Tumor Size on Microscopy, CT, and MRI Assessments Versus Pathologic Gross Specimen Analysis of Pancreatic Neuroendocrine Tumors.

The purpose of the present study was to assess the consistency of measurements of pancreatic neuroendocrine tumor (PNET) tumor size obtained using pre-operative imaging, pathologic gross specimen analysis, and microscopic examination of large pathologic sections; evaluate the impact of differences in pathologic and radiologic measurements of size on T categorization; and investigate the exact relationships among tumor size measurements obtained from microscopic analysis, CT, MRI, and pathologic gross specimen analysis. We enrolled 64 patients with pathologically confirmed PNETs who underwent radiologic examination between December 2016 and September 2019. Tumor sizes were measured by CT, MRI, pathologic gross specimen analysis, and microscopic examination. The relationship between the tumor sizes calculated by MRI and microscopy was analyzed using univariate and multivariate logistic regression models. The measurements of tumor sizes calculated by pathologic and radiologic assessments and CT and MRI assessments showed good concordance, but measurements calculated by microscopic analysis and other methods showed poor concordance. When T categories from pathologic gross specimen analysis were considered the reference, alterations in T category were found in the microscopic assessments of 12 of 64 patients (18.75%), CT assessments of 15 of 64 patients (23.44%), and MRI assessments of 13 of 64 patients (20.31%). In the fully adjusted model, microscopic size (β, 1.05; 95% CI, 0.98-1.12; < .001), CT size (β, 0.90; 95% CI, 0.78-1.02; < .001), and MRI size (β, 0.92; 95% CI, 0.81-1.04; < .001) were significantly correlated with gross tumor size. Tumor sizes measured by microscopy, CT, and MRI were significantly associated with the gross size of PNETs. This finding provides physicians with new tools for rapid identification of gross tumor size.