سامانه همایش‌ها TSTA | ثبت‌نام و ارسال مقاله کنفرانس‌ها و کنگره‌های ملی

Home / سی و دومین کنفرانس ملی و دهمین کنفرانس بین المللی مهندسی زیست پزشکی ایران

MRI to SPECT Image Translation for Parkinson's Disease Diagnosis

Authors :

Pegah Zandian PourEsfahani¹ Abolfazl Adib Almojahedi² Seyyede Zohreh Seyyed Salehi³

1- دانشگاه علوم پزشکی آزاد تهران 2- دانشگاه علوم پزشکی آزاد تهران 3- دانشگاه علوم پزشکی آزاد تهران

Keywords :

CNN،CycleGAN،Data Augmentation،DaTSPECT،MRI،Parkinson's disease

Abstract :

Parkinson's disease, the second most common neurological disorder worldwide, results from the degeneration of dopaminergic neurons in the substantia nigra, making early diagnosis complex and challenging. Although MRI images are widely accessible, they lack the sensitivity required to detect early changes. Conversely, DaTscan images provide higher accuracy by directly visualizing dopaminergic activity, but their use is limited by cost and complexity. This study employs a deep learning-based generative model, CycleGAN, to convert MRI images of patients and healthy individuals into DaTSPECT-like images. Subsequently, a Convolutional Neural Network (CNN) classifies patients and healthy subjects using the transformed images. This approach combines the advantages of both MRI and DaTscan modalities while addressing the scarcity of DaTscan data. Data were obtained from the PPMI dataset, which includes DaTSPECT and MRI images of patients and healthy controls. Results demonstrate that incorporating synthetic images with the original data increases classification accuracy to 85% and recall to 82.5%, reflecting improvements of 8.75% and 17.5%, respectively, over using only original DaTscan images. These findings confirm the effectiveness of the proposed method in enhancing Parkinson’s disease diagnosis.