Predictors of treatment failure for adenocarcinoma in situ of the uterine cervix: Up to 14 years of recorded follow-up

The incidence of adenocarcinoma-in-situ (AIS) of the uterine cervix is rising, with invasive adenocarcinoma becoming increasingly common relative to squamous cell carcinoma. The present study reviewed a cohort of 84 patients first-time treated by conization for histologically-confirmed AIS from January 2001 to January 2017, to identify risk factors associated with recurrent/persistent AIS as well as progression to invasive cervical cancer. Nearly 80% of the patients were age 40 or younger at conization. Endocervical and ectocervical margins were deemed clear in 42 of the patients. All but two patients had ≥1 follow-up, with post-conization high-risk human papilloma virus (HPV) results documented in 52 patients. Altogether, 12 histopathologically-confirmed recurrences (14.3%) were detected; two of these patients had microinvasive or invasive carcinoma. In three other patients cytology showed AIS, but without recorded histopathology. Eight patients underwent hysterectomy for incomplete resection very soon after primary conization; they were not included in bivariate or multivariate analyses. Having ≥1 post-follow-up positive HPV finding yielded the highest sensitivity for histologically-confirmed recurrence: 87.5 [95% confidence interval (CI) 47.4-99.7]. Current or historical smoking status provided highest specificity: 94.4 (95% CI 72.7-99.9) and overall accuracy: 88.0 (95% CI 68.8-97.5) for histologically-confirmed recurrence. With multiple logistic regression (MLR), adjusting for age at conization and abnormal follow-up cytology, positive HPV18 was the strongest predictor of histologically-confirmed recurrence (P<0.005). Having ≥2 positive HPV results also predicted recurrence (P<0.02). Any unclear margin yielded an odds ratio 7.21 (95% CI 1.34-38.7) for histologically-confirmed recurrence adjusting for age, but became non-significant when including abnormal cytology in the MLR model. The strong predictive value of HPV, particularly HPV18 and persistent HPV positivity vis-à-vis detected recurrence indicated that regular HPV testing for patients treated for AIS is imperative. In conclusion, furthering a participatory approach, including attention to smoking with encouragement to attend needed long-term follow-up, can better protect these patients at high risk for cervical cancer.