Optimizing Skin Cancer Survival Prediction with Ensemble Techniques.

Area of Science:

• Computational biology and bioinformatics
• Artificial intelligence in oncology
• High-throughput data analysis

Background:

• Cancer research is advancing with high-throughput technology and artificial intelligence (AI) for improved diagnosis and targeted therapy.
• Complex, imbalanced, and high-dimensional data present significant challenges for computational approaches and multi-omics analysis in oncology.
• Accurate prediction and survival analysis are crucial for effective skin cutaneous melanoma (SKCM) management.

Purpose of the Study:

• To develop and evaluate machine learning (ML)-based ensemble methods for predicting skin cancer (SKCM) and analyzing overall survival probability.
• To address the challenges posed by high-dimensional and imbalanced datasets in computational cancer research.

Main Methods:

• Utilized a publicly available skin cutaneous melanoma (SKCM) dataset from the ICGC Data Portal.
• Employed eight baseline classifiers: random forest (RF), decision tree (DT), gradient boosting (GB), AdaBoost, Gaussian naïve Bayes (GNB), extra tree (ET), logistic regression (LR), and light gradient boosting machine (LGBM).
• Developed and evaluated four ensemble methods (stacking, bagging, boosting, voting) alongside survival analysis using Kaplan-Meier and Cox proportional hazards regression.

Main Results:

• The proposed ML-based ensemble methods demonstrated superior performance compared to traditional approaches.
• The random forest (RF) classifier achieved outstanding precision, while the voting ensemble method reached 99% accuracy.
• The RF classifier also achieved an excellent 99% accuracy, indicating its superior performance in SKCM prediction.

Conclusions:

• The developed ML ensemble techniques significantly improve the accuracy of skin cutaneous melanoma (SKCM) diagnosis.
• The random forest classifier and voting ensemble method show high potential for clinical application in SKCM.
• This research contributes to more accurate and efficient SKCM diagnosis, outperforming existing state-of-the-art techniques.