Graduation Date

Spring 5-4-2024

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Programs

Medical Sciences Interdepartmental Area

First Advisor

Michael J Baine

Second Advisor

Chi Lin

Third Advisor

Christopher Deibert

Fourth Advisor

Jane Meza

Abstract

Multi-parametric magnetic resonance imaging (MP-MRI)-derived radiomics have been shown to capture sub-visual patterns for the quantitative characterization of prostate cancer (PC) phenotypes. The present dissertation seeks to develop, evaluate, and compare the performance of an MRI-derived radiomic model for the prediction of PC recurrence following definitive treatment with radical prostatectomy (RP).

MP-MRI was obtained from 339 patients who had a minimum of 2 years follow-up following RP at three institutions. The prostate was manually delineated as the region of interest and 924 radiomic features were extracted. All features were evaluated for stability via intraclass correlation coefficient (ICC) and image normalization was conducted via histogram matching.

Eighteen important and non-redundant features were found to be predictors of PC recurrence at a mean ± SD of 3.4±1.9 years and were aggregated into a radiomic model. Five-fold, ten-run cross-validation yielded a receiver-operator characteristic area under the curve (ROC-AUC) of 0.82±0.04 in the training set (n=290). In comparison, the University of California San Fransisco Cancer of the Prostate Risk Assessment score (UCSF-CAPRA) and Memorial Sloan Kettering Cancer Center (MSKCC) Pre-Radical Prostatectomy nomograms yielded AUC of 0.66±0.05 and 0.64±0.04, respectively (p<0.01). Finally, when the radiomic model was applied to the test set (n=49), ROC-AUC was 0.72 and sensitivity, specificity, positive predictive value, and negative predictive value were 88%, 56%, 28% and 96%, respectively. Finally, correlational analysis revealed multiple significant correlations between radiomic and clinical features, but high potential of multicollinearity between variables must be considered before causal relationships could be established.

Comments

2024 Copyright, the authors

Download

Share

COinS