Digital Twin-Assisted Deep Learning with Model Fusion for Detecting Multiple Sclerosis in MRI Modalities

Abstract

The advancement of technologies such as IoT, Big Data, Data Science, Augmented Reality/Virtual Reality, and cloud computing has transformed the manufacturing sector through the Digital Twin (DT), serving as a potent instrument for simulating concepts into practice. Recently, artificial intelligence has been merged with digital technology to conduct research in the healthcare sector, facilitating judgments on the planning of patients’ clinical courses and the allocation of available medical resources. DTs can be utilized in the medical industry to facilitate clinical decision-making, providing prognoses and personalized treatment for patients. Multiple sclerosis (MS) is an autoimmune neurological disorder that impacts the central nervous system, potentially resulting in neurological impairment and mortality if left untreated. [X1]. In this research study, magnetic resonance imaging (MRI) samples collected from the E-Health lab and IXI databases are used to construct the DT empowered with artificial intelligence for the diagnosis of MS in a robust manner. We propose a hybrid CNN-RNN model for detecting MS in two phases. In the first phase, the deep features of the MRI modalities are extracted by two transfer learning models, m-InceptionV3 and m-DenseNet121. In the second phase, the classification of extracted features to either healthy or MS is performed by the Long Short-Term Memory RNN model. Deep learning metrics like precision, recall, F1 score, and accuracy are used to validate the performance of the proposed model. The proposed model outperformed the other state-of-the-art models achieving a good performance of 99.67% in validation accuracy. This healthcare DT pipeline may assist in clinical decision-making for MS detection and planning post-MS rehabilitation.

Received: 19 September 2024 | Revised: 26 November 2024 | Accepted: 2 January 2025 

Conflicts of Interest

The authors declare that they have no conflicts of interest to this work.

Data Availability Statement

The data that support the findings of this study are openly available in IXI Dataset at https://brain-development.org/ixi-dataset/.

Author Contribution Statement

Ramya Palaniappan: Conceptualization, Methodology, Software, Validation, Formal analysis, Writing – original draft, Writing – review & editing. Siva Rathinavelayutham: Investigation, Resources, Data curation, Visualization, Supervision, Project administration.