A novel predictive model of lymphovascular space invasion in early-stage endometrial cancer

Area of Science:

• Gynecologic Oncology
• Pathology
• Medical Informatics

Background:

• Endometrial cancer (EC) staging relies on accurate assessment of prognostic factors.
• Lymphovascular space invasion (LVSI) is a critical indicator of recurrence risk in early-stage EC.
• Predictive models can improve risk stratification and treatment planning for EC patients.

Purpose of the Study:

• To develop and validate a predictive model for lymphovascular space invasion (LVSI) positivity in early-stage (stages 1-2) endometrial cancer (EC).
• To identify key clinicopathological prognostic factors significantly associated with LVSI in EC.

Main Methods:

• Retrospective analysis of 461 patients with early-stage EC treated with hysterectomy and lymphadenectomy.
• Histopathological examination of surgical specimens for LVSI status.
• Statistical analysis using the Loess algorithm and penalized maximum likelihood estimation to build and evaluate the predictive model, calculating the C-index.

Main Results:

• LVSI positivity was significantly associated with older age, menopause, type 2 EC, advanced histological grade, malignant peritoneal cytology, cervical involvement, and >50% myometrial invasion depth.
• The strongest predictors for LVSI were >50% myometrial invasion (OR: 3.78), advanced histological grade (OR: 1.98), and malignant peritoneal cytology (OR: 3.06).
• The developed model achieved 87% classification accuracy with a C-index of 0.876.

Conclusions:

• A robust predictive model incorporating key prognostic factors can effectively predict LVSI in early-stage EC.
• Myometrial invasion depth (>50%), advanced histological grade, and malignant peritoneal cytology are crucial determinants of LVSI risk.
• This model can aid clinicians in better predicting LVSI and tailoring management strategies for EC patients.