Title
Multimodal AI for pneumonia and lung cancer classification using x-ray and HRCT
Authors
Chauhan Pradip*, 1, Chauhan Girish2, Chauhan Bhoomika3, Vaza Jayesh4, Ratanpara Lalit1 & Mehra Simmi1
Affiliation
1Department of Anatomy, All India Institute of Medical Sciences, Rajkot, Gujarat, India; 2Department of Oral Pathology and Maxillofacial Surgery, Government Dental College, Jamnagar, Gujarat, India; 3Department of Obstetrics and Gynaecology, Bhagyoday Medical College, Kadi, Gujarat, India; 4Department of Orthopaedics, Narendra Modi Medical College and Hospital, Ahmedabad, Gujarat, India; *Corresponding author
Chauhan Pradip - E-mail: prajjawalitresearch@gmail.com; Phone: +91 8866199560
Chauhan Girish - E-mail: drgirishchauhan@gmail.com
Chauhan Bhoomika - E-mail: drbhoomikachauhanchauhan@gmail.com
Vaza Jayesh - E-mail: jayesh.vaza@gmail.com
Ratanpara Lalit - E-mail:draksharphc@gmail.com
Mehra Simmi - E-mail: docsims27july@gmail.com
Article Type
Research Article
Date
Received January 1, 2026; Revised January 31, 2026; Accepted January 31, 2026, Published January 31, 2026
Abstract
Chest X-ray and HRCT are essential for diagnosing pneumonia and lung cancer, but their accuracy is limited. Hence, DeepScan, a multimodal AI combining CNNs trained on both imaging types, was developed using public datasets. The architecture included resnet-50 for X-rays, densenet-121 for HRCT and a late-fusion network. DeepScan outperformed single-modality models, achieving 94.6% accuracy, 95.2% sensitivity, 93.9% specificity and an AUC of 0.97 on 2,000 test patients. Multimodal integration reduced false negatives for early-stage lung cancer and improved differentiation from pneumonia, supporting earlier intervention and potentially enhancing clinical workflows.
Keywords
Artificial intelligence, deep learning, multimodal imaging, pneumonia, lung cancer, Chest X-Ray, HRC
Citation
Pradip et al. Bioinformation 22(1): 605-609 (2026)
Edited by
P Kangueane
ISSN
0973-2063
Publisher
License
This is an Open Access article which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. This is distributed under the terms of the Creative Commons Attribution License.
