• Sang Il Kim, Joo Hee Yoon, Sung Jong Lee, Min Jong Song, Jin Hwi Kim, Hae Nam Lee, Gyul Jung, and Ji Geun Yoo.
• Department of Obstetrics and Gynecology, St. Vincent's hospital, College of Medicine, The Catholic University of Korea, Seoul, Republic of Korea.
• Int J Med Sci. 2021 Jan 1; 18 (13): 2828-2834.

AbstractObjective: Predict the presence of lymphovascular space invasion (LVSI), using uterine factors such as tumor diameter (TD), grade, and depth of myometrial invasion (MMI). Develop a predictive model that could serve as a marker of LVSI in women with endometrial cancer (EC). Methods: Data from 888 patients with endometrioid EC who were treated between January 2009 and December 2018 were reviewed. The patients' data were retrieved from six institutions. We assessed the differences in the clinicopathological characteristics between patients with and without LVSI. We performed logistic regression analysis to determine which clinicopathological characteristics were the risk factors for positive LVSI status and to estimate the odds ratio (OR) for each covariate. Using the risk factors and OR identified through this process, we created a model that could predict LVSI and analyzed it further using receiver operating characteristic curve analysis. Results: In multivariate logistic regression analysis, tumor size (P = 0.027), percentage of MMI (P < 0.001), and presence of cervical stromal invasion (P = 0.002) were identified as the risk factors for LVSI. Based on the results of multivariate logistic regression analysis, we developed a simplified LVSI prediction model for clinical use. We defined the "LVSI index" as "TD×%MMI×tumor grade×cervical stromal involvement." The area under curve was 0.839 (95% CI= 0.809-0.869; sensitivity, 74.1%; specificity, 80.5%; negative predictive value, 47.3%; positive predictive value, 8.6%; P < 0.001), and the optimal cut-off value was 200. Conclusion: Using the modified risk index of LVSI, it is possible to predict the presence of LVSI in women with endometrioid endometrial cancer. Our prediction model may be an appropriate tool for integration into the clinical decision-making process when assessed either preoperatively or intraoperatively.© The author(s).

33 keywords...

hide…