Early and accurate cancer diagnosis is critical for improving treatment outcomes and reducing the mortality rate. This study presents a deep learning-based multimodal cancer detection framework designed for the analysis of ultrasound, MRI, and CT scan images of the breast, liver, and thyroid organs. The proposed system utilizes EfficientNet-based architectures for the binary classification of cancerous and noncancerous medical images. EfficientNetB0 was employed for smaller ultrasound datasets to achieve computational efficiency and reduce overfitting, whereas EfficientNetB3 was utilized for more complex MRI and CT scan datasets to enhance feature extraction capability and classification performance. To improve model generalization, EfficientNetB3 was pretrained on approximately 117,000 medical images across 135 classes before fine-tuning on the target datasets. The framework was implemented using Python and PyTorch, along with supporting libraries, including OpenCV, NumPy, Pillow, and Torchvision. The experimental evaluation demonstrated promising results, achieving accuracies of 94% for breast ultrasound, 90% for liver ultrasound, 85% for thyroid ultrasound, 85% for breast MRI, 93% for liver MRI, and 80% for liver CT scan images. The results indicate that the proposed multimodal approach combined with transfer learning can effectively improve medical image classification and support early cancer diagnosis.

This research presented a deep learning-based multi-modal cancer detection framework using ultrasound, MRI, and CT scan images for the early diagnosis of breast, liver, and thyroid cancers. The proposed system utilized EfficientNetB0 and EfficientNetB3 architectures to perform binary classification of cancerous and non-cancerous medical images across multiple imaging modalities. By integrating transfer learning with EfficientNet models, the framework achieved effective feature extraction and improved classification performance on heterogeneous medical datasets. Experimental results demonstrated promising accuracy across different modalities, with breast ultrasound and liver MRI datasets achieving the highest classification performance. The adaptive selection of EfficientNet architectures based on dataset size and image complexity contributed to improved computational efficiency and reduced overfitting. Furthermore, pretraining EfficientNetB3 on a large-scale medical image dataset enhanced model generalization capability and supported reliable performance across diverse imaging conditions. The findings of this study indicate that multi-modal deep learning systems can significantly support early cancer diagnosis and assist healthcare professionals in medical image interpretation. The proposed framework highlights the potential of artificial intelligence in developing accurate, scalable, and clinically supportive computer-aided diagnostic systems. Future work can focus on integrating additional imaging modalities, explainable AI techniques, and larger standardized datasets to further improve the robustness and real-world applicability of intelligent cancer detection systems.

[1] P. Kaushik and P. Sharma, “A MobileNetV3-Based Approach to Breast Cancer Detection Through Transfer Learning and Histopathological Image Analysis,” 2024 International Conference on Advances in Computing, Communication and Materials (ICACCM), Dehradun, India, 2024, pp. 1-6, doi: 10.1109/ICACCM61117.2024.11059095. [2] P. Kaushik and P. Sharma, “Breast Cancer Detection using MobileNetV3: A Deep Learning Approach for Ultrasound Image Classification,” 2025 Fourth International Conference on Power, Control and Computing Technologies (ICPC2T), Raipur, India, 2025, pp. 1-5, doi: 10.1109/ICPC2T63847.2025.10958661. [3] A. K. Dwivedi, A. Yadav and R. S. Gamad, “Breast Cancer Disease Prediction using Convolutional Neural Networks with Ultrasound Image Dataset for Healthcare-IOT Application,” 2025 4th OPJU International Technology Conference (OTCON) on Smart Computing for Innovation and Advancement in Industry 5.0, Raigarh, India, 2025, pp. 1 6, doi: 10.1109/OTCON65728.2025.11070903. [4] M. I. Rosadi, C. Fatichah and A. Yuniarti, “Breast Cancer Ultrasound Images Classification Using Hybrid Pre-Trained CNN and SVM,” 2025 17th International Conference on Knowledge and Smart Technology (KST), Bangkok, Thailand, 2025, pp. 231-236, doi: 10.1109/KST65016.2025.11003377. [5] R. Sivakumar, “Deep Learning-Based Automated Breast Cancer Ultrasound Image Classification: A Study,” 2025 Eleventh International Conference on Bio Signals, Images, and Instrumentation (ICBSII), Chennai, India, 2025, pp. 1-5, doi: 10.1109/ICBSII65145.2025.11013993. [6] F. I. Mir, G. Raj, K. Lata and A. S. Verma, “Deep Data Analysis for Liver Cancer Prediction Using Machine Learning Approaches,” 2024 1st International Conference on Advances in Computing, Communication and Networking (ICAC2N), Greater Noida, India, 2024, pp. 864-870, doi: 10.1109/ICAC2N63387.2024.10895272. [7] Z. Naaqvi, S. Akbar, S. A. Hassan and Q. Ul Ain, “Detection of Liver Cancer through Computed Tomography Images using Deep Convolutional Neural Networks,” 2022 2nd International Conference on Digital Futures and Transformative Technologies (ICoDT2), Rawalpindi, Pakistan, 10.1109/ICoDT255437.2022.9787429. 2022, pp. 1-6, doi: [8] L. Basavarajappa, J. Li, H. Tai, J. Song, K. J. Parker and K. Hoyt, “Early Detection of Liver Steatosis using Multiparametric Ultrasound Imaging,” 2021 IEEE International Ultrasonics Symposium (IUS), 10.1109/IUS52206.2021.9593500. Xi’an, China, 2021, pp. 1-4, doi: [9] W. G. Negassa, S. Mishra and H. D. Bizuneh, “Liver Tumor Detection and Classification Using GWO-ELM Model,” 2023 International Conference in Advances in Power, Signal, and Information Technology (APSIT), Bhubaneswar, India, 2023, pp. 526 531, doi: 10.1109/APSIT58554.2023.10201687. [10] H. Zhang et al., “Multi-Source Transfer Learning Via Multi-Kernel Support Vector Machine Plus for B-Mode Ultrasound-Based Computer-Aided Diagnosis of Liver Cancers,” in IEEE Journal of Biomedical and Health Informatics, vol. 25, no. 10, pp. 3874-3885, Oct. 2021, doi: 10.1109/JBHI.2021.3073812. [11] D. Yang et al., “Automatic Thyroid Nodule Detection in Ultrasound Imaging With Improved YOLOv5 Neural Network,” in IEEE Access, vol. 12, pp. 22662-22670, 2024, doi: 10.1109/ACCESS.2024.3359367. [12] T. A. Vu, N. A. Huyen, H. Q. Huy and P. T. V. Huong, “Enhancing Thyroid Cancer Detection Through Machine Learning Approach,” 2023 12th International Conference on Control, Automation and Information Sciences (ICCAIS), Hanoi, Vietnam, 2023, pp. 188 193, doi: 10.1109/ICCAIS59597.2023.10382297. [13] C. Kaushik Viknesh, S. Kanimozhi and R. Thirumalai Selvi, “Investigation of Thyroid Nodule Detection Using Ultrasound Images with Deep Learning,” 2024 Tenth International Conference on Bio Signals, Images, and Instrumentation (ICBSII), Chennai, India, 2024, pp. 1-7, doi: 10.1109/ICBSII61384.2024.10564034. [14] Z. İLKILIÇ AYTAÇ, İ. İŞERİ and B. DANDIL, “Thyroid Cancer Detection Using Convolutional Neural Networks,” 2023 2nd International Engineering Conference on Electrical, Energy, and Artificial Intelligence (EICEEAI), Zarqa, Jordan, 2023, pp. 1-4, doi: 10.1109/EICEEAI60672.2023.10590377. [15] X. Yang and G. Zhu, “Ultrasound-based discrimination of thyroid cancer based on deep learning,” 2024 5th International Conference on Artificial Intelligence and Electromechanical Automation (AIEA), Shenzhen, China, 2024, pp. 939-943, doi: 10.1109/AIEA62095.2024.10692904. [16] S. R, S. R. Navaneethakrishnan, R. B, K. S. Balamurugan and M. Sudhagar, “Machine Learning in Oncology: SVM-Based Classification of Lung, Breast and Liver Cancer from MRI Scans,” 2024 IEEE 9th International Conference on Engineering Technologies and Applied Sciences (ICETAS), Bahrain, Bahrain, 2024, pp. 1-6, doi: 10.1109/ICETAS62372.2024.11119794. [17] C. Rokde, P. Kharwandikar, A. Kungwani and A. R. Dudhe, “Predictive Modeling for Early Cancer Detection: A Machine Learning Approach to Prostate, Lung, and Breast Cancer,” 2024 2nd DMIHER International Conference on Artificial Intelligence in Healthcare, Education and Industry (IDICAIEI), Wardha, India, 2024, pp. 1-6, doi: 10.1109/IDICAIEI61867.2024.10842741.

© 2025 International Journal of Engineering and Techniques (IJET).

Digital Object Identifier (DOI)

IJET assigns a unique DOI to every accepted article via Zenodo for persistent linking and citation. Your article’s DOI: https://doi.org/{{doi}}

Open DOI About Zenodo DOI

Close