Predicting platinum resistance in primary advanced ovarian cancer patients with an in vitro resistance index

Abstract

We aimed to identify primary platinum resistance in epithelial ovarian cancer (OC) patients with FIGO stage III-IV disease by an in vitro drug-response assay and to correlate the findings with clinical response. We considered whether neoadjuvant chemotherapy or anatomic sample site and tumor heterogeneity would influence the results. We combined the ATP-based tumor-chemosensitivity and the extreme drug resistance assays for testing of 85 biopsies from 58 patients. Tumors were classified as sensitive or resistant by a resistance index (RI). We did separate analyses of primary tumors and metastases and compared chemo-naïve samples with samples obtained after neoadjuvant chemotherapy. Results were analyzed for association with clinical platinum resistance, progression-free survival (PFS), and overall survival (OS). RI · 250 predicted primary platinum resistance, without misclassification of sensitive patients. The test sensitivity for primary tumors was 15/15, specificity 3/10, negative predictive value 3/3, and positive predictive value 15/22. Patients with in vitro platinum-resistant samples had shorter PFS compared with patients with sensitive samples (3.4 vs. 10.0 months, p = 0.02). Comparing patient-matched primary and metastatic samples, there was about 1/3 mismatch in resistance. RI for platinum was lower in primary tumors exposed to neoadjuvant chemotherapy than in chemo-naïve tumors (p < 0.01). This in vitro assay predicted primary platinum resistance, without misclassification of sensitive OC patients, and the results were significantly associated with PFS. We suggest that samples from primary tumor and metastatic samples have different responses to chemotherapy and that exposure to chemotherapy might induce in vitro platinum resistance.