Application of the triple diagnosis technique to endoscopic ultrasound–guided fine-needle aspiration of solid pancreatic lesions: Impact on diagnostic accuracy and positive and negative predictive values

Abstract

Introduction

Endoscopic ultrasound–guided fine-needle aspiration (EUS-FNA) is an accepted method for the investigation of pancreatic solid mass lesions/solid masses. Reports have shown it to have high but imperfect diagnostic sensitivity and specificity. Sensitivity ranges from 77% to 98% with specificity varying from 25% to 100%. Negative predictive value (NPV) has been reported to be 72.7%, meaning a benign result does not exclude malignancy in a quarter of cases. We investigated if the correlation of cytology, imaging, and clinical evaluation would improve diagnostic accuracy.

Materials and Methods

An electronic search was carried out for EUS-FNA cases performed at our institution between 2018 and 2022. Each case was correlated with corresponding surgical pathology results or >2 years clinical follow-up. Cases with adequate follow-up were correlated with results of imaging and clinical diagnosis. Accuracy statistics were calculated for each method and for a triplet composed of clinical, imaging, and cytologic diagnoses.

Results

The search documented 198 EUS-FNAs, of which 181 had adequate follow-up; 140 cases had clinical diagnoses and 141 had imaging diagnoses. The sensitivity, specificity, positive predictive value (PPV), and NPV of cytology was 72%, 100%, 1.00, and 0.79, respectively. The diagnostic triplet had a sensitivity of 93% with a specificity of 100%. The PPV and NPV were 1.00 and 0.96, respectively.

Conclusions

The sensitivity and NPV of EUS-FNA cytology were low, indicating that a benign diagnosis did not exclude malignancy with nearly a quarter of benign specimens being false negatives. The triple diagnosis method improved diagnostic accuracy substantially with sensitivity, specificity, PPV, and NPV being 93%, 100%, 1.00, and 0.96, respectively.