Laparoscopic predictability of minimally invasive interval debulking in advanced ovarian cancer: the MIID-SOC trial

Abstract

<h3>Objectives:</h3> We sought to create a laparoscopic-based model to predict the ability to perform a minimally invasive (MIS) cytoreductive surgery in advanced epithelial ovarian cancer patients who have received neoadjuvant chemotherapy (NACT). <h3>Methods:</h3> A total of 50 women with at least a partial response by RECIST 1.1 criteria to NACT were enrolled in a multi-institutional prospective pilot study (<i>MIID-SOC trial- NCT03378128</i>). Each patient underwent laparoscopic evaluation of 43 abdominopelvic sites followed by primary surgeon dictated surgical approach, either continue laparoscopically (MIS) or laparotomically. However, if the procedure was to continued MIS, the placement of a hand-assist port for manual palpation was mandated as to emulate a laparotomic approach and all 43 sites were re-evaluated. A total of 2 patients could not be evaluated by laparoscopy because of dense adhesions, 2 patients did not undergo cytoreductive surgery as the laparoscopy deemed the patient unresectable, and 1 patient withdrew consent prior to surgery. Sensitivity, specificity, positive predictive value, negative predictive value, and overall accuracy were calculated for each site to predict resectability via the MIS approach. Based on statistical probability of each factor predicting cytoreductive approach, 10 abdominopelvic sites were selected for inclusion in the final model. Each parameter was assigned a numeric value based on the strength of statistical association and a total predictive index score (PIV) was assigned for each patient. Receiver operating characteristic curve analysis (ROC-AUC) was used to assess the ability of the model to predict the MIS surgical approach. Statistical significance was evaluated using Fisher's exact test. <h3>Results:</h3> A total of 28 patients (61%) underwent MIS cytoreductive surgery. All patients had an optimal cytoreductive surgery (<1 cm residual disease) regardless of approach. The presence of disease on the following abdominopelvic sites were most strongly associated with predicting an MIS surgical approach: gastrosplenic ligament, rectum, left mesocolon, transverse colon, right colon, cecum, appendix, liver capsule, intrahepatic fossa/gallbladder, ileum/jejunum. Using the PIV, a ROC was generated with an AUC=0.699. In the final model, a PIV <2 identified patients able to undergo an optimal MIS cytoreductive surgery with an accuracy of 68.9%. The specificity, or ability to identify patients who would be able to undergo an optimal MIS interval cytoreductive surgery was 70.6%. <h3>Conclusions:</h3> In this model, a PIV of <2 was able to identify patients who were likely to undergo an optimal MIS interval cytoreductive surgery. This predictive index model may help to guide future inclusion criteria in randomized studies evaluating the MIS approach in advanced epithelial ovarian cancer.