
out of the report.
Noncardiac findings—For this study, a car-
diovascular imaging fellowship-trained radiolo-
gist with 7 years of experience independently re-
viewed all of the images of all 495 cardiovascular
MRI studies, with particular focus on noncardiac
findings and without knowledge of the prior clini-
cal report or the patient’s medical history. Those
findings were then reviewed for accuracy by a
body MR fellowship trained radiologist with 16
years of experience.
For each scan reviewed, a database entry was
created, which included the patient’s date of birth,
sex, examination date, indication for examination,
presence or absence of noncardiac findings, pres-
ence or absence of a radiologist at the cardiovas-
cular MRI readout, cardiovascular MRI sequenc-
es performed, cardiovascular MRI sequences that
depicted the noncardiac finding (if present), and
classification of the noncardiac finding (if pres-
ent). Noncardiac findings were classified as be-
nign, indeterminate, or worrisome on the basis of
their MR appearance and without knowledge of
the patient’s medical history.
Benign findings were categorized as those
thought not to require follow-up assessment and
were deemed unlikely to change patient manage-
ment. These included such findings as gyneco-
mastia, small hiatal hernia, simple renal cyst, and
benign liver cyst or hemangioma (Fig. 1).
Indeterminate findings were those considered
more concerning, for which some form of fol-
low-up would have been suggested. These includ-
ed such findings as a pleural effusion or enlarged
thoracic lymph nodes as well as liver and kidney
masses that could not be definitively characterized
as cysts from the cardiovascular MR images (Fig.
2). Contrast-enhanced cardiac images, if avail-
able, were also reviewed; however, these images
did not always include the incidental noncardiac
finding in the field of view.
Worrisome findings were considered most like-
ly to represent serious disease; these might in-
clude lung nodules (Fig. 3), aortic dissection, aor-
tic ulcer, and mediastinal soft-tissue mass (Fig. 4).
Fig. 1—Benign liver lesions, cysts, or hemangiomas (arrows) in 54-year-old man referred for assessment of
hypertrophic obstructive cardiomyopathy.
A and B, Coronal (A) and axial (B) scout images (TR/TE, 1.85/0.77; flip angle, 50°). Combined readout was
performed and finding was not reported.
TABLE 1: Indications for
Pulmonary vein anatomy
Coronary artery assessment
Congenital heart disease
Cardiac mass assessment
Evaluation of great vessels
Note—An individual study may have more than one
indication. Data in parentheses are percentages.
W382 AJR:196, April 2011
Khosa et al.
We characterized both indeterminate and worri-
some noncardiac findings as potential manage-
Radiologic findings that were not considered
noncardiac findings included expected sequelae
of known prior surgeries; for example, breast im-
plants, sternal wires, bypass graft markers, etc.
However, if a complication from a prior surgery
was identified, this was included in our data as
a management changing finding (e.g., ruptured
breast implant (Fig. 5).
The medical center’s electronic medical re-
cord of all patients with indeterminate and worri-
some noncardiac findings was reviewed to deter-
mine whether the noncardiac findings represented
new findings and to determine what follow-up, if
any, was performed. For patients who had been re-
ferred from an outside institution, we sent corre-
spondence to the referring physician to request such
data. Follow-up was not deemed necessary for pa-
tients whose findings were previously known or for
patients whose new findings represented expected
sequelae from previously known conditions (e.g., a
pleural effusion in a patient with congestive heart
failure required no further follow-up).
Data are presented as mean ± SD. Comparisons
were performed by chi-square test with a statisti-
cal significance level of p ≤ 0.05.
Independent review of the cardiovascu-
lar MR images identified a total of 295 non-
cardiac findings in 212 (43%) of 495 studies.
Of these noncardiac findings, 50% (148/295)
were benign, 45% (133/295) were indetermi-
nate, and 5% (14/295) were worrisome. The
most common benign noncardiac findings
were gynecomastia (n = 41) and hiatal her-
nia (n = 22), whereas the most common in-
determinate noncardiac findings were pleu-
ral effusion (n = 29) and renal mass (n = 27).
Lung nodules (n = 11) were the most com-
mon worrisome noncardiac finding. The
more common indeterminate and worrisome
noncardiac findings as well as how many of
these represented new previously unknown
noncardiac findings are presented in Table
2. Forty-seven percent of indeterminate and
57% of worrisome findings were not previ-
ously known (Table 2).
The age of those with noncardiac find-
ings was higher than those without (56 vs
48 years, p < 0.001). When the benign (and
common) finding of gynecomastia was ex-
cluded in men, there was no significant dif-
ference in frequency of noncardiac findings
between men and women (36% vs 38%, re-
spectively, p = 0.85).
Cardiovascular MRI Sequences and
Noncardiac findings were most often iden-
tified on SSFP scout images (depicting 63% of
all noncardiac findings) and axial T1-weight-
ed fast spin-echo images (60% of all noncar-
diac findings). Only 14 (5%) noncardiac find-
ings were not visualized on either of these two
Fig. 2—Complex renal cysts seen in 69-year-old woman with atrial fibrillation referred for evaluation of
pulmonary vein anatomy before percutaneous pulmonary vein isolation.
A, Short-axis cine image (TR/TE, 3.05/1.52; flip angle, 60°) shows large septated complex renal cyst (arrow).
B, Coronal scout image (1.85/78; flip angle, 50°) in same patient. Combined readout was performed and finding
was listed in clinical report. Subsequent abdominal CT revealed complete replacement of left kidney by dilated
collecting system with very minimal peripheral renal parenchyma (i.e., obstructed atrophied kidney).
Fig. 3—Lung nodule in 67-year-old man referred for evaluation of left ventricular function and myocardial viability.
A, Axial T1-weighted image (TR/TE, 1091/20; flip angle, 90°) shows right lower lobe lung nodule (arrow).
B and C, Coronary MR survey image (B) (5.11/2.55; flip angle, 90°) from same study and coronal scout image (C) (1.96/82; flip angle, 50°) also show nodule (arrow).
Combined readout was performed, and finding was not listed in clinical report. Later workup revealed right lower lobe adenocarcinoma.
AJR:196, April 2011 W383
Noncardiac Findings on Cardiovascular MRI
sequences, including 12 benign findings and
two indeterminate renal masses. Thus, 95%
(281/295) of all noncardiac findings and 99%
(155/157) of potential management changing
noncardiac findings were visualized on either
thoracic SSFP scout or axial T1-weighted tho-
racic fast spin-echo images.
Noncardiac Finding Follow-Up
Seventy-one indeterminate and worrisome
noncardiac findings (48% of all indetermi-
nate and worrisome noncardiac findings) rep-
resented new findings (i.e., disease that was
not known before the patient’s cardiovascu-
lar MRI study). Ten of these findings needed
no further follow-up because they represent-
ed expected sequelae of a known prior condi-
tion. Table 3 summarizes the new noncardiac
findings and their follow-up in the remaining
61 patients with indeterminate or worrisome
findings. Follow-up on new unreported non-
cardiac findings (i.e., new findings that were
not discovered until our review of the imag-
es) occurred serendipitously in many cases
(finding discovered incidentally on a later ra-
diologic study or physical examination).
Follow-up on all new findings revealed 14
major management-changing findings in 11
patients, whereas 11 noncardiac findings in
10 patients were determined to be benign or
were not appreciable on follow-up studies.
Management-changing diseases included
lung cancer (n = 2); typical pulmonary car-
cinoid (n = 1); pulmonary nodules (n = 2);
mediastinal lymphadenopathy (n = 1), still
being followed at the time of this writing;
cryptogenic organizing pneumonitis (n = 1);
multifocal pneumonia (n = 1); intracapsular
breast implant rupture (n = 1); obstructed, at-
rophied kidney (n = 1); and recurrent abdom-
inal aortic aneurysm (n = 1). Pneumonitis
and pneumonia were treated with antibiotics
after the cardiovascular MRI findings were
reported to the clinician. The clinician was
informed of the intracapsular breast implant
rupture and it was then removed. The clini-
cian was informed of the obstructed atro-
phied kidney and lithotripsy was attempted.
The recurrent abdominal aortic aneurysm is
being followed-up with serial imaging. Thus
3% (14/495) of clinical cardiovascular MRI
reports contained management-changing pa-
thology. More than one half (14/25) of new
indeterminate or worrisome findings that re-
ceived appropriate follow-up turned out to be
Report Findings and the Combined Readout
Versus Later (Serial) Readout by the Radiologist
Overall, 36% (105/295) of all noncardiac
findings had been listed in the original clinical
report, including 21% (30/148) of benign find-
ings and 51% (75/147) of indeterminate and
worrisome findings. Eleven noncardiac find-
ings noted in the original clinical report were
not identified on our retrospective image re-
view of the studies, including eight benign non-
cardiac findings, two indeterminate noncardiac
findings (atelectasis and an enlarged thoracic
lymph node), and one worrisome noncardiac
finding (lung mass). Six (43%) of 14 manage-
ment-changing findings were included in the
patient’s clinical cardiovascular MRI report.
A combined (concurrent radiologist and
cardiology) readout was used for 78% (384/495)
of all cardiovascular MRI studies and was as-
sociated with increased reporting of all noncar-
diac findings (42% vs 15% for serial readout,
p < 0.001) as well as combined indeterminate
and worrisome (i.e., potential management-
changing) noncardiac findings (56% vs 33%
for serial readout, p = 0.03) (Table 4).
Fig. 4—Mediastinal soft-tissue mass in 22-year-old man with mediastinal B-cell lymphoma who was referred
for evaluation of left ventricular ejection fraction and pericardial disease.
A and B, Axial T1-weighted image (A) (TR/TE, 811/20; flip angle, 90°) and axial T1-weighted gadolinium-
enhanced image with fat suppression (B) (811/20; flip angle, 90°) shows mediastinal soft-tissue mass (arrow)
lateral to aortic arch. This finding was consistent with patient’s known history of lymphoma. Combined readout
was performed, and finding was listed in clinical report.
Fig. 5—Intracapsular breast implant rupture of right
breast in 55-year-old woman referred for evaluation
of pericardial constriction.
A and B, Sagittal (A) and axial (B) scout images
(TR/TE, 1.8/0.77; flip angle, 50°) show implant
rupture (arrow). Serial readout was performed and
finding was not listed in clinical report. Patient
later underwent removal of implant due to breast
contracture noted on physical examination.
W384 AJR:196, April 2011
Khosa et al.
In this study of 495 consecutive clinical
cardiovascular MRI studies performed in a
single calendar year at an academic institu-
tion, we found that noncardiac findings are
common, identified in 43% of all clinical
cardiovascular MRI studies. Although 50%
were benign, the other 50% (147/295) rep-
resented indeterminate or worrisome find-
ings, with 10% (14/147) of these (3% of 495
studies) leading to changes in patient man-
agement. The vast majority of management-
changing noncardiac findings were visu-
alized on thoracic SSFP scout and axial
T1-weighted thoracic fast spin-echo images.
Prior studies that have examined noncar-
diac findings in this population have gener-
ally focused on noncardiac findings included
in clinical reports. Dewey et al.  inves-
tigated the prevalence of noncardiac find-
ings in patients undergoing coronary MRI
only. They found “insignificant” noncardiac
findings in 6% and “significant” noncardiac
findings in 2% of their subjects, values much
lower than our study but similar to a cardio-
vascular MRI report review from our center
by Chan et al. , who found “minor” non-
cardiac findings in 4.6% and “major” non-
cardiac findings in 3.3% of subjects. Simi-
larly, a preliminary report by Burchell et al.
 that extracted noncardiac findings from
cardiovascular MRI clinical reports in 714
examinations found 24% of subjects with
noncardiac findings. The majority were liv-
er or renal cysts with a very small number of
major noncardiac findings (lung, mediasti-
nal, and renal masses or lymphadenopathy).
More extensive data exist for noncardi-
ac findings identified by cardiac CT: Horton
et al.  identified noncardiac findings on
electron-beam CT (EBCT) of coronary ar-
teries in 7.8% of studies. Hunold et al. 
identified noncardiac findings on EBCT of
coronary arteries in 11% of studies. Schragin
et al.  identified noncardiac findings on
EBCT of coronary arteries in 4.2% of stud-
ies. Haller et al.  had a rate of 4.8% sig-
nificant findings on MDCT of coronary ar-
teries. Onuma et al.  found a higher rate
of 22.7% of significant findings on MDCT of
McKenna et al.  reported on noncardi-
ac findings among a group of 107 predomi-
nantly male subjects who underwent screen-
ing cardiovascular MRI for research using
similar cardiovascular MRI sequences. Simi-
lar to our study, the images were specifically
reviewed for noncardiac disease. They found
81% of their subjects had noncardiac findings,
including 17% with potentially significant dis-
ease . No data were reported by McKenna
et al. as to whether the disease was “previous-
ly known,” nor was follow-up provided to dis-
cover the true impact on patient care.
Nearly all (99%) noncardiac findings identi-
fied in our study were seen on either initial scout
TABLE 2: Noncardiac Findings in 495 Consecutive Clinical Cardiovascular
FindingsTotal No. of New Findings
Liver cyst or hemangioma
Enlarged thoracic nodes < 1.5 cm
Enlarged thoracic nodes > 1.5 cm
Large hiatal hernia
Lung nodule or mass
Mediastinal soft-tissue mass
Note—New findings are those not previously known according to review of the subject’s electronic medical
record or physician contact. Data in parentheses are percentages.
TABLE 3: Follow-Up Data on New Noncardiac Findings
FindingsPatients With New Findings Needing Follow-UpFollow-Up Performed
Note—Data in parentheses are percentages.
aReferring physicians were notified but the suggested follow-up was not performed in three cases.
AJR:196, April 2011 W385
Noncardiac Findings on Cardiovascular MRI
SSFP images or axial T1-weighted fast spin-
echo thoracic images. This highlights the im-
portance of focused attention on these images
for noncardiac findings. Although additional se-
quences were also effective at depicting noncar-
diac findings when performed, in most cases,
these examinations facilitated further character-
ization of a noncardiac finding by comparing its
appearance on multiple sequences.
In our study, the reporting of noncardiac
findings was aided by a combined readout at-
tended by both a radiologist and cardiologist.
Our intent is for a cardiologist and radiologist
to both participate in the interpretation of
clinical cardiovascular MRI studies at a com-
bined readout. Aside from education of our
fellows and residents, an important reason for
a radiologist’s participation is to aid in the de-
tection and characterization of noncardiac
findings. Our data validate the utility of this
approach. The reason for the improved identi-
fication of noncardiac findings in a combined
readout cannot be discerned from our study.
Possible considerations when interpreting the
serial study format data include the presump-
tion that another reader had already identified
noncardiac findings leading to a more cursory
review or a potential influence from the satis-
faction-of-search phenomenon when certain
findings were encountered .
Even with a board-certified body MR-
trained radiologist present for the combined
cardiovascular MRI readout, only 42% of all
noncardiac findings identified on image review
were reported. Such a low reporting rate may be
partially explained by certain benign findings
that were considered by the interpreting physi-
cians to be clinically unimportant for reporting
purposes. However, only 51% of indeterminate
and worrisome noncardiac findings were re-
ported. The cause of their omission cannot be
determined by our study and remains to be fur-
ther explored by future investigations.
Our study was limited by our reliance on a
single radiologist reviewing all 495 cardiovas-
cular MRI studies, with subsequent pathology
review or confirmation by a second radiolo-
gist. We recognize that there can be disagree-
ment among radiologists reading the same
images. Indeed, although 64% of noncardiac
findings identified on our review were not in-
cluded in the clinical report, 10% of noncar-
diac findings in the original clinical report
were not identified by our reviewer.
As expected, follow-up on reported noncar-
diac findings was performed more frequent-
ly than on unreported noncardiac findings but
was still only performed in 57% of the cases in
which further workup was suggested. This is
likely due to possible failure at any of the many
links in the chain of follow-up: the initial com-
munication of the finding to the referring phy-
sician, the referring physician’s communication
of the finding to the patient, the scheduling of
a follow-up study, and the patient’s response to
the follow-up request. Additionally, the refer-
ring physician may have deemed the findings
not worthy of further workup for other reasons.
We did not examine the health care cost im-
pact of noncardiac finding follow-up. There is a
TABLE 4: Impact of Combined Versus Serial Cardiovascular MRI Readout on
Reporting of Noncardiac Findings
ReportedNo. ReportedTotalReported (%)pa
Indeterminate and worrisome
Note—Data in parentheses are number/total.
ap value calculated for combined readout versus serial readout.
trade-off between expending resources on be-
nign findings and detecting meaningful man-
agement-altering findings. In about one half of
patients (11/21) with a new indeterminate or
worrisome finding who received appropriate
follow-up, that finding turned out to be man-
agement altering, suggesting that follow-up in
this cohort is worthwhile. Our findings regard-
ing the value of a combined (vs serial) readout
were also not based on prospective randomized
reading assignment. Nor do we know the po-
tential differential impact of a cardiac-trained
radiologist vs body radiologist. Finally, our re-
sults may not be applicable to all institutions
with varying cardiovascular MRI expertise or
those lacking a structured cardiovascular MRI
scanning protocol that always includes axial
thoracic T1-weighted fast spin-echo imaging.
In summary, noncardiac findings on clinical
cardiovascular MRI are common although
only a small minority of studies contain man-
agement-changing noncardiac findings. The
vast majority of management-changing non-
cardiac findings are visualized on thoracic
SSFP scout and axial T1-weighted thoracic fast
spin-echo images. Cardiovascular MRI practi-
tioners and trainees should focus attention on
these images for noncardiac findings.
1. Finn JP, Nael K, Deshpande V, Ratib O, Laub G.
Cardiac MR imaging: state of the technology. Ra-
diology 2006; 241:338–354
2. Manning WJ and Pennell DJ. Cardiovascular
magnetic resonance. Philadelphia PA: Churchill-
3. Higgins CB, DeRoos A, eds. MRI and CT of the
cardiovascular system. Philadelphia, PA: Lippin-
cott Williams & Wilkins, 2005
4. Maizlin ZV, Barnard SA, Gourlay WA, Brown
JA. Economic and ethical impact of extrarenal
findings on potential living kidney donor assess-
ment with computed tomography angiography.
Transpl Int 2007; 20:338–342
5. Liu W, Mortele KJ, Silverman SG. Incidental ex-
traurinary findings at MDCT urography in pa-
tients with hematuria: prevalence and impact on
imaging costs. AJR 2005; 185:1051–1056
6. Ginnerup Pedersen B, Rosenkilde M, Christians-
en TE, Laurberg S. Extracolonic findings at com-
puted tomography colonography are a challenge.
Gut 2003; 52:1744–1747
7. Hara AK, Johnson CD, MacCarty RL, Welch TJ.
Incidental extracolonic findings at CT colonogra-
phy. Radiology 2000; 215:353–357
8. Hellstrom M, Svensson MH, Lasson A. Extraco-
lonic and incidental findings on CT colonography
(virtual colonoscopy). AJR 2004; 182:631–638
W386 AJR:196, April 2011
Khosa et al.
9. Glockner JF. Incidental findings on renal MR an-
giography. AJR 2007; 189:693–700
10. Dewey M, Schnapauff D, Teige F, Hamm B. Non-
cardiac findings on coronary computed tomogra-
phy and magnetic resonance imaging. Eur Radiol
11. Haller S, Kaiser C, Buser P, Bongartz G, Bremer-
ich J. Coronary artery imaging with contrast-en-
hanced MDCT: extracardiac findings. AJR 2006;
12. Horton KM, Post WS, Blumenthal RS, Fishman
EK. Prevalence of significant noncardiac findings
on electron-beam computed tomography coronary
artery calcium screening examinations. Circula-
tion 2002; 106:532–534
13. Hunold P, Schmermund A, Seibel RM, Grone-
meyer DH, Erbel R. Prevalence and clinical sig-
nificance of accidental findings in electron-beam
tomographic scans for coronary artery calcifica-
tion. Eur Heart J 2001; 22:1748–1758
14. Mueller J, Jeudy J, Poston R, White CS. Cardiac CT
angiography after coronary bypass surgery: preva-
lence of incidental findings. AJR 2007; 189: 414–419
15. Onuma Y, Tanabe K, Nakazawa G, et al. Noncar-
diac findings in cardiac imaging with multidetec-
tor computed tomography. J Am Coll Cardiol
16. Schragin JG, Weissfeld JL, Edmundowicz D,
Strollo DC, Fuhrman CR. Non-cardiac findings
on coronary electron beam computed tomography
scanning. J Thorac Imaging 2004; 19:82–86
17. Kim RJ, de Roos A, Fleck E, et al. Guidelines for
training in cardiovascular magnetic resonance
(CMR). J Cardiovasc Magn Reson 2007; 9:3–4
18. Chan PG, Rofsky NM, Yeon SB, et al. Non-cardiac
findings on clinical cardiac magnetic resonance im-
aging are rare. (abstr) Circulation 2007; 116:II; 773
19. Burchell TR, Locca D, Mathur A, Davies C, West-
woo MA. Prevalence of non-cardiac incidental
findings during routine clinical CMR assessment.
J Cardiovasc Magn Reson 2009; 11[suppl 1]:48
20. McKenna DA, Laxpati M, Colletti PM. The prev-
alence of incidental findings at cardiac MRI.
Open Cardiovasc Med J 2008; 2:20–25
21. Samuel S, Kundel HL, Nodine CF, Toto LC.
Mechanism of satisfaction of search: eye position
recordings in the reading of chest radiographs.
Radiology 1995; 194:895–902