251P - A prognostic index model for predicting long-term recurrence of uterine leiomyoma after initial myomectomy in women aged 18-44 years

Abstract

BackgroundUterine leiomyoma (UL) is the most common female pelvic benign tumor with recurrence rates of 16.7%-52.8%. We aimed to propose a prognostic index (PI) model for predicting the long-term (>5 years) risk of UL recurrence. MethodsWomen aged 18–44 years who initially underwent myomectomy for UL at one of the three study hospitals between April 2012 and October 2013 were enrolled. The data collected included patients’ demographics, leiomyomas’ characteristics, relapses, subsequent contraceptive methods, and postoperative gravidity and parity. A PI model was proposed based on the β-coefficients in the results of multivariate Cox regression analysis in prediction model group. The differences among the PI-based risk groups were tested by Kaplan–Meier analysis (using paired log-rank test) and univariate Cox regression analysis (using the Forward: LR) in prediction model, internal validation and external validation group. ResultsThere were 725 patients included in this study. PI formula =1.5(if 3-5 leiomyomas)+2(if > 5 leiomyomas)+1(if residue)+1(if not submucosal)+1(if combined endometriosis). The PI value (0-5) was divided into low-risk group, intermediate-risk group, and high-risk group by cut-off values 1.25 and 3.75. In the prediction model group, the high-risk group had a significantly 4.55 times greater recurrence risk of UL than that in the low-risk group [cumulative recurrence rate (CR): 82.1% vs 29.5%, HR = 4.55, 95% CI; 2.821-7.339]; the intermediate-risk group had a significantly 2.81 times greater recurrence risk of UL than that in the low-risk group (CR: 62.3% vs 29.5%, HR = 2.81, 95% CI; 2.035-3.878). The differences between any two risk groups were also statistically significant (P < 0.05) in both internal and external validation groups. ConclusionsThe PI model was proved to be effective in distinguishing low-risk, intermediate-risk, and high-risk groups for long-term recurrence of UL after initial myomectomy in women aged 18-44 years, allowing it to be an objective tool aid in clinical decision making. In the future, prospective researches should be carried out to confirm the predictive ability of this PI model, and its practical value in clinical decision making. Legal entity responsible for the studyZheng Yu Li. FundingHas not received any funding. DisclosureAll authors have declared no conflicts of interest.