EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

A computer‐aided diagnosis (CAD) system based on convolutional neural networks for lung cancer diagnosis from 2D [ 18 F]‐ PET/CT images

Mohammad Karimpour iD, Neda Taghinezhad, Alireza Mehdizadeh, Mehrosadat Alavi, Tahereh Mahmoudi iD

Journal of Applied Clinical Medical Physics · 2025

DOI: 10.1002/acm2.70285 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Applied Clinical Medical Physics, 2025. Verfasst von Mohammad Karimpour und 4 weiteren Autor:innen. Beteiligte Einrichtung: Research Center for Neuromodulation and Pain, Shiraz University of Medical Sciences, Shiraz, Iran. Hauptthemen laut MEDLINE: Positron Emission Tomography Computed Tomography, Lung Neoplasms, Fluorodeoxyglucose F18, Carcinoma, Non-Small-Cell Lung. Die Arbeit nennt 32 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Objective

This study aims to automatically classify lung conditions into normal, non‐small cell lung cancer (NSCLC), and small cell lung cancer (SCLC) using [ 18 F] FDG PET/CT images and deep learning.

Methods

PET/CT scans from 146 patients (1974 scans) were retrospectively analyzed using two strategies: (1) transfer learning with pre‐trained CNNs, and (2) a custom CNN (Res‐SE Net) incorporating residual and squeeze‐and‐excitation (SE) modules. A patient‐based data splitting approach was used to avoid data leakage. Models were trained and validated at the scan level and evaluated at the patient level using majority voting. Grad‐CAM was employed to generate lesion‐localization heatmaps.

Results

Among the seven evaluated CNN models, the proposed Res‐SE Net demonstrated superior performance, achieving an accuracy of 91.67% and a sensitivity of 92.00% in detecting NSCLC, and an accuracy of 90.14% with a sensitivity of 90.00% for distinguishing SCLC cases. When tested on an external dataset, the model attained an accuracy of 98.00% in binary classification (Normal vs. Cancer). In the three‐class classification task, the model achieved an accuracy of 73.02% for NSCLC and 66.26% for SCLC.

Conclusion

These findings demonstrate the potential of Res‐SE Net architecture for accurate multi‐class lung cancer classification using [18F] FDG PET/CT images.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Positron Emission Tomography Computed TomographyLung NeoplasmsFluorodeoxyglucose F18Carcinoma, Non-Small-Cell LungNeural Networks, ComputerDiagnosis, Computer-AssistedImage Processing, Computer-AssistedSmall Cell Lung CarcinomaHumansRetrospective StudiesMaleFemaleMiddle AgedRadiopharmaceuticalsAgedDeep LearningAlgorithmsConvolutional Neural Networks

Stichwörter der Autor:innen

(18F) fluorodeoxyglucosePET/CTcomputer aided diagnosis (CAD)convolutional neural network (CNN)lung cancer

Substanzen

Fluorodeoxyglucose F18Radiopharmaceuticals

Autor:innen und Institutionen

  • Research Center for Neuromodulation and Pain, Shiraz University of Medical Sciences, Shiraz, Iran.

ORCID-Profile: Mohammad Karimpour, Tahereh Mahmoudi

Zitieren

APA

Karimpour, M., Taghinezhad, N., Mehdizadeh, A., Alavi, M., & Mahmoudi, T. (2025). A computer‐aided diagnosis (CAD) system based on convolutional neural networks for lung cancer diagnosis from 2D [ 18 F]‐ PET/CT images. Journal of Applied Clinical Medical Physics. https://doi.org/10.1002/acm2.70285

Vancouver

Karimpour M, Taghinezhad N, Mehdizadeh A, Alavi M, Mahmoudi T. A computer‐aided diagnosis (CAD) system based on convolutional neural networks for lung cancer diagnosis from 2D [ 18 F]‐ PET/CT images. Journal of Applied Clinical Medical Physics. 2025. doi:10.1002/acm2.70285

BibTeX
@article{Karimpour2025,
  author = {Karimpour, Mohammad and Taghinezhad, Neda and Mehdizadeh, Alireza and Alavi, Mehrosadat and Mahmoudi, Tahereh},
  title = {A computer‐aided diagnosis (CAD) system based on convolutional neural networks for lung cancer diagnosis from 2D [ 18 F]‐ PET/CT images},
  journal = {Journal of Applied Clinical Medical Physics},
  year = {2025},
  doi = {10.1002/acm2.70285}
}
Kostenlos starten →