Objective:

To investigate whether molecular information from primary tumor regions can identify prostate cancer likely to spread, potentially reducing unnecessary lymph node testing.

Approach:

• Study Design: Analyzed tissue from 51 prostate cancer patients using spatial transcriptomics to assess gene activity in primary tumor regions and matched lymph node metastases.
• Machine Learning: Trained machine learning models on molecular profiles to recognize patterns linked to lymph node involvement.

Key Findings:

• Primary tumor regions associated with lymph node metastases exhibited a distinct molecular profile.
• Machine learning models demonstrated promising performance in identifying patients at higher risk of lymph node metastasis.

Limitations:

• The work is exploratory and requires further validation before clinical application.
• The study does not yet incorporate morphology-based machine learning approaches.

Sources:

• ECP 2026 Presentation