Differential Diagnosis between Chronic Pancreatitis and Pancreatic Cancer: A Prospective Study of 156 Patients

Summary . The aim of this study was to evaluate prospectively the efficacy of different imaging methods in differentiating between chronic pancreatitis and cancer in focal pancreatic lesions and staging of adenocarcinoma. Material and Methods. patients were enrolled into prospective clinical The patients were randomized into two groups. Ultrasonoscopy and endoscopic ultrasound were performed in both the groups. Group A patients were additionally examined by computed tomography and tumor marker assay, while in the group B, intraoperative ultrasonoscopy with biopsy and urgent histological examination were done. Results of each imaging technique regarding differential diagnosis and cancer stage were compared with the ﬁ ndings of surgical and histological examination. and patients, The lesions and both patients older Preoperatively accuracy differential diagnosis ultrasonography in lymph node involvement distant metastases imaging the of differential and distant metastases (100%). Conclusions. In the differential diagnosis between chronic pancreatitis and adenocarcinoma, preoperative ultrasonography and intraoperative ultrasound are the best imaging methods. When ultrasonography is nondiagnostic, computed tomography and endoscopic ultrasound are alternative techniques.


Introduction
Differential diagnosis of pancreatic masses can be challenging in clinical practice because the patients with chronic pancreatitis may present with clinical symptoms and radiographic fi ndings that can be diffi cult to distinguish from those of pancreatic adenocarcinoma (1)(2)(3). Up to 6% of the cases suspected to be malignant were found to be benign at surgery; this was associated with a postsurgical complication rate of up to 21% (2). Although ultrasonography (US) and computed tomography (CT) remain to be the initial tests of choice for the differential diagnosis of pancreatic masses, improved imaging technologies, including endoscopic ultrasonography (EUS) and intraoperative ultrasound (IOUS), provide opportunities to detect early morphologic changes (4)(5)(6)(7)(8)(9). The literature data of the sensitivity of various imaging methods in the differential diagnosis of chronic pancreatitis and pancreatic adenocarcinoma is presented in Table 1 (10)(11)(12)(13)(14)(15)(16)(17). How-ever, no consensus about the best approach to assess differential diagnosis, tumor stage, and resectability has been reached so far (1,3,8,9). In fact, reliable data are limited to very few prospective studies, whereas most of noncomparative analyses only refl ect the expertise of a group in one particular technique, thus making it diffi cult to extrapolate their results (18). Finally, an appropriate gold standard based on the surgical and histopathological fi ndings is lacking in many studies.
The aim of this prospective study was to evaluate the ability of different imaging techniques to establish the diagnosis of focal pancreatic lesion in assessing the differential diagnosis between chronic pancreatitis and pancreatic cancer and staging of adenocarcinoma.

Materials and Methods
Study Design. The imaging entrance criterion for the study was the presence of a pancreatic mass on transabdominal US examination in clinically selected patients. Patients with pancreatic focal lesion detected on US were prospectively studied to compare two different protocols of diagnostic tools. Given that the gold standard for all comparisons was histopathologic and surgical fi ndings, patients judged unfi t for either curative or palliative surgery due to impaired physical condition or associated severe diseases and older than 80 years were not included. All imaging techniques as well as surgical examination were performed following a pre-established protocol. The study protocol was approved by the Lithuanian Bioethics Committee (No. 28, May 27, 2005) in accordance with the ethical guidelines of the World Medical Association Declaration of Helsinki. Written informed consent was obtained from each patient. After signing informed consent and agreement to participate in the study, the patients were randomized in two groups. In group A patients, US, EUS, CT, and tumor marker (carbohydrate antigen , carcinoembryonic antigen [CEA]) assays were performed before the surgery, while the group B patients underwent preoperative US, EUS, and IOUS, intraoperative biopsy (IOB) with urgent histological examination. After the surgery, the postoperative diagnosis (POD) was established by the operating surgeon for all the patients in both groups with adenocarcinoma stage according to the TNM staging system. The results of every diagnostic method were compared with the surgical fi ndings, which were complemented by histopathologic examination of the resected specimens (n=142). For the patients who were judged in consensus to have unresectable pancreatic adenocarcinoma, diagnosis was histologically proved by percutaneous biopsy (n=14). Specifi c technical details of each imaging technique are summarized below. The patients underwent the following diagnostic procedures at the same hospital and with the same equipment: -CA 19-9 and CEA assay was performed with a chemiluminescent microparticle immunoassay technique ( Of the 156 patients, 142 underwent surgery. A conclusive diagnosis was made in these 142 patients on histologic examination of the surgical specimens. Fourteen patients were not operated on -histological examination of the transabdominal Tru-cut biopsy specimens under US control was performed to confi rm clinical diagnosis (12 adenocarcinomas, 1 chronic pancreatitis, and 1 B-cell lymphoma). The follow-up period for the patient with chronic pancreatitis was more than 3 years. Nonoperated patients were excluded for the evaluation of pancreatic adenocarcinoma stage according to the TNM staging system. Baseline demographic characteristics of patients included into the study are shown in Table 2. Both groups (A and B) were compared by patients' demographic data; the comparison showed no signifi cant differences.
Statistical Analysis. Concerning the diagnosis obtained by imaging techniques, sensitivity, specifi city, positive and negative predictive values, and accuracy were calculated with a 2×2 table. Continuous variables were expressed as mean values and ranges, and comparative analysis between them was performed by using the Mann-Whitney test. The signifi cance level was 5% for all statistical procedures. Numerical variables were expressed as mean and standard deviation, and comparative analysis between them was performed by using the Student t test. All categorical data were expressed as absolute frequencies and percentages and were analyzed by using the chi-square and Fischer exact tests. Data were analyzed by means of the SAS 9.1 (SAS Institute Inc., Charlotte, NC, USA).

Results
The mean diameter of the lesions detected with US was 42.2 mm (range, 20-80 mm). The lesions were located in the head of the pancreas in 102 (65.4%), in the head and body in 4 (2.6%), in the body of the pancreas in 20 (12.8%), in the body and tail in 5 (3.2%), and in the tail of the pancreas in 23 patients (14.7%). In 2 patients (1.3%), the lesion involved the head, body, and tail of the pancreas. The conclusive diagnosis of patients included into the study is outlined in Table 3. Of the 156 patients enrolled into the study, 78 (50%) were histologically proven to have pancreatic adenocarcinoma, and 58 patients (37.2%) were proven to have chronic pancreatitis. Both groups (A and B) were compared by fi nal clinical diagnosis; no signifi cant differences were found. The size of the pancreatic lesions and clinical presentation were similar in patients with malignant and benign lesions (Table 4). However, the patients with cancer were signifi cantly older than the patients with chronic pancreatitis. Pancreatic adenocarcinoma was found in 78 patients. Only 1 tumor (1.3%) was staged as T1, 3 (3.8%) as T2, and 35 (44.9%) as T3. The distribution of patients with adenocarcinoma according to the TNM stating system is presented in Table 5. More than half of our patients had locally advanced or metastatic disease: 46 (59.0%) of them had stage III or IV cancer. Only 4 patients (5.1%) had early-stage cancer (I A and I B).
Sensitivity, specifi city, positive and negative predictive values, and diagnostic accuracy of every diagnostic method with respect to each evaluated parameter in both the groups separately and together were evaluated. Tumor marker CA 19-9 and CEA assays were always considered diagnostic. US, EUS, CT, IOUS, IOB with urgent histological examination, and POD were considered diagnostic in 93.6%,  *Three patients with gastrointestinal stroma tumors, one patient with B-cell lymphoma, one patient with brachiogenic cyst, and one patient with multicystic benign mesothelioma. Table 3. Conclusive Diagnosis of Patients Included in the Study 95.5%, 77.6%, 98.6%, 100%, and 100% of the patients, respectively. Data on the accuracy of various diagnostic procedures when a diagnosis was rendered are presented in Table 6. It shows the diagnostic values of differential diagnosis between pancreatic adenocarcinoma and chronic pancreatitis of US, EUS, CT, CA 19-9, CEA, IOUS, IOB with urgent histological examination, and POD. The patients with benign pancreatic tumors (n=14) and other pathology (n=6) were excluded from the analysis. US, EUS, and POD were performed in both groups. Table 6 presents diagnostic values in each group separately (group A and group B) and both groups together (group A+B). The best results in preoperative differential diagnosis of pancreatic cancer and chronic pancreatitis were achieved with EUS and US. These diagnostic methods were more accurate as compared with CT, but the differences were not signifi cant (P>0.05). IOUS showed the highest diagnostic values for differential diagnosis of pancreatic cancer and chronic pancreatitis. IOB with urgent histological examination failed to detect pancreatic adenocarcinoma in 4 patients. Differences in the POD diagnostic values comparing the groups A and B were not signifi cant (P>0.05). The mean levels of tumor markers in patients with pancreatic adenocarcinoma and chronic pancreatitis are shown in Table 7 Table  8 shows the values of diagnostic methods only for pancreatic adenocarcinoma and T, N, and M staging. EUS and US showed the best preoperative diagnostic accuracy for pancreatic cancer. The accuracy of CT was lower, but the differences were not signifi cant. IOUS was found to be the best imaging method for detection of adenocarcinoma.
CT and US showed the highest accuracy for preoperative T staging, with the differences being not signifi cant (P>0.05). IOUS yielded the highest overall accuracy of T staging. US had the best accuracy and specifi city for the preoperative evaluation of lymph node metastases (N1). The accuracy, sensitivity, and specifi city of EUS and CT were lower compared with US. The highest overall diagnostic values for staging lymph node metastases were achieved with IOUS. US had the highest preoperative accuracy and specifi city for staging metastases (M1), but IOUS showed the highest overall diagnostic values for staging metastases.

Discussion
In this prospective study, we compared two different protocols of diagnostic tools to establish clinical diagnosis of chronic pancreatitis or pancreatic adenocarcinoma and obtained summary estimates of US, EUS, CT, CA 19-9, CEA, IOUS, IOB with UH, and POD for diagnosis and staging of pancreatic cancer. The strength of this study lies in several factors. First, this was a prospective study in which the most widely available imaging techniques were compared. Second, histopathologic examination constituted the gold standard for all comparisons in a large series of patients with chronic pancreatitis and pancreatic cancer. Third, although the effi cacy of individual tests was completely evaluated with respect to each parameter, this study specially focused on the potential advantage of combining different imaging techniques. On the other hand, we are also aware of the limitations of the study. First, although special attention was paid to avoid a selection bias, it cannot be ruled out. Indeed, about 60% of the patients had advanced pancreatic cancer. Second, the unambiguous comparison among imaging techniques is hampered by both the rapid advances in technology, with signifi cant variances depending on which apparatus generation was used, and their dependence on the operator's expertise, a circumstance especially important for EUS and CT.
Although weight loss and jaundice were seen more commonly in patients with cancer, no signifi cant difference was detected in relation to the clinical presentation of both diseases. However, patients with cancer were older than those suffering from chronic pancreatitis (P<0.001). The results of the present study indicate that US is the mainstay for differential diagnosis of chronic pancreatitis and pancreatic cancer preoperatively, with the best values in the evaluation of the primary diagnosis, tumor size, and metastatic spread in malignancy. This imaging technique was proved to be a safe and most successful procedure for differential diagnosis in pancreatic lesions and cancer staging in most of the patients (Fig.). For differential diagnosis and cancer staging, the diagnostic values of CT were lower compared with US. CT was the best preoperative imaging method in determining T stage pancreatic cancer, but the differences were not signifi cant comparing it with US (P>0.05). These results deserve some comment. First, although CT was said to have potential advantages, it seems not to surpass US and EUS in differential diagnosis and pancreatic cancer staging. This fact, along with the greater availability, low cost, and noninvasiveness of US, favors the use of this procedure. Second, US have serious technical limitations, mainly because of the diffi culty of visualizing the pancreas if a large amount of intestinal gas is present. These limitations were responsible for 10 nondiagnostic US cases (6.4%) in our study. When the pancreas is well-defi ned, it is usually possible to establish the nature of the lesion. US would therefore seem important in the differential diagnosis of chronic pancreatitis and pancreatic cancer when the lesion is visualized. CT is less affected by intestinal gas. However, 17 nondiagnostic cases (22.4%) with CT were identifi ed because it was impossible to determine the exact nature of the lesion, due to the absence of specifi c criteria for diagnosis. Third, microscopic involvement of lymph nodes by tumor cells is not uncommon, even in small pancreatic cancer, thus explaining the relatively low sensitivity of all imaging techniques in its detection,    (20,21). CA 19-9 and CEA assays were not valuable for differential diagnosis of pancreatic cancer and chronic pancreatitis, but the levels of CA 19-9 were signifi cant higher in patients with advanced malignancy (T4, N1, and M1). Serum CA 19-9 level was proved to be more useful in indicating prognosis for pancreatic cancer, and this corresponds to literature data (22). IOUS gave the best overall diagnostic values both in differential diagnosis, pancreatic cancer detection and the staging of malignancy. IOB with urgent histological examination was thought to be the method of choice for intraoperative histological diagnosis of pancreatic focal lesion. The sensitivity, however, is never 100%, and a review of false-negative diagnoses shows most to be caused by sampling error and rarely because of interpretive results. In our study, UH gave four false-negative cases; all the patients had a fi nal diagnosis of pancreatic adenocarcinoma. Technological advances in body imaging will probably change the proposed approach to assess differential diagnosis and pancreatic cancer stage in upcoming years. Among others, the local application of ultrasound devices through laparoscopy, portal vein, or biliary and pancreatic ducts, combination of CT, MRI, and PET, use of elastography, contrast-enhanced color Doppler sonography, and Tru-cut needle biopsy guided by EUS show promise and await stringent evaluation. Since then, a combination of US as preoperative imaging techniques, followed by IOUS as confi rmatory test intraoperatively, seems to be the most reliable strategy for improving the overall accuracy of preoperative and intraoperative differential diagnosis of the focal pancreatic lesions and pancreatic cancer staging.

Conclusions
High-quality preoperative imaging of chronic pancreatitis and cancer of the pancreas is essential in selecting patients who are candidates for surgery. No consensus exists with regard to the optimal imaging test for evaluation of the pancreas. Analysis of these data based on histopatological fi ndings showed Intraoperative Ultrasound Surgery that ultrasonography is the most useful preoperative diagnostic technique for differential diagnosis of chronic pancreatitis and pancreatic adenocarcinoma and staging of malignancy. If the pancreas is not well visualized on ultrasonography, computed tomography should be performed, and then, if the focal lesion is not evident, endoscopic ultrasonography can also be done. Serum CA 19-9 is useful in indicating prognosis for pancreatic cancer. Patients scheduled for surgery should be assessed by using intraoperative ultrasound. Application of the Trucut biopsy technique to intraoperative ultrasound has not specifi cally addressed its role in diagnosing malignancy: intraoperative biopsy with urgent histological examination does not improve the diagnostic values of intraoperative ultrasound, but can be used in the cases, classifi ed as chronic pancreatitis.