Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2025 Jun 14:80:100703.
doi: 10.1016/j.clinsp.2025.100703. Online ahead of print.

Optimizing malignancy prediction: A comparative analysis of transfer learning techniques on EBUS images

Ali Erdem Ozcelik  1 , Neslihan Ozcelik  2 , Emre Bendes  3 , Gizem Ozcibik Isik  4 , Omer Topaloglu  5
Affiliations

Optimizing malignancy prediction: A comparative analysis of transfer learning techniques on EBUS images

Ali Erdem Ozcelik et al. Clinics (Sao Paulo). .

Abstract

Background: Improving diagnostic accuracy in EBUS image analysis using machine learning is a current challenge. This study aimed to identify the most effective transfer learning model for predicting lymph node malignancy.

Methods: EBUS images collected between 2020-2023 were retrospectively analyzed. Demographic data, sampled lymph nodes, and pathology results were retrospectively collected from the files. Eight pre-trained CNN models (VGG, ResNet, InceptionNet, Xception, MobileNet, DenseNet, NasNet, EfficientNet) were evaluated.

Results: The study shows that the VGG19, EfficientNetV2L and DenseNet201 models have the highest performance in malignancy prediction, achieving areas under the curve of 0.96, 0.96 and 0.95 respectively, with consistent training and testing accuracy, indicating successful models without overfitting. In contrast, the ResNet152V2, Xception, and NasNet models show lower performance with areas under the curve of 0.88, 0.85, and 0.84 respectively, indicating overfitting due to discrepancies between training and test data. The MobileNetV2 model, with an area under the curve of 0.50, fails to discriminate between benign and malignant cases, resulting in an accuracy of only 0.51.

Conclusions: The application of transfer learning to the analysis of EBUS images offers significant potential for improving diagnostic accuracy in thoracic medicine, particularly in lung cancer.

Keywords: EBUS; InceptionNet; Lung cancer; Machine learning; ResNet; Transfer learning; VGG.

PubMed Disclaimer

Conflict of interest statement

Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Figures

Fig. 1:
Fig. 1
Representative image from an Endobronchial Ultrasound (EBUS) procedure showing the mediastinal Lymph Node (LN) adjacent to the airway wall. The image was captured during a real-time Transbronchial Needle Aspiration (TBNA) session.
Fig. 2:
Fig. 2
Workflow of the deep learning-based malignancy prediction process. The diagram illustrates key steps from EBUS image acquisition and augmentation, through training and testing data separation, to model training and evaluation using transfer learning methods.
Fig. 3:
Fig. 3
Patient flow diagram.
Fig. 4:
Fig. 4
ROC curve graph of the models. The evaluation of the models' success relied on the area under the curve. ROC, Receiver Operating Characteristic; AUC, Area Under the Curve.
Fig. 5:
Fig. 5
(a) Accuracy and loss graph of the Visual Geometry Group (VGG19) model. (b) Accuracy and loss graph of the EfficentNetV2L model. (c) Accuracy and loss graph of the DenseNet201 model.
Fig. 6:
Fig. 6
(a) Accuracy and loss graphs of the ResNet152V2 model, (b) Accuracy and loss graphs of the Xception model, (c) Accuracy and loss graphs of the NasNet model.
Fig. 7:
Fig. 7
(a) Accuracy and loss graphs of the InceptionResNetV2 model, (b) Accuracy and loss graphs of the MobileNetV2 model.
Fig. 8:
Fig. 8
The performance comparison of transfer learning models based on their Area Under the Curve (AUC) scores.
See this image and copyright information in PMC

References

    1. Philip B., Jain A., Wojtowicz M., Khan I., Voller C., Patel R.S.K., et al. Current investigative modalities for detecting and staging lung cancers: a comprehensive summary. Indian J Thorac Cardiovasc Surg. 2023;39(1):42–52. - PMC - PubMed
    1. Vilmann P., Clementsen P.F., Colella S., Siemsen M., De Leyn P., Dumonceau J.M., et al. Combined endobronchial and esophageal endosonography for the diagnosis and staging of lung cancer: european Society of Gastrointestinal Endoscopy (ESGE) Guideline, in cooperation with the European Respiratory Society (ERS) and the European Society of Thoracic Surgeons (ESTS) Endoscopy. 2015;47(6):c1. - PubMed
    1. So C., Matsumoto Y., Imabayashi T., Uchimura K., Ohe Y., Furuse H., et al. Identifying factors causing failure of nodal staging by endobronchial ultrasound-guided transbronchial needle aspiration in non-small cell lung cancer. Transl Lung Cancer Res. 2023;12(11):2169–2180. - PMC - PubMed
    1. Liu E., Bhutani M.S., Sun S. Artificial intelligence: the new wave of innovation in EUS. Endosc Ultrasound. 2021;10(2):79–83. - PMC - PubMed
    1. Lin C.K., Wu S.H., Chua Y.W., Fan H.J., Cheng Y.C. TransEBUS: the interpretation of endobronchial ultrasound image using hybrid transformer for differentiating malignant and benign mediastinal lesions. J Formos Med Assoc. 2025;124(1):28–37. - PubMed

LinkOut - more resources