Abstract LB-307: Computational modeling of serous ovarian carcinoma dynamics: Implications for screening and therapy

Abstract

Abstract High-grade serous ovarian carcinoma (HG-SOC) is the most common subtype of ovarian cancer and has the worst prognosis. There is intense controversy on whether the temporal order of cytoreductive surgery and chemotherapy affects treatment outcome, with the main options being primary debulking surgery with adjuvant chemotherapy (PDS) versus neo-adjuvant chemotherapy with interval debulking surgery (NACT). Although some studies report that PDS-treated patients survive significantly longer than those receiving NACT, other reports showed no significant difference in patient outcome. To address this question in an unbiased way, we used computational modeling to simulate HG-SOC progression dynamics with different treatments. We adapted a mathematical framework to predict the evolution of chemotherapy resistance, and populated our model with survival data from >300 patients receiving PDS or NACT. After estimating the mutation rate of carcinoma cells and cancer cell number at diagnosis, we determined that most HG-SOC patients likely harbor chemotherapy-resistant cancer cells at diagnosis. Furthermore, we predicted the effects of PDS and NACT on the number of sensitive and resistant cells, as well as patient survival following treatment, and found that our model closely recapitulated clinical observations in both training and test sets. Based on our results, we predict that PDS with optimal debulking (<1mm residual tumor) has the potential to be curative because surgery can sometimes remove all chemo-resistant cells, while adjuvant chemotherapy depletes the remaining chemo-sensitive cells. By contrast, NACT is unlikely to cure the disease because neo-adjuvant chemotherapy depletes chemo-sensitive cells that can mark the location of accompanying “passenger” chemo-resistant cells. As a result, it leads to extensive enrichment of chemo-resistant cells before surgery, but the deposits of such cells are typically too small to be visualized at eventual surgery. Our model also predicts that PDS should have a better outcome than NACT, when controlled for residual tumor size. Finally, we evaluated the potential benefits of early diagnosis of naïve or relapsed HG-SOC. We recapitulated the clinical finding that CA125-based earlier diagnosis of relapsed cancer does not improve survival. We also predict that more sensitive detection methods (such as ctDNA-based diagnosis) are unlikely to improve survival post-relapse with current chemotherapy, because earlier diagnosis does not decrease the number of chemo-resistant cells, which are already enriched at recurrence. By contrast, our model predicts that with sufficiently sensitive assays, early detection could improve survival time and increase chances of cure. Note: This abstract was not presented at the meeting. Citation Format: Shengqing Gu, Liat Hogen, Stephanie Lheureux, Azin Sayad, Iryna Vyarvelska, Paulina Cybulska, Marcus Bernardini, Barry Rosen, Amit Oza, Benjamin G. Neel. Computational modeling of serous ovarian carcinoma dynamics: Implications for screening and therapy. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr LB-307. doi:10.1158/1538-7445.AM2015-LB-307