USMF logo

Institutional Repository in Medical Sciences
of Nicolae Testemitanu State University of Medicine and Pharmacy
of the Republic of Moldova
(IRMS – Nicolae Testemitanu SUMPh)

Biblioteca Stiintifica Medicala
DSpace

University homepage  |  Library homepage

 
Please use this identifier to cite or link to this item: http://hdl.handle.net/20.500.12710/34293
Title: THE ROLE OF MACHINE LEARNING ALGORITHMS BASED ON T2-WEIGHTED MRI IN PREDICTING BRCA MUTATIONS IN PROSTATE CANCER
Authors: Cazacu, Oleg
Scutari, Adrian
Leşan, Octavian
Borcoman, Marian
Keywords: Prostate cancer;BRCA1;BRCA2;Prostate MRI;T2-weighted imaging;Radiomics;Machine learning;KNN;Random Forest
Issue Date: 2026
Publisher: Centrul Editorial-Poligrafic Medicina
Citation: Cazacu, Oleg; Scutari, Adrian; Leşan, Octavian; Borcoman, Marian. THE ROLE OF MACHINE LEARNING ALGORITHMS BASED ON T2-WEIGHTED MRI IN PREDICTING BRCA MUTATIONS IN PROSTATE CANCER. In: MedEspera 2026, p. 477. ISBN 978-9975-82-486-6.
Abstract: Introduction. Prostate cancer is one of the most common malignant diseases among men, and BRCA1 and BRCA2 germline mutations are associated with an increased risk of occurrence, early onset, and aggressive tumor progression. At the same time, multiparametric MRI has become an indispensable method for diagnosis and risk assessment, and T2 sequences are an important element in the morphological evaluation of the prostate. Radiomics techniques and machine learning algorithms represent a promising approach in the non-invasive identification of imaging biomarkers with important prognostic value. Aim of study. To evaluate current evidence and assess the role of KNN (K-Nearest Neighbors) and Random Forest machine learning algorithms applied to T2-weighted MRI images for predicting BRCA mutations in prostate cancer. Methods and materials. A structured review of recent medical literature from the last years and urological clinical guidelines was conducted, analyzing BRCA germline mutations in prostate cancer, the role of prostate MRI, and the application of machine learning in T2 image analysis. Studies on imaging features derived from T2-weighted sequences and on the use of KNN, Random Forest, and support vector machine algorithms for assessing tumor characteristics were analyzed. Results. Current data confirm that BRCA1/2 mutations, especially BRCA2 alterations, are of major clinical relevance in prostate cancer and justify the growing interest in non-invasive methods for identifying patients at increased genetic risk. In this context, radiomics offers a promising approach by extracting imaging features related to tumor texture and heterogeneity that can be integrated into machine learning models. The KNN model supports classification based on similarities between imaging profiles, whereas Random Forest can simultaneously analyze multiple variables in order to identify the most informative features. Current evidence suggests that machine learning analysis of prostate MRI may provide information beyond conventional visual assessment and represents a promising direction for improving prostate cancer characterization. Conclusions. Machine learning algorithms applied to T2-weighted MRI images represent a modern approach with potential for the development of non-invasive imaging biomarkers capable of estimating the presence of BRCA mutations in prostate cancer. Although this approach complements genetic testing, rapid advances in radiomics and machine learning support the prospect of future integration of such models into risk assessment and the selection of patients requiring additional molecular investigations.
metadata.dc.relation.ispartof: MedEspera 2026. The 11th International Medical Congress for Students and Young Doctors, 6-8 May 2026
URI: https://medespera.md/documents/abstract-book-4.pdf
https://repository.usmf.md/handle/20.500.12710/34293
ISBN: 978-9975-82-486-6
Appears in Collections:MedEspera 2026

Files in This Item:
File Description SizeFormat 
M_477.pdf466.73 kBAdobe PDFView/Open


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.

 

Valid XHTML 1.0! DSpace Software Copyright © 2002-2013  Duraspace - Feedback