AJR 2000; 174:135-139
© American Roentgen Ray Society
Is It Necessary to Biopsy the Obvious?
Michael Y. M. Chen1,
David W. Gelfand1,
Robert E. Bechtold1,
Steven A. Cremer1,
Bradley J. Casolo1 and
Paul D. Savage2
1
Department of Radiology, Wake Forest University School of Medicine, Medical
Center Blvd., Winston-Salem, NC 27157-1088.
2
Department of Hematology-Oncology, Wake Forest University School of Medicine,
Winston-Salem, NC 27157-1088.
Received April 19, 1999;
accepted after revision June 25, 1999.
Address correspondence to M. Y. M. Chen.
Abstract
OBJECTIVE. The radiologist and oncologist are often confident that
biopsy will confirm their suspicion of recurrent disease, but a biopsy is
performed to confirm the histologic diagnosis before beginning or altering
therapy. We have examined data to determine how often the biopsied lesion
represents recurrent disease from the primary tumor or is an instance of new
cancer, and whether recurrent disease can be predicted.
MATERIALS AND METHODS. We reviewed the medical and imaging records
of 253 patients who underwent CT-guided biopsy of an abdominal or pelvic
lesion between 1993 and 1996. Sixty-nine of the 253 patients had a previously
diagnosed primary tumor and were being examined for possible tumor recurrence
or metastasis. The images of these 69 patients were analyzed to determine if
the pattern of disease was typical of recurrence or metastasis.
RESULTS. In 55 of the 69 patients, the pattern was judged to be
typical of metastatic or recurrent disease. Biopsy confirmed this suspicion in
all 55 patients. In 14 of the 69 patients, the pattern of spread was judged
not to be typical of recurrence or metastasis. These 14 patients were found to
have a new primary tumor (n = 4), benign processes (n = 2),
and recurrences (n = 8).
CONCLUSION. Of the patients for whom radiographic findings suggested
recurrence, we found no patients in whom a new primary tumor would have been
missed if biopsy had been avoided. Data should now be acquired prospectively
to determine whether it may be prudent to make treatment decisions on the
basis of imaging findings alone, without histologic confirmation.
Introduction
Biopsies in the abdomen or pelvis performed with CT guidance are
often requested to confirm recurrent or metastatic tumor in patients with a
known primary cancer [1]. A
common example is percutaneous biopsy of a liver lesion in a patient with
known colonic cancer. The radiographic appearance may be highly suggestive of
metastatic disease, but a biopsy is performed anyway to remove any uncertainty
before beginning or altering therapy. Percutaneous biopsy, however, is an
expensive invasive procedure. Avoiding unnecessary biopsies when the
radiographic appearance strongly suggests recurrent tumor could reduce health
care costs and prevent potential complications. We retrospectively examined
data from CT-guided biopsies of abdominal and pelvic processes performed at
our institution to determine how often the biopsied lesion represents a second
primary tumor and whether recurrent disease can be predicted with such
certainty that biopsy can be avoided in some circumstances.
Materials and Methods
We reviewed the medical and imaging records of 253 patients who underwent
CT-guided biopsy of abdominal or pelvic lesions at our institution between
June 1993 and October 1996. Biopsies of chest lesions and bony lesions were
not included in this series primarily because the authors are abdominal
radiologists who do not perform interventional procedures outside their area
of expertise. A total of 177 patients had no known primary tumor at the time
of the biopsy and were excluded from the study. Seven other patients were
excluded either because the biopsy was nondiagnostic (the specimen was graded
unsatisfactory or marginal [n = 4]) or because the imaging or
pathology records were not available (n = 3). The remaining 69
patients (39 women, 30 men), having an average age of 62 years (range, 32-91
years), were consecutively selected for this study.
The distribution of disorders and the location of primary carcinomas are
shown in Table 1. During
initial diagnosis the primary carcinomas were staged as follows: eight stage
I, 26 stage II, 24 stage III, and 11 stage IV. The average time from the
initial surgery to CT-guided biopsy was 44 months (range, 1-300 months). The
distribution of biopsy sites was in the liver, 20; in the lymph nodes, 11; in
the pelvis, 11; in the retroperitoneum, 10; in the peritoneum and abdominal
wall, 15; and in the pancreas, two. The size of the biopsied lesion was
recorded for 44 patients (64%). The lesion size was less than 5 cm in 34
patients, ranged from 5 to 10 cm in eight patients, and was larger than 10 cm
in two patients.
Two radiologists retrospectively analyzed the abdominal CT scan of each
patient to determine if it was typical of recurrent disease. Chest CT was not
included in this evaluation. The investigators had access to all pertinent
information one would expect to be available on request for a biopsy of a
potential recurrent mass—namely, knowledge of the original tumor
histology and anatomic location and all imaging studies available at the time
of biopsy. Both reviewers assessed the lesion to determine whether it was
typical or atypical for lesions categorized as recurrent, but the reviewers
were not informed of who originally interpreted the report. The time between
assessing potential for recurrence and performing the biopsy was 2-4 years.
The investigators were not informed of the biopsy results.
Each case was then categorized, by consensus of the investigators, as being
either suggestive of recurrent tumor or not suggestive of recurrent tumor on
the basis of the investigators' experience and knowledge of known metastatic
pathways [2]. For example, if
multiple areas of low attenuation in the liver associated with multiple
enlarged lymph nodes were shown in a patient with known colonic carcinoma, the
liver lesion was considered suggestive of a recurrence. Cases were considered
not to be suggestive of recurrent tumor for any of the following reasons: a
new lesion developed at an unusual location away from the normal metastatic
pathway; a new lesion occurred at the normal pathway but without intervening
lesions between the primary and new lesions; a new lesion occurred with
unusual timing; or there were two or more primary neoplasms and it was not
known which one was causative. If a new lesion appeared 20 years after the
diagnosis of primary breast cancer, the lesion was categorized as not
suggestive of recurrent tumor because of unusual timing. If a patient with
known colonic carcinoma had a pancreatic mass with no enlarged lymph nodes
between the original tumor and the pancreatic lesion, this mass was also
categorized as not suggestive of recurrent tumor. Categorizing by the
investigators was then compared with the biopsy results to determine the
accuracy of the investigators in predicting recurrent tumor on the basis of
imaging findings.
Results
On the basis of the results of the 69 biopsies performed in patients with a
known primary tumor, 63 patients were found to have a recurrence of the
primary tumor, four had a new (second) primary tumor, and two had benign
processes. The size of the biopsied lesion in one benign process and two new
(second) primary tumors was recorded. The size of one benign lesion was
recorded as between 5 and 10 cm. The size of two primary tumors was recorded
as less than 5 cm for one tumor and between 5 and 10 cm for the other tumor.
Correlation between mass size and incidence of recurrence was not performed
because of our limited sample size. The investigators categorized 55 (80%) of
69 cases as suggestive of recurrent tumor
(Figs. 1 and
2) and 14 (20%) as not
suggestive of recurrent tumor (Figs.
3, 4,
5). The correlation between
their category and the biopsy results is shown in
Table 2.
View larger version (129K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1. —75-year-old man with colonic cancer suggestive of recurrent tumor.
Patient had colonic cancer found at surgery 2 years earlier. CT scan shows
multiple poorly defined low-attenuation lesions (arrow) in liver.
None of lesions suggests a second primary tumor. Because patient was not
febrile, multifocal liver abscesses were not considered in differential
diagnosis. Metastatic adenocarcinoma was confirmed at biopsy.
|
|
View larger version (132K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2. —55-year-old woman with ovarian cystadenocarcinoma suggestive of
recurrent tumor. Patient had ovarian cystadenocarcinoma diagnosed 4 years
earlier and treated with chemotherapy. CT scan shows new pelvic soft-tissue
mass measuring 4 x 5 cm (arrow). Because lesion is near primary
carcinoma and developed within reasonable time interval, metastasis from
ovarian carcinoma was diagnosed and was confirmed by biopsy.
|
|
View larger version (124K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3. —76-year-old woman with colonic adenocarcinoma not suggestive of
recurrent tumor. Patient had colonic adenocarcinoma diagnosed 8 years
previously. CT scan shows diffuse ascites and mass in pancreas
(arrow). No adenopathy intervenes between site of original tumor and
newer pancreatic mass, and no other evidence of metastatic disease is
present—a pattern considered atypical of colonic cancer metastases.
Biopsy showed chronic inflammation but no tumor.
|
|
View larger version (122K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4. —59-year-old woman with breast carcinoma not suggestive of recurrent
tumor. Patient had carcinoma of right breast and underwent resection 23 years
earlier. Carcinoma in left breast was resected 5 years previously. CT scan
shows large renal lesion (arrow). Because kidney is not typical
metastatic pathway for breast carcinoma, because multiple primary neoplasms
are present, and because too much time has elapsed to suggest metastatic
lesion, tumor was categorized as not suggestive of recurrence.
Well-differentiated oncocytoid renal cell adenocarcinoma in right kidney was
confirmed during biopsy.
|
|
View larger version (121K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 5. —65-year-old man with prostate carcinoma not suggestive of recurrent
tumor. Patient had prostate carcinoma diagnosed 3 years earlier. CT scan shows
innumerable areas of decreased density in liver (arrow). No
adenopathy or intervening metastatic lesions are present between site of
original tumor, prostate gland, and liver to suggest recurrent neoplasm. New
primary small cell carcinoma in liver was confirmed by biopsy. Large gallstone
(S) was incidental finding.
|
|
Recurrent tumor was confirmed in all 55 patients whose tumors were
categorized as suggestive of recurrence. No patients with either a new
(second) primary tumor or a benign process were incorrectly categorized as
having a recurrence from the primary tumor. Therefore, the categorizing of the
investigators produced a sensitivity for predicting recurrent tumor of 87%
(55/63), a specificity of 100% (6/6), a positive predictive value of 100%
(55/55), and a negative predictive value of 43% (6/14)
(Table 2).
The reasons for categorizing a tumor as not suggestive of recurrence are
listed in Table 3. In two
patients with benign processes, tumors were categorized as not suggestive of
recurrence, in one patient because of unusual timing and in the other patient
because of unusual timing and location. In four patients with new (second)
primary tumors, the reasons for categorizing as not suggestive of recurrent
tumor included no intervening lesions between primary and new lesions, unusual
location, and multiple primary lesions.
In 69 biopsies, the complication rate was 4% (3/69) and included hematoma
in two patients and pain requiring pain control treatment in one. The total
charge for each percutaneous biopsy in our institution was approximately
$1000.
Discussion
Many CT-guided biopsies are performed to establish unequivocally the cause
of newly discovered lesions in patients with known primary cancers. The
radiographic appearance is often completely suggestive of recurrent tumor, but
a biopsy is performed anyway. Although the biopsy may seem unnecessary, it is
performed to remove any doubt about the nature of the lesion, thereby avoiding
a number of errors. For example, if a lesion is assumed to be a recurrence of
the primary tumor but is in fact a new (second) primary tumor, the
continuation of prior treatment or institution of a specific alternative
treatment may be inappropriate. Alternatively, if a lesion is assumed to be a
recurrence but is in fact a benign process, the patient may receive
unnecessary chemotherapy or radiation. Because the consequences of these
errors could be significant, traditional thought has dictated that any doubt
about the nature of the lesion in question must be removed. Nevertheless, our
study shows that experienced radiologists can accurately identify 80% of
lesions that represent recurrent tumor. More important, this relatively small
survey suggests that it can be done without neglecting to biopsy a new primary
neoplasm.
Most recurrent tumors occur at sites along the lymphatic and vascular
circulation when the primary tumor cells lodge in the first capillary bed
encountered during their circulation. Nicolson
[3] and Morgan-Parkes
[4] describe in detail the
pathways and sites for recurrent tumors from different types of cells and
different sites of primary tumors. Some recurrent tumors cannot be explained
by simple anatomic or mechanical hypotheses and may follow the "seed and
soil" hypothesis [5,
6,
7,
8,
9]. Paget
[5] proposed that metastases
form when a particular tumor cell (seeds) has a suitable environment (soil) in
which to grow. The likelihood that colonic cancer will metastasize to the
liver is 50-60% [2]; however,
colonic cancer has a 6% chance of metastasizing to the spleen or a 3% of
chance of metastasizing to the myocardium, according to autopsy records
[2]. In the clinical setting, a
new liver lesion in a patient with colonic carcinoma is categorized as
suggestive of recurrent tumor, but a new lesion in the spleen or myocardium in
a patient with a history of colonic carcinoma is categorized as not suggestive
of recurrent tumor because of low frequency. Our categorizing system, which
considers metastatic pathways and timing of recurrence, depends on common
clinical experiences.
In no instance was a lesion that was judged to represent recurrent tumor
actually benign or a second primary tumor. Therefore, with our categorizing
system, 80% of biopsies may be avoided, specifically if the new lesion is
located along a typical metastatic pathway and if the new lesion occurs within
a reasonable time, such as 1-2 years. Not all lesions that represented
recurrent tumor were judged by the investigators to be so; that is, we had
some false-negative results. These results occurred because the investigators
were not trying to correctly categorize all lesions a priori, but rather to
determine if a subgroup of patients could be identified in whom the
radiographic diagnosis was so certain that biopsy could be safely
obviated.
In our categorizing system, the major factors for judging a lesion as not
suggestive of recurrent tumor are unusual timing or metastatic pathway,
absence of intervening lesions, and multiple primary neoplasms. The positive
predictive value in our study was 100%, and biopsy was avoidable in 55 of 63
patients with recurrent tumor. Each of the four patients who were found to
have a second primary tumor had a pattern of disease that was judged to be
atypical of recurrent neoplasms. For example, a 65-year-old man underwent a
biopsy for multiple areas of low attenuation in the liver. The man had
prostatic adenocarcinoma treated with radiation 3 years earlier, and there was
no intervening lymphadenopathy between the original prostate lesion and the
new lesion in the liver. Therefore, the lesion was categorized as not
suggestive of recurrent tumor. The pathology result from biopsy confirmed a
new neoplasm of neuroendocrine small cell carcinoma in the liver. The tumor
was strongly immunoreactive for synaptophysin and failed to stain for
prostate-specific antigen and prostatic acid phosphatase.
At our institution, the combined radiology, pathology, hospital, and
professional fees for a percutaneous biopsy total approximately $1000. If
biopsy had been avoided in the 55 patients with recurrent neoplasms judged
suggestive of recurrent tumor, a substantial reduction in health care
cost—$55,000—would have been realized. This analysis
underestimates the true financial cost because it does not take into
consideration travel costs for outpatients and their families or inpatient
hospital stays longer than would otherwise be necessary because biopsy cannot
be performed immediately.
We had four patients with nondiagnostic results because of unsatisfactory
biopsy specimens. The percentage of nondiagnostic biopsies is another factor
that increases health care costs. Costs related to care or treatment of
complications also are not factored into this analysis. One minor
complication, a hematoma, would have been avoided if biopsy had not been
performed.
Several limitations of this study bear mention. Because no absolute rules
exist concerning how and where cancer can spread, a completely objective
assessment of the lesion is not possible. It may also be argued that even if
the radiographic appearance is typical of recurrent tumor, as with liver
lesions in a patient with known colonic cancer, coexisting lung cancer is a
possibility and could be the source of the liver metastases. Some biopsies are
performed exclusively to satisfy clinical treatment protocols that require
biopsies for proof despite the obvious nature of the lesion. Those protocols
are beyond the control of both clinician and radiologist now, but perhaps data
of this type may alter even those practice patterns. At our institution,
biopsy of lesions is customary for all patients. However, some patients who
have a disease such as lymphoma and persistent lymphadenopathy do not undergo
a biopsy. If used in a large population sample, our categorizing system may
have a positive predictive value even lower than our data in this study
indicate.
Our limited sample size may lead to bias. Study of a larger population with
common types of cancer represented would help increase confidence in our
results. Furthermore, because ours was a retrospective study, we are now
conducting a prospective analysis of our ability to identify lesions that
represent recurrent tumor. Our sample selection may be biased because
reviewers who examined the CT scans knew all patients had a lesion to biopsy.
Finally, the categorizing of a lesion as suggestive or not suggestive of
recurrent tumor was based on the judgment of the investigators and their
knowledge of known metastatic pathways and lesion appearance.
No one would argue that blind adherence to cost-cutting measures warrants
making a mistake in even one instance. However, the reality of today's medical
environment is that attempts are being made to save money on all possible
fronts. It may be appropriate to reconsider the need to perform biopsies of
potential recurrent masses that follow expected behavior patterns if, as our
study suggests, these biopsies can be safely eliminated with the only error
being that a few atypical recurrent masses may still be biopsied.
In summary, our data suggest that most recurrent abdominal and pelvic
masses can be recognized on CT, a fact that can help in the selection of
masses to be biopsied and result in cost savings and avoidance of some
complications. In any instance in which doubt exists, the conservative
choice—biopsy—is always available.
References
-
Gazelle GS, Haaga JR. Imaging-guided percutaneous abdominal biopsy.
Oncology
1991;5:27-46
-
Weiss L. Principles of metastasis. Orlando,
FL:Academic Press,
1985:160-199
-
Nicolson GL. Organ specificity of tumor metastasis: role of
preferential adhesion, invasion and growth of malignant cells at specific
secondary sites. Cancer Metastasis Rev
1988;7:143-188[Medline]
-
Morgan-Parkes JH. Metastases: mechanisms, pathways, and cascades.
AJR
1995;164:1075-1082[Abstract/Free Full Text]
-
Paget S. The distribution of secondary growths in cancer of the
breast. Lancet
1889;1:571-573
-
McDaniel KP, Charnsangavej C, DuBrow RA, Varma DGK, Granfield CAJ,
Curley SA. Pathways of nodal metastasis in carcinomas of the cecum, ascending
colon, and transverse colon: CT demonstration. AJR
1993;161:61-64[Abstract/Free Full Text]
-
Shirkhoda A. Computed tomography of perirenal metastases. J
Comput Assist Tomogr
1986;10:435-438[Medline]
-
Charnsangavej C, Whitley NO. Metastases to the pancreas and
peripancreatic lymph nodes from carcinoma of the right side of the colon: CT
findings in 12 patients. AJR
1993;160:49-52[Abstract/Free Full Text]
-
Granfield CAJ, Charnsangavej C, Dubrow RA, et al. Regional lymph
node metastases in carcinoma of the left side of the colon and rectum: CT
demonstration. AJR
1992;159:757-761[Abstract/Free Full Text]
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
Top
Abstract
Introduction
