AJR 2004; 182:393-398
© American Roentgen Ray Society
Role of CT in the Management of Recurrent Ovarian Cancer
Stacey A. Funt1,
Hedvig Hricak1,
Nadeem Abu-Rustum2,
Madhu Mazumdar3,
Howard Felderman1 and
Dennis S. Chi2
1 Department of Radiology, Memorial Sloan-Kettering Cancer Center, 1275 York
Ave., New York, NY 10021.
2 Department of Surgery, Gynecology Service, Memorial Sloan-Kettering Cancer
Center, New York, NY 10021.
3 Department of Epidemiology and Biostatistics, Memorial Sloan-Kettering Cancer
Center, New York, NY 10021.
Received April 11, 2003;
accepted after revision August 19, 2003.
Presented at the 2001 annual meeting of the Radiological Society of North
America, Chicago, IL.
Address correspondence to H. Hricak.
Abstract
OBJECTIVE. The purpose of this study was to evaluate the potential
role of preoperative CT in patients with recurrent ovarian cancer who undergo
secondary cytoreductive surgery.
MATERIALS AND METHODS. Preoperative CT examinations of 36
consecutive patients (age range, 30–75 years; mean age, 55 years) were
reviewed retrospectively. Patients had recurrent ovarian cancer and secondary
cytoreduction within a mean CT–surgery interval of 22 days (range,
2–69 days). The CT findings recorded were upper abdominal metastases
(e.g., peritoneal carcinomatosis; perihepatic, perisplenic, gastrohepatic or
gastrosplenic ligaments; gallbladder fossa; falciform ligament; lesser sac),
lymphadenopathy (above or below the renal hilum), liver metastasis, large- and
small-bowel obstruction, hydronephrosis, ascites, and the presence of a pelvic
mass. CT findings and cancer antigen–125 (CA-125) levels were correlated
with surgical resectability.
RESULTS. At surgery, tumors in 27 patients were optimally debulked
(residual disease of 
1 cm) and in nine patients were nonresectable. Using
multivariate analysis, hydronephrosis (odds ratio = 19.4, p = 0.03)
and invasion of pelvic sidewall (odds ratio = 35.6, p = 0.006) were
found to be most indicative of tumor nonresectability. The presence of
small-bowel obstruction; nodal or perihepatic liver metastasis; ascites;
peritoneal carcinomatosis; bladder, rectum, sigmoid colon, or vaginal
involvement; or infrarenal paraaortic adenopathy; and the level of CA-125 were
not strong indicators of tumor nonresectability.
CONCLUSION. In patients with recurrent ovarian carcinoma considered
for secondary cytoreductive surgery, preoperative CT can be helpful in
identifying the extent of the disease and can be used as an adjunct to
treatment planning and management decisions.
Introduction
CT in the preoperative evaluation of patients with ovarian cancer is
controversial. Although some surgeons advocate the routine use of imaging for
the staging and assessment of disease before primary surgical debulking
[1], others believe imaging has
no added value in treatment planning
[2]. In spite of a lack of
consensus, the use of imaging in ovarian cancer is increasing, and the use of
CT to evaluate primary ovarian cancer is gaining acceptance. However, the role
of imaging in recurrent ovarian cancer has received little attention and has
not been investigated.
Understanding the terminology used to differentiate stages of ovarian
cancer and its relationship to patient management is essential to appreciate
the role of imaging. The initial treatment for patients with ovarian cancer is
primary surgical debulking followed by chemotherapy. Surgical cytoreduction is
considered optimal when the greatest diameter of residual disease is 1 cm or
less [3]. Specific terms are
used to define growth patterns of ovarian cancer after initial treatment.
Patients with platinum-sensitive disease have a complete clinical response to
primary platinum-based chemotherapy that lasts at least 6 months from the
completion of the initial therapy. Platinum-resistant disease is defined as
tumor recurrence detected less than 6 months after completion of initial
treatment. Refractory, unresponsive, or progressive disease is tumor that
progresses during initial therapy
[4,
5]. Recurrent ovarian cancer is
defined as tumor recurrence after a complete initial response to first-line
therapy, negative findings at a second-look operation, if performed, and a
disease-free interval greater than 6 months
[5]. The treatment of recurrent
ovarian cancer depends on tumor bulk and extent and ranges from surgical
debulking and palliation to medical therapy. In this patient group, a survival
benefit is seen when cytoreduction is optimal despite the significant
morbidity associated with surgical intervention
[6–8];
therefore, pretreatment identification of nonresectable recurrent cancer is
clinically relevant and can assist in patient management.
The potential role of imaging to identify nonresectable disease in primary
ovarian cancer has been shown
[1,
9]. However, cross-sectional
imaging has not been evaluated for its role in the triage, planning, and
monitoring of treatment response in persistent, resistant, refractory,
progressive, or recurrent ovarian cancer. Specifically, in a patient with
recurrent ovarian cancer, the pattern of CT findings has not been described,
CT correlation with surgical outcome has not been performed, and no imaging
guidelines exist to assist in identification of tumor nonresectability. Tumor
biology of primary and recurrent disease is often different. Ancillary
findings such as ascites have varying incidence, and distortion of anatomy
after surgery or the presence of surgical complications such as adhesions may
have a profound impact on the role of imaging in assessing the presence and
extent of the disease.
The purpose of this study was to evaluate the potential role of imaging in
recurrent ovarian cancer. Specifically, the aim was to describe CT findings in
recurrent ovarian cancer and to correlate CT findings with surgical outcome in
patients undergoing secondary cytoreduction.
Materials and Methods
Patients
In a study approved by our institutional review board, medical histories of
patients who were admitted with a diagnosis of recurrent ovarian cancer from
January 1, 1996, to June 1, 2002, were reviewed. Of 291 patients, 169
underwent surgery. Patients undergoing palliation for bowel obstruction
(n = 19) or surgery for treatment of other complications (n
= 28) were excluded. Of the 122 patients who were taken to surgery for
secondary cytoreduction and optimal debulking, 36 had preoperative CT scans
obtained at a single institution that were available for review. CT slice
thickness ranged from 5 to 10 mm. All 36 patients received an oral contrast
agent, and 32 of 36 patients received an IV contrast agent (150 mL).
A gynecologic oncologist reviewed preoperative and operative notes to
confirm the initial extent of optimal debulking and surgical outcome.
Preoperative CT was performed an average of 22 days before surgery (range,
2–69 days). Two radiologists who were unaware of the surgical outcome
retrospectively interpreted the CT scans. Imaging findings were evaluated by
guidelines described in studies for primary ovarian carcinoma
[1]. Tumor metastases were
recorded if present in intraperitoneal locations in the abdomen (peritoneal
carcinomatosis including omental, mesenteric, peritoneal, and serosal
deposits; perihepatic, perisplenic, gastrohepatic, and gastrosplenic
ligaments; and metastases in the gallbladder fossa, the falciform ligament,
the lesser sac, and the liver). Abdominal retroperitoneal enlarged nodes (>
1 cm in the short axis) were recorded separately for above and below the renal
hilum; and the presence of large- and small-bowel obstruction, hydronephrosis,
and ascites was also recorded. The pelvis was evaluated for masses, and we
recorded the size, definition, and extension to the bladder, bowel, vagina,
and pelvic sidewall as well as lymphadenopathy. Invasion of the pelvic
sidewall was diagnosed when tumor was inseparable from at least half of the
circumference of an iliac vessel or within 3 mm of muscle
[10]
(Fig. 1). CT findings were
correlated with surgical outcome. Cancer antigen–125 (CA-125) values
were recorded. All surgery reports were reviewed. Tumors were considered to be
optimally resected when postoperative residual tumor measured 1 cm or less in
greatest diameter.
View larger version (145K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1. —Axial contrast-enhanced CT scan in 53-year-old-woman with
recurrent ovarian cancer and pelvic sidewall invasion that was suboptimally
cytoreduced shows bilateral masses invading pelvic side walls (piriform
muscles) (arrows).
|
|
Statistical Methods
Using univariate logistic regression, the binary dependent variable of
optimal resectability is correlated to various binary covariates of interest
such as the presence of disease in pelvic nodes or the presence of ascites and
CA-125 in continuous form
[11]. These analyses were then
followed by a multivariate analysis restricted to the variables with
p values of less than 0.1. Two variables (pelvic mass and disease in
the paraaortic region) with null cell frequency were analyzed univariably
using Fisher's exact test [12]
but could not be included in the multivariate analysis because of a
convergence problem in regression analysis; the odds ratio and the 95%
confidence interval are reported for all variables except these two. The
distribution of CA-125 values is shown in a box-and-whisker plot
[13]
(Fig. 2) for the suboptimally
and optimally resected groups. All analyses were performed using SAS software
(version 8, SAS Institute, Cary, NC).
View larger version (8K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2. —Box-and-whisker plot shows cancer antigen–125 (CA-125)
values according to resectability status. Whiskers represent entire range and
box represents first and third quartiles. Line in middle of box represents
median, and plus sign represents mean.
|
|
Results
Patient Demographics
The average patient age was 55 years (range, 30–75 years). Of the 36
patients, 28 patients (78%) had stage III or IV disease at the time of primary
diagnosis. Three patients had stage II disease, two patients had stage IB
disease, and three patients did not have staging information available from
their initial presentation. At secondary cytoreductive surgery, tumors in 27
patients (75%) were optimally cytoreduced, and tumors in nine patients (25%)
had suboptimal cytoreduction.
Correlative Findings
Frequency of imaging findings and their correlation to resectability status
in terms of odds ratios and the corresponding p values are presented
in Figure 2 and
Table 1. Variables most
indicative of tumor nonresectability include the presence of a pelvic mass
extending to the adjacent tissue (p = 0.08), invasion of the pelvic
sidewall (odds ratio, 43.8; p = 0.0001), hydronephrosis (odds ratio,
25.0; p = 0.001), and the presence of disease in the paraaortic area
(p = 0.06). Hydronephrosis was a result of direct involvement of the
ureter by the pelvic mass in all cases. An odds ratio of 25.0 for the variable
hydronephrosis means that a patient having this feature is at 25 times greater
risk of having her tumor suboptimally resected than a patient who does not
have the feature.
In nine patients with no pelvic mass, all tumors were optimally resected.
Similarly, only two patients had disease in the paraaortic region, and both
tumors were suboptimally resected. This gives rise to null frequency in a 2
x 2 table, and multivariate regression including these variables is not
possible. Therefore, we fitted only the remaining two variables—invasion
of pelvic sidewall and hydronephrosis—multivariately and found them to
be statistically significant, with odds ratio of 35.6 (p = 0.006) and
19.4 (p = 0.03), respectively.
Large- or small-bowel obstruction; nodal or perihepatic liver metastasis;
ascites; peritoneal carcinomatosis; bladder, rectum, sigmoid colon, or vaginal
involvement; or infrarenal paraaortic adenopathy were not strong indicators of
tumor nonresectability. No significant difference was seen in the size of the
pelvic mass in either the maximum (p = 0.33) or the minimum
(p = 0.75) dimension between patients whose tumors were optimally
resected and those whose tumors were suboptimally resected.
Four (15%) of 27 patients whose tumors were optimally resected and five
(56%) of nine patients whose tumors were suboptimally resected had recurrent
tumor in the abdomen. Abdominal disease included carcinomatosis, perihepatic
metastases, and hepatic parenchymal metastases
(Fig. 3). In our patient
cohort, no disease was found in the gastrohepatic ligament, gastrosplenic
ligament, gallbladder fossa, falciform ligament, and lesser sac or perisplenic
location. Eighteen (67%) of 27 optimally cytoreduced tumors and nine (100%) of
nine suboptimally cytoreduced tumors were in patients who had a pelvic mass.
Pelvic mass contour was either sharply or ill defined. Five (28%) of 18
patients whose tumors were optimally cytoreduced and four (50%) of eight
patients whose tumors were suboptimally resected had an ill-defined pelvic
mass. Two patients with suprarenal lymphadenopathy, one patient with hepatic
parenchymal metastases, and two patients with portal adenopathy all had
nonresectable disease but were not subjected to any statistical analysis
because of their small number (Fig.
4A,
4B).
View larger version (120K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3. —Axial contrast-enhanced CT scan in 61-year-old woman with
recurrent ovarian cancer that was suboptimally cytoreduced shows multiple
intraabdominal tumor metastases (arrows) and mild ascites
(arrowheads).
|
|
View larger version (115K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4A. —67-year-old woman with recurrent ovarian cancer that was
suboptimally cytoreduced. Axial contrast-enhanced CT scans reveal calcified
tumor metastasis in porta hepatis (arrow, A) and calcified
interaortocaval lymph node (arrow, B).
|
|
View larger version (114K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4B. —67-year-old woman with recurrent ovarian cancer that was
suboptimally cytoreduced. Axial contrast-enhanced CT scans reveal calcified
tumor metastasis in porta hepatis (arrow, A) and calcified
interaortocaval lymph node (arrow, B).
|
|
CA-125 values were available in 33 patients (24 with optimal resections and
nine with suboptimal resections). The patients with optimal and suboptimal
resections had median CA-125 values of 39.5 U/mL (range, 5.0–464.0 U/mL)
and 74.0 U/mL (range, 4.0–1,659.0 U/mL), respectively, with no
significant difference (p = 0.6).
Discussion
Abundant knowledge exists about the surgical treatment of primary ovarian
cancer, with reported success rates of optimal cytoreduction varying from 42%
to 98% [1,
14,
15]. Much less has been
published about recurrent disease. The literature on this subject does support
prolonged survival after complete surgical debulking
[2,
16,
17] and optimal (
1–2 cm residual tumor) surgical debulking
[5,
8,
18]. Survival also correlates
with a disease-free interval before recurrence and residual disease after
primary surgery [5,
8,
19]. No survival benefit has
been shown for suboptimal reduction
[17]. Suboptimal debulking
after attempts at secondary cytoreduction has been reported at rates of
17–62%.
To spare patients unnecessary surgery, attempts at preoperative prediction
of resectability have been made. Chi et al.
[20] predicted suboptimal
results from primary cytoreduction with a sensitivity of 78% using CA-125
levels. CA-125 level as a predictor of patient outcome in recurrent ovarian
cancer has not been studied. In our small series, CA-125 level was not a
predictor of optimal cytoreduction.
Imaging studies to predict surgical outcome in primary ovarian cancer have
been performed. Nelson et al.
[21] predicted surgical
outcome for primary cytoreduction with a sensitivity of 92%, specificity of
79%, positive predictive value of 67%, and negative predictive value of 96%.
Suboptimal resection for primary cytoreduction was suspected on CT when
evidence was seen of hepatic parenchymal metastasis, or when significant
disease was seen in the upper abdomen, including the undersurface of the
diaphragm, the porta hepatis, the intersegmental fissure, and the root of the
small-bowel mesentery, lesser sac, and suprarenal paraaortic lymph nodes
[1,
9,
10,
21].
No data exist on the use of imaging in the pretreatment evaluation of
recurrent ovarian cancer. In our study, the significant indicators of tumor
nonresectability in recurrent ovarian cancer were the presence of
hydronephrosis and pelvic sidewall invasion. These radiologic findings may be
indicators of tumors that are encasing the iliac vessels, which would explain
their strong association with nonresectability. In the patient cohort studied,
no disease was present in the gastrohepatic ligament, gastrosplenic ligament,
gallbladder fossa, falciform ligament, and lesser sac or perisplenic
locations. This finding most likely reflects patient selection bias because
patients with diffuse upper abdominal disease on pretreatment CT may not have
been thought to be appropriate candidates for secondary cytoreduction.
In this study, most patients selected for secondary cytoreduction had the
bulk of their tumor burden in the pelvis. The size of a pelvic mass was not an
indicator of surgical outcome, but pelvic mass extension to the pelvic
sidewall was. Sacral or major neural involvement was noted before surgery on
CT, and these patients were not considered surgical candidates. Thus, our
cases were preselected, and only pelvic muscle involvement or vascular
involvement was seen with pelvic sidewall invasion. Although the findings of
hepatic metastases, portal adenopathy, and suprarenal lymphadenopathy resulted
in tumor nonresectability (Figs.
5 and
6A,
6B), our sample size was too
small for statistical analysis. Similarly, the presence of either small- or
large-bowel obstruction was not a strong indicator of nonresectability. In
this series, the number of patients with bowel obstructions was small because
the goal of surgery in patients with recurrent ovarian cancer and bowel
obstruction is generally palliation and not cytoreduction.
View larger version (90K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 5. —Axial contrast-enhanced CT scan in 46-year-old woman with
recurrent ovarian cancer that was optimally cytoreduced shows
well-circumscribed complex cystic mass (arrow) that is inseparable
from vagina, rectum, and bladder.
|
|
View larger version (98K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 6A. —55-year-old woman with recurrent ovarian cancer that was
suboptimally cytoreduced. Axial contrast-enhanced CT scan reveals infiltrative
mass with central necrosis (arrow) that is inseparable from sigmoid
colon, vagina, and bladder.
|
|
View larger version (110K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 6B. —55-year-old woman with recurrent ovarian cancer that was
suboptimally cytoreduced. Axial contrast-enhanced CT scan reveals left-sided
hydronephrosis (arrow) and hepatic metastasis in segment VI
(arrowhead).
|
|
The surgeon's assessment of a patient for optimal secondary cytoreduction
and outcome is a subjective process. None of our patients was considered for
pelvic exenteration. The two tumors of pelvic sidewall invasion that were
optimally cytoreduced required resection of the piriform and psoas muscles.
These muscle resections are considered radical procedures and are not
undertaken routinely (Fig. 7);
therefore, ultimate patient outcome depends on the style and aggressiveness of
the surgeon.
View larger version (140K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 7. —Axial contrast-enhanced CT scan in 49-year-old woman with
recurrent ovarian cancer and pelvic sidewall invasion that was optimally
cytoreduced shows heterogeneous mass (arrow) invading left pelvic
sidewall (< 3 mm from piriform muscle) (arrowhead). Surgery
required resection of piriform muscle and bone scraping.
|
|
Our study has limitations. Its retrospective design limits detailed
correlation of imaging and surgical findings. Because this study was designed
to evaluate patients with recurrent tumor who were taken to surgery for
attempted optimal secondary cytoreduction, a separate study assessing patient
outcome in the cohort not taken to surgery needs to be performed. Not all
patients presenting for surgery during the study period had in-house
preoperative CT, which increases the need for a prospective study.
In conclusion, in patients with recurrent ovarian carcinoma who are
considered for secondary cytoreductive surgery, preoperative CT identifies the
presence and extent of disease and points out potential areas of
nonresectability that can affect patient management.
References
- Bristow RE, Duska LR, Lambrou NC, et al. A model for predicting
surgical outcome in patients with advanced ovarian carcinoma using computed
tomography. Cancer2000; 89:1532
–1540[Medline]
- Eisenkop SM, Friedman RL, Spirtos NM. The role of secondary
cytoreductive surgery in the treatment of patients with recurrent epithelial
ovarian carcinoma. Cancer2000; 88:144
–153[Medline]
- Chi DS, Liao JB, Leon LF, et al. Identification of prognostic
factors in advanced epithelial ovarian carcinoma. Gynecol
Oncol 2001;82:532
–537[Medline]
- Sabbatini P, Spriggs D. Salvage therapy for ovarian cancer.
Oncology 1998;12:833
–843[Medline]
- Bristow RE, Lagasse LD, Karlan BY. Secondary surgical cytoreduction
for advanced epithelial ovarian cancer: patient selection and review of the
literature. Cancer1996; 78:2049
–2062[Medline]
- Williams L. The role of secondary cytoreductive surgery in
epithelial ovarian malignancies. Oncology1992; 6:25
–32
- Rose PG. Surgery for recurrent ovarian cancer. Semin
Oncol 2000;27[3 suppl 7]:17
–23[Medline]
- Segna RA, Dottino PR, Mandeli JP, Konsker K, Cohen CJ. Secondary
cytoreduction for ovarian cancer following cisplatin therapy. J
Clin Oncol 1993;11:434
–439[Abstract/Free Full Text]
- Meyer JI, Kennedy AW, Friedman R, Ayoub A, Zepp RC. Ovarian
carcinoma: value of CT in predicting success of debulking surgery.
AJR 1995;165:875
–878[Abstract/Free Full Text]
- Forstner R, Hricak H, Occhipinti KA, Powell CB, Frankel SD, Stern
JL. Ovarian cancer: staging with CT and MR imaging.
Radiology1995; 197:619
–626[Abstract/Free Full Text]
- Hosmer DW, Lemeshow S. Applied logistic
regression. New York, NY: John Wiley and Sons,1989
- Hollander M, Wolfe DA. Nonparametric statistical
methods. New York, NY: John Wiley and Sons,1973
- Cleveland WS. Visualizing data. Summit, NJ:
Hobart Press, 1993
- Curtin JP, Malik R, Venkatraman ES, Barakat RR, Hoskins WJ. Stage
IV ovarian cancer: impact of surgical debulking. Gynecol
Oncol 1997;64:9
–12[Medline]
- Eisenkop SM, Friedman RL, Wang HJ. Complete cytoreductive surgery
is feasible and maximizes survival in patients with advanced epithelial
ovarian cancer: a prospective study. Gynecol Oncol1998; 69:103
–108[Medline]
- Gadducci A, Iacconi P, Cosio S, et al. Complete salvage surgical
cytoreduction improves further survival of patients with late recurrent
ovarian cancer. Gynecol Oncol2000; 79:344
–349[Medline]
- Janicke F, Holscher M, Kuhn W, et al. Radical surgical procedure
improves survival time in patients with recurrent ovarian cancer.
Cancer 1992;70:2129
–2136[Medline]
- Zang RY, Zhang ZY, Li ZT, et al. Effect of cytoreductive surgery on
survival of patients with recurrent epithelial ovarian cancer. J
Surg Oncol 2000;75:24
–30[Medline]
- Zang RY, Zhang ZY, Li ZT, et al. Impact of secondary cytoreductive
surgery on survival of patients with advanced epithelial ovarian cancer.
Eur J Surg Oncol2000; 26:798
–804[Medline]
- Chi DS, Venkatraman ES, Masson V, Hoskins WJ. The ability of
preoperative serum CA-125 to predict optimal primary tumor cytoreduction in
stage III epithelial ovarian carcinoma. Gynecol Oncol2000; 77:227
–231[Medline]
- Nelson BE, Rosenfield AT, Schwartz PE. Preoperative abdominopelvic
computed tomographic prediction of optimal cytoreduction in epithelial ovarian
carcinoma. J Clin Oncol1993; 11:166
–172[Abstract]
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
Top
Abstract
Introduction
