Value of Dual-Energy Computed Tomography for Detecting Small Pancreatic Ductal Adenocarcinoma
Yosuke Fujisaki, Yoshihiko Fukukura, Yuichi Kumagae, Fumitaka Ejima, Ryoji Yamagishi, Shinya Nakamura, Junki Kamizono, Hiroshi Kurahara, Shinichi Hashimoto, Takashi Yoshiura
Abstract
OBJECTIVE: The aim of the study is to evaluate the usefulness of virtual monoenergetic imaging (VMI) generated from dual-energy computed tomography (DECT) in detecting small pancreatic ductal adenocarcinomas (PDACs). METHODS: This study included 82 patients pathologically diagnosed with small PDAC (≤30 mm) and 20 without pancreatic tumors who underwent triple-phase contrast-enhanced DECT. To assess diagnostic performance for small PDAC detection via a receiver operating characteristic analysis, 3 observers reviewed 2 image sets (conventional computed tomography [CT] set and combined image set [conventional CT + 40-keV VMI from DECT]). The tumor-to-pancreas contrast-to-noise ratio was compared between conventional CT and 40-keV VMI from DECT. RESULTS: The area under the receiver operating characteristic curve of the 3 observers were 0.97, 0.96, and 0.97 in conventional CT set and 0.99, 0.99, and 0.99 in combined image set (P = 0.017-0.028), respectively. The combined image set yielded a better sensitivity than the conventional CT set (P = 0.001-0.023), without a loss of specificity (all P > 0.999). The tumor-to-pancreas contrast-to-noise ratios of 40-keV VMI from DECT were approximately threefold higher than those of conventional CT at all phases. CONCLUSIONS: The addition of 40-keV VMI from DECT to conventional CT had better sensitivity for detecting small PDACs without compromising specificity.