Flow cytometric evaluation of early invasive cervical cancer

Abstract

Objective: To define the role of flow cytometry as a prognostic indicator in early cancers of the uterine cervix.

Methods: Flow cytometry was used to determine ploidy, DNA index, and S-phase fraction on 141 samples from the tumors of 53 women with stage IB cancers of the cervix treated by radical hysterectomy and pelvic lymphadenectomy as primary therapy. Multiple samples of the same tumor were analyzable for 47 (89%) of the subjects. One-way analysis of variance for the multiple samples was used to compare the heterogeneity of flow cytometry data, both within each tumor and between patients. Flow cytometry results, as well as previously described clinical and pathologic prognostic factors, were correlated to recurrence and survival using Cox regression hazard ratios and Kaplan-Meier estimates.

Results: We found DNA aneuploidy in 25 (47%) of the cancers, with a mean (+/- standard error) DNA index of 1.52 +/- 0.07. The mean S-phase fraction was 7.6 +/- 0.4% for diploid tumors and 9.2 +/- 0.4% for aneuploid tumors. The cancers from 24 women (45%) were homogeneously diploid, 18 (34%) were consistently aneuploid, and five (9%) had mixed diploid/aneuploid samples. Analysis of variance of the multiple samples for each woman revealed a greater standard deviation (SD) between patients than within any individual tumor for both DNA index (ratio of between SD to within SD 2.1; P < .0001) and S-phase fraction (ratio 1.6; P < .0001). Of the previously described clinical and pathologic prognostic factors, only depth of invasion, expressed as either percent of cervical wall thickness or as thirds, was correlated with recurrence or survival. Neither the DNA index nor S-phase fraction correlated significantly with recurrence or survival.

Conclusions: These results suggest that alterations in the DNA content or proliferative activity of early invasive cancers of the uterine cervix do not reflect biologic behavior in terms of recurrence or survival, and that this behavior is not due to tumor heterogeneity.