Trust-aware clinical decision support in IoMT using hesitant fuzzy sets and deep learning-based medical image analysis
8 affiliations
- 1Department of Computer Science and Technology, Manav Rachna University
- 2Electronics and Computer Science Engineering, St. John College of Engineering and Management
- 3School of Computer Science and Engineering, RV University
- 4Computer Application and Information Technology, St. Xavier’s University
- 5Department of Computer Science and Engineering, Dayananda Sagar University
- 6School of Cyber Security and Forensic Science, Faculty of Engineering and Technology, Sri Sri University
- 7Department of Computer Science and Business Systems, Dayananda Sagar College of Engineering
- 8Department of Electrical and Computer Engineering, Dambi Dollo University
Abstract
The Internet of Medical Things promises continuous monitoring for patients and intelligent clinical analysis; yet, an important challenge is how to make reliable decisions from uncertain, noisy, and possibly non-trustworthy data. Current methods typically focus on either medical image analysis or integrity assessment of connected medical devices, with limited reliability of automated clinical decision support. In this work, we propose a trust-aware framework using deep learning medical image analysis, reliability-adaptive hesitant fuzzy fusion, and multi-criteria decision ranking to overcome this limitation. The framework was assessed with the BraTS 2021 multimodal magnetic resonance imaging (MRI) dataset for patient level tumour burden classification. The deep image features were combined with controlled device-integrity attributes, and the interdependent decision criteria were globally weighted and then adapted on the patient level based on predictive uncertainty, temporal hesitation and image–integrity agreement. The proposed framework obtained 99.1% accuracy, 98.6% precision, 98.2% recall and 98.4% F1-score under mixed operating conditions. The performance was measured using the same protocol as the strongest deep-learning baseline evaluated, and found to be 4.3% higher in accuracy, 4.7% higher in precision, 5.0% higher in recall, and 4.9% higher in F1-score. Controlled communication and device degradation resulted in a 96.2% decision reliability, a 95.7% trust-index accuracy, and a 5.2% false diagnostic rate. The results demonstrate that the adaptive fusion mechanism enhances the classification of the tumor burden, as well as an operational level of trustworthiness in connected healthcare environments.
Related
- Real-world evaluation of a multilingual agentic retrieval-augmented clinical AI framework for evidence-grounded medical decision support
- Development and usability testing of a machine learning clinical decision support tool for suggesting etiologies and treatment of clinical deterioration
- Design and Evaluation of a Source-Grounded Medical LLM for Clinical Decision Support and Patient Care in Trustworthy Diagnostic Systems.
- Fairness in multimodal machine learning applications in clinical decision support: a systematic review
- Machine Learning-Augmented Precision Dosing: Integrating Pharmacokinetic Modeling With Clinical Decision Support.
BibTeX
@article{mohammad2026trustaware,
title = {{Trust-aware clinical decision support in IoMT using hesitant fuzzy sets and deep learning-based medical image analysis}},
author = {Awwab Mohammad and Subhasini Shukla and Basavaraj Patil and Sibun Parida and Savitha Hiremath and Gyana Ranjana Panigrahi and Anil D and Kiran Ramaswamy},
journal = {Scientific Reports},
year = {2026},
month = jul,
doi = {10.1038/s41598-026-63609-6},
url = {https://doi.org/10.1038/s41598-026-63609-6},
note = {CC BY},
}
RIS
TY - JOUR
TI - Trust-aware clinical decision support in IoMT using hesitant fuzzy sets and deep learning-based medical image analysis
AU - Awwab Mohammad
AU - Subhasini Shukla
AU - Basavaraj Patil
AU - Sibun Parida
AU - Savitha Hiremath
AU - Gyana Ranjana Panigrahi
AU - Anil D
AU - Kiran Ramaswamy
PY - 2026
DA - 2026/07/01
JO - Scientific Reports
DO - 10.1038/s41598-026-63609-6
UR - https://doi.org/10.1038/s41598-026-63609-6
AB - The Internet of Medical Things promises continuous monitoring for patients and intelligent clinical analysis; yet, an important challenge is how to make reliable decisions from uncertain, noisy, and possibly non-trustworthy data. Current methods typically focus on either medical image analysis or integrity assessment of connected medical devices, with limited reliability of automated clinical decision support. In this work, we propose a trust-aware framework using deep learning medical image analysis, reliability-adaptive hesitant fuzzy fusion, and multi-criteria decision ranking to overcome this limitation. The framework was assessed with the BraTS 2021 multimodal magnetic resonance imaging (MRI) dataset for patient level tumour burden classification. The deep image features were combined with controlled device-integrity attributes, and the interdependent decision criteria were globally weighted and then adapted on the patient level based on predictive uncertainty, temporal hesitation and image–integrity agreement. The proposed framework obtained 99.1% accuracy, 98.6% precision, 98.2% recall and 98.4% F1-score under mixed operating conditions. The performance was measured using the same protocol as the strongest deep-learning baseline evaluated, and found to be 4.3% higher in accuracy, 4.7% higher in precision, 5.0% higher in recall, and 4.9% higher in F1-score. Controlled communication and device degradation resulted in a 96.2% decision reliability, a 95.7% trust-index accuracy, and a 5.2% false diagnostic rate. The results demonstrate that the adaptive fusion mechanism enhances the classification of the tumor burden, as well as an operational level of trustworthiness in connected healthcare environments.
SN - 2045-2322
ER -