Cell proliferation in ovarian carcinoma: superior accuracy of S-phase fraction (SPF) by DNA labeling index versus flow cytometric SPF, lack of independent prognostic power for SPF and DNA ploidy, and limited effect of SPF on tumor growth rate

Abstract

Objectives: The goal of this work was to test the hypotheses that S-phase fraction (SPF) by DNA labeling index (SPF-LI) would predict the course of the disease for ovarian/peritoneal carcinomas and that SPF-LI would correlate better with pathologic classification and outcome than SPF by DNA flow cytometry (SPF-F).

Methods: Tritiated thymidine (1985-1988) and bromodeoxyuridine (1988-1999) DNA labeling (SPF-LI) was evaluated in vitro on 178 tumors. Cellular DNA and SPF-F were measured flow cytometrically. During this time, 90% of ovarian/peritoneal tumors accessioned in surgical pathology were studied.

Results: Tumors of low malignant potential (LMP, "borderline") had low SPF-LI (median = 1.2%). High-grade invasive carcinomas of various types and carcinosarcomas all had high SPF-LI (medians = 11.2-23.4%). Serous low-grade invasive carcinomas (median = 1.05) resembled LMP tumors. SPF-LI of ovarian carcinomas other than LMP tumors increased slightly as FIGO stage increased (P = 0.07). Survival of patients with high-grade ovarian carcinomas was not predicted by SPF-LI or SPF-F, nor was DNA ploidy predictive. SPF-LI produced tighter distributions for various tumor types than did SPF-F. Neither SPF nor DNA ploidy contributed to prediction of outcome when tumor type and stage were included in multivariate models. We calculated the mean cell loss rate of high-grade carcinomas to be 94%.

Conclusions: LMP ovarian/peritoneal tumors have low proliferation rates in contrast to high-grade carcinomas. Proliferation correlated with tumor type and stage, but neither it nor DNA ploidy predicted survival independently. Proliferation rate is growth limiting only when low. At higher levels cell loss limits growth. SPF-LI measures proliferation more accurately than SPF-F; SPF-F is not sufficiently reliable for clinical use.