Data mining and machine learning in cancer survival research: An overview and future recommendations.

Data mining and machine learning techniques are transforming the decision-making process in the medical world. From using nomograms and expert advice, scientists are now moving towards machine learning and deep learning techniques to make informed decisions for patients. The change in this aspect is mainly attributed to large amounts of digital data stored in hospitals. This study is focused on the transformation of cancer survival research in the past few years. A road map based on seven different aspects has been provided in this study utilizing various machine learning techniques, presenting a review of 62 articles published in the past 15 years. It was found that researchers are now moving to more clinical data even with less number of instances. Though most of the studies used traditional machine learning techniques for predicting cancer survival, researchers are now moving towards deep learning and hybrid approaches to gain some insights into survival prediction. Finally, this study presents ten new open research issues and possible future research plans to focus on for better results in cancer survival research. It is hoped that this review will be viewed by both apprentice and expert researchers as a valuable resource to understand the currently used practices and possible future recommendations to work.

An Integrated Approach for Cancer Survival Prediction Using Data Mining Techniques.

Ovarian cancer is the third most common gynecologic cancers worldwide. Advanced ovarian cancer patients bear a significant mortality rate. Survival estimation is essential for clinicians and patients to understand better and tolerate future outcomes. The present study intends to investigate different survival predictors available for cancer prognosis using data mining techniques. Dataset of 140 advanced ovarian cancer patients containing data from different data profiles (clinical, treatment, and overall life quality) has been collected and used to foresee cancer patients' survival. Attributes from each data profile have been processed accordingly. Clinical data has been prepared corresponding to missing values and outliers. Treatment data including varying time periods were created using sequence mining techniques to identify the treatments given to the patients. And lastly, different comorbidities were combined into a single factor by computing Charlson Comorbidity Index for each patient. After appropriate preprocessing, the integrated dataset is classified using appropriate machine learning algorithms. The proposed integrated model approach gave the highest accuracy of 76.4% using ensemble technique with sequential pattern mining including time intervals of 2 months between treatments. Thus, the treatment sequences and, most importantly, life quality attributes significantly contribute to the survival prediction of cancer patients.

Time-range based sequential mining for survival prediction in prostate cancer.

BACKGROUND AND OBJECTIVE: Metastatic prostate cancer has a higher mortality rate than localized cancers. There is a need to investigate the survival outcome of metastatic prostate cancers separately. Also, the treatments undertaken by the patients affect their overall survival. The present study tries to analyze the sequence of treatments given to the patients, along with the time intervals between each set of treatments. The time when medication needs to be changed may provide some useful insights into the survival outcome of the patients. MATERIALS AND METHODS: A total of 407 metastatic prostate cancer patients' data was collected and analyzed from an Indian tertiary care center. Popular sequence mining algorithms with exact order constraint have been applied to the treatment data. Appropriate time intervals were added in the resulted frequent sequences and fed to machine learning techniques along with other clinical data. RESULTS: The study suggests that the proposed methodology of the time range based sequence mining approach gave better results than the existing methods with 84.5% accuracy and 0.89 AUC. The time intervals in the existing sequence mining algorithms can give the clinicians some useful insights into the survival analysis and in determining the best lines of treatments for a particular patient.