CT-based radiomics for predicting Ki-67 expression in lung cancer: a systematic review and meta-analysis

Area of Science:

• Radiomics and medical imaging analysis
• Oncology and cancer biomarker research
• Diagnostic accuracy studies

Background:

• Lung cancer is a leading cause of cancer mortality globally.
• Accurate Ki-67 assessment is vital for predicting non-small cell lung cancer (NSCLC) aggressiveness and treatment response.
• Radiomics offers a novel approach for non-invasive biomarker prediction in solid tumors.

Approach:

• Systematic review and meta-analysis adhering to PRISMA-DTA guidelines.
• Literature search of PubMed, Embase, and Web of Science for radiomics studies predicting Ki-67 in lung cancer.
• Quality assessment using QUADAS-2 and Radiomics Quality Score (RQS); statistical analysis using STATA 14.2.

Key Points:

• Ten retrospective studies were meta-analyzed, demonstrating encouraging diagnostic performance of CT-based radiomics for Ki-67 prediction.
• Pooled sensitivity, specificity, and AUC were 0.78, 0.81, and 0.85 (training), and 0.78, 0.70, and 0.81 (validation), respectively.
• Study quality was generally acceptable; heterogeneity was noted, and publication bias was detected in the training cohort.

Conclusions:

• CT-based radiomics shows significant potential for predicting Ki-67 expression in lung cancer.
• The findings suggest radiomics could serve as a less invasive alternative to biopsy for Ki-67 assessment.
• Further research is recommended to improve diagnostic accuracy and facilitate clinical integration of radiomics.