Orysya Svystun
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Orysya Svystun has a total of 8 co-authors. Here are the top co-authors by number of shared publications:
Atul B. Shinagare
Division of Abdominal Imaging and Intervention, Department of Radiology Brigham and Women's Hospital Boston Massachusetts USA
1 shared publication- Jessie L. Chai
Division of Abdominal Imaging and Intervention, Department of Radiology Brigham and Women's Hospital Boston Massachusetts USA
1 shared publication Matthew Mossanen
Department of Urology Brigham and Women's Hospital Boston Massachusetts USA
1 shared publication- Michael D. Weintraub
Division of Abdominal Imaging and Intervention, Department of Radiology Brigham and Women's Hospital Boston Massachusetts USA
1 shared publication - Michelle S. Hirsch
Brigham and Women’s Hospital, Harvard Medical School
1 shared publication - Mohd Javed Saifullah Shaikh
Department of Radiology, Brigham and Women’s Hospital, Boston, Mass
1 shared publication
Latest Publications
- 14 October 2026
- Published by Wiley in Journal of Medical Imaging and Radiation Oncology
MR Imaging and Radiomics Features for Predicting Variant Histology of Bladder Cancer: A Single Centre Retrospective Analysis
Introduction: Compare MRI features of bladder variant histology (VH) and pure urothelial carcinoma (UC) and determine if MRI radiomics features can predict VH. Methods: In this retrospective study, cystectomy pathology database search yielded 37 patients with tumours with more than 50% VH and preoperative MRI's. Cystectomy database search yielded 38 consecutive patients with pure UC and preoperative MRI's. Three radiologists characterized MRI T2, diffusion‐weighted, apparent diffusion coefficient (ADC) maps and postcontrast features of bladder tumours. Inter‐reader agreement was assessed using Gwet's agreement and intra‐class coefficients. Another radiologist segmented bladder tumours and adjacent perivesical fat on T2‐weighted and ADC images using radiomics software. Least absolute shrinkage and selection operator (LASSO) was used to determine radiomics features that distinguished VH from UC. Fifty datasets were prepared using multiple imputation approach, and variables were normalized for LASSO variable selection. Results: Patients with VH (mean age = 68, 54 (72%) men) included squamous cell ( n = 17), sarcomatoid ( n = 6), plasmacytoid ( n = 6), adenocarcinoma ( n = 4) and small cell ( n = 4) variants. Inter‐reader agreement was similar for VH and UC ( p > 0.44). All readers characterized all squamous cell carcinoma as T2 intermediate. For two of three readers, 92%–100% of hypoenhancing tumours were VH at pathology whereas 74%–78% of hyperenhancing masses were UC ( p = 0.001). Combination of 11 selected radiomics features had AUC of 0.98 for training and 0.95 for validation data for prediction of VH. Conclusion: Squamous cell carcinoma was T2 intermediate, and VH was more likely to hypoenhance compared to UC. Our radiomic model had excellent performance for predicting VH.
- 14 October 2026
- Published by Wiley in Journal of Medical Imaging and Radiation Oncology
MR Imaging and Radiomics Features for Predicting Variant Histology of Bladder Cancer: A Single Centre Retrospective Analysis
Introduction: Compare MRI features of bladder variant histology (VH) and pure urothelial carcinoma (UC) and determine if MRI radiomics features can predict VH. Methods: In this retrospective study, cystectomy pathology database search yielded 37 patients with tumours with more than 50% VH and preoperative MRI's. Cystectomy database search yielded 38 consecutive patients with pure UC and preoperative MRI's. Three radiologists characterized MRI T2, diffusion‐weighted, apparent diffusion coefficient (ADC) maps and postcontrast features of bladder tumours. Inter‐reader agreement was assessed using Gwet's agreement and intra‐class coefficients. Another radiologist segmented bladder tumours and adjacent perivesical fat on T2‐weighted and ADC images using radiomics software. Least absolute shrinkage and selection operator (LASSO) was used to determine radiomics features that distinguished VH from UC. Fifty datasets were prepared using multiple imputation approach, and variables were normalized for LASSO variable selection. Results: Patients with VH (mean age = 68, 54 (72%) men) included squamous cell ( n = 17), sarcomatoid ( n = 6), plasmacytoid ( n = 6), adenocarcinoma ( n = 4) and small cell ( n = 4) variants. Inter‐reader agreement was similar for VH and UC ( p > 0.44). All readers characterized all squamous cell carcinoma as T2 intermediate. For two of three readers, 92%–100% of hypoenhancing tumours were VH at pathology whereas 74%–78% of hyperenhancing masses were UC ( p = 0.001). Combination of 11 selected radiomics features had AUC of 0.98 for training and 0.95 for validation data for prediction of VH. Conclusion: Squamous cell carcinoma was T2 intermediate, and VH was more likely to hypoenhance compared to UC. Our radiomic model had excellent performance for predicting VH.