Multi-Modal Lung Cancer Detection Using Pyramidal Cascade Neuro-Fuzzy Fractional Network

Area of Science:

• Medical Imaging
• Artificial Intelligence in Healthcare
• Computational Biology

Background:

• Early lung cancer detection (LCD) is critical for patient survival and treatment efficacy.
• Computed Tomography (CT) and Positron Emission Tomography (PET) are key imaging modalities for lung cancer screening.
• Existing LCD techniques often face limitations and uncertainties.

Purpose of the Study:

• To introduce a novel Pyramidal Cascade Neuro-Fuzzy Fractional Network (PCNFFN) for enhanced lung cancer detection.
• To improve the accuracy and reliability of lung cancer diagnosis using integrated CT and PET imaging.
• To address the uncertainties in current lung cancer detection methods.

Main Methods:

• Pre-processing of CT and PET images using a Bilateral Filter (BF).
• Segmentation of lung lobes via Dual-Attention V-Network (DAV-Net).
• Tumor segmentation using Black Hole Entropic Fuzzy Clustering (BHEFC).
• Feature extraction and final lung cancer detection using the PCNFFN, a hybrid of PyramidNet, NFN, and Fractional Calculus (FC).

Main Results:

• The PCNFFN model demonstrated a high accuracy of approximately 91.002%.
• Achieved a True Negative Rate (TNR) of about 90.504%.
• Achieved a True Positive Rate (TPR) of about 92.571% in lung cancer detection.

Conclusions:

• The developed PCNFFN offers a robust and accurate approach for lung cancer detection.
• The integration of advanced image processing and deep learning techniques significantly improves diagnostic performance.
• This method holds promise for more effective early diagnosis and management of lung cancer.