TP53 Mutation Estimation Based on MRI Radiomics Analysis for Breast Cancer

Area of Science:

• Radiology and Oncological Imaging
• Biomarker Discovery
• Machine Learning in Medicine

Background:

• Noninvasive detection of TP53 mutations aids breast cancer molecular stratification.
• TP53 mutations are crucial in understanding breast cancer progression and treatment response.

Purpose of the Study:

• To identify MRI radiomics features indicative of TP53 mutations in breast cancer.
• To develop a classifier for noninvasively detecting TP53 mutations using MRI data.

Main Methods:

• Retrospective analysis of 139 breast cancer patients.
• Extraction of 944 radiomics and 7 clinicopathological features from T1-weighted DCE-MRI.
• Application of various machine learning classifiers including Logistic Regression (LR), Support Vector Machine (SVM), and Random Forest (RF).

Main Results:

• The radiomics model achieved a maximum Area Under the Curve (AUC) of 0.74 with LR.
• The combined clinicopathological-radiomics model demonstrated superior performance, with SVM reaching an AUC of 0.86.
• Random Forest (RF) showed high accuracy (AUCs 0.83 and 0.94) in triple-negative and luminal breast cancer subtypes.

Conclusions:

• A clinicopathological-radiomics combined model using SVM shows promise as a noninvasive biomarker for TP53 mutations.
• Random Forest (RF) is recommended for TP53 mutation detection in triple-negative and luminal breast cancers.