Clinical Report: Combined Approach Predicts Metastasis in Prostate Cancer

Overview

A study presented at ECP 2026 explores the use of molecular information from primary tumor regions to predict lymph node metastasis in prostate cancer. The findings indicate that machine learning models trained on spatial transcriptomics data may help identify patients at higher risk for metastasis.

Background

Prostate cancer is a prevalent malignancy in men, with a significant number of patients undergoing pelvic lymph node dissection to assess metastatic spread. However, not all patients benefit from this procedure, highlighting the need for improved predictive methods. Understanding the molecular characteristics of primary tumors could enhance risk stratification and patient management.

Data Highlights

The study analyzed tissue from 51 prostate cancer patients, comparing primary tumor regions with matched lymph node metastases to identify molecular profiles associated with metastasis.

Key Findings

• Primary tumor regions linked to lymph node metastases exhibited distinct molecular profiles.
• Machine learning models demonstrated promising performance in recognizing patterns associated with lymph node involvement.
• The study utilized spatial transcriptomics to maintain tissue context while analyzing gene activity.
• Further validation is needed before clinical application of the findings.
• Combining molecular data with morphological assessments may enhance predictive accuracy.

Clinical Implications

The research indicates that molecular profiling of primary tumors could refine the selection of patients for lymph node dissection.

Conclusion

This exploratory study highlights the potential of molecular information in predicting lymph node metastasis in prostate cancer.

Related Resources & Content

1. Wasinger G, ECP 2026 -- Combined Approach Predicts Metastasis in Prostate Cancer
2. Frontiers in Oncology, 2026 -- A multivariable prediction model combining 18F-PSMA PET/CT and mpMRI for clinically significant prostate cancer: development and validation
3. npj Digital Medicine, 2025 -- Combining Single-Cell and Spatial Transcriptomics with Explainable AI Uncovers Critical Prognostic Factors in Prostate Cancer
4. npj Digital Medicine, 2026 -- A new multi-omics machine learning approach reveals rapid progression in prostate cancer classified as clinically low risk.
5. the pathologist — Can Digital Pathology Improve Risk Stratification After Prostatectomy?
6. EAU-EANM-ESTRO-ESUR-ISUP Guidelines on Prostate Cancer 2026
7. Pelvic Lymph Node Dissection in Prostate Cancer: Update from a Randomized Clinical Trial
8. Guideline of guidelines: pelvic lymph node dissection at time of radical prostatectomy
9. Impact of Genomic Classifiers on Risk Stratification and Treatment Intensity in Patients With Localized Prostate Cancer : A Systematic Review.
10. Decipher
11. Validation of a Digital Pathology–Based Multimodal Artificial Intelligence Biomarker in a Prospective, Real-World Prostate Cancer Cohort Treated with Prostatectomy - PMC
12. AI-based prediction of molecular aberrations in prostate cancer using digital pathology: A systematic review.
13. 491MO Combinatorial multimodal AI (MMAI) pathology and decipher genomic classifier (GC) enhance NCCN prognostication in localized prostate cancer (PCa) - Annals of Oncology
14. Can pelvic lymph node dissection be spared in intermediate-risk prostate cancer patients with negative PSMA PET scan? A systematic review and diagnostic meta-analysis - PubMed
15. Prognostic value of cribriform pattern and intraductal carcinoma of the prostate after radical prostatectomy: A systematic review and meta-analysis using contemporary consensus recommendations - PubMed