Machine Learning in Spine Oncology: A Narrative Review

Area of Science:

• Oncology
• Artificial Intelligence
• Medical Imaging

Background:

• Machine learning (ML) represents a significant advancement in artificial intelligence (AI) with profound implications for medicine and surgery.
• In spine oncology, ML is increasingly used for analyzing medical imaging and classifying tumors with high accuracy.

Purpose of the Study:

• This narrative review specifically addresses the application of machine learning (ML) in the field of spine oncology.
• The review aims to synthesize current knowledge on ML's role in diagnosing, prognosing, and treating spinal tumors.

Main Methods:

• A systematic literature review was conducted using PRISMA methodology across major databases (PubMed, EMBASE, Web of Science, Scopus, Cochrane Library).
• Search terms included "Machine Learning" OR "Artificial Intelligence" AND "Spine Oncology" OR "Spine Cancer".
• Studies were categorized by tumor type (primary, metastatic, intradural) and data extracted on algorithms, sample sizes, and outcomes.

Main Results:

• Out of 480 references, 45 studies met the inclusion criteria.
• ML studies in spine oncology primarily use clinical and imaging features for risk stratification, mortality, and frailty prediction.
• ML demonstrated utility in tumor detection, differentiation, segmentation, and predicting survival and readmission rates.

Conclusions:

• Specialized neural networks and deep learning algorithms show high efficacy in predicting malignancy and aiding diagnosis.
• ML algorithms can predict tumor recurrence or progression using imaging and clinical data, optimizing treatment planning.
• ML has the potential to significantly improve healthcare accuracy, efficiency, and patient outcomes in spine oncology.