An enhanced Genetic Folding algorithm for prostate and breast cancer detection

Area of Science:

• Oncology and Bioinformatics
• Computational Biology and Machine Learning

Background:

• Genomic classification is vital for targeted cancer therapy due to increasing genomic complexity.
• Prostate and breast cancers exhibit significant heterogeneity, necessitating accurate diagnostic tools.
• Artificial Intelligence (AI) and Machine Learning (ML) offer potential for improved cancer diagnosis accuracy.

Purpose of the Study:

• To develop and validate an AI/ML model for accurate genomic classification of prostate and breast cancers.
• To evaluate the efficacy of the Genetic Folding (GF) algorithm in predicting cancer status.
• To identify key clinical features influencing diagnostic accuracy for prostate cancer.

Main Methods:

• The Genetic Folding (GF) algorithm was employed for prostate cancer status prediction.
• A comprehensive pipeline including exploratory data analysis (EDA), label encoding, feature standardization, decomposition, log transformation, outlier removal (Z-score), and BAGGINGSVM was used for breast cancer classification.
• Model performance was assessed using metrics like accuracy and Area Under the Curve (AUC).

Main Results:

• The GF algorithm achieved 96% accuracy in prostate cancer diagnosis, a new benchmark.
• The BAGGINGSVM approach attained 95.96% accuracy for breast cancer classification.
• Integrating the rate of change of Prostate-Specific Antigen (PSA) and age improved the AUC for prostate cancer by 6.8%, while BMI and race had no significant impact.

Conclusions:

• The Genetic Folding algorithm represents a significant advancement in accurate prostate cancer diagnosis.
• The developed breast cancer classification model demonstrates high accuracy and identifies key predictive factors.
• AI/ML models, particularly the GF algorithm, hold substantial promise for enhancing oncological diagnostics and personalized treatment strategies.