CT and MR Imaging of Pancreatic Cancer Part 2

 

2.3     Differential Diagnosis
2.3.1    Solid Tumors of the Pancreas
The accuracy of CT for the differentiation of solid tumors of the pancreas varies widely and can be explained by different techniques, different study populations, and the degree of awareness of the investigators assessing the images. In addition, it seems to be of crucial importance whether or not the investigators assessing the lesions have knowledge of the clinical picture (Lemke et al.
2004). Catalano et al. (2003) studied a total of 46
patients and reported a sensitivity of 97.0% and a specificity of 80.0%. In contrast, the sensitiv¬ity was 76.6% and the specificity was 63.9% in a study with 100 patients examined by Lemke et al. (2004) without clinical information. Most of the studies published so far do not break down the lesions according to size, but several studies have demonstrated a low specificity of CT in pa-tients with lesions with a diameter of less than 2 cm (Taoka et al. 1999; Baum et al. 1999; Freeny
1999).


2.3.1.1 Magnetic Resonance Imaging
The vast majority of pancreatic adenocarcinomas are generally slightly hypointense relative to the pancreas on T2 weighted images. However, the tumors are difficult to visualize on plain MR im-ages unless there is substantial necrosis.
When compared with normal pancreatic tis-sue, pancreatic adenocarcinomas enhance to a lesser extent than normal pancreatic tissue. This effect is often transient and may be best visual-ized on early post-contrast images during the ar-terial phase after bolus injection of nephrotropic contrast media. Many tumors, especially when substantial necrosis is apparent, demonstrate a more or less thin rim of greater enhancing pan-creatic tissue and may underline the focal nature of a pancreatic lesion. Vascular encasement due to the typical perivascular growth of this malig-nant tumor is equivocal or better delineated by MRI when compared with dynamic contrast-
enhanced CT (Martin and Semelka 2000). MR
angiography with fat suppression is superior to other sequences in delineating regional vascular encasement or occlusion as well as regional vas-cular anatomy.
MRCP is the method of choice to visualize the characteristic features of pancreatic head adeno-carcinoma including encasement and obstruc¬tion of the pancreatic duct or bile duct. The well-known "double duct sign" occurs in 77% of the cases, while biliary duct stenosis alone may be observed in about 9% as well as pancreatic duct stenosis alone (about 12%) (Fulcher et al. 1998). While the detection of even small peripancreatic lymph nodes is now possible on MRI, the accu¬rate differentiation between malignant and reac¬tive enlarged lymph nodes is still a diagnostic challenge that remains unresolved until today. In contrast, MR using gadolinium DTPA has greater accuracy in the detection and character¬ization of liver metastasis compared with helical
CT (Martin and Semelka 2000; Freeny 1999).

2.3.2   Cystic Pancreas Tumor
Cystic tumors of the pancreas are rare but must be considered in every patient with cystic lesions of the pancreas. In contrast to solid tumors of the pancreas where adenocarcinomas are predomi-nant, there is a wide variation of histologic find-ings, dignity, and prognosis of cystic tumors. In addition, known entities have been better clas-sified as new ones are described. For example, many pancreatic neoplasms, including those pre-viously termed papillary carcinomas, ductectatic mucinous cystadenomas, villous adenomas, and mucin-producing tumors of the pancreas, are now classified as intraductal papillary mucinous neoplasms (IPMN) of the pancreas. Once con-sidered a rare tumor, now they are increasingly recognized at CT and MR imaging. Therefore, radiologic differentiation between benign and malignant lesions is important in the determina-tion of the appropriate treatment.
The presence of mural nodules, mural thick-ening, and contrast enhancement is suggestive of malignancy; however, the absence of mural nodules or enhancement does not indicate that the tumor is benign (Fig. 2.5). A maximum main pancreatic duct diameter of greater than 15 mm and diffuse dilatation of the main pancreatic duct are suggestive of malignancy in main duct-

type tumors. Among branch duct-type tumors, malignant tumors tend to be larger than benign tumors; however, this finding is variable. The presence of main pancreatic duct dilatation may be helpful in determining malignancy of branch duct-type tumors.
The treatment decision with regard to cystic neoplasms often is based on the patient's age at presentation, the lesion location, the sex and the presence or absence of symptoms, and malig¬nant features (Kosmahl et al. 2004; Sahani et al.
2006).
In contrast to other cystic neoplasms, IPMN have a communication to the pancreatic duct but should be differentiated from pseudocysts by CT or MRI. MRI using the MRCP technique may show whether a pancreatic cystic lesion com-municates with the main pancreatic duct (MPD) and demonstrate the extent of ductal involve¬ment (Irie et al. 2000; Sugiyama et al. 1998). Due to improvements in CT technology including image post-processing such as curved reforma¬tion, the capability of CT for the evaluation of the pancreatic parenchyma and the pancreatic ducts in patients with IPMN is enhanced (Fig. 2.1).

2.4     Resectability
Resectability of pancreatic cancer usually is as-sessed according to presence of infiltration of adjacent tissue or vessels. In some studies, lymph node involvement and distant metasta-ses are used as additional parameters, resulting in a strong influence of the results of the stud¬ies (Bluemke et al. 1995; Warshaw et al. 1990). CT has still shown discouraging results with lymph node assessment in pancreatic cancer. De¬spite advances in CT techniques, differentiation based on a morphologic parameter is not suffi¬cient because of the small size (<1.2 cm) of many malignant lymph nodes, which is next to the commonly used cutoff size (>1 cm) used to dif-ferentiate between benign and malignant lymph nodes. (Robinson and Sheridan 2000; Taoka et al. 1999; Freeny et al. 1993).
The grading of tumor involvement of the portal and superior mesenteric veins and the celiac, hepatic, and superior mesenteric arteries based on circumferential contiguity of tumor to vessel as proposed by Lu et al. (1997) is a prac-ticable tool for daily clinical practice (Fig. 2.6). The splenic artery has not usually been consid-ered as critical for surgical resection. Based on a five-point scale (grade 0, no contiguity of tumor to vessel; grade 1, tumor contiguity of less than one-quarter circumference; grade 2, between one-quarter and one-half circumference; grade 3, between one-half and three-quarters circum-ference; and grade 4, greater than three-quarters circumferential involvement or any vessel con-striction), Lu et al. obtained a sensitivity of 84%



Fig. 2.6a-e All-in-one MRI in a patient with irresectable adenocarcinoma of the pancreas. a Contrast-enhanced T1 weighted images with fat suppression demonstrates the necrotic center of the tumor located in the pancreatic corpus but fails in the determination of the extension of the tumor. b MRCP shows the typical double duct sign, indicating that the tumor involves the pancreatic head. c Arterial MR angiography (single slice) shows aberrant origin of the hepatic artery from the superior mesenteric artery (SMA). The shifting of the SMA indicates as well a large tumor of the head. d Augmented view of arterial MR angiography: note the small hypodense structure over the liver artery and between celiac trunk and SMA, which was confirmed as a malignant infiltration intraoperatively. e Portal venous MR angiography (3D reconstruction) shows only a minor reduction of the vessel caliber around the confluens region



and a specificity of 98% for unresectability when a threshold between group 2 and group 3 was chosen. With regard to the sensitivity, specificity, and accuracy in the detection of vascular infiltra-tion, Furukawa et al. (1998) had similar results (83.0%, 100.0%, and 89.0%). However, this score only provides a statistical probability whether a patient is resectable or not and may not substi-tute for surgical exploration in many cases today (Varadhachary et al. 2006).
When compared with MRI, CT still faces the problem of low contrast between the lesion and the surrounding tissue. Although some publica-tions demonstrate that combined arterial and ve-nous phase CT scanning will detect even small lesions, the probability of detection is reduced.
For MRI, the positive and negative predictive values for cancer nonresectability of unenhanced and contrast-enhanced MR were 90% and 83%, respectively, and the accuracy, sensitivity, and specificity were reported to be 85%, 69%, and 95%, respectively (Lopez Hanninen et al. 2002). Malignant encasement of the vessels may be op-timally visualized using contrast-enhanced dy-namic images.

2.5     Conclusions
In conclusion, a combination of several phases, or at least an arterial or pancreatic parenchymal phase and a portal venous phase, is essential for an optimal multiphasic CT protocol for the com-prehensive evaluation of pancreatic adenocarci-nomas. The addition of multiplanar and curved reformations may increase the sensitivity of CT and improves its agreement with surgical find¬ings. Beyond abdominal MR imaging, techniques such as magnetic resonance cholangiopancreati-cography (MRCP) and MR angiography should be integrated in the imaging protocol whenever possible.
Cross-sectional imaging of the pancreas en-ables a reliable detection rate of pancreatic tu-mors and is useful for the differentiation between benign and malignant lesions. Vascular infil-tration as a main predictor of resectability may be visualized or excluded in most of the cases. While the differentiation of lymph node involve¬ment is still an unsolved diagnostic challenge in imaging of pancreatic cancer, distant metastases in the liver and other organs may be detected ac¬curately by both imaging methods.