Sensitivity and specificity of sonographically guided fine-needle puncture (FNP) in the diagnosis of focal lesions of the pancreas

A476 AGA ABSTRACTS G1936 SENSITIVITY AND SPECIFICITY OF SONOGRAPHICALLY GUIDED FINE-NEEDLE PUNCTURE (FNP) IN THE DIAGNOSIS OF FOCAL LESIONS OF THE PANCREAS. W. Kratzer, A. Schmidt, S. Hagel, M. Orth, D. Mtiller, G. Adler, K. Beckh. Dept. of Internal Medicine I, University of Ulm, Germany. Introduction: Studies of FNP in the diagnosis of focal lesions of the pancreas have reported a sensitivity of 57-93% and a specificity of 95-100% for the method. The objective of the present investigation was to evaluate the sensitivity and specificity of FNP in the diagnosis of focal lesions of the pancreas. Patients and Methods: Between May, 1995 and June, 1996 we performed FNP on 25 patients (14 females, 11 males; average age 61.4 + 10.91 years, range 31-77 years). In nine cases, confirmation of findings was surgical, in the remaining 16 patients based on a one-year clinical follow-up period. Results: Aspiration of diagnostically adequate material was possible in 18 cases (72%) (malignancy, 10; chronic pancreatitis, 4; benign lesions, 4). For the other seven patients (28%), clinical follow-up revealed pancreatic carcinoma in two patients, pancreatitis in four patients, while in one case even the surgical specimen obtained for histologic examination failed to provide a diagnosis. For patients with adequate cytology findings, one-year follow-up revealed false-negative findings in three cases: three of four patients with chronic pancreatitis developed carcinoma. Discussion.: FNP in the diagnosis of focal pancreatic lesions showed a sensitivity of 77% and a specificity of 100%. There were no complications observed in our group. Conclusion: Negative findings at FNP in chronic pancreatitis does not exclude the possibility of carcinoma. In the case of equivocal findings, criteria for surgery should be broad. • G1937 DIRECT EVIDENCE FOR PREMATURE PROTEASE ACTIVATION IN LIVING PANCREATIC ACINAR CELLS. B. Krtiger, W. Tessenow, Department of Medical Biology, Universitiit Rostock; M.M. Lerch, Department of Medicine B, Universit~tt Mtinster, Germany. Whether, when and why a premature and intrapancreatic protease activation might occur during the course of acute pancreatitis has remained controversial. We have developed an assay that allows observation of and localize the intracellular activation of serine-proteases directly in living, functionally intact pancreatic acinar ceils. Dispersed acini from rat pancreas were prepared by collagenase digestion, adjusted to a biovolume of 1-2mm3/ml, and incubated in oxygenated (pH 7.5) cell culture medium. After an equilibration of 30 rain the secretagogue caerulein was added at different concentrations, from submazimai (10pM) to supramaximal (I~M), for up to 60 rain. Acini were then transferred to 96-well microtiter plates and the cell-permeable and specific serine protease substrate CBZ-Ile-Pro-Arg (101aM), bis-substituted with the fluorochrome rhodamine110, was added. Readings from intact acini were taken on a fluorescence cytofluorometer (excitation 485nm, emission 530nm). Intracellular protease activation was observed directly when the substrate was added at the start of the incubation and acini were placed under a fluorescence microscope (excitation 485nm) Incubation with increasing concentrations of caerulein resulted in a time and dose dependent protease activation in living acini together with a parallel decline in physiological amylase secretion. Maximal activation was achieved after 40rain with lpM caerulein. Cleavage of the fluorogenic substrate was completely abolished by preincubation with lmM of the cell permeable serine-protease inhibitor Pefabloc SC. Protease activation could also be observed microscopically because substrate cleavage resulted in bright fluorescence in membrane-confined vesicles in the apical portion of the cell and was, again, abolished by preincubation with the protease inhibitor. To characterize these vesicles we prepared subcellular fractions by densitygradient-centrifugation in Percoll. Protease activity was abundant in the 1.08g/ml fraction containing partially condensed cytoplasmic vesicles on EM but was almost absent in the zymogen granule fraction (1A5g/ml). The above technique allowed, for the first time, observation and localization of intracellular activation of proteases in living acinar cells. As opposed to measuring proteolytic activity in pancreatic lysates this novel tool detects activation directly where it occurs and does not entail the contact of catalytic enzymes from different subcellular compartments that are known to crossactivate each other. • G1938 COMPARISON OF ENDOSCOPIC ULTRASOUND AND SECRETIN TESTING FOR THE DIAGNOSIS OF MINIMAL CHANGE CHRONIC PANCREATITIS. L.R. Lambiase. B. Pollack, S. Amann, C.E. Forsmark. Tulane University School of Medicine, New Orleans, LA and the University of Florida, Gainesville, FL. Introduction: The diagnosis of chronic pancreatitis (CP) can be particularly difficult in its early stages, when obvious structural changes are not present. Hormonal stimulation tests have traditionally been used to diagnose this group of patients. Secretin test (ST) is a hormonal function test that has been shown to correlate positively with histologic changes of chronic pancreatitis.

GASTROENTEROLOGY Vol. 114, No. 4

Endoscopic ultrasonography (EUS) allows much better definition of pancreatic parenchyma than standard imaging studies, and some have suggested this technique is able to diagnose CP in its early stages. No direct comparisons exist comparing EUS to standard hormonal stimulation testing. Aim: To compare EUS with ST in patients with suspected CP without easily identifiable structural changes. Methods: Eleven non-alcoholic patients with abominal pain and normal conventional imaging studies (CT and ERCP) underwent both EUS and ST. The EUS was performed with the Olympus GF-UM20 and the pancreas was imaged in the head, body, and tail. The EUS results were read by examiners who were blinded to the results of ST. As previously published, EUS criteria for CP included focal decrease in echogenicity of pancreatic parenchyma, stones, accentuated lohular pattern, cysts, ductal irregularity, ductal dilation, or increased echogenicity of duct wall. Greater than three criteria is considered to be diagnostic of CP. ST was performed as previously described and considered abnormal if the peak (HCOj) was < 80 mEq/L in each of the four- 15 minute collection periods. Results: 6/11 patients had CP by ST. Two patients were positive for CP by EUS. All patients had at least 1 EUS finding but none had greater than 3 findings. The table compares the results of EUS and ST: EUS+ EUS-

ST + 2 4

ST 2 3

Assuming that ST is the gold standard for diagnosing CP in this group of patients the sensitivity of EUS in diagnosing minimal change CP was 20% while the specificity was 66% Conclusions: ST and EUS correlated poorly in this group of patients. Assuming that ST is the gold standard in diagnosing minimal change CP, EUS performs poorly in this group of patients with normal conventional imaging. Perhaps the criteria for diagnosing CP by EUS need modification. G1939 RISK FACTORS FOR DEVELOPING CHRONIC PANCREATITIS AFTER A FIRST ATTACK OF ACUTE ALCOHOLIC PANCREATITIS. P. G. Lankisch. C. Assmus, D. Pflichthofer, Ltineburg, Germany; P. Maisonneuve, Milan, Italy; A. B. Lowenfels, Valhalla, N.Y. According to the Marseille classification, chronic pancreatitis rarely develops in patients with acute pancreatitis. We studied this problem in patients admitted to the Ltineburg County Hospital. Patients and Methods. We studied 220 patients admitted with a first attack of acute pancreatitis from 1988 to 1995. The diagnosis was based on characteristic signs and symptoms and on contrast-enhanced computed tomography (CT) obtained within 72 h after admission. Alcohol abuse was the etiology in 69 (31%) patients with acute pancreatitis. The diagnosis of chronic pancreatitis was based on an abnormal secretin-pancreozymin test (SPT) and on at least one abnormal imaging procedure (ultrasound, CT, ERCP, endoscopic ultrasound). Univariate und multivariate (Cox proportional hazards model) analyses were performed. Results. Sixteen (23%) of the 69 alcoholics, hut only 3 (2%) of patients with non-alcoholic acute pancreatitis developed chronic pancreatitis (p < 0.001). In univariate analysis, admission on a weekend or a holiday was associated with a marginal (p < 0.20) increased risk of developing chronic pancreatitis, whereas a milder course (no operation, no artificial ventilation) was associated with alcoholic acute pancreatitis. None of 14 patients who either needed surgery or ventilatory support developed chronic pancreatitis, whereas 16 (42%) of 39 unoperated, non-ventilated patients developed chronic pancreatitis (p=0.01; follow-up 3-8 yrs). Based on these two variables, a prediction of progression from alcoholic acute pancreatitis to chronic pancrcatitis was possible with a sensitivity and specificity of 84% and 81%, respectively. Conclusion. Alcohol is a risk factor for devloping chronic pancreatitis. In patients with alcoholic acute pancreatitis, mild attacks predicted development of chronic pancreatitis. This research was partially funded by Solvay Pharmaceuticals GmbH, Hannover, Germany. G1940 WHICH ETIOLOGY CAUSES THE MOST SEVERE ACUTE PANCREATITIS? P. G. Lankisch, C. Assmus, D. Pfiichthofer, K. Struckmann, Ltineburg, Germany; D. Lehnick, GOttingen, Germany The question which etiology leads to the most severe acute pancreatitis is still open. Patients and Methods. This study involved 208 patients admitted with a first attack of acute pancreatitis from 1988 to 1995. As parameters for the severity of the disease, days spent in intensive care unit (ICU), total hospital stay (THS), amylase and lipase levels, Ranson, lmrie and Balthazar (contrastenhanced computed tomography within 72 h after admission) scores, indication for artificial ventilation, dialysis or surgery, development of pancreatic pseudocysts, and mortality, were compared within the different etiological groups.