Linear-array EUS improves detection of pancreatic lesions in high-risk individuals: a randomized tandem study.

BACKGROUND: Studies comparing linear and radial EUS for the detection of pancreatic lesions in an asymptomatic population with increased risk for pancreatic cancer are lacking.

OBJECTIVES: To compare pancreatic lesion detection rates between radial and linear EUS and to determine the incremental diagnostic yield of a second EUS examination.

DESIGN: Randomized controlled tandem study.

SETTING: Five academic centers in the United States.

PATIENTS: Asymptomatic high-risk individuals (HRIs) for pancreatic cancer undergoing screening EUS.

INTERVENTIONS: Linear and radial EUS performed in randomized order.

MAIN OUTCOME MEASUREMENTS: Pancreatic lesion detection rate by type of EUS, miss rate of 1 EUS examination, and incremental diagnostic yield of a second EUS examination (second-pass effect).

RESULTS: Two hundred seventy-eight HRIs were enrolled, mean age 56 years (43.2%), and 90% were familial pancreatic cancer relatives. Two hundred twenty-four HRIs underwent tandem radial and linear EUS. When we used per-patient analysis, the overall prevalence of any pancreatic lesion was 45%. Overall, 16 of 224 HRIs (7.1%) had lesions missed during the initial EUS that were detected by the second EUS examination. The per-patient lesion miss rate was significantly greater for radial followed by linear EUS (9.8%) than for linear followed by radial EUS (4.5%) (P = .03). When we used per-lesion analysis, 73 of 109 lesions (67%) were detected by radial EUS and 99 of 120 lesions (82%) were detected by linear EUS (P < .001) during the first examination. The overall miss rate for a pancreatic lesion after 1 EUS examination was 47 of 229 (25%). The miss rate was significantly lower for linear EUS compared with radial EUS (17.5% vs 33.0%, P = .007).

LIMITATIONS: Most detected pancreatic lesions were not confirmed by pathology.

CONCLUSION: Linear EUS detects more pancreatic lesions than radial EUS. There was a "second-pass effect" with additional lesions detected with a second EUS examination. This effect was significantly greater when linear EUS was used after an initial radial EUS examination.