Artificial Intelligence for Risk Stratification in Diffuse Large B-Cell Lymphoma: A Systematic Review of...

Area of Science:

• Hematology
• Oncology
• Medical Informatics

Background:

• Diffuse large B-cell lymphoma (DLBCL) is a heterogeneous cancer with variable outcomes.
• Current prognostic tools like the International Prognostic Index (IPI) lack individual patient precision.
• Artificial intelligence (AI), machine learning (ML), and deep learning (DL) offer potential for improved DLBCL prognostic models.

Purpose of the Study:

• To systematically review the literature on AI/ML applications in DLBCL outcome prediction and risk stratification.
• To categorize studies by data modality and computational approach.
• To identify trends, knowledge gaps, and translation opportunities for AI/ML in DLBCL.

Main Methods:

• Structured literature search of PubMed/MEDLINE, Scopus, and Cochrane Library databases.
• Inclusion of original studies applying AI/ML for DLBCL survival prediction, risk classification, or subtype identification.
• Categorization of studies by data input: clinical, PET/CT, CT, histopathology, transcriptomics, genomics, ctDNA, and multi-omics.

Main Results:

• 91 studies met inclusion criteria, utilizing diverse data modalities including PET/CT imaging (n=30) and gene expression profiling (n=19).
• Common AI/ML techniques included ensemble learning, CNNs, and LASSO-Cox models.
• AI/ML models frequently surpassed IPI performance (AUC > 0.80), with multi-omics and ctDNA showing strong translation potential.

Conclusions:

• AI/ML methods are increasingly utilized to enhance DLBCL prognostic accuracy using diverse data inputs.
• These approaches provide superior risk stratification, aiding early identification of high-risk patients and personalized therapy.
• Future research should prioritize external validation, model interpretability, and integration into clinical workflows.