A Pilot Randomized Crossover Trial of Wet Suction and Conventional Techniques of Endoscopic Ultrasound-Guided Fine-Needle Aspiration for Upper Gastrointestinal Subepithelial Lesions

Table 4

Univariate and multivariate analyses of factors influencing the moderate and high cellularity of EUS-FNA specimens.

(a) Analysis of the specimen cellularity obtained by DRY

Variable factor	Univariate analysis; value	Multivariate analysis; OR (95% CI), value
Study arm (dry-first arm vs. wet-first arm)	0.322
Tumor size (<20 vs. ≧20 mm)	0.509
SEL location (others vs. upper stomach)	0.183	0.514 (0.056–4.697), 0.555
Final diagnosis (GIST vs. others)	0.028	9.079 (1.012–81.485), 0.049

(b) Analysis of the specimen cellularity obtained by WET

Variable factor	Univariate analysis; value	Multivariate analysis; OR (95% CI), value
Study arm (dry-first arm vs. wet-first arm)	0.107	4.614 (0.662–32.141), 0.123
Tumor size (<20 vs. ≧20 mm)	0.412
SEL location (others vs. upper stomach)	0.032	0.125 (0.018–0.858), 0.034
Final diagnosis (GIST vs. others)	0.230

EUS-FNA: endoscopic ultrasound-guided fine-needle aspiration; SEL: subepithelial lesion; GIST: gastrointestinal stromal tumor; DRY: dry technique; WET: wet technique; OR: odds ratio; CI: confidence interval. Chi-squared test or Fisher’s exact test. Multivariate logistic regression analysis (acellular and poor cellularity; cellularity scores of 0 and 1 vs. moderate and high cellularity; scores of 2 and 3). Including cases which were not diagnosed by EUS-FNA.