# Non‐Invasive Assessment of Complete Regression in Endometrial Cancer Patients Undergoing Fertility Preservation Using MRI‐Based Radiomics and Immune Heterogeneity

## TL;DR

This study develops a non-invasive MRI-based model to predict treatment success in young women with early endometrial cancer, helping personalize fertility-preserving therapies.

## Contribution

A novel non-invasive radiomics model combining MRI and immune profiling to predict complete regression in fertility-preserving endometrial cancer treatment.

## Key findings

- A radiomics-clinical nomogram achieved high accuracy (AUC 0.963-0.986) in predicting complete regression.
- High radiomics score groups showed lower complete regression rates and an immunosuppressive tumor environment.
- Upregulated oxidative phosphorylation and lipid metabolism pathways were observed in high-score groups.

## Abstract

Fertility‐preserving treatment (FPT) offers a critical option for young women diagnosed with atypical endometrial hyperplasia (AEH) or early‐stage endometrial cancer (EC), however, the commonly used methods for evaluating complete regression (CR) are invasive. This study aimed to develop a non‐invasive tool to predict treatment outcomes using radiomics and molecular profiling. We retrospectively analyzed 146 patients with AEH or early EC receiving FPT. Radiomic features extracted from MRI were used to construct a radiomics signature predictive of CR through a machine‐learning approach. A radiomics‐clinical nomogram integrating radiomics scores with clinical variables demonstrated excellent predictive performance, with area under the curve values of 0.963 and 0.986 in the training and validation cohorts, respectively. Patients stratified into high‐ and low‐score groups based on radiomics scores showed significantly different CR rates, with the high‐score group exhibiting a lower likelihood of CR. Single‐cell RNA sequencing further confirmed immune alterations in the high‐score group, including reduced CD8+ T‐cells, and elevated levels of M2 macrophages. Bulk RNA sequencing revealed upregulation of oxidative phosphorylation and lipid metabolism pathways, suggesting a metabolically active and immunosuppressive tumor microenvironment. This radiomics‐based approach holds promise for guiding individualized FPT strategies for AEH and early EC patients.

Fertility preservation therapy allows young women with early EC to preserve fertility. We developed a non‐invasive MRI radiomics model to predict treatment success. It accurately identifies patients less likely to achieve remission, linking this to an immunosuppressive tumor environment. This aids personalized treatment planning.