In:
Journal of Clinical Gastroenterology, Ovid Technologies (Wolters Kluwer Health), Vol. 53, No. 7 ( 2019-08), p. e261-e268
Abstract:
Pancreatic solid masses (PSM) are difficult to assess; endoscopic ultrasound with fine-needle aspiration (FNA) enables tissue acquisition, but has high false-negative rates. Quantitative elastography (QE) predicts diagnosis on the basis of the strain ratio (SR). We aimed to compare both methods to evaluate PSM. Materials and Methods: This prospective study, carried out between January and December 2016, included suspected PSM cases; those with advanced disease and cystic components were excluded. Both procedures were performed; histologic information was obtained for the final diagnoses. Diagnostic tests and receiver-operating characteristic curve were calculated. P 〈 0.05 was considered statistically significant. Results: We included 134 patients (53% women; mean, 53±16.2 y). The median tumor size was 30 (10 to 78) mm, with 69.4% and 30.6% malignant and benign tumors (median SR: 19.5 vs. 7.5; P =0.000), respectively, and 87% were pancreatic adenocarcinoma. QE with SR cutoff ≥10 showed similar parameters to FNA in both PSM types: sensitivity, 94% in both; specificity, 85% versus 87%; positive predictive value, 93% versus 94%; negative predictive value, 87% in both; and accuracy, 92% for malignant and sensitivity, 85% versus 87%; specificity, 94% in both; positive predictive value, 87% in both; negative predictive value, 93% versus 94%; and accuracy, 92% for benign. The area under the curve was 0.96 ( P 〈 0.000; 95% confidence interval, 0.940-0.995). New classifications on the basis of positive likelihood ratio were grouped as follows: A ≤8.7 (benign tumor); B 〉 8.7 to 〈 15.5 (indeterminate); and C ≥15.5 (malignant). Conclusions: QE has similar capacity to FNA in PSM evaluation. However, the former can be used potentially as a substitute of the latter in certain cases on the basis of these new SR cutoff-based classifications.
Type of Medium:
Online Resource
ISSN:
0192-0790
DOI:
10.1097/MCG.0000000000001017
Language:
English
Publisher:
Ovid Technologies (Wolters Kluwer Health)
Publication Date:
2019
detail.hit.zdb_id:
2041558-8
Permalink