New Diagnostic Imaging Modalities for Pancreatic Disease

Cyrus Piraka; James M. Scheiman


Curr Opin Gastroenterol. 2011;27(5):475-480. 

In This Article

Pancreatic Cancer

In 2010, multiple studies attempted to address some of the more challenging aspects in the management of pancreatic cancer: differentiation of cancer from benign tumors, determination of resectability, and assessment of prognosis.

Differentiating Cancer From Chronic Pancreatitis or Autoimmune Pancreatitis

Distinguishing pancreatic adenocarcinoma from nonmalignant masses remains challenging with current imaging techniques. In a prospective study by Napoleon et al.,[15•] contrast-enhanced harmonic endoscopic ultrasound (CEH-EUS) was employed in 35 patients presenting with solid pancreatic lesions. In addition to conventional B mode and power Doppler EUS, CEH-EUS was performed with an Olympus prototype echoendoscope (xGF-UCT 180) after an intravenous bolus injection of 2.4 ml of an ultrasound contrast agent (SonoVue). There were 18 adenocarcinomas, nine neuroendocrine tumors, seven cases of chronic pancreatitis, and one stromal tumor. Power Doppler failed to display microcirculation, whereas harmonic imaging demonstrated it in all cases. Out of 18 lesions with a hypointense signal on CEH-EUS, 16 were adenocarcinomas. The sensitivity, specificity, negative predictive value (NPV), positive predictive value (PPV), and accuracy of hypointensity for diagnosing pancreatic adenocarcinoma were 89, 88, 88, 89, and 88.5%, compared with corresponding values of 72, 100, 77, 100, and 86% for EUS-FNA. Of five adenocarcinomas with false-negative results at EUS-FNA, four had a hypointense echo signal at CEH-EUS. These pilot data are exciting and do have the potential to influence clinical management in lesions when malignancy is suspected but cannot be confirmed by FNA.

Kamisawa et al. [16•] evaluated the clinical utility of DWI in 13 patients with autoimmune pancreatitis (AIP) and 40 with pancreatic cancer. DWI offers a quantitative measurement of the diffusivity of water described by the apparent diffusion coefficient (ADC). ADC represents microcirculation of blood perfusion as well as the molecular diffusion of water. Decreased ADC values may occur in tumors as increased tumor cellularity may restrict water diffusion (malignant tumors generally have higher cellularity than benign lesions). As pancreatic cancer has been reported to have lower ADC values than normal pancreas, this study evaluated if DWI could help differentiate cancer from AIP. In addition to differing patterns of high-signal intensity areas with DWI (cancer more frequently had solitary high intensity areas with a nodular shape, whereas AIP more frequently had a longitudinal shape with different patterns: diffuse, solitary, and multiple areas), ADC values were significantly lower in AIP (1.012 ± 0.112 × 10−3 mm2/s) than in pancreatic cancer (1.249 ± 0.113 × 10−3 mm2/s) and normal pancreas (1.491 ± 0.162 × 10−3 mm2/s) (P < 0.001). After steroid therapy, high-intensity areas on DWI disappeared or were markedly decreased, and the ADC values of the reduced pancreatic lesions increased almost to the values of normal pancreas. This study offers some promise that DWI could be another tool to make the diagnosis of AIP.

Yamada et al.[17] retrospectively evaluated the ability of triple phase helical CT scanning in differentiating between chronic pancreatitis and pancreatic adenocarcinoma. A total of 172 patients were studied (42: chronic pancreatitis, 85: pancreatic adenocarcinoma, 45: normal pancreas). Two distinct protocols for CT image acquisition were used: protocol A had scan delays of 30, 60, and 150 s after administration of 300 mg I/ml of contrast vs. protocol B, with delays of 40, 70, and 150 s after 370 mg I/ml of contrast. Time attenuation curves were evaluated and in the normal pancreas, contrast enhancement peaked in the first phase, an early washout pattern. Contrast enhancement from chronic pancreatitis displayed a peak in the second phase (delayed washout pattern). In pancreatic cancer, enhancement gradually rose (increasing pattern). Protocol B performed better, with a sensitivity of 94.1% for pancreatic cancer, 83% specificity, and accuracy of 90.4%. Obtaining time attenuation curves, therefore, may help in differentiating between chronic pancreatitis and cancer, but tissue remains 'the issue' in clinical practice.

Lee et al. [18•] describe the periductal hypoechoic sign (PHS), patchy 1–2 mm hypoechoic areas adjacent to an obstructed main pancreatic duct, as seen on EUS. This is a finding they propose may help differentiate malignant vs. benign pancreatic duct obstruction at the time of EUS. PHS is seen around the main pancreatic duct upstream from the level of the obstruction. They retrospectively studied 427 patients who underwent pancreatic EUS and found that the PHS was 73.8% sensitive for pancreatic CA, with a specificity of 85.7%. Adjusting for age, patients with PHS were 17 times more likely to have a malignancy. The reason for this finding could not be correlated with pathology, but the authors postulate that the finding may be due to patchy acute pancreatitis from obstruction and as these findings are more prominent adjacent to the tumor obstruction, a pressure gradient may account for the changes. This may not occur in chronic pancreatitis due to chronicity and/or prevention of these changes due to parenchymal scarring. Alternatively, the hypoechoic areas may represent infiltration of the parenchyma with tumor cells.


Several studies addressed whether we can accurately determine cancer resectability preoperatively. Grieser et al.[19] retrospectively evaluated the accuracy of multidetector CT (MDCT) and multiplanar image reconstructions in the preoperative evaluation of pancreatic masses for resectability. In 105 patients, 70 malignant pancreatic tumors and 335 benign diseases were found. Two independent blinded readers reviewed transverse sections and additional 3D vascular reconstructions. Accuracy differentiating benign vs. malignant lesions was 93% for one and 91% for the other. In terms of arterial invasion of tumor, the receiver operating characteristics analysis averaging the results from both showed an area under the curve of 0.931 for transverse sections and 0.986 for 3D reconstructions. Ultimately, the PPV for determining nonresectability was 97% for both transverse and 3D images, NPV was 84 and 89%, respectively for the 1st reader and 86 and 91%, respectively for the 2nd. MDCT has shown itself to be highly accurate for the evaluation of resectability of pancreatic masses, although there are some limitations in the specificity of the evaluation of arterial infiltration. Although CT has been the traditional imaging choice in determining the nature and resectability of pancreatic lesions, a retrospective study by Tapper et al.[20] suggests that MRI may be a reasonable alternative. Of 124 patients treated for a radiographic diagnosis of pancreatic head adenocarcinoma, sensitivity was 100% in resectability determination, and 73.2–78.9% specific. Diagnosis was correct 96.0% of the time. MRI could detect venous and arterial involvement with 95 and 95.9% accuracy. MRI missed six metastases.

Larena-Avellaneda et al.[21] report a feasibility study of intraarterial ultrasound in pancreatic CA, to determine celiac and SMA invasion. In this case series of five patients with suspected cancer, technical success was achieved in all. One had CT evidence of vascular infiltration and two were uncertain. In those two cases, vascular ultrasound correctly predicted tumor infiltration in one and freedom of tumor in the other. Of note, one case had a short dissection of the mesenteric artery managed nonsurgically. This may become a useful tool in equivocal cases where T4 staging is uncertain.


Schellenberg et al.[22] evaluated the prognostic value of PET/CT in 55 unresectable pancreatic cancer patients undergoing stereotactic body radiotherapy. Prior to radiation therapy, PET/CT was performed and maximum standardized uptake value (SUVmax) and metabolic tumor burden (MTB) were calculated. On the basis of the median SUVmax, patients were divided into high-SUVmax and low-SUVmax subgroups as well as high-MTB and low-MTB subgroups. Median survival overall was 12.7 months, but was lower in the high SUVmax group (9.8 months) than the low SUVmax group (15.3 months, P < 0.01). The high MTB subgroup similarly had a lower survival than the low MTB subgroup (10.1 vs. 18.0 months, P < 0.01). For patients with SUVmax greater than 10, median survival was 6.4 months, at SUVmax 5.0–10.0, survival was 9.5 months, and at SUVmax less than 5, survival was 17.7 months (P < 0.01). On multivariate analysis, SUVmax was an independent predictor for overall survival (P = 0.03) and progression-free survival (P = 0.03).

In 39 patients with pancreatic cancer (14 days prior to resection), Lauenstein et al.[23] classified MRI contrast enhancement in the arterial phase as low, moderate, or high. As compared with histological grading after resection (12 poorly differentiated cancers, two poorly to moderately differentiated, 22 moderately differentiated, and three well differentiated adenocarcinomas), there was agreement between MRI arterial enhancement and grading in 30 of 39 cases.

These studies may have implications for prognostication and in the future may have implications for therapy as new therapies evolve for this devastating disease.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.