Meeting Banner
Abstract #5098

Deep Learning and Clinical Data Fusion in Prostate Cancer: Diagnosis of Clinically Significant Lesions Using Multiparametric MRI

Gelareh Valizadeh1, Farzan Moodi1,2, Fereshteh Khodadadi Shoushtari1, Mahmoud Morafegh1, Mahyar Ghafoori3, and Hamidreza Saligheh Rad 1,4
1Quantitative MR Imaging and Spectroscopy Group (QMISG), Tehran University of Medical Sciences, Tehran, Iran (Islamic Republic of), 2School of Medicine, Iran University of Medical Sciences, Tehran, Iran (Islamic Republic of), 3Department of Radiology, School of Medicine, Hazrat Rasoul Akram Hospital, Iran University of Medical Sciences, Tehran, Iran (Islamic Republic of), 4Department of Medical Physics and Biomedical Engineering, Tehran University of Medical Sciences, Tehran, Iran (Islamic Republic of)

Synopsis

Keywords: Diagnosis/Prediction, Machine Learning/Artificial Intelligence, multiparametric magnetic resonance imaging (mpMRI), ProstateX dataset, computer-aided diagnosis (CAD), deep learning, ensemble learning, clinical data, transfer learning

Motivation: Prostate cancer diagnosis requires accurate differentiation between clinically significant (csPCa) and non-significant (non-csPCa) lesions. Traditional methods, such as PSA and TRUS biopsy, lack specificity, often resulting in unnecessary biopsies.

Goal(s): This study aims to enhance diagnostic accuracy and reduce interventions by developing an automated deep learning framework for lesion classification.

Approach: Using the PROSTATEx dataset, an ensemble of DenseNet-121, ResNet-50, and ConvMixer integrates mpMRI and clinical data (lesion zone, PSAD).

Results: The model outperformed individual networks and prior methods, showing high AUC, sensitivity, specificity, and accuracy, promising improved diagnostic efficiency and patient outcomes.

Impact: This study’s integration of AI, multiparametric MRI, and clinical data refines prostate cancer diagnostics, equipping clinicians with a robust tool for precise lesion classification. This approach fosters personalized patient management, reduces overtreatment, and encourages further advancements in AI-driven oncology.

How to access this content:

For one year after publication, abstracts and videos are only open to registrants of this annual meeting. Registrants should use their existing login information. Non-registrant access can be purchased via the ISMRM E-Library.

After one year, current ISMRM & ISMRT members get free access to both the abstracts and videos. Non-members and non-registrants must purchase access via the ISMRM E-Library.

After two years, the meeting proceedings (abstracts) are opened to the public and require no login information. Videos remain behind password for access by members, registrants and E-Library customers.

Click here for more information on becoming a member.

Keywords