F. Gress

Endoscopic Ultrasonography–Guided Fine-Needle Aspiration Biopsy of Suspected Pancreatic Cancer

Pancreatic mass lesions discovered on abdominal imaging in older patients are most often caused by pancreatic adenocarcinoma. Each year in the United States, carcinoma of the pancreas is diagnosed in more than 29 000 persons, and nearly as many die of the disease. Pancreatic cancer is the fourth leading cause of cancer-related death in both men and women (1). When a patient is already symptomatic at diagnosis, survival is poor. Nevertheless, a variety of benign or malignant neoplasms can arise in the exocrine and endocrine tissues of the pancreas. Because this has important therapeutic implications, a tissue diagnosis is paramount (2). Pancreatic mass lesions can pose vexing problems to the clinician. Obviously, patients who have a neoplasm with a more favorable prognosis than adenocarcinoma must be identified quickly and accurately. For patients with resectable adenocarcinoma, surgery may provide the best chance for a cure, but palliative care should also be considered. Even for patients who are not surgical candidates, a tissue diagnosis needs to be established to guide further management. At many institutions, computed tomography (CT)guided percutaneous fine-needle aspiration (FNA) has become the biopsy procedure of choice for pancreatic mass lesions. However, this method yields a considerable false-negative rate, approximately 20% (3). Furthermore, tumor seeding after transcutaneous biopsies has been reported (4). Endoscopic retrograde cholangiopancreatography (ERCP) sampling of malignant-appearing stricture is fraught with problems. Even in experienced hands and with use of multimethod sampling, the false-negative rate is higher than 30% (5). Endoscopic ultrasonographyguided FNA biopsy of the pancreas has recently been shown to be a feasible, accurate, and safe alternative (6-8). No studies are currently investigating the role of endoscopic ultrasonographyguided FNA biopsy of pancreatic masses as a secondary method of biopsy. In this paper, we report our experience at Indiana University (Indianapolis, Indiana) with endoscopic ultrasonographyguided FNA biopsy of pancreatic mass lesions in patients who had negative results on previous CT-guided FNA biopsy or ERCP sampling. Methods All patients referred to Indiana University Medical Center for further evaluation of suspected pancreatic cancer were eligible for this study. However, only patients who had negative results on either CT-guided FNA biopsy or ERCP with sampling performed at Indiana University were prospectively identified and consecutively enrolled. Our study, which was approved by the institutional review board of Indiana University, consisted of the following participant groups: 1) Patients who had negative ERCP sampling and subsequently underwent CT-guided biopsy, 2) patients with negative results on CT-guided biopsy who subsequently underwent ERCP with sampling, and 3) patients with negative results on CT-guided biopsy who did not have sampling at ERCP. After giving written, informed consent to undergo the procedure and participate in the study, patients were consecutively scheduled for endoscopic ultrasonography with endoscopic ultrasonographyguided FNA biopsy. The endosonographer was aware of the patients previous biopsy results. Computed tomography of the pancreas, bile ducts, and the region of the papilla of Vater was performed by using helical (spiral) CT (High Speed Advantage, General Electric, New York, New York) with a 3- to 5-mm collimation and a pitch between 1 and 2. All participating radiologists were well experienced in abdominal CT and CT-guided FNA biopsy. Experienced endoscopists performed the ERCP tissue sampling by using a combination of brush cytology, forceps, and FNA methods (9). Endoscopic ultrasonography was performed by using radial scanning endosonography (EUM-20, Olympus America, Inc., Melville, New York) or linear-array endosonography (FG32UA, Pentax Instrument Corp., Orangeburg, New York). A single experienced endosonographer performed all endoscopic ultrasonography examinations and obtained the biopsy specimens. (Endoscopic ultrasonographyguided FNA biopsies of pancreatic mass lesions can be performed competently by gastroenterologists who have appropriate training in endoscopic ultrasonography and endoscopic ultrasonographyguided FNA.) After identification of a pancreatic mass lesion, the lesion was targeted for biopsy by using linear-array endosonography. Fine-needle aspiration biopsy was performed by using either a 23-gauge, 4-cm needle (Wilson-Cook, Winston Salem, North Carolina) or a 22-gauge, 10-cm needle (GIP/Mediglobe, Tempe, Arizona). A transgastric biopsy approach was used for mass lesions in the body and tail regions of the pancreas, and a transduodenal approach was used for lesions in the head region. Samples from endoscopic ultrasonographyguided FNA biopsy were prepared in the same manner as samples from CT-guided biopsy and were interpreted by the same cytopathologist. Data were collected on the size of the lesion, the number of needle passes, the diagnosis rendered by the cytopathologist, complications, surgical pathologic findings, and ultimate outcome on follow-up. Patients underwent surgery if findings on endoscopic ultrasonographyguided FNA biopsy were positive for cancer, the mass was thought to be resectable, and the surgical risks were acceptable to the patient and surgeon. Some patients with negative biopsy results had surgery on the basis of a strong clinical suspicion that the biopsy result was false negative. Patients with negative biopsy results who did not have surgery were followed clinically and, if indicated, underwent periodic imaging studies. Data for our analysis included cytologic results on FNA biopsy and pathologic findings on surgery. For patients who had negative findings on cytologic examination of the FNA biopsy sample and declined surgery, we also used observations at long-term clinical follow-up. A similar composite gold standard has been used in other clinical studies (6). Pancreatic cancer was defined by the following criteria: 1) The histologic examination of the surgical specimen revealed pancreatic cancer; 2) the results of the FNA cytology of a pancreatic mass lesion, visualized by and biopsied under the guidance of endoscopic ultrasonography, disclosed cells that appeared malignant; 3) clinical follow-up, including imaging studies, was consistent with pancreatic cancer; or 4) when these criteria did not apply, the investigators determined by consensus that the clinical course was consistent with pancreatic cancer. Surgical pathologic findings were determined from a review of the resected specimen by a pathologist with particular competence in pathologicanatomical diagnosis, and disease stage was determined from the features of the resected specimen. Clinical follow-up data were gathered from repeated office visits with a medical oncologist or gastroenterologist and repeated imaging studies, such as chest radiography, abdominal CT, and repeated ERCP for stent exchanges in patients with obstructive jaundice. Follow-up studies were individualized by the patients needs. At the end of the study period, the investigators (all of whom are clinicians experienced in the diagnosis, treatment, and follow-up of patients with pancreatic cancer) reviewed the cases in which diagnosis was made on clinical grounds alone. Decisions on the presence or absence of pancreatic cancer at the end of the follow-up period were unanimous. Likelihood ratios for positive and negative test results were calculated with 95% CIs (9, 10). Posterior probabilities of pancreatic cancer after positive and negative test results were calculated by applying Bayes theorem on the basis of the prevalence of pancreatic cancer in our study group (11). We used SAS statistical software, version 6.12 (SAS Institute, Inc., Cary, North Carolina), to calculate the confidence intervals. Results From August 1992 to December 1996, 102 patients (58 men and 44 women) with pancreatic mass lesions who had negative cytologic results by the primary sampling method were enrolled. The mean and median ages, which were similar for men and women, were 63.6 years and 65 years, respectively. The median size (SD) of all mass lesions was 3.5 1.2 cm for the long axis and 2.7 0.9 cm for the short axis. The correlation coefficient between the long-axis and short-axis diameter was 0.80 (P�<�0.001). Most of the masses were located in the pancreatic head (83%). The size of the mass lesions in the head region and the size elsewhere were not significantly different. The median number of passes of the biopsy needle into the mass on endoscopic ultrasonographyguided FNA biopsy was 3.4 (range, 2 to 9). Among the 102 participants, all of whom had negative cytologic results on CT or ERCP sampling, 57 patients had positive cytologic findings on endoscopic ultrasonographyguided FNA biopsy. Results on endoscopic ultrasonographyguided FNA biopsy were negative in 37 patients and inconclusive or nondiagnostic in 8 patients; of these 45 patients, 4 were ultimately found to have adenocarcinoma of the pancreas at surgery (Figure). Of the 102 total patients, 30 underwent surgery, and the remainder had clinical and imaging follow-up for a median of 24 months (range, 5 to 48 months). By June 1998, all of the 57 patients with positive results on FNA biopsy had died of pancreatic cancer. By contrast, as of the writing of this manuscript, 41 of the 45 patients with negative, inconclusive, or nondiagnostic results on FNA biopsy had no evidence of pancreatic malignancy. Two of the 4 nonsurviving patients in this group died of causes unrelated to pancreatic cancer. Figure. Diagnostic results on computed tomographyguided percutaneous fine-needle aspiration biopsy ( CT ) or endoscopic retrograde cholangiopancreatography with sampling ( ERCP ), endoscopic ultrasonographyguided fine-needle aspiration biopsy, surgery, and follow-up. The 95% CI for the likelihood rat

Endoscopic ultrasound guided fine needle aspiration biopsy versus mediastinoscopy in the staging of non-small cell lung cancer: a cost effective analysis

Endoscopic ultrasound guided fine needle aspiration biopsy versus mediastinoscopy in the staging of non-small cell lung cancer: a cost effective analysis