Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{202802,
        author = {Sahil Yadav and Dr. Priyanka Makkar},
        title = {EAMDLF: Multimodal Deep Learning Framework for Lung Cancer},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {12},
        pages = {7641-7655},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=202802},
        abstract = {Lung cancer continues to be one of the most serious health concerns worldwide, with approximately 2.2 million new cases and nearly 1.8 million deaths reported annually. Early detection plays a crucial role in improving the effectiveness of treatment and increasing patient survival rates. In recent years, advances in artificial intelligence (AI) and multimodal deep learning have provided new possibilities for developing automated and reliable diagnostic systems. Motivated by these advancements, this study proposes an Explainable Attention-Based Multimodal Deep Learning Framework (EAMDLF) for lung cancer diagnosis. The proposed framework combines several deep learning components, including Convolutional Neural Networks (CNNs), DenseNet-201 transfer learning, attention mechanisms, and multimodal feature fusion techniques, to improve diagnostic performance. The model utilizes different types of medical information collected from publicly available datasets, such as LIDC-IDRI, TCIA, LC25000, and ChestX-ray14, integrating CT scans, chest X-ray images, histopathological images, and clinical information for a more comprehensive diagnostic process. One of the common challenges in medical data analysis is the presence of data imbalances and variations among different data modalities. To address these issues, preprocessing methods such as image normalization, data augmentation, feature scaling, and synthetic oversampling techniques were applied. Feature extraction is performed using CNN and DenseNet architectures, whereas an attention-based fusion mechanism combines useful information from multiple sources. In addition, Explainable Artificial Intelligence (XAI) techniques, particularly Grad-CAM, were incorporated to improve model transparency and support clinical interpretation. The experimental findings indicate that the proposed framework achieved an accuracy of 99.12%, precision of 99.35%, recall of 98.91%, and F1-score of 99.13%, outperforming conventional single-modality and standard deep learning approaches. The results demonstrated fewer classification errors and improved diagnostic reliability. Overall, the proposed framework offers a practical and scalable solution that may assist healthcare professionals in improving lung cancer diagnosis and supporting personalized clinical decision making.},
        keywords = {Index Terms—Lung Cancer Diagnosis, Multimodal Deep Learning, Explainable Artificial Intelligence (XAI), Attention Mechanism, CNN, DenseNet-201, Feature Fusion, CT Imaging, Histopathology, Grad-CAM, Medical Imaging, Healthcare Analytics, Clinical Decision Support System, Artificial Intelligence.},
        month = {May},
        }