Factors Affecting The Diagnostic Performance Of EUS-Guided Sampling In Solid Pancreatic Lesions By Trainee Endoscopists

Aims EUS is associated with a long learning curve. Both the technical and cognitive skills of the operator are needed to reach the necessary level of competence. We aimed to assess the factors which may influence the accuracy and sensitivity for malignancy in EUS-guided sampling of solid pancreatic lesions by trainees.

Methods Our study included EUS-FNA/FNB of solid pancreatic masses identified in a prospectively collected database of two Austrian centers. Examinations were performed by five endosonographers (three trainees). An experienced endosonographer was defined as having performed at least 225 EUS examinations, including 50 interventions (25 on pancreatic tumors). The accuracy (Acc) was calculated as the proportion of true positive and true negative in all evaluated cases, while sensitivity (Se) for malignancy represented the proportion of true positive in all malignant cases. The final diagnosis was determined by cyto-histopathology or clinical follow-up with a combination of tumor-markers and radiological findings.

Results 201 EUS-FNA/FNB of solid pancreatic lesions (73.1 % malignant) in 179 patients (median age 69 years, 57.7 % males) were enrolled. Trainees performed 73/201 (36.3 %) interventions.Overall accuracy and sensitivity tend to increase with endoscopists’ experience: 83.6 %vs. 71.2 % (p = 0.11), and 79.1 %vs.66.7 % (p = 0.07), respectively. More experienced endoscopists achieved better performance in sampling via the transduodenal approach (83.7 %vs. 70 %, p = 0.09). The tumor size seems to play a role in the sample adequacy independent of the operator’s experience.However, sensitivity in the trainee group for small lesions ( < 20mm) was poor (60 %). Trainees more frequently decided to use larger FNA needles and showed lower sensitivity for malignancy when sampling using 22 G needles: 33.3 %vs. 86 %, p = 0.01 (Tab. 1).

Tab. 1

Variables	Expert (Acc)	Trainee (Acc)	Expert (Se)	Trainee (Se)
Puncture
(A) Gastral	(A) 83.3 % (30/36)	(A) 73.9 % (17/23)	(A) 73.9 % (17/23)	(A) 68.4 % (13/19)
(B) Duodenal	(B) 83.7 % (77/92)	(B) 70 % (35/50)	(B) 80.9 % (55/68)	(B) 65.8 % (25/38)
TU Type
(A) Maligne	(A) 79.1 % (72/91)	(A) 66.7 % (38/57)	(A) 79.1 % (72/91)	(A) 66.7 % (38/57)
(B) Benigne	(B) 94.6 % (35/37)	(B) 87.5 % (14/16)
TU Size (mm)
(A) < 20	(A) 84.6 % (11/13)	(A) 62.5 % (5/8)	(A) 81.8 % (9/11)	(A) 60 %     (3/5)
(B) 20-40	(B) 74.6 % (44/59)	(B) 69.5 % (41/59)	(B) 70.6 % (36/51)	(B) 64.6 % (31/48)
(C) > 40	(C) 94.4 % (17/18)	(C) 100 % (5/5)	(C) 92.3 % (12/13)	(C) 100 %   (4/4)
Needle Type
(A) FNA	(A) 76 % (38/50)	(A) 70.5 % (43/61)	(A) 69.2 % (27/39)	(A) 66 %   (31/47)
(B) FNB	(B) 88.5 % (69/78)	(B) 75 %  (9/12)	(B) 86.5 % (45/52)	(B) 70 %    (7/10)
Needle Size
(A) 19 G	(A) 77.3 % (41/53)	(A) 73 % (46/63)	(A) 70.7 % (29/41)	(A) 70.6 % (36/51)
(B) 22 G	(B) 88 % (66/75)	(B) 60 % (6/10)	(B) 86 %  (43/50)	(B) 33.3 %  (2/6)

Conclusions During EUS training, it would be reasonable to perform punctures of smaller lesions via transduodenal approach by more advanced trainees. The type and size of the needle chosen by trainees may influence the accuracy and sensitivity of EUS-guided sampling of solid pancreatic lesions.

Citation Razpotnik M, Bota S, Kutilek M et al. eP406 FACTORS AFFECTING THE DIAGNOSTIC PERFORMANCE OF EUS-GUIDED SAMPLING IN SOLID PANCREATIC LESIONS BY TRAINEE ENDOSCOPISTS. Endoscopy 2021; 53: S229.

Publication History

Article published online:
19 March 2021