Developing an Efficient Cancer Detection and Prediction Tool Using Convolution Neural Network Integrated with Neural...

Area of Science:

• Computational biology
• Bioinformatics
• Medical diagnostics

Background:

• Current cancer diagnosis methods are becoming outdated due to rising cancer incidence.
• Accurate cancer-type prediction is crucial for effective treatment strategies.
• Gene expression data holds significant potential for improving cancer diagnosis.

Purpose of the Study:

• To develop and evaluate a novel computational model for accurate cancer-type prediction.
• To integrate Convolutional Neural Networks (CNN) with Neural Pattern Recognition (NPR) for enhanced diagnostic capabilities.
• To leverage high-dimensional gene expression data for predicting cancer type based on tissue of origin.

Main Methods:

• A novel architecture combining 1-D Convolutional Neural Networks (CNN) with Neural Pattern Recognition (NPR) was developed.
• The model was trained and tested on gene expression profiles from a 5000-person sample.
• The model accounts for the tissue of origin in its cancer type prediction.

Main Results:

• The proposed CNN-NPR model achieved a prediction accuracy of 94%.
• The model demonstrated efficiency by requiring fewer parameters for training.
• The integration of CNN and NPR proved effective for cancer type prediction using gene expression data.

Conclusions:

• The developed CNN-NPR model shows significant promise for long-term cancer diagnosis and detection.
• This computational approach offers an effective and efficient methodology for cancer-type prediction.
• The model's high accuracy highlights the potential of integrating advanced computational techniques with genomic data in oncology.