Comparison of AI-Driven Method with Semi-Automated Segmentation for Total Kidney Volume (TKV) Estimation in ADPK

Comparison of AI-Driven Method with Semi-Automated Segmentation for Total Kidney Volume (TKV) Estimation in ADPK

This poster, which was presented at the American Society of Neurology (ASN) 2025 Kidney Week meeting, compares an artificial intelligence (AI)-driven method with semi-automated expert radiologist segmentation for estimating Total Kidney Volume (TKV) in patients with Autosomal Dominant Polycystic Kidney Disease (ADPKD). Using MRI data from major clinical datasets, the study demonstrates that the AI model achieves highly consistent results with expert radiologists, both in single timepoint and longitudinal TKV assessments. The findings show strong correlation and minimal mean difference between methods, supporting the use of AI-driven segmentation as a reliable, reproducible, and efficient tool for TKV measurement in clinical trials. The poster concludes that integrating automated segmentation into clinical trial operations can enhance accuracy, reproducibility, and workflow efficiency, while also supporting expert review when needed. 

Why download this poster?

• Regulatory Alignment: Demonstrates use of TKV as an FDA-accepted prognostic biomarker in ADPKD trials.
• AI Validation: Provides evidence that AI-driven segmentation matches expert radiologist performance for TKV estimation.
• Longitudinal Monitoring: Shows reliable detection of physiological changes over time, critical for tracking disease progression and treatment effects.
• Operational Efficiency: Highlights how automated segmentation can streamline workflows, reduce manual workload, and improve reproducibility.
• Expert Integration: Supports a hybrid approach where AI can initialize segmentation for radiologist overread, enhancing both speed and accuracy.
• Data-Driven Insights: Includes quantitative comparisons, regression analysis, and visual examples to inform best practices in imaging-based endpoints.