Related Experiment Video
Updated: May 6, 2026
Tracking the Mammary Architectural Features and Detecting Breast Cancer with Magnetic Resonance Diffusion Tensor Imaging
Published on: December 15, 2014
Diagnostic performance of MDCT for predicting important prognostic factors in pancreatic cancer
Jung Hoon Kim1, Hyo Won Eun, Kyung Won Kim
1From the *Department of Radiology and Institute of Radiation Medicine, Seoul National University College of Medicine; and †Health Promotion Center, Asan Medical Center, University of Ulsan, Songpa-gu, Seoul, Korea.
Objectives:
Tumor stage, node metastasis, tumor size, vascular invasion, and perineural invasion are the most important prognostic factors that might be determined by preoperative multidetector computed tomography (MDCT) in pancreatic cancer. The purpose of our study is to investigate diagnostic accuracy of MDCT for determining these prognostic factors.
Methods:
For 6 years, 111 patients with surgically resected pancreatic cancer underwent preoperative MDCT. Two radiologists retrospectively assessed tumor stage, node metastasis, tumor size, and vascular invasion. They also graded perineural invasion using a 3-point scale focused on 5 routes. Statistical analyses were performed using the receiver operating characteristic analysis, McNemar test, and paired t test.
Results:
Statistically, tumor size on specimens (3.4 ± 1.46 cm) is larger than tumor size on MDCT (3.2 ± 1.41 cm; P = 0.001). The diagnostic accuracy rates for tumor stage were 82.9% and 77.5%, with moderate agreement (κ = 0.732). The accuracy rates for node metastasis were 59.5% and 55.0%, with fair agreement (κ = 0.597). The diagnostic accuracy rates for vascular invasion were 94% and 92%. The areas under the curve for perineural invasion were 0.733 and 0.66 (P = 0.069), with moderate agreement (κ = 0.77).
Conclusions:
Multidetector computed tomography is very useful for the preoperative evaluation of tumor stage, perineural invasion, and vascular invasion of pancreatic cancer, but it has limited evaluation of node metastasis and tumor size.
