Combined Use of Tissue Morphology, Neural Network Analysis of Chromatin Texture and Clinical Variables to Predict Prostate Cancer Aggressiveness from Biopsy Material
Abstract
We proposed to combine standard prognostic methods (clinical stage, PSA, Gleason score, and biopsy information) with Neural Network analysis of chromatin texture and computer derived tissue morphology prospectively to predict pathologic stage. We also intended to retrospectively investigate in prostatectomy specimens using a similar combination of clinical, histologic, and computer derived characteristics to predict disease recurrence following surgery. This resulting technology and nuclear analysis would then be applied to a study group of men with long term follow up after surgery to develop and validate this technology in predicting recurrence following surgery. Lastly, we intended to use this methodology to develop and validate an accurate model for predicting time to metastatic progression/death after biochemical recurrence. The scope of this project involved prospective enrollment of 500 men who were scheduled to undergo radical retropubic prostatectomy (year 01), development of an artificial neural network model (year 02), and prospective validation of this model (projected year 03). All models will be tested and developed for biopsy and prostatectomy material. To date, we have completed prospective enrollment of 557 men. Tissue, serum and clinical pathologic information have been collected on all 557 cases (100%). Of these cases, 409 (83%) have retained enough cancer material on the archival biopsy specimens for image analysis. Image analysis is complete on all 409 cases evaluated. A predictive model of organ-confined prostate cancer was defined in January 2001. This model utilized a combination of unique biomarkers to address the contemporary challenge of patients in this cohort that presented with clinical stage Tlc disease (69.4%). We are presently completing our evaluation of this technology on a cohort of 300 men (task 5) who have had biochemical recurrence following radical retropubic prostatectomy.
Document Details
- Document Type
- Technical Report
- Publication Date
- Apr 01, 2001
- Accession Number
- ADA398173
Entities
People
- Alan W. Partin
Organizations
- Johns Hopkins University