Predicting Colorectal Cancer Using Residual Deep Learning with Nursing Care

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Oncology

Background:

• Colorectal cancer is a leading cause of cancer-related deaths, necessitating improved early detection methods.
• Current image analysis and earlier cancer prediction (IAECP) systems face limitations in extracting crucial high- and low-level features from MRI scans.
• Effective IAECP is vital for reducing colorectal cancer mortality rates.

Purpose of the Study:

• To develop and evaluate a deep learning system (DLS) for enhanced image analysis and earlier cancer prediction (IAECP) of colorectal cancer using MRI.
• To overcome the limitations of conventional feature extraction methods in identifying cancerous regions.
• To improve the accuracy and efficiency of colorectal cancer detection.

Main Methods:

• A deep learning system (DLS) was employed to analyze entire bowel MRI images for cancer identification, feature extraction, and training.
• Residual convolution networks were utilized for training extracted bowel features, minimizing prediction errors.
• The Sum which is more Absolute to the Cross-correlation Template Feature Matching (SACC) algorithm was applied for comparing test images with trained data.

Main Results:

• The DLS-based IAECP system demonstrated a significant improvement in overall colorectal cancer identification accuracy.
• The deep learning features, unlike generic features, enhanced the IAECP prediction rate.
• Lab-scale analysis predicted an impressive IAECP rate of 99.8% using 100,000 histological datasets.

Conclusions:

• The proposed DLS-based IAECP system offers a highly accurate and effective method for early colorectal cancer prediction from MRI data.
• The use of deep learning features and advanced algorithms like residual convolution networks and SACC significantly boosts prediction accuracy.
• This approach holds promise for reducing colorectal cancer mortality through earlier and more precise diagnosis.