Desarrollo y validación de un sistema de apoyo diagnóstico para la detección de nódulos pulmonares en tomografía computarizada mediante aprendizaje profundo

Camilo Cifuentes; Gabriel Jaime  Duarte López; Oscar Mauricio Gelvez Lizaraso; John Monzaide Álvarez Cely

doi:10.22490/25394088.11796

Vol. 20 No. 2 (2026)

Published 2026-09-29

license

This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.

When the Publicaciones e Investigaciones Journal receives an original study or article from its author(s), whether by email, postal service, or the platforms available for said purpose, know that it may be published in physical or electronic formats in national or international archives, databases, or SIRES. As such, Publications and Research authorizes the reproduction and citation of said material, provided that the description of information is carried out in conformity with bibliographic norms, and mention the corresponding names, authors, article, issue, and pages. Publications and Research, in advance, expresses that the information, concepts, and methods are the responsibility of the author(s). As such, the UNAD does not have any influence whatsoever over that expressed in the manuscript.

Article

Development and validation of a deep learning-based diagnostic support system for pulmonary nodule detection in computed tomography

DOI: https://doi.org/10.22490/25394088.11796

Camilo Cifuentes Universidad Santo Tomás

Gabriel Jaime Duarte López Universidad Santo Tomás

Oscar Mauricio Gelvez Lizaraso Universidad Santo Tomás

John Monzaide Álvarez Cely Universidad Santo Tomás

Pre-print (Spanish)

Abstract
References

Lung cancer remains the leading cause of cancer-related mortality worldwide, with approximately 2.5 million new cases and 1.8 million deaths in 2022. Early detection of pulmonary nodules on computed tomography (CT) significantly improves survival rates, yet manual review of hundreds of slices per study imposes a considerable workload and carries a risk of missing small lesions. This article aims to describe the development and validation of a computer-aided diagnostic support system for pulmonary nodule detection in CT images using deep learning techniques. The research was conducted as an applied and experimental study using the public LUNA16 and LIDC-IDRI datasets (888 CT studies). An ETL pipeline was implemented, including image normalization, slice preparation, and patient-level partitioning to prevent data leakage. Classification architectures (ResNet18, EfficientNet-B0) and a final YOLOv8 Triple Ensemble system with three-dimensional non-maximum suppression were comparatively evaluated. Results showed that patient-level analysis provides greater clinical utility than isolated slice-level evaluation. At patient level, the final system achieved 93.26% accuracy, 94.59% sensitivity, 91.04% specificity, 94.59% F1-score, and 0.928 AUC, outperforming conventional CAD systems and approaching the performance of specialized three-dimensional architectures. It is concluded that the developed tool shows potential as an experimental support system for prioritizing suspicious studies; however, external clinical validation is required before its use in healthcare settings.

keywords: Pulmonary nodules, Computed tomography, Deep learning, YOLOv8, Computer-aided diagnosis (CAD), Early lung cancer detection

[1] Organización Mundial de la Salud. (2026). Lung cancer. World Health Organization. Disponible en línea en el sitio oficial de la OMS. Consultado: 4 de junio de 2026.

[2] Organización Mundial de la Salud. (2026). Cancer. World Health Organization. Disponible en línea en el sitio oficial de la OMS. Consultado: 4 de junio de 2026.

[3] Armato, S. G., McLennan, G., Bidaut, L., McNitt-Gray, M. F., Meyer, C. R., Reeves, A. P. et al. (2011). “The Lung Image Database Consortium (LIDC) and Image Database Resource Initiative (IDRI): A completed reference database of lung nodules on CT scans”. Medical Physics, 38(2), pp. 915–931. DOI: 10.1118/1.3528204.

[4] Clark, K., Vendt, B., Smith, K., Freymann, J., Kirby, J., Koppel, P. et al. (2013). “The Cancer Imaging Archive (TCIA): Maintaining and operating a public information repository”. Journal of Digital Imaging, 26(6), pp. 1045–1057. DOI: 10.1007/s10278-013-9622-7.

[5] Setio, A. A. A., Traverso, A., de Bel, T., Berens, M. S. N., van den Bogaard, C., Cerello, P. et al. (2017). “Validation, comparison, and combination of algorithms for automatic detection of pulmonary nodules in computed tomography images: The LUNA16 challenge”. Medical Image Analysis, 42, pp. 1–13. DOI: 10.1016/j.media.2017.06.015.

[6] Ardila, D., Kiraly, A. P., Bharadwaj, S., Choi, B., Reicher, J. J., Peng, L. et al. (2019). “End-to-end lung cancer screening with three-dimensional deep learning on low-dose chest computed tomography”. Nature Medicine, 25, pp. 954–961. DOI: 10.1038/s41591-019-0447-x.

[7] Gu, Y., Lu, X., Yang, L., Zhang, B., Yu, D., Zhao, Y. et al. (2021). “A survey of computer-aided diagnosis of lung nodules from CT scans using deep learning”. Computers in Biology and Medicine, 137, 104806. DOI: 10.1016/j.compbiomed.2021.104806.

[8] Gao, C., Wu, L., Wu, W., Huang, Y., Wang, X., Sun, Z., Xu, M. y Gao, C. (2025). “Deep learning in pulmonary nodule detection and segmentation: A systematic review”. European Radiology, 35, pp. 255–266. DOI: 10.1007/s00330-024-10907-0.

[9] Ding, J., Li, A., Hu, Z. y Wang, L. (2017). “Accurate pulmonary nodule detection in computed tomography images using deep convolutional neural networks”. arXiv preprint arXiv:1706.04303.

[10] Dou, Q., Chen, H., Jin, Y., Lin, H., Qin, J. y Heng, P. A. (2017). “Automated pulmonary nodule detection via 3D ConvNets with online sample filtering and hybrid-loss residual learning”. arXiv preprint arXiv:1708.03867.

[11] He, K., Zhang, X., Ren, S. y Sun, J. (2016). “Deep residual learning for image recognition”. En: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas: IEEE, pp. 770–778. DOI: 10.1109/CVPR.2016.90.

[12] Tan, M. y Le, Q. V. (2019). “EfficientNet: Rethinking model scaling for convolutional neural networks”. En: Proceedings of the 36th International Conference on Machine Learning. Proceedings of Machine Learning Research, 97, pp. 6105–6114.

[13] Lin, T. Y., Goyal, P., Girshick, R., He, K. y Dollár, P. (2017). “Focal loss for dense object detection”. En: Proceedings of the IEEE International Conference on Computer Vision. Venice: IEEE, pp. 2980–2988. DOI: 10.1109/ICCV.2017.324.

[14] Jocher, G., Chaurasia, A. y Qiu, J. (2023). Ultralytics YOLOv8. Computer software. Disponible en línea en el repositorio oficial de Ultralytics. Consultado: 4 de junio de 2026.

[15] Paszke, A., Gross, S., Massa, F., Lerer, A., Bradbury, J., Chanan, G. et al. (2019). “PyTorch: An imperative style, high-performance deep learning library”. En: Advances in Neural Information Processing Systems, 32.

[16] Cardoso, M. J., Li, W., Brown, R., Ma, N., Kerfoot, E., Wang, Y. et al. (2022). “MONAI: An open-source framework for deep learning in healthcare”. arXiv preprint arXiv:2211.02701.

[17] Ronneberger, O., Fischer, P. y Brox, T. (2015). “U-Net: Convolutional networks for biomedical image segmentation”. En: Medical Image Computing and Computer-Assisted Intervention – MICCAI 2015. Cham: Springer, pp. 234–241. DOI: 10.1007/978-3-319-24574-4_28.

[18] Van Ginneken, B., Setio, A. A. A., Jacobs, C. y Ciompi, F. (2015). “Off-the-shelf convolutional neural network features for pulmonary nodule detection in computed tomography scans”. En: IEEE 12th International Symposium on Biomedical Imaging. New York: IEEE, pp. 286–289. DOI: 10.1109/ISBI.2015.7163869.

license

How to Cite

Cifuentes, C., Duarte López, G. J. ., Gelvez Lizaraso, O. M., & Álvarez Cely, J. M. (2026). Development and validation of a deep learning-based diagnostic support system for pulmonary nodule detection in computed tomography. Publicaciones E Investigación, 20(2). https://doi.org/10.22490/25394088.11796

Download Citation

Almétricas