An Effective Method for Lung Cancer Diagnosis from CT Scan Using Deep Learning-Based Support Vector Network.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Oncology

Background:

• Early-stage lung cancer diagnosis is difficult due to asymptomatic presentation and limitations of computed tomography (CT).
• Manual analysis of lung CT scans for pulmonary nodules is time-consuming and error-prone.
• Existing diagnostic methods face challenges in accuracy and efficiency.

Purpose of the Study:

• To develop and validate a novel deep learning-enabled support vector machine (SVM) model for early-stage lung cancer detection.
• To create a computer-aided design (CAD) system that identifies subtle pathological changes in lung CT images.
• To improve the accuracy and efficiency of pulmonary nodule detection in lung cancer screening.

Main Methods:

• A deep learning-assisted SVM model was developed and trained using a dataset of 888 annotated lung CT scans from the LIDC/IDRI database.
• The model was trained to measure and compare profile values in CT images from patients and control subjects.
• The model's performance was tested and validated on unseen CT scan data.

Main Results:

• The proposed deep learning-assisted SVM model achieved 94% accuracy in detecting pulmonary nodules indicative of early-stage lung cancer.
• The model demonstrated superior performance compared to existing complex deep learning, simple machine learning, and hybrid techniques.
• The computer-aided design (CAD) model effectively identified physiological and pathological changes in soft tissues within lung lesions.

Conclusions:

• The developed deep learning-assisted SVM model offers a highly accurate and efficient tool for early lung cancer detection.
• This computer-aided design (CAD) approach can significantly assist radiologists in identifying pulmonary nodules.
• The findings suggest potential for improved patient management through earlier and more reliable lung cancer diagnosis.