An Efficient and Lightweight Convolutional Neural Network Model for Lung Cancer Diagnosis

Öz

Lung cancer stands as a major driver of mortality among cancer patients on a global scale, poses a substantial public health challenge due to its late diagnosis and the ambiguity of its early symptoms. This study proposes a novel, lightweight, and optimized Convolutional Neural Network architecture for detecting lung cancer types from Computed Tomography (CT) images. Initially, CT images of various lung cancer patients were combined for binary classification against normal patients. Experiments using different architectures and systematic hyperparameter optimizations resulted in the best model achieving a 93% accuracy rate. In the next phase, multi-class classification tasks were performed on the original dataset, and the performance of optimization algorithms such as Adamax, Adam, RMSprop, SGD, Adadelta, Nadam, and Adagrad were compared to determine the best optimizer. In these experiments, Adam algorithm achieved an 87% accuracy. The same study was repeated on an augmented dataset using data augmentation methods, reaching an accuracy of up to 96% with the RMSprop optimization algorithm. Furthermore, fine-tuned models were used and their test accuracies were evaluated using transfer learning methods (VGG16, ResNet50, InceptionV3, DenseNet121) on the available dataset. The primary contribution of this study is demonstrating that the proposed lightweight, optimized model achieves higher performance and efficiency on task-specific datasets, outperforming both established transfer learning methods and existing literature in accuracy. The proposed model will help healthcare professionals make quick and reliable decisions in lung cancer diagnoses.

Anahtar Kelimeler

• Convolutional Neural Networks
• Lung Cancer
• Transfer Learning
• Hyperparameter Optimizations
• Computed Tomography

Kaynakça

1. [1]. Retico A, Delogu P, Fantacci ME, Gori I, Preite Martinez A. Lung nodule detection in low-dose and thin-slice computed tomography. Comput Biol Med. 2008 Apr ;38(4):525–34.
2. [2]. Akciğer Kanseri Belirtileri Nelerdir? Nedenleri ve Tedavisi | Anadolu Sağlık Merkezi Hastanesi [Internet]. [cited 2024 Now 21]. Available from: https://www.anadolusaglik.org/saglik-rehberi/akciger-kanseri
3. [3]. World Health Organization. Lung cancer [Internet]. [cited 2024 Apr 21]. Available from: https://www.who.int/news-room/fact-sheets/detail/lung-cancer
4. [4]. Temurtaş F, Öztekin M, Yazdani M, Yörük YE, Aydemir F, Yonar D, et al. Akciğer Kanseri Tanısı İçin Yeni Bir Yöntem. Mühendislik Bilimleri ve Araştırma Dergisi. 2019;35–48.
5. [5]. Xie Y, Xia Y, Zhang J, Song Y, Feng D, Fulham M, et al. Knowledge-based Collaborative Deep Learning for Benign-Malignant Lung Nodule Classification on Chest CT. IEEE Trans Med Imaging. 2019 Apr 1;38(4):991–1004.
6. [6]. Munir K, Elahi H, Ayub A, Frezza F, Rizzi A. Cancer Diagnosis Using Deep Learning: A Bibliographic Review. 2019 Aug 23 ;11(9):1235.
7. [7]. Kaya M, Cetin-Kaya Y. A novel ensemble learning framework based on a genetic algorithm for the classification of pneumonia. Eng Appl Artif Intell. 2024 Jul 1; 133:108494
8. [8]. Krizhevsky A, Sutskever I, Hinton GE. ImageNet classification with deep convolutional neural networks. Commun ACM. 2017 May 24;60(6):84–90.

9. [9]. Çetin-Kaya Y, Kaya M. A Novel Ensemble Framework for Multi-Classification of Brain Tumors Using Magnetic Resonance Imaging. Diagnostics 2024;14(4):383.
10. [10]. Akila Agnes S, Alex Pandian Immanuel S, Anitha J, Arun Solomon A. Classification of Lung nodules using Convolutional long short-term Neural Network. Proceedings - 5th International Conference on Computing Methodologies and Communication, ICCMC 2021. 2021 Apr 8;1349–53.
11. [11]. Rehman A, Kashif M, Abunadi I, Ayesha N. Lung Cancer Detection and Classification from Chest CT Scans Using Machine Learning Techniques. In: 2021 1st International Conference on Artificial Intelligence and Data Analytics, CAIDA 2021. Institute of Electrical and Electronics Engineers Inc.; 2021. p.101–4.
12. [12]. Lei Y, Shan H, Zhang J. Meta ordinal weighting net for improving lung nodule classification. In: ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings. Institute of Electrical and Electronics Engineers Inc.; 2021. p. 1210–4.
13. [13]. Alakwaa W, Nassef M, Badr A. Lung Cancer Detection and Classification with 3D Convolutional Neural Network (3D-CNN). Int J Adv Comput Sci Appl. 2017;8(8):1- 6.
14. [14]. Liao F, Liang M, Li Z, Hu X, Song S. Evaluate the Malignancy of Pulmonary Nodules Using the 3-D Deep Leaky Noisy-OR Network. IEEE Trans Neural Netw Learn Syst. 2019 Nov 1;30(11):3484–95.
15. [15]. Anthimopoulos M, Christodoulidis S, Ebner L, Christe A, Mougiakakou S. Lung Pattern Classification for Interstitial Lung Diseases Using a Deep Convolutional Neural Network. IEEE Trans Med Imaging. 2016 May 1;35(5):1207–16.
16. [16]. Song QZ, Zhao L, Luo XK, Dou XC. Using Deep Learning for Classification of Lung Nodules on Computed Tomography Images. J Healthc Eng. 2017 Jan 1;2017(1):8314740.
17. [17]. Ashhar SM, Mokri SS, Rahni AAA, Huddin AB, Zulkarnain N, Azmi NA, et al. Comparison of deep learning convolutional neural network (CNN) architectures for CT lung cancer classification. International Journal of Advanced Technology and Engineering Exploration. 2021;8(74):126–34.
18. [18]. Mohamed TIA, Oyelade ON, Ezugwu AE. Automatic detection and classification of lung cancer CT scans based on deep learning and ebola optimization search algorithm. PLoS One. 2023 Aug 17 ;18(8): e0285796.
19. [19]. Tan H, Bates JHT, Matthew Kinsey C. Discriminating TB lung nodules from early lung cancers using deep learning. BMC Med Inform Decis Mak. 2022 Dec 1 ;22(1):1–7.
20. [20]. Cheyi, J., & Kaya, Y. Ç. (2024). Advanced CNN-Based Classification and Segmentation for Enhanced Breast Cancer Ultrasound Imaging. Gazi University Journal of Science Part A: Engineering and Innovation, 11(4), 647-667.
21. [21]. Güneş, A., & Kaya, Y. Ç. (2024). Evrişimsel Sinir Ağları ile Görüntülerde Gürültü Türünü Saptama. Bilgisayar Bilimleri ve Mühendisliği Dergisi, 17(1), 75-89.
22. [22]. Saraçoğlu, Y. D., & Kaya, Y. Ç. (2025). Efficient Hyperparameter-Tuned Convolutional Neural Network for Waste Classification. Gazi University Journal of Science Part A: Engineering and Innovation, 12(3), 815-835.
23. [23]. Kaya M, Cetin-Kaya Y. A Novel Deep Learning Architecture Optimization for Multiclass Classification of Alzheimer’s Disease Level. IEEE Access. 2024; 12:46562–81.
24. [24]. Chest CT-Scan images Dataset [Internet]. [cited 2023 Dec 11]. Available from: https://www.kaggle.com/datasets/mohamedhanyyy/chest-ctscan-images
25. [25]. Simonyan K, Zisserman A. Very Deep Convolutional Networks for Large-Scale Image Recognition. 3rd International Conference on Learning Representations, ICLR 2015 - Conference Track Proceedings. 2014 Sep 4; Available from: https://arxiv.org/abs/1409.1556v6
26. [26]. He K, Zhang X, Ren S, Sun J. Deep Residual Learning for Image Recognition. Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition. 2015 Dec 10 ;2016-December:770–8. Available from: https://arxiv.org/abs/1512.03385v1
27. [27]. Szegedy C, Vanhoucke V, Ioffe S, Shlens J, Wojna Z. Rethinking the Inception Architecture for Computer Vision. Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition. 2015 Dec 2;2016- December:2818–26. Available from: https://arxiv.org/abs/1512.00567v3
28. [28]. Ioffe S, Szegedy C. Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift. 32nd International Conference on Machine Learning, ICML 2015. 2015 Feb 11; 1:448–56. Available from: https://arxiv.org/abs/1502.03167v3
29. [29]. Huang G, Liu Z, Van Der Maaten L, Weinberger KQ. Densely Connected Convolutional Networks. Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017. 2016 Aug 25;2017- January:2261–9. Available from: https://arxiv.org/abs/1608.06993v5
30. [30]. Goutte C, Gaussier E. A Probabilistic Interpretation of Precision, Recall and F1- skore, with Implication for Evaluation. Conference: Advances in Information Retrieval, 27th European Conference on IR Research. 2005 March 21-23.
31. [31]. Saito T, Rehmsmeier M. The Precision-Recall Plot Is More Informative than the ROC Plot When Evaluating Binary Classifiers on Imbalanced Datasets. PLoS One. 2015 Mar 4;10(3): e0118432.
32. [32]. Memiş S, Enginoğlu S, Erkan U. A Data Classification Method in Machine Learning Based on Normalised Hamming Pseudo-Similarity of Fuzzy Parameterized Fuzzy Soft Matrices. Bilge International Journal of Science and Technology Research. 2019; 3:1–8.
33. [33]. Krstinić D, Braović M, Šerić L, Božić-Štulić D. Multi-label classifier performance evaluation with confusion matrix. 2020 Jan 14;1–14.
34. [34]. Kaya M. Bayesian optimization based CNN framework for automatic detection of brain tumors. Balkan Journal of Electrical and Computer Engineering. 2023;11(4):395-404.
35. [35]. Ba JL, Kiros JR, Hinton GE. Layer Normalization. 2016 Jul 21 Available from: https://arxiv.org/abs/1607.06450v1
36. [36]. Gal Y, Ghahramani Z. Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning. 33rd International Conference on Machine Learning, ICML 2016. 2015 Jun 6; 3:1651–60.
37. [37]. Jeczmionek E, Kowalski PA. Flattening Layer Pruning in Convolutional Neural Networks. Symmetry 2021. 2021 Jun 27;13(7):1147.