Combining Molecular, Imaging, and Clinical Data Analysis for Predicting Cancer Prognosis

Area of Science:

• Oncology
• Bioinformatics
• Medical Informatics

Background:

• Cancer remains a leading global health concern, necessitating accurate patient prognosis prediction.
• Predictive modeling is crucial for effective cancer patient management and treatment planning.

Purpose of the Study:

• To review state-of-the-art scientific papers on cancer prognosis predictive models using multimodal data.
• To analyze the evolution, current issues, and future trends in multimodal cancer prognosis prediction.

Main Methods:

• Systematic review of 43 scientific papers published within the last six years.
• Categorization of studies based on modeling approaches.
• Analysis of data modalities: clinical, anatomopathological, molecular, and medical imaging.

Main Results:

• Significant evolution from traditional statistical models to data-driven approaches integrating multimodal data.
• Machine Learning and Deep Learning techniques are increasingly applied for complex prognosis prediction.
• Multimodal models demonstrate enhanced patient stratification capabilities compared to single-modality approaches.

Conclusions:

• Multimodal data integration significantly improves cancer prognosis prediction accuracy.
• These models facilitate personalized medicine and enhance clinical decision-making.
• Further research can yield deeper insights into cancer biology and progression.