Journal DOI 10.68956/journal.human-performance-futures

Creative Commons source reading · Data quality and algorithm evaluation

First steps in clinical implementation of a lung nodule classification method by artificial intelligence

Ajertil A; Farahat Z; Bouallou A; Le Tertre E; Elbher MJB; El Quessar A; Kabbaj N; Cherkaoui Malki M; Chat L; Hammani L; Ngote N

Originally published in Front Radiol · 2026-07-21

From the article

Selected source text. JournalOS removes markup and shortens excerpts where indicated. The full article provides the complete context.

Abstract

AbstractLung cancer remains one of the leading causes of cancer-related mortality worldwide, and early detection is essential to improve patient survival. Computed tomography (CT) is currently the reference imaging modality for lung cancer screening. Recent advances in Artificial Intelligence (AI), particularly Deep Learning (DL), have significantly improved automated medical image analysis and diagnostic accuracy. In this study, we propose a Convolutional Neural Network (CNN)-based approach for pulmonary nodule classification. The model was trained using a merged dataset composed of public CT image databases (IQ-OTH/NCCD and SPIE-AAPM) and real-world CT scans collected at Cheikh Zaid Hospital, Morocco. The experimental dataset included 1,103 malignant images, 508 benign images, and 427 normal images. Contrast Limited Adaptive…

Results and discussion

Despite the excellent performance achieved, several limitations must be acknowledged. The private clinical dataset remains relatively small and does not yet represent full-scale prospective validation. However, this deliberate phased approach allowed controlled integration of real-world clinical data while ensuring methodological rigor. The present work should therefore be considered as an initial translational step toward clinical implementation rather than a definitive deployment. Future studies will focus on expanding the clinical dataset, incorporating multicenter data, and conducting prospective validation studies to further confirm robustness and generalizability. The clinical dataset was used as part of a controlled validation strategy rather than for large-scale training, allowing preliminary assessment under real acquisition…

Conclusion

This study presents a high-performance Convolutional Neural Network (CNN) model for pulmonary nodule classification into three clinically relevant categories: malignant, benign, and normal. The model demonstrated outstanding diagnostic performance, achieving an accuracy of 99.84%, sensitivity of 100%, and specificity of 99.9%, establishing a strong benchmark for AI-assisted diagnosis in thoracic imaging. These results stem from an integrative approach combining carefully selected public datasets with real clinical CT images from Cheikh Zaid Hospital, enabling the model to adapt to the variability commonly encountered in routine medical practice. The clinical implications of this work are significant. The model's perfect sensitivity minimizes the risk of missed malignant nodules, while its high specificity reduces false positives and…

Editorial reading guide

  • What data created and tested the model, and was evaluation independent?
  • What is the cost of a false positive, false negative or out-of-distribution decision?