Search

Screening for Pancreatic Neoplasia in Asymptomatic High Risk Individuals: A Prospective, Controlled Study Marcia Canto, Michael Goggins, Sanjay Jagannath, Sergey Kantsevoy, Constance Griffin, Jennifer Axilbund, Kieran Brune, Gloria Petersen, Elliot Fishman, Charles Yeo, Ralph Hruban, Francis Giardiello, Anthony Kalloo Individuals from familial pancreatic cancer (PC) kindreds and persons with Peutz Jeghers syndrome (PJS) are at increased risk for PC. EUS and ERCP abnormalities associated with chronic pancreatitis (CP have been described in relatives from familial PC kindreds.AIM: To screen for early pancreatic neoplasia and to evaluate the radiologic pancreatic abnormalities in high-risk individuals compared to controls. METHODS:We prospectively evaluated high risk individuals with either: 1) > 1 first-degree relative with PC from kindreds with familial PC with > 2 affected members, or 2) Peutz-Jeghers syndrome. Each patient had screening with a dual phase, thin section CT scan and EUS, each interpreted in a blinded fashion. If EUS was abnormal, EUS-guided fine needle aspiration (FNA) and ERCP were performed. ERCP was performed by an endoscopist unaware of the clinical and radiologic findings. EUS and ERCP findings of high risk subjects were compared with those of normal controls. RESULTS:We studied 63 high risk patients (mean age 57 yrs, 47%male); 58 were from familial PC kindreds and 5 had Peutz-Jeghers syndrome. We also evaluated 130 normal controls (mean age 56 yrs, 44% male). No patients had symptoms referable to the pancreas, a suspicion of pancreatic cancer, or a history of pancreatitis. In the high risk group, a 1.5 cm pancreatic mass was detected in a PJS patient by both EUS and CT, which was surgically resected and found to be an intraductal papillary mucinous neoplasm with carcinoma-in-situ. Screening also detected 2 asymptomatic non-pancreatic cancers in the high risk group: a renal cell carcinoma (EUS and CT) and an ovarian cancer (CT only); both cancers were treated operatively. EUS showed CP (3 or more of 9 features) in 72% of high risk patients, which was significantly more common than in controls (14%, p

Lesions found at CT scan or ERCP that are indeterminate for cancer (CA) are a diagnostic dilemma. AIM: To compare the diagnostic yield, accuracy, and long-term outcome of EUS/FNA for the diagnosis of indeterminate pancreatico-biliary (PB) lesions. METHODS: Patients (pts)with indeterminate lesions (mass or fullness by CT scan and/or ductal stricture at ERCP with no or atypical symptoms) were prospectively evaluated by EUS/FNA and compared to negative and positive controls. Negative controls had a low risk of CA (normal CT with symptoms, abnormal CT, or stricture without symptoms). Positive controls had a high risk of CA (mass ± invasion or metastasis by CT, ductal stricture(s), pain, jaundice, weight loss, and/or anorexia). EUS/FNA was repeated for indeterminate lesions if initially nondiagnostic or surgery not performed. Final diagnoses was determined by FNA, surgery, or minimum follow-up of 6 months. RESULTS: 200 pts (21 low, 171 indeterminate, 8 high) were studied over 3 years. Follow-up was available on all but 9 pts. Mean follow-up was 14 months (range 7-37). 97 pts had EUS/FNA and 71 pts had surgery. 0%, 30%, and 88% of low, indeterminate, and high risk lesions, resp, were diagnosed by EUS/FNA as tumor (52 adenoCA, 4 islet cell, 1 lymphoma). The sensitivity, specificity, and accuracy of EUS for diagnosis of tumor in all pts was 96%, 95%, and 96%, resp. EUS/FNA provided a histologic diagnosis of CA in 3/5 and 3/6 pts with nondiagnostic surgery/surgical biopsies and percutaneous FNA, resp. EUS led to a change in diagnosis or CA stage, diagnostic or treatment plan in 86% of pts with indeterminate lesions. 32 patients had 1-3 repeat EUS/FNA for persistent lesions, which led to diagnosis of 3 more CA and showed stability, improvement, or complete resolution of all benign masses. Pseudotumors were due to focal pancreatitis and pancreas divisum. Compared to CT, EUS was more likely to visualize a tumor (65% vs 17%, resp, p =.004) and diagnose earlier stage, resectable indeterminate lesions (41% vs 10% Stage I, resp, p=.02). Significantly more patients with CA diagnosed by EUS alone were alive at follow-up (62% vs. 24%, p=.001). CONCLUSION: EUS/FNA is the diagnostic modality of choice for indeterminate PB lesions. It accurately differentiates tumor from pseudotumor, diagnoses early lesions, impacts on patient care, and improves outcome. EUS/FNA can diagnose CA when percutaneous biopsy and surgical exploration are nondiagnostic. Repeat EUS/FNA improves diagnostic yield.

EUS/FNA is an accurate method of diagnosing pancreatico-biliary cancer (CA) but missed lesions have been reported. AIMS: 1) To evaluate the factors associated with non-diagnostic EUS/FNA and missed CA. 2) To determine the accuracy of EUS/FNA in diagnosing CA in chronic pancreatitis (CP) and the yield of adding FNA to EUS. METHODS: The history, CT scan, cytology, and pathology results were prospectively evaluated in patients (pts) with a suspicion of a pancreatico-biliary CA. The EUS diagnoses and technical aspects of FNA were recorded. EUS/FNA was performed with a 22-gauge needle and on-site cytologist. Final diagnoses were determined by surgery, repeat EUS/FNA (21 pts), or clinical follow-up. RESULTS: 232 pts underwent EUS, 129 had EUS/FNA, and 79 had surgery. 81 tumors were diagnosed (71 adenoCA, 1 acinar cell CA, 6 islet cell, 1 lymphoma). 1 CA with no mass by CT or EUS was found by random EUS/FNA passes. The EUS false negative rate for tumor was 3.7% (96% sensitivity). There were 2 missed adenoCA without a mass and 1 islet cell tumor misdiagnosed as pancreas divisum and CP. All were not evident by prior radiologic tests. 1 adenoCA was missed by surgical exploration and diagnosed only by random biopsies at a second operation. The EUS false positive rate was 9% (96% specificity) due to focal CP (n=5), primary sclerosing cholangitis (n=1), and pancreatic hamartoma (n=1). FNA gave a correct histologic diagnosis in all but 9 adenoCA (84% sensitivity). 3 of these had prior nondiagnostic CT-guided FNA. 7 were unresectable. 6 had firm or hard masses and all but 1 had inadequate specimens despite 6-12 passes. 1 mass was not seen well during EUS/FNA. 1 neuroendocrine tumor was misdiagnosed by EUS/FNA as an adenoCA (98% specificity). Nondiagnostic EUS/FNA for CA was not significantly associated with acute or chronic pancreatitis, pancreas divisum, mass location, tumor differentiation, or difficulty penetrating the lesion. 48 pts had CP;38% had a mass at EUS (12 with CA). 27% of EUS/FNA specimens from tumors in CP were inadequate. CP lowered the sensitivity of EUS/FNA for CA(87% to 80%). Although specificity was high (97%), FNA did not improve EUS accuracy for CA in CP (92%) or in its absence. CONCLUSION: EUS alone is highly accurate for the diagnosis of CA. FNA provides a specific histologic diagnosis but does not increase EUS accuracy in CP. Cancers missed by EUS are infiltrating lesions without a distinct mass and FNA should be performed if suspicion is high.

Thermal therapy is the cornerstone of endoscopic treatment of mucosal disease of the gastrointestinal tract.We have recently described results of animal studies with a novel endoscopic device that delivers a freezing spray based on the Joules-Thompson effect (Gastrointes Endosc 1999;49:627-31). This device is simple, portable, inexpensive and uses compressed gas. We report the results of a pilot study using an improved version of the device to treat mucosal vascular lesions in humans. AIM: To determine the safety and efficacy of endoscopic cryotherapy to treat patients with mucosal vascular lesions and recurrent gastrointestinal bleeding. METHODS: Patients with recurrent GI bleeding from mucosal vascular lesions such as AVM's, watermelon stomach and radiation gastritis/proctitis who had failed conventional thermal therapy were recruited. Cryotherapy was applied to affected areas in sessions until all areas were treated. Follow-up endoscopy was performed 2-4 weeks after the final session in all patients. Serial hematocrits and blood transfusions requirements following therapy were recorded. RESULTS: 8 patients were recruited with gastric AVM's (n=5), watermelon stomach(n=1), radiation gastritis(n=1), radiation proctitis (n=1). There mean number of sessions for therapy was 3.25(mean range of 1-6 sessions). 6 patients had no further bleeding after a mean follow of 6.7±4.6 weeks. Follow-up endoscopy in these patients revealed normally regenerated mucosa. 2 patients continued to have bleeding despite cryotherapy, one had a watermelon stomach with refractory post-chemotherapy thromboctopenia and the second patient had radiation gastritis and maked debilitation from metastatic cancer. There were no complications noted in any of the 8 patients enrolled. CONCLUSIONS: This pilot study demonstrates that cryotherapy is safe and effective treatment for recurrent gastrointestinal bleeding from mucosal vascular lesions. Further controlled trials are needed.

Because of the frequency of colorectal disease and the effectiveness of early intervention in the polyp-cancer sequence, surveillance for this problem should be a routine part of health-maintenance examinations. The general use of flexible sigmoidoscopy in conjunction with fecal occult blood screening is an extremely important part of healthcare delivery. Flexible sigmoidoscopy can easily be learned by most physicians. If use of the procedure becomes widespread, the effects of colorectal disease, in terms of both patient suffering and healthcare costs, can be reduced.