Microenvironment at a Distance: Multi-Endocrine-Organ Radiomics to Identify Systemic Signatures in PSMA-Negative...

Area of Science:

• Radiology
• Oncology
• Medical Imaging

Background:

• Prostate cancer (PCa) is a leading cause of cancer mortality in men.
• Predicting clinical progression in PSMA-negative recurrent PCa is challenging.
• Radiomic features from normal endocrine organs may offer novel biomarkers.

Purpose of the Study:

• To assess if radiomic features from endocrine organs, combined with clinical data, can predict clinical progression in PSMA-negative prostate cancer.
• To evaluate the predictive performance of multimodal fusion models.
• To determine the incremental value of endocrine organ radiomics over clinical variables.

Main Methods:

• Retrospective analysis of 101 men with biochemically recurrent, PSMA-negative PCa.
• Extraction of radiomic features from adrenal glands, thyroid, hypothalamus-pituitary complex, and testes.
• Development and evaluation of predictive models using stratified train/test splits and AUC analysis.

Main Results:

• Multimodal models integrating CT/PET radiomics and clinical variables showed the highest performance (AUC up to 0.758).
• Clinical-only models were competitive (AUC up to 0.727).
• Endocrine organ radiomics numerically improved specificity but did not yield statistically significant improvements over clinical variables alone.

Conclusions:

• Endocrine organ radiomics may offer complementary biomarkers for PSMA-negative PCa, reflecting tumor-host interactions.
• The incremental clinical value of endocrine organ radiomics over clinical variables alone was modest and not statistically significant.
• Further validation in larger cohorts is required to confirm the utility of these imaging biomarkers.