A model for predicting surgical outcome in patients with advanced ovarian carcinoma using computed tomography

BACKGROUND. A reliable model for predicting the outcome of primary cytoredu ctive surgery may be a useful tool in the clinical management of patients w ith advanced epithelial ovarian carcinoma. METHODS. Forty-one women with a preoperative computed tomographic (CT) scan of the abdomen and pelvis and a histologic diagnosis of Stage III or IV ep ithelial ovarian carcinoma following primary surgery performed by one of ni ne gynecologic oncologists were identified from tumor registry databases. A ll CT scans were analyzed retrospectively using a panel of 25 radiographic features without knowledge of the operative findings. Patient demographics, surgical findings and outcome, Gynecologic Oncology Group performance stat us, and pre-operative serum CA125 values were collected from patient medica l records. Residual disease measuring less than or equal to 1 cm in maximal diameter was considered an optimal surgical result. Sensitivity, specifici ty, positive predictive value (PPV), and negative predictive value (NPV) we re calculated for each radiographic feature and clinical characteristic. Ba sed on statistical probability of each factor predicting cytoreductive outc ome, 13 radiographic features, in addition to performance status, were sele cted for inclusion in the final model. Each parameter was assigned a numeri c value based on the strength of statistical association, and a total Predi ctive Index score was tabulated for each patient. Receiver operating charac teristic (ROC) curve analysis was used to assess the ability of the model t o predict surgical outcome. Statistical significance was evaluated using th e Fisher exact test. RESULTS. Twenty of 41 patients (48.8%) underwent optimal cytoreduction to l ess than or equal to 1 cm residual disease. CT features of peritoneal thick ening, peritoneal implants (greater than or equal to 2 cm), bowel mesentery involvement (greater than or equal to 2 cm), suprarenal paraaortic lymph n odes (greater than or equal to 1 cm), omental extension (spleen, stomach, o r lesser sac), and pelvic sidewall involvement and/or hydroureter were most strongly associated with surgical outcome. Using the Predictive Index scor es, a receiver operating characteristic curve was generated with an area un der the curve = 0.969 +/- 0.023. In the final model, a Predictive Index sco re greater than or equal to 4 had the highest overall accuracy at 92.7% and identified patients undergoing suboptimal surgery with a sensitivity of 10 0% (21/21). The specificity, or ability to identify patients undergoing opt imal surgery, was 85.0% (17/20). The PPV of a Predictive Index score greate r than or equal to 4 was 87.5% (21/24), and the NPV was 100%. The ability o f this model to correctly predict surgical outcome was statistically signif icant (P < 0.001). CONCLUSIONS. In this model, a Predictive Index score greater than or equal to 4 demonstrated high sensitivity, specificity, PPV, and NPV, and was high ly accurate in identifying patients with advanced epithelial ovarian carcin oma unlikely to undergo optimal primary cytoreductive surgery. The Predicti ve Index model may have clinical utility in guiding the management of patie nts with ovarian carcinoma. Cancer 2000;89: 1532-40. (C) 2000 American Canc er Society.