Purpose: To validate the prognostic value of the endometrial carcinoma prognostic index (ECPI; combined myometrium invasion, flow cytometric DNA ploidy, and morphometric mean shortest nuclear axis [MSNA]) versus classic prognosticators.
Patients and methods: Prospective multicenter ECPI analysis was conducted in 463 endometrial carcinomas with a median of 6.5 years (range, 1 to 10 years) follow-up, review of pathology features, and univariate (Kaplan-Meier) and multivariate (Cox) analyses.
Results: Initial routine and review diagnoses varied considerably (invasion depth, 11%; type, 20%; grade, 34%; vessel invasion, 72%); the review diagnoses were stronger prognostically. In International Federation of Gynecology and Obstetrics stage 1 (after histopathologic examination; pFIGO-1; n = 372; 38 deaths occurred as a result of disease [10.2%]), DNA ploidy was prognostic in hysterectomies (P <.00001) but not in curettages (P =.06). ECPI was a stronger prognostic indicator than other features. ECPI, MSNA, and DNA ploidy were also prognostic in pFIGO-1B and -1C subgroups. Multivariate analysis in pFIGO-1 showed that uterine MSNA < or = versus > 7.93 microm (hazard ratio [HR], 3.4) and grade (as 1 + 2 v 3; HR, 2.6) added to the ECPI (HR, 32), but only in patients with an unfavorable ECPI of > 0.87. Adjuvant radiotherapy was not an independent prognostic factor in any of the subgroups. In pFIGO-2 (n = 46), ECPI, DNA-ploidy, and age (< or = 64, > 64 years) were significant. In FIGO-3 (n = 31) and FIGO-4 (n = 14), none of the classic or other features analyzed was of prognostic value, which explains why in previous studies using different mixtures of FIGO stages, DNA ploidy prognostic results varied.
Conclusion: In endometrial carcinoma, DNA-ploidy is prognostic in hysterectomy and not in curettage samples. The ECPI is prognostically much stronger than the classic features widely used for therapy triage in pFIGO-1 and -2.