Pancreatic adenocarcinoma is a devastating disease and is projected by 2030 to be the second leading cause of cancer mortality [2–6]. Because survival can be improved with surgical intervention, particularly for small or lower stage tumors (Fig. 6.3), early detection and characterization is very important. Multiphasic multidetector CT (MDCT) plays a major role in this evaluation [4, 5], but there are still recognized shortcomings. Despite state of the art MDCT imaging, the sensitivity for detection of pancreatic ductal adenocarcinoma for lesions less than 2 cm is decreased [7]. In addition, four to 27 % [8–10] of pancreatic cancers are isoattenuating (Fig. 6.4) to the adjacent pancreatic parenchyma, especially when small [9, 10]. These lesions may be discovered only by detection of secondary signs on imaging, such as segmental dilation of the pancreatic duct [10, 11].

Pancreatic adenocarcinoma is characteristically hypoattenuating relative to the adjacent parenchyma and is best depicted during the pancreatic parenchymal phase MDCT [12]. Complete assessment for vascular invasion and distant metastatic disease requires venous phase images [4, 13]; thus, state of the art single energy or standard pancreatic MDCT involves at least two acquisition image sets. Preliminary investigations utilizing dual energy CT [1, 14] and single energy low kVp imaging [15, 16] have demonstrated improved conspicuity of hypovascular pancreatic adenocarcinomas at lower viewing or acquisition energy levels. In our practice at present, DECT is acquired through the upper abdomen during the pancreatic parenchymal phase of a multiphasic exam, and single energy DECT is acquired through the abdomen and pelvis in portal venous phase. In efforts to reduce radiation exposure, Brook et al. [17] recently described a split bolus spectral DECT protocol that takes advantage of low (60 keV) energy DECT simulated monenergetic image viewing during hepatic venous phase while providing parenchymal attenuation values similar to a standard MDCT pancreatic parenchymal phase in patients with a variety of pancreatic lesions. Other investigators have reported improved detection and staging of focal pancreatic lesions using other types of images available only with DECT, such as iodine material density [18] images, or during novel modes of DECT image acquisition such as CT perfusion [19] and portal venography [20]. Although DECT acquisitions have been applied during a single venous phase [1], during the pancreatic parenchymal phase of a multiphasic abdominal protocol [14, 18], during a single split bolus single acquisition [17], and during alternative timing schemes, the principles resulting in production of advantageous image data apply no matter the timing. These will be described based on the category of dual energy image and the various types of pancreas neoplasms in more detail below.

Macari et al. [1] first showed that conspicuity of pancreatic adenocarcinoma was greater using 80 kVp compared to blended 120 kVp images on dual source DECT obtained in portal venous phase. In this study, the weighting for the blended images was 0.3 for 80 kVp and 0.7 for 140 kVp. Chu et al. [21] identified that blended images with a weighting factor of 0.5 were preferred for evaluation of pancreatic diseases. Patel et al. [14] first showed the potential of rapid kV switching DECT acquired during pancreatic parenchymal phase to evaluate pancreatic adenocarcinoma lesion conspicuity in a cohort of 65 patients, and found a statistically significant increase of lesion contrast, a near doubling of the Hounsfield unit (HU) difference between tumoral and nontumoral tissue, on CNR-optimized keV images compared to the 70 keV image, the image energy typically used for routine PACS viewing (Fig. 6.5). The CNR-optimized keV calculation feature on the independent workstation and client server of the rapid kV switching DECT system workstation generates this value for each individual subject based on ROIs within tumoral and nontumoral tissue, and the mean optimized keV for the population in that early study was 52 keV. Note that with the rapid kV switching type of dual energy acquisition, there is no blended image and no generation of separate 80 and 140 kVp image sets for diagnosis. In a second study by the same group, a similar gain in conspicuity using low viewing energy was found for small pancreatic adenocarcinomas, an important and clinically relevant improvement, as smaller lesions might not alter the contour of the gland [18]. The use of lower simulated monoenergetic viewing energy (52 keV) in this study provided the best objective image quality measures (CNR, lesion contrast) for multiple readers. Also in that study, four of five lesions that were suspected clinically based on secondary findings of main pancreatic duct dilation were identified using low energy and iodine MD images combined with 70 keV [18] (Fig. 6.6). Others have also shown an improvement in pancreatic ductal adenocarcinoma detection by combining low keV imaging with a split bolus IV contrast administration technique [17] or by combining dsDECT 80 kVp images with a CT perfusion technique [19].