Tradeoff Analysis of Needle Passes vs. Adequacy for Endoscopic Ultrasound-Guided Fine-Needle Aspiration Cytology: A Mathematical Modeling Approach

Abstract

Purpose: Endoscopic Ultrasound-Guided Fine-Needle Aspiration Cytology (EUS-FNAC) involves a tradeoff between the number of needle passes and the probability of obtaining an adequate sample. There are many possible sampling protocols that differ by the number of needle passes, the use of rapid onsite evaluation (ROSE) and the accuracy of the ROSE assessor. EUS-FNAC adequacy has been investigated in clinical studies; however, these studies are only able to compare two different methods. Sampling is a statistical process that can be analyzed by mathematical methods. We therefore developed mathematical models to compare the performance characteristics of a wide range of sampling protocols. Methods: We use simple probability models to develop formulae for the probability of per-case sampling success (at least one adequate sample), number of needle passes, and the incremental cost/benefit ratio expressed as needle passes per change in probability of success. Results: The probability of success increased with increasing needle passes. Sampling protocols that used ROSE required fewer needle passes to acheive the same probability of success as sampling based on a fixed number of needle passes. For ROSE-based methods, the probability of success and number of needle passes depended on the accuracy of the ROSE assessor. A less accurate assessor (e.g. cytotechnologist) achieved similar rates of success as a high accuracy assessor (e.g., pathologist) if the stopping point (number of observed adequate samples) was increased. Conclusion: ROSE sampling methods require fewer needle passes to achieve a target probability of success. A cytotechnologist can acheive success rates similar to that of a pathologist if the stopping point is adjusted.