Pancreatic adenocarcinoma: variability in measurements of tumor size among computed tomography, magnetic resonance imaging, and pathologic specimens.

Abstract

To compare the tumor size measurements assessed by computed tomography (CT) and magnetic resonance imaging (MRI) versus measurements of resected pathologic specimens from patients with pancreatic ductal adenocarcinoma (PDAC). This study included 114 patients with histologically confirmed PDAC who underwent contrast-enhanced CT and MRI before surgery. The tumor sizes from CT, MRI, and pathologic specimens were compared by using Bland-Altman analyses and intraclass correlation coefficients (ICCs). The discrepancies in PDAC size between CT/MRI and pathologic specimens were calculated and contributing factors for the discrepancies, including tumor locations (pancreatic head/neck, body, or tail), T stages (T1, T2, or T3), and N stages (N0, N1, or N2), were analyzed with Pearson's correlation coefficients and multivariable linear regression analyses. There was significant difference among the mean tumor sizes of three measurements (P < 0.001). The difference in mean tumor size between the pathologic sizes for PDAC was 4.3mm (ICC 0.67) on CT and 5.8mm (ICC 0.65) on MRI. Both CT and MRI showed wide ranges of limits of agreement (LOAs) between the pathologic specimens for tumor size measurements (LOAs, - 28.9 to 21.4 and - 29.4 to 17.8, respectively). The tumor size on CT or MRI was estimated to be smaller than that on pathology when the tumor was > 30mm. The discrepancies in the tumor size estimated between CT/MRI and pathologic specimens were significantly different for tumors of different T stages (P < 0.001). Both contrast-enhanced CT and MRI underestimate the mean tumor size by 4.3mm and 5.8mm, respectively, compared to the size of pathologic specimens. A larger tumor size indicates a greater discrepancy in the PDAC size measurements between CT/MRI and pathologic specimens.