Araştırma Makalesi

Evaluation of Deep Learning Architectures in Lung Cancer Diagnosis Using Computed Tomography Images

Cilt: 7 22 Temmuz 2026
Emine Gümüş *, Furkan Göz
PDF İndir
OMÜ Mühendislik Bilimleri ve Teknolojisi Dergisi Kapak OMÜ Mühendislik Bilimleri ve Teknolojisi Dergisi
TR EN

Evaluation of Deep Learning Architectures in Lung Cancer Diagnosis Using Computed Tomography Images

Öz

Lung cancer remains one of the leading causes of cancer-related mortality worldwide, and early detection is critical for improving patient survival rates. Computed tomography (CT) imaging is regarded as the gold standard in the diagnostic workflow due to its high spatial resolution; however, manual interpretation by radiologists continues to be prone to inter-observer variability and human error, particularly in the early stages of the disease. A vast majority of existing studies in the literature using the IQ-OTH/NCCD dataset either evaluate a single architecture in isolation or attempt to improve model performance through hybrid approaches, additional classifiers, or data augmentation techniques, which makes it difficult to objectively demonstrate the original contribution of the architectures themselves. In this study, a systematic and controlled comparative evaluation of five state of the art transfer learning architectures (DenseNet121, MobileNet, ResNet50, VGG16, and EfficientNet-B0) is presented using the open access IQ-OTH/NCCD dataset, which consists of 1,097 high quality CT images categorized into three classes: normal, benign, and malignant. In contrast to existing literature, all models were trained without the application of hybridization, additional classifiers, or data augmentation, utilizing the same preprocessing steps, hyperparameter configurations, and evaluation protocols; thus, ensuring that the observed performance differences stem solely from the architectural characteristics of each model. To validate the generalization ability of the models, a stratified 5 fold cross validation method was preferred. Experimental results revealed that among all evaluated models, the MobileNet architecture exhibited the highest performance with %99.52 accuracy and a %98.72 macro F1 score. This was followed by ResNet50 with %98.18 accuracy. The findings prove that a well-structured standard transfer learning approach can outperform hybrid and complex architectures, and that lightweight models such as MobileNet, in particular, hold high potential for real-time clinical diagnostic systems due to their low computational cost.

Anahtar Kelimeler

Lung Cancer Diagnosis, Deep Learning, Image Processing

Destekleyen Kurum

This study was supported by TÜBİTAK (The Scientific and Technological Research Council of Turkey) under the 2209-A Research Project Support Programme for University Students, Project No. 1919B012467956.

Proje Numarası

1919B012467956

Kaynakça

  1. L. Kalinke, R. Thakrar, and S. M. Janes, “The promises and challenges of early non-small cell lung cancer detection: patient perceptions, low-dose ct screening, bronchoscopy and biomarkers,” Molecular Oncology, vol. 15, no. 10, pp. 2544–2564, 2021.
  2. M. M. Lell and M. Kachelrieß, “Recent and upcoming technological developments in computed tomography: high speed, low dose, deep learning, multienergy,” Investigative radiology, vol. 55, no. 1, pp. 8–19, 2020.
  3. R. Indumathi and R. Vasuki, “Lung cancer detection using cad system,” in 2023 3rd International Conference on Advance Computing and Innovative Technologies in Engineering (ICACITE). IEEE, 2023, pp. 206–210.
  4. P. M. Shakeel, M. A. Burhanuddin, and M. I. Desa, “Lung cancer detection from ct image using improved profuse clustering and deep learning instantaneously trained neural networks,” Measurement, vol. 145, pp. 702–712, 2019.
  5. H. Alyasriy and M. AL-Huseiny, “The iq-oth/nccd lung cancer dataset,” Mendeley Data, V3, 2023, dataset. [Online]. Available: https://doi.org/10.17632/bhmdr45bh2.3
  6. S. Hangaragi, N. Neelima, V. Venugopal, S. Ganguly, J. Mudi, and J.-H. Choi, “Cae synthimggen: Revolutionizing cancer diagnosis with convolutional autoencoder-based synthetic image generation,” Alexandria Engineering Journal, vol. 115, pp. 343–354, 2025.
  7. S. Dev, P. S. Roy, N. Chakraborty, and R. Sarkar, “Lung cancer identification from ct scans using a soft-attention enabled deep transfer learning model,” in 2025 3rd International Conference on Intelligent Systems, Advanced Computing and Communication (ISACC). IEEE, 2025, pp. 254–259.
  8. A. S. Akbari, A. Kumar, B. R. Reddy, K. K. Singh, and M. Takei, “Vision transformer based automated model for enhancing lung cancer classification,” in 2024 IEEE International Conference on Imaging Systems and Techniques (IST), 2024, pp. 1–6.
  9. N. S. Jozi and G. A. Al-Suhail, “Lung cancer detection: The role of transfer learning in medical imaging,” in 2024 International Conference on Future Telecommunications and Artificial Intelligence (IC-FTAI), 2024, pp. 1–6.
  10. S. Lad, B. Chafekar, and P. Bide, “Lung cancer classification using deep learning: A comprehensive approach with modified convolutional neural networks,” in 2024 International Conference on Computational Intelligence and Network Systems (CINS), 2024, pp. 1–6.

Ayrıntılar

Birincil Dil

İngilizce

Konular

Görüntü İşleme

Bölüm

Araştırma Makalesi

Yazarlar

Yayımlanma Tarihi

22 Temmuz 2026

Gönderilme Tarihi

17 Mart 2026

Kabul Tarihi

20 Mayıs 2026

Yayımlandığı Sayı

Yıl 2026 Cilt: 7

IZ

https://izlik.org/JA76GF99XP

Kaynak Göster

RIS / Bibtex
APA
Gümüş, E., & Göz, F. (2026). Evaluation of Deep Learning Architectures in Lung Cancer Diagnosis Using Computed Tomography Images. OMÜ Mühendislik Bilimleri ve Teknolojisi Dergisi, 7, 17-32. https://izlik.org/JA76GF99XP
AMA
1.Gümüş E, Göz F. Evaluation of Deep Learning Architectures in Lung Cancer Diagnosis Using Computed Tomography Images. OMUJEST. 2026;7:17-32. https://izlik.org/JA76GF99XP
Chicago
Gümüş, Emine, ve Furkan Göz. 2026. “Evaluation of Deep Learning Architectures in Lung Cancer Diagnosis Using Computed Tomography Images”. OMÜ Mühendislik Bilimleri ve Teknolojisi Dergisi 7 (Temmuz): 17-32. https://izlik.org/JA76GF99XP.
EndNote
Gümüş E, Göz F (01 Temmuz 2026) Evaluation of Deep Learning Architectures in Lung Cancer Diagnosis Using Computed Tomography Images. OMÜ Mühendislik Bilimleri ve Teknolojisi Dergisi 7 17–32.
IEEE
[1]E. Gümüş ve F. Göz, “Evaluation of Deep Learning Architectures in Lung Cancer Diagnosis Using Computed Tomography Images”, OMUJEST, c. 7, ss. 17–32, Tem. 2026, [çevrimiçi]. Erişim adresi: https://izlik.org/JA76GF99XP
ISNAD
Gümüş, Emine - Göz, Furkan. “Evaluation of Deep Learning Architectures in Lung Cancer Diagnosis Using Computed Tomography Images”. OMÜ Mühendislik Bilimleri ve Teknolojisi Dergisi 7 (01 Temmuz 2026): 17-32. https://izlik.org/JA76GF99XP.
JAMA
1.Gümüş E, Göz F. Evaluation of Deep Learning Architectures in Lung Cancer Diagnosis Using Computed Tomography Images. OMUJEST. 2026;7:17–32.
MLA
Gümüş, Emine, ve Furkan Göz. “Evaluation of Deep Learning Architectures in Lung Cancer Diagnosis Using Computed Tomography Images”. OMÜ Mühendislik Bilimleri ve Teknolojisi Dergisi, c. 7, Temmuz 2026, ss. 17-32, https://izlik.org/JA76GF99XP.
Vancouver
1.Emine Gümüş, Furkan Göz. Evaluation of Deep Learning Architectures in Lung Cancer Diagnosis Using Computed Tomography Images. OMUJEST [Internet]. 01 Temmuz 2026;7:17-32. Erişim adresi: https://izlik.org/JA76GF99XP