Radiogenomics Map-Based Molecular and Imaging Phenotypical Characterization in Localised Prostate Cancer Using...

Area of Science:

• Radiology
• Genomics
• Oncology

Background:

• Prostate cancer (PCa) diagnosis and grading rely on histopathology.
• Radiogenomics offers a non-invasive approach to link imaging phenotypes with molecular characteristics.

Purpose of the Study:

• To develop a radiogenomics map for localized prostate cancer.
• To assess correlations between molecular (gene expression) and imaging (radiomic) phenotypes.
• To evaluate the predictive accuracy of integrated radiogenomic models.

Main Methods:

• Radiomic features extracted from T2-weighted (T2WI) and Apparent Diffusion Coefficient (ADC) MRI images.
• Gene expression analysis (androgen receptor, apoptosis, hypoxia) from FFPE samples.
• Pearson's correlation and ROC analysis to assess relationships and predictive accuracy.

Main Results:

• Significant correlations found between radiomic texture features and gene copy number variations (CNV) for apoptosis, hypoxia, and androgen receptor (p ≤ 0.05).
• Specific radiomic features (e.g., Sum Entropy ADC, Entropy T2WI) showed potential for predicting cancer grade and biological processes.
• An integrated radiogenomic model achieved an Area Under the Curve (AUC) of 0.95 for predicting clinically significant prostate cancer.

Conclusions:

• Radiomic texture features are significantly correlated with genotypes related to apoptosis, hypoxia, and androgen receptor expression in localized PCa.
• Combining radiomics and genomics substantially enhances the prediction accuracy of clinically significant prostate cancer.
• This integrated approach holds promise for improved non-invasive prostate cancer grading.