Sep 29

SS 22 - Radiation and Cancer Physics 2: Imaging Biomarkers for Response Monitoring

231 - Prediction of Metastasis-Free Survival in Patients with Prostate Adenocarcinoma Using Primary Tumor and Lymph Node Radiomics from Pre-Treatment PSMA PET/CT Scans

11:05am - 11:15am PT

Room 22/23

Presenter(s)

Apurva Singh, PhD - University of Maryland, Baltimore, MD

A. Singh¹, W. Mendes¹, S. Oh¹, O. C. C. Guler², A. Sawant³, P. T. Tran⁴, H. C. Onal⁵, and L. Ren⁴; ¹Department of Radiation Oncology, University of Maryland School of Medicine, Baltimore, MD, ²Baskent University Faculty of Medicine, Department of Radiation Oncology, Adana, Turkey, ³University of Maryland, School of Medicine, Baltimore, MD, ⁴University of Maryland School of Medicine, Baltimore, MD, ⁵Baskent University Faculty of Medicine, Department of Radiation Oncology, Ankara, Turkey

Purpose/Objective(s): To predict metastasis-free survival (MFS) for patients with prostate adenocarcinoma (PCa) treated with androgen deprivation therapy (ADT) and external radiotherapy using clinical factors and radiomics extracted from primary tumor and node volumes in pre-treatment PSMA PET/CT scans. We hypothesize that addition of information from node radiomics to primary tumor radiomics and prognostic clinical variables will result in an improvement in performance in the prediction of MFS.

Materials/Methods:

Our cohort includes 134 PCa patients (nodal involvement in 28 patients-with 6 having local metastasis). Gross tumor volumes of primary tumor (GTVp) and nodes (GTVn) on CT and PET scans were segmented. A 5mm expansion ring area outside primary tumor was defined. Z-score normalization was applied to radiomics features extracted from tumor and ring; dimensionality reduced using Principal Components Analysis (PCA). For patients with only primary tumor, we took 3 principal components (PCs) from GTVp and one ring PC as representative radiomics components from CT and PET scans. For patients with nodes, we calculated weighted average (by volume) of radiomics from primary tumor and nodes, computed first 3 PCs and combined it with 1^st PC from the ring. Radiomics PCs and clinical variables (age, Gleason score, initial prostate specific antigen value (i PSA), PSA_relapse) formed the predictors. Due to MFS data imbalance (metastasis-24, no metastasis-110), we performed 70:30 train-test split and applied imbalance correction to train data. Univariate Cox-regression was used to select top predictors (logistic regression p < 0.05). Multivariate Cox-regression was performed on imbalance-corrected train data and fit on test (using predictors selected from train). Model 2 was built using clinical variables and radiomic PCs from primary tumors (GTVp, ring). Model 3 was built using clinical variables only. Binary classification analysis for prediction of five-year MFS was also performed.

Results:

Results of time-to-event analysis (MFS) were: Cox-regression c-scores: model1: train- 0.77 [0.72, 0.78]; test- 0.69 [0.64, 0.70]; model2: train- 0.72 [0.66, 0.73]; test- 0.63 [0.58, 0.64]; model3: train- 0.62 [0.57, 0.63]; test- 0.54 [0.51, 0.56]. The results of 5 year MFS classification analysis were [sensitivity, specificity, AUC]: model 1: train- [81.4%, 89.1%, 0.87]; test- [75.1%, 81.4%, 0.79]; model 2: train- [76.2%, 83.4%, 0.82]; test- [70.2%, 77.3%, 0.74]; model 3: train- [68.5%, 76.2%, 0.75]; test- [63.2%, 70.4, 0.66]. Integration of node with primary tumor-radiomics provides the best prognostic performance in MFS prediction.

Conclusion:

This is one of the first studies to explore the prognostic value of pre-treatment PSMA-PET, a relatively recent advancement in the care of prostate adenocarcinoma patients. Results showed that using PSMA PET/CT radiomics information from primary tumor and nodes improves MFS prediction, compared to using primary tumor radiomics and clinical variables only.