© Georg Thieme Verlag KG Stuttgart · New York
04 December 2008 (eFirst)
Histopathologic correlates of noncalcific chronic pancreatitis by EUS: a prospective tissue characterization study (Varadarajulu et al., Gastrointest Endosc 2007 )
Significant amount of attention in the literature has been focused on the question of whether endoscopic ultrasonography (EUS) is capable of reliably diagnosing chronic pancreatitis earlier than other tests (endoscopic retrograde cholangiopancreatography [ERCP] or pancreatic functional assays, i. e. secretin testing). Studies have shown variable results, where good sensitivities were sometimes contrasted by weaker specificities . EUS-imaging parameters have been linked to a scoring system to make the diagnosis of early chronic pancreatitis more or less likely , but studies correlating these criteria with histology are retrospective and mostly include only patients with severe disease or calcific pancreatitis.
In this prospective study, Varadarajulu et al. correlated EUS findings in patients diagnosed with noncalcifying chronic pancreatitis (NCCP) with histopathology of surgical specimens. A total of 42 patients undergoing EUS for pancreaticobiliary indications and subsequent pancreatic surgery were evaluated; patients with calcific pancreatitis were excluded. Of the 42 patients, 40 had neoplastic disease. NCCP was diagnosed histologically in 21 patients (50 %). None of them had chronic pancreatitis diagnosis by computed tomography (CT). Histopathology assessment was performed by a single pathologist who was blinded to EUS findings (fibrosis total score, 12; ≥ 6, unequivocal chronic pancreatitis). Receiver operating characteristic curve analysis revealed that four or more EUS criteria provided best sensitivity (90.5 %), specificity (85.7 %), and accuracy (88.1 %) for diagnosing NCCP. EUS features significantly associated with histopathologic NCCP were foci (P < 0.0001), stranding (P < 0.001), and lobulations (P = 0.04); ductal features that were significantly associated with histopathologic NCCP were a dilated (P < 0.0001) or irregular main pancreatic duct (P < 0.0001), side branches (P < 0.001), and hyperechoic duct margins (P = 0.03). There was also significant correlation between the number of EUS criteria and severity of NCCP on histology (P < 0.0001).
Previous, retrospective analyses have also supported a correlation between endosonographic and histopathologic findings in early chronic pancreatitis. For example, in a recent study, 37 cases with negative results on CT and secretin function testing were nevertheless characterized as “early chronic pancreatitis” by EUS between 1993 and 1998. After a mean of 8.5 years, signs of chronic pancreatitis were found in 67 % of these patients on CT, secretin testing, or both . In another previous study including 32 patients with normal ERCP but slightly abnormal EUS, 69 % developed signs of chronic pancreatitis during a mean follow-up period of 18 months, which included repeated pancreatic tests . Selection bias in such retrospective studies could however be substantial, as patients presenting again for assessment in retrospective studies are probably those with a higher likelihood of having the disease. Thus, the importance of the paper by Varadarajulu et al. lies in the fact that similar data were derived from a prospective study. These data support the use of EUS for early diagnosis of early noncalcifying chronic pancreatitis when other modalities are uninformative.
Despite its importance, the study also has some limitations. The most important one (acknowledged by the authors) is the small number of patients, although one must admit that it is unlikely that studies from a single center can gather greater numbers of patients with suspected mild/early chronic pancreatitis with subsequent surgical confirmation; a multicenter design might be more appropriate for such studies but is also prone to increased interobserver variability, which was also addressed in this single-center study. Here, an initial phase of 50 examinations (performed by both EUS experts) to achieve good interobserver agreement regarding EUS criteria for chronic pancreatitis preceded patient enrollment; however, how this actually augmented interobserver agreement is not further clarified. Another weakness is that the vast majority of the patients in the study had neoplastic disease, which could possibly affect EUS findings. Finally, the fact that in spite of its improvements EUS imaging is still not perfect (as proven by some false-positive and false-negative findings) is another limitation of the paper. However, these limitations do not undermine the value of the study, which offers good evidence that changes demonstrated by EUS are actually also histologically present.