Habitat Radiomics Based on Diffusion-weighted MRI for Predicting PARP Inhibitor Resistance in Advanced High-grade...

Area of Science:

• Radiomics
• Medical Imaging
• Oncology

Background:

• Ovarian cancer is a leading cause of gynecologic cancer deaths.
• PARPi resistance is a significant challenge in treating advanced high-grade serous ovarian cancer (HGSOC).
• Predictive biomarkers for PARPi resistance are crucial for personalized treatment strategies.

Purpose of the Study:

• To develop a habitat radiomics model using diffusion-weighted MRI (DW-MRI) to predict PARPi resistance in advanced HGSOC.
• To evaluate the performance of the radiomics model in predicting progression-free survival (PFS).

Main Methods:

• Retrospective study of 157 advanced HGSOC patients from two centers.
• Gaussian mixture model for habitat identification and feature extraction (851 features).
• Lasso-Cox regression for feature selection; construction of clinical, habitat, and combined models to predict PFS.

Main Results:

• Habitat radscore independently predicted PARPi resistance in advanced HGSOC.
• The combined model (radiomics + clinical factors) demonstrated superior performance compared to the clinical model alone.
• The combined model achieved AUCs of 0.763 and 0.749 for predicting 18- and 24-month disease progression, respectively.

Conclusions:

• A combined model integrating habitat radiomics and clinical factors can predict PARPi resistance in advanced HGSOC.
• This model may aid in selecting patients for first-line PARPi maintenance therapy.
• The findings support the use of DW-MRI radiomics for personalized treatment of HGSOC.