Influence of in Utero Diethylstilbestrol Exposure on the Prognosis and Biologic Behavior of Vaginal Clear-Cell Adenocarcinoma

Abstract

Most young women with clear-cell adenocarcinoma (CCA) of the vagina have a history of prenatal exposure to diethylstilbestrol (DES). Some, however, develop vaginal CCA without a prior history of DES exposure. We hypothesized that the natural history of DES-exposed vaginal CCA and DES-unexposed vaginal CCA may differ. Cases were identified from the Registry for Research on Hormonal Transplacental Carcinogenesis which maintains information on cases of CCA of the lower genital tract occurring in women born after 1940. Four hundred and thirty-one cases satisfied FIGO criteria for primary vaginal carcinoma of which 318 had prenatal, hospital, obstetrician, or pharmacy records available for review. Of these, 80% (255/318) had written documentation of prenatal exposure to DES (DES+) and 20% (63/318) had no evidence of DES exposure (DES-) in their medical records. DES exposure was undetermined in 113 cases due to lack of appropriate medical records. Among cases with documentation, DES exposure was not associated with mean age at diagnosis, (DES+, 20.3 years, DES-, 21.1 years), stage (stage I: DES+, 59%, DES-, 54%; stage II: DES+, 32%, DES-, 33%; stage III: DES+, 7%, DES-, 11%; stage IV: DES+, 2%, DES-, 2%, mean tumor diameter or surface area, grade, histology, cell type, or initial therapy. Among cases which underwent pelvic and paraaortic lymph node sampling (DES+, 63%; DES-, 56%; P = NS) the prevalence of pelvic node involvement was similar (DES+, 18.6%; DES-, 17.1%). However, only 1.2% (2/161) of DES+ cases had positive paraaortic lymph nodes compared to 8.6% (3/35) of DES- cases (P = 0.041). Survival differed significantly between the two groups. Probability of survival at 5 and 10 years for DES+ cases was 84 and 78%, respectively, compared to 69 and 60%, respectively, for DES- cases (5 years, P = 0.007, and 10 years, P = 0.008). Presently, 21% (53/255) of DES+ cases are known to have died, compared to 37% (23/63) of DES- cases (P = 0.008). Sites of disease recurrence also differed. DES- cases were more likely than DES+ cases to present with or to later develop distant tumor to the lungs (24% vs 9%; P = 0.002) or metastases to supraclavicular lymph nodes (8% vs 1.6%; P = 0.017). Among the 113 cases with an uncertain history of DES exposure, survival was intermediate between the well-documented cases (79% at 5 years and 65% at 10 years with 35/113 or 31% known dead), as was frequency of metastases to the lungs (13%) or supraclavicular lymph nodes (5.3%). Thus, the prognosis and metastatic behavior of CCA of the vagina appear to be influenced by DES exposure. DES negative cases have a worse prognosis and higher rate of distant metastasis than cases associated with exposure to DES. These observations do not appear to be due to differences in clinical prognostic factors such as tumor stage or diameter, but instead suggest differences in tumor behavior for as of yet undetermined reasons.