Accurate laboratory predictions of the clinical response of patients with advanced ovarian cancer to treatment with cyclophosphamide, doxorubicin, and cisplatin.

Seventy-six patients with advanced ovarian cancer treated with cyclophosphamide, doxorubicin, and cisplatin (CAP) at 3-week intervals were tested for the response of their tumors to treatment with CAP in the subrenal capsule tumor implant assay. Thirty-four of the patients' tumors were assayed prospectively before clinical treatment and 33 were assayed retrospectively, after clinical treatment with CAP. Nine of the patients' tumors were assayed both prospectively and retrospectively. All of the patients underwent a tumor debulking laparotomy. Of the patients with clinically measurable residual disease, 17 had a partial response of at least 50% regression of disease, and 11 had a progression of disease. Of the patients with known residual but nonmeasureable disease, 7 had surgically verified complete responses, 8 at least 50% regression, and 23 had progression of disease: 10 had no evidence of disease clinically but had not had surgical confirmation. Twenty-six of the tumors were adenocarcinomas not otherwise specified (2 grade I, 2 grade II, and 22 grade III), 39 were serous adenocarcinomas (7 grade I, 9 grade II, and 23 grade III), 7 were endometrioid adenocarcinoma (all grade III), 3 were mucinous adenocarcinomas (1 each of grade I, II, and III) and 1 was an adenosquamous carcinoma (grade III). Thirty-four of the patients failed the therapy. The subrenal capsule (SRC) assay predicted 21 of these failures (4 prospective and 17 retrospective). Thirty-two of the patients responded to CAP chemotherapy. The SRC assay accurately predicted the clinical regression of the tumors of 22 of the patients (15 prospective and 7 retrospective). Second-look laparotomy confirmed 7 patients with no evidence of disease, 5 patients with minimal disease, and 5 patients with a greater than 50% reduction of their disease. The SRC assay predicted the response of all these patients except 2 with partial responses to chemotherapy. Thus, while the overall positive predictive value of the SRC assay in this study is 65%, it is 100% for those patients whose tumors respond completely and for those who have minimal residual disease after CAP chemotherapy.