Dose Efficient Low-Contrast Detectability In CT: Multi-Vendor Benchmarking and Time-Efficient Workflow In Clinical Set-Ups
Abstract
Purpose
Use of task-based detectability index (d’) enables comparison of phantom-patient image quality in a clinically meaningful and vendor-agnostic manner, incorporating resolution details and noise characteristics beyond CNR. Measuring and optimizing d’ during equipment testing requires knowledge of d’ benchmarks and a robust workflow. The aims are 1) to develop a clinically-feasible ( 1.54, abdomen > 1.25) and d’ (> 2) to evaluate optimization availability. Viability of evaluations was assessed to optimize trade-off for time-efficient workflow.
Results
From derived CNR/CHO-d’ and CNR/CTDIvol relationships, reference dose levels at the clinically minimum IQ were established across vendors for adult abdomen and head protocols. Differences between clinical and dose-efficient levels quantified a potential range of mA-based dose optimization and respective d’/CNR change. Kernel‑dependent and iterative reconstruction dependent percent changes in CNR/d’ were consistent across repeated measurements. Early multi‑vendor comparisons suggest promising consistency in the underlying CNR/CHO-d’ and CNR/CTDIvol relationships.
Conclusion
This methodology provides a practical, quantitative framework for LCD‑based CT protocol optimization using CNR and d’. Early results support the feasibility of cross‑vendor standardization and the development of scalable, vendor‑agnostic reference tables for clinical protocol refinement.