15151 Background: Phase I and Phase II trials of induction Cisplatin (C) and Irinotecan (I) followed by concurrent C+I and radiotherapy (RT) have produced encouraging clinical and pathologic response rates in esophageal cancer. However, to date published data detailing clinical outcome with this regimen is limited. Methods: A retrospective review of 17 patients with locally advanced esophageal cancer was performed to characterize toxicity and outcomes for this regimen in our institutional experience. Patients were treated during the period from August 2003 to July 2006 and staged comprehensively with EUS, CT and PET. Induction therapy with weekly I (65mg/m2) and C (30mg/m2) was administered (wks1,2,4,5). Those without progression received weekly I (65mg/m2) and C (30mg/m2) (wks 8,9,11,12) concurrent with RT. Patients considered surgical candidates based on pre- treatment clinical stage, medical co-morbidities, and pre-operative re-staging were offered an esophagectomy. Results: The patient demographics were as follows: 15 men, 2 women, median age 63 (45–76), white 53%, black 47%, adenocarcinoma 59%, squamous cell carcinoma 41%, 47% stage T3N1, 94% node positive, 6% M1a, and 29% T4 primary tumors. The mean RT dose was 50.4 Gy and median number of elapsed days of RT was 43 days. All patients completed induction C+I. Grade 3–4 hematologic toxicity (CTC v3.0)neutropenia (47%), anemia (6%), thrombocytopenia(18%). One patient died of a TE fistula 22 days after completing therapy. Two patients (12%) progressed during therapy. An esophagectomy was performed on 41% (7/17) of patients; 6 of 7 were R0, 71% (5/7) were downstaged, 14% (1/7) pCR. At median follow-up of 15 months, 8/17 (47%) pt are alive NED, 3/17 are alive with disease, 3/17 dead of disease and 3/17 are dead of other causes. Median DFS is 7 months and median local RFS is 7 months. Median OS has not been reached at 15 months. Conclusions: In this population with very advanced disease, I+C+RT have produced promising OS and local control outcomes with greater toxicity than in previously reported experiences with this regimen. No significant financial relationships to disclose.

Background and Study Aims: During pancreatobiliary imaging by endoscopic ultrasound (EUS) at the authors' institution, it is customary to attempt to obtain the "stack sign", where the bile duct and the pancreatic duct can be seen to run in parallel through the pancreatic head, We suspected that such a view may not be attainable in patients with pancreas divisum because of the short ventral pancreatic duct, The aim of the study was to investigate whether the presence of pancreas divisum could be suspected on the basis of EUS findings,Patients and Methods: The stack sign is obtained by positioning the echo endoscope in the long scope position with the transducer in the duodenal bulb. The balloon is then inflated and advanced snugly into the apex of the bulb, From this position, the bile duct (closest to the transducer) and the pancreatic duct can be seen to run in parallel through the pancreatic head. We attempted to obtain a stack sign during EUS examinations of six patients with pancreas divisum, EUS; was done in these patients to look for evidence of chronic pancreatitis and the pancreas divisum was confirmed by endoscopic retrograde pancreatography, An attempt to obtain the stack sign was also made in 30 patients who had EUS for pancreatobiliary indications but did not have pancreas divisum,Results: In only two out of six patients with pancreas divisum (33 %) were we able to obtain a stack sign, This was significantly different from the rate of observation of a stack sign in 83.3 % (25/30) of patients who did not have pancreas divisum (P=0.04), Of the two patients with pancreas divisum in whom a stack sign was seen, the ventral duct was markedly dilated (6.6 mm) in one, and the other patient had an unusually large ventral pancreas,Conclusions: The absence of a stack sign during pancreatobiliary imaging by EUS may suggest the diagnosis of pancreas divisum,

Apart from dedicated fellowships, endoscopic ultrasound (EUS) training comprises ”informal” methods, including EUS courses and the use of animal models. We have tried to determine the usefulness of a ”hands-on” experience with a live animal model as a teaching strategy for EUS training. In 1997 and 2000 the American Society for Gastrointestinal Endoscopy (ASGE) sponsored hands-on EUS workshops using a live porcine model and participants in these courses were asked to complete a questionnaire about the course (2 years after the 1997 workshop and immediately after the 2000 workshop). The main outcome measurements were the usefulness and quality of the overall course and of the hands-on component in particular, and the effect on EUS practice patterns. Twenty of 38 attendees (53 %) responded to the survey administered 2 years after the 1997 course: 95 % thought the overall course was useful and 85 % valued the hands-on portion; 60 % of respondents had either started performing EUS or, if already experienced in EUS, had increased the number of EUS examinations and/or the complexity of EUS procedures they performed. Of the 34 participants who were at the 2000 ASGE course, over 90 % thought that the course enhanced their EUS skills and 88 % believed that they would be likely to perform EUS in the future. The study was limited in that it did not account for other methods of EUS training that could affect EUS practice patterns and also by the fact that not all the 1997 course participants responded to the questionnaire. In conclusion, live animal models may be a useful adjuctive method for learning EUS skills for the novice or the less experienced endosonographer.

Background: Outcomes and complications of endoscopic management of D/AA have not been recently described. Aim: To determine the short and long term outcome and complication rates of endoscopic snare removal of D/AA. Methods: A retrospective review of our EUS and endoscopy databases was performed. Between 1999 and 2004, 52 patients (pts) (25M/27F, mean age = 57.9 ± 14.2 years, range 27-80) including 12 FAP pts with adenomas of the duodenal wall and/or major papilla underwent endoscopic snare excision (ESE). Follow-up (fup) was performed with endoscopy at 6-8 weeks after ESE and thereafter repeated every 2-3 months until no residual lesion was present. When no lesion was present in the 1st endoscopy after ESE, fup was recommended in 1 year. Results: Complete removal of adenoma on 1st endoscopic treatment was achieved in 39/46 (85%) pts. Histopathology revealed 27 (51.9%) tubular, 19 (36.5%) tubulovillous, 3 (5.8%) villous adenomas, and 3 (5.8%) invasive carcinomas. Complications occurred in 9/52 (17.3%) pts. Immediate complications included mild bleeding (4/52, 7.7%), successfully managed endoscopically and mild pancreatitis (4/35, 11.4%). The presence (n=27) or absence (n=8) of a pancreatic stent was not related with post procedural pancreatitis (3/27 vs. 1/8, p=NS). Delayed bleeding occurred in 3/52 (5.7%) pts. During a mean fup 12.4 months, 7/46 (15.2%) pts had recurrence [4/11 (36.7%) with FAP and 3/35 (8.6%) of sporadic cases] after a median period of 10.7 months. All recurrences were managed successfully with endoscopic therapy (1.5 sessions; range 1-6). Only one patient was referred for surgery (1 of the 3 pts with invasive carcinoma; the other two are on fup, with no evidence of recurrence). Conclusion: Endoscopic snare excision, especially single enbloc resection, is a safe and effective modality for the management of adenomas of D/AA.

OBJECTIVES:The recurrence of disease after the complete resection of early stage non-small cell lung cancer (NSCLC) indicates that undetected metastases were present at the time of surgery. Quantitative real-time reverse transcriptase-polymerase chain reaction (RT-PCR) is a highly sensitive technique for detecting rare gene transcripts that may indicate the presence of cancer cells, and endoscopic ultrasound (EUS)-guided fine-needle aspiration (FNA) is a minimally invasive technique for the nonoperative sampling of mediastinal lymph nodes. The aim of this study was to determine whether these two techniques could enhance the preoperative detection of occult metastases. METHODS:Patients with NSCLC were evaluated with chest CT and positron emission tomography scans. Those patients without evidence of metastases (87 patients) underwent EUS-guided FNA. Lymph nodes from levels 2, 4, 5, 7, 8, and 9 were sampled and evaluated by standard cytopathology and real-time RT-PCR. Normal control FNA specimens were obtained from patients without cancer who were undergoing EUS for benign disease (17 control specimens). For each sample, messenger RNA was extracted and real-time RT-PCR was used to quantitate the expression of six lung cancer-associated genes (ie, CEA, CK19, KS1/4, lunx, muc1, and PDEF) relative to the expression of an internal control gene (beta(2)-microglobulin). RESULTS:Clinical thresholds of marker positivity were set at 100% specificity, as determined by the receiver operating characteristic curve analysis. Of the cytology-positive lymph nodes (27 lymph nodes), the expression of the KS1/4 gene was above its respective clinical threshold in 25 of 27 samples (93%), making this the most sensitive marker for the detection of metastatic NSCLC. At least one of the six lung cancer-associated genes was overexpressed in 18 of 61 cytology-negative patients (30%), of which KS1/4 was overexpressed in 15 of 61 patients (25%). CONCLUSIONS:Based on the high accuracy of EUS-guided FNA/RT-PCR, we predict that some of the patients in the cytology-negative/marker-positive category will have high NSCLC recurrence rates. Among the genes used in our marker panel, KS1/4 appears particularly useful for the detection of overt or occult metastatic disease.

The utility of EUS in staging of lung cancer and evaluation of unexplained mediastinal adenopathy is well established. There are regions which are missed by EUS which may be accessed via the trachea during bronchoscopic US. (EBUS). We have shown that combined procedures can be done in a timely fashion, safely, and with significant yield. However, coordinating the presence of a pulmonologist and a gastroenterologist can be difficult. Aim: To develop a training program to include EUS and EBUS-FNA. Methods: The decision to train the EUS fellows and GI attendings was based upon the endoscopic experience, the established referral pattern, and the ability of EUS to allow screening for metastatic disease. Three attendings and 3 fellows participated in the training. Each fellow trainee was required to perform 25 bronchoscopic US FNAs using a model under the tutelage of the EBUS-FNA certified pulmonologist. The trainees in fellowship were also required to complete 25 linear array bronchoscopic FNAs using a phantom model. Every trainee was required to observe 5 bronchoscopies before performing their first live exam. Results: Each of the fellows completed the model examinations and FNAs in a manner deemed acceptable by the pulmonologist. This was done in 2 separate afternoon sessions. One attending has finished the observation period and is performing supervised EBUS. Conclusion: It is feasible to provide training in EUS/EBUS. EUS fellows and GI endosonographers appear to have a short learning time for EBUS-FNA. Whether appropriate credentialing will be achievable remains to be seen.

Background: EUS plays an important role in local staging of rectal cancer and may help in the evaluation of perineal Crohn's disease. The role of EUS-FNA in the evaluation of pelvic masses however has not been well defined. Aim: To evaluate the diagnostic yield, safety and accuracy of EUS-FNA in patients (pts) with pelvic lesions of unknown etiology. Methods: A retrospective review of our EUS database identified 24 pts (14 women; mean age 58 years, range 28-83) referred for EUS-FNA to assess pelvic masses extrinsic to the rectum or distal colon over a 10-year period (1994-2004). Pts underwent radial followed by linear array EUS; FNA utilized a 22G needle. An on-site cytopathologist reviewed all samples and classified specimens as benign (inflammatory/infectious) or malignant. EUS-FNA diagnostic accuracy was confirmed by surgical pathology or clinical follow-up. Results: FNA was trans-rectal (n = 17), trans-sigmoid (n = 6) or through the descending colon (n = 1); samples were considered adequate for diagnosis in 21/24 (87.5%) pts. There were 12/21 (57%) EUS-FNA diagnoses of malignancy (n = 11) or atypical cells (n = 1) and 9/21 (43%) of benign lesions. Mean number of needle passes was 3.9 for benign vs. 3.2 for malignant lesions (range 1-7). In the 21 pts with adequate specimens, EUS-FNA accurately predicted the presence of malignant disease (sensitivity 92%, specificity 100%, PPV 100%, NPV 88%, accuracy 95%), with only one false negative diagnosis. One infectious complication occurred that was successfully managed with medical therapy. Conclusions: EUS-FNA is an accurate and safe method to diagnose pelvic masses. Prophylactic antibiotics should be used to reduce infectious complications.

Background: The safety, accuracy and cost-effectiveness of endoscopic ultrasound-guided fine needle aspiration (EUS-FNA) has been established for the evaluation of mediastinal lymph nodes for staging of lung carcinoma (ca). However, little data exists on the role of EUS-FNA in the primary diagnosis of lung masses accessible through the esophagus. The objectives of this study were to assess the utility of on-site cytologic assessment (OCA) of these masses in patients with suspected lung ca and to evaluate the feasibility and safety of this procedure in such patients. Methods: A single-center computerized search for EUS-FNA of lung masses over an 8-year period (1996-2004) was undertaken. Linear EUS-FNA was performed using a 22-gauge needle. OCA was performed by an experienced cytopathologist utilizing microscopic evaluation of air-dried Diff-Quik®-stained slides. Diagnostic accuracy of OCA was assessed and compared to the final cytologic diagnosis, subsequent histology or clinical history (hx). Follow-up to evaluate for adverse patient effects was 6 weeks. Results: Fifteen patients (12 males, 3 females; mean age 62.7 yrs, range 20-80 yrs) underwent EUS-FNA of a lung mass. Mass size ranged from 20 × 24 mm to 45 × 47 mm. A mean of 4 passes (range 1-5) provided adequate cytologic sample in 100% of the cases. OCA showed malignancy in 12 cases: primary (1°) non-small cell lung ca, 10; metastases, 2 (colonic, 1 and uterine ca, 1) and negative, 3 cases. Final cytologic diagnosis was the same. Histologically, one of the three negative cases was squamous cell ca (false-negative). Sensitivity and specificity of OCA was 92% and 100% respectively. OCA also staged 7/10 cases of 1° lung ca (including one case with liver metastasis) with 100% accuracy as determined by subsequent histology and clinical hx. No complications occurred. Conclusions: EUS-FNA of lung masses is a safe and useful procedure for obtaining diagnostic tissue. OCA provided an accurate preliminary diagnosis in most cases and at the same time accurately staged some of the 1° malignant tumors of the lung, allowing for appropriate patient triage.

Esophageal adenocarcinoma (EA) is increasing faster than any other cancer in the U.S. In this report, we first show that EA can be distinguished from normal esophagus (NE) and esophageal squamous cell carcinoma by plotting expression values for EpCam, TFF1, and SBEM in three-dimensional Euclidean space. For monitoring progression of Barrett's esophagus (BE) to EA, we developed a highly sensitive assay for limited quantities of tissue whereby 50 ng of RNA are first converted to cDNA using 16 gene-specific primers. Using a set of training tissues, we developed a novel quantitative three-tiered algorithm that allows for accurate (overall accuracy = 61/63, 97%) discrimination of BE versus EA tissues using only three genes. The gene used in the first tier of the algorithm is TSPAN: samples not diagnosed as BE or EA by TSPAN in the first tier are then subjected to a second-tier analysis using ECGF1, followed by a third-tier analysis using SPARC. Addition of TFF1 and SBEM to the first tier (i.e., a five-gene marker panel) increases the overall accuracy of the assay to 98% (62/63) and results in mean molecular diagnostic scores (+/- SD) that are significantly different between EA and BE samples (3.19 +/- 1.07 versus -2.74 +/- 1.73, respectively). Our results suggest that relatively few genes can be used to monitor progression of BE to EA.

Purpose: To assess the risk of colon cancer among individuals with diabetes within a nationally representative population. Previous studies have yielded evidence that diabetes mellitus may be a risk factor for colon cancer. Both hyperinsulinemia and hyperglycemia have been noted in vitro to be promoters of colon cancer growth. Furthermore, insulin and insulin-like growth factor-1 (IGF-1) receptors have been found on colon cancer tissue. Also, high levels of circulating IGF-1 are associated with an elevated risk of colorectal adenomas and cancer. This study further considers the relationship between diabetes mellitus and colon cancer. Methods: Data collected by the 1997–2003 National Health Interview Survey (NHIS) was analyzed to assess the risk of colon cancer among individuals with diabetes. The NHIS is a comprehensive nationally representative survey weighted to represent the U.S. adult population. There were 226,953 subjects represented in the combined seven years of the NHIS, of which 161,235 had complete and comparable data which could be evaluated. Multiple logistic regression was performed probing the relationship between diabetes mellitus and colon cancer while controlling for age, race, gender, obesity, alcohol use, tobacco use, and physical activity. SAS statistical software was used to account for the complex survey design of the NHIS. Results: Among the 161,235 subjects in this study, 16,664 (10.3%) revealed a history of diabetes mellitus. Adjusted for potential confounders, individuals with diabetes were siginficantly more likely to have colon cancer than persons without diabetes (Odds ratio (OR) = 1.59, 95% Confidence interval (CI) = 1.3 to 1.9). Age > 50 years (OR = 14.48, 95% CI = 10.8 to 19.4), caucasians (OR = 1.91, 95% CI = 1.1 to 3.2) and smokers (OR = 1.55, 95% CI = 1.2 to 2.0) were also revealed to be significant risk factors. Conclusions: Caution should be taken with interpretation of this analysis, given the cross-sectional nature of this design. Commonly accepted risk factors for colon cancer including inflammatory bowel disease and family history were not clearly defined in the database. Furthermore, methods of treatment or duration of diabetes were not included in this analysis. Nevertheless, the results from this large cross-sectional survey analysis confirm and expand on previous research which indicates that diabetes mellitus is a significant risk factor for colon cancer.

Background: The finding of “dilated duct(s) of unknown cause” in asymptomatic individuals usually prompts further evaluation. EUS has been advocated as a useful tool to rule out an obstructive lesion, but there is insufficient data to fully evaluate the utility of EUS in this clinical scenario. Aim: To determine the diagnostic yield and accuracy of EUS in patients with asymptomatic biliopancreatic ductal dilation and normal liver enzymes. Methods: A retrospective review of our EUS database identified over a 10-year period (1994-2004) 30 asymptomatic patients with normal liver biochemistry referred for EUS to evaluate bile duct (CBD) and/or pancreatic duct (PD) dilation detected by US, CT or MRI. Mean age: 69 ± 9 (range 50-84); sex: 26F/4M. 18/30 (60%) had isolated CBD dilation, 4/30 (13%) isolated PD dilation, and 8/30 (27%) CBD and PD dilation. 10/30 (33%) patients had a previous cholecystectomy. EUS diagnoses were confirmed with pathology (FNA or surgery) or extended clinical/imaging follow-up. Results: Mean duct diameters at EUS: CBD = 12.2 mm (7-22), PD = 4.4 mm (2-9). Follow-up: Patients with significant EUS findings (n = 4); pancreatic masses, died of cancer (1), on chemoradiotherapy (1); IPMN (1) and focal narrowing of PD (1) alive, respectively, after 42 and 11 months of follow-up; EUS without significant findings (n = 26); died of unrelated causes (3), lost for follow-up (3), and asymptomatic (20) after a mean follow-up of 22 ± 24 months (2-102). Conclusion: The incidental finding of isolated bile duct dilatation in asymptomatic individuals with normal liver enzymes may not warrant further investigation, but EUS accurately rules out significant pathology. The finding of unexplained pancreatic duct dilation should not be ignored and EUS accurately defines the pathology.

Background: The diagnosis of early chronic pancreatitis (CP) can be difficult. Endoscopic ultrasound (EUS) has been proposed as a non-invasive and accurate test for CP, with good correlation compared to ERCP, the most commonly used test. However, there is lack of comparative data between EUS findings and the gold standard of histopathology. Aims: The aim of this study was to compare EUS criteria for CP with histopathology from surgical specimens. Methods: We retrospectively identified all patients with partial pancreatectomies for presumed CP between 1995-2001 from the MUSC EUS database.

Background: Complete pharyngoesophageal disconnect is a rare complication of radiation therapy for head and neck malignancies. Surgery has been the mainstay of treatment for this condition. Lew et al (Head and Neck, Feb 2004) described a combined endoscopic/laryngoscopic approach to establishing luminal patency using a stiff guidewire. However, repeated dilation may be required to achieve maximum diameter, and the restenosis rate for radiation-induced strictures has been high. Mitomycin-C has been used with success in the prevention of recurrent laryngeal and tracheal stenosis and is thought to have an antiproliferative effect. We describe a novel one-step method of reconnection, dilation, and treatment for pharyngoesophageal disconnect. Methods: Retrograde endoscopy through an existing gastrostomy tube tract is performed to the level of the disconnected proximal esophagus. Direct laryngoscopy is performed at the same time to visualize the light source below. An EUS needle (Echotip, Wilson Cook Corp) is advanced by the laryngoscopist with endoscopic guidance. A .035 mm ERCP guidewire (Jagwire, Microvasive corp) is passed through the EUS needle into the esophagus and grasped with a forcep. The wire is pulled back thru the gastrostomy and fixed in place. Savary style dilators are then passed per os, serially from the smallest to a minimum of 51 Fr. Before removing the guidewire, mitomycin-C pledgets are applied via laryngoscopy. Results: Four patients with pharyngoesophageal disconnect after radiation therapy for head and neck malignancies underwent the combined procedure. Luminal patency was achieved in all patients, and all were able to progress to teaching for self dilation. One patient suffered a minimal localized esophageal perforation for which conservative management was used. Luminal patency was maintained in all patients with follow-up to 7 months after the initial procedure. Conclusions: This combined one-step approach offers a less invasive and technically superior option to surgery. Dilation in one setting to 51 Fr can be performed with minimal complications. The EUS needle allows for a controlled puncture and passage of a guidewire. Despite their prior disconnect, these patients can be taught to perform self-dilation to prevent or delay recurrence without forming false tracts. Mitomycin-C topical treatments may provide an antiproliferative effect for radiation-induced strictures, however further long term studies are warranted.

BACKGROUND AND STUDY AIMS:Endoscopic ultrasound-guided fine-needle aspiration (EUS-FNA) is a minimally invasive and highly accurate method of detecting mediastinal lymph-node metastases in gastrointestinal and lung cancer. Little information is available regarding the use of EUS-FNA to stage tumors in the head and neck region. This study reports experience with EUS in the diagnosis and staging of these tumors and their mediastinal spread.PATIENTS AND METHODS:The records of patients who underwent EUS for diagnosis and/or staging of head and neck tumors were reviewed. Referral criteria were suspected invasion of the esophagus by a lower-neck mass on cervical computed tomography (CT) or magnetic resonance imaging (MRI), or mediastinal lymphadenopathy > 10 mm on a chest CT.RESULTS:Thirty-two patients (23 men, nine women; mean age 65 years, range 44 - 80) were referred and underwent 35 EUS examinations. In one patient, EUS was not possible due to a benign esophageal stricture. In 17 patients with suspected esophageal invasion on CT scans, EUS demonstrated invasion of the esophagus in four cases and of the pleura in one; 12 tumors showed no visible invasion of adjacent structures. The other 17 examinations were carried out for suspected mediastinal metastatic disease. In eight cases, EUS-FNA confirmed metastatic disease, whereas only benign changes were shown in the other nine cases. EUS-FNA also provided the first tissue diagnosis in two primary tumors and identified malignancy in one patient with no CT suspicion of positive mediastinal lymph nodes. EUS avoided the need for more invasive investigations in all patients with mediastinal lymphadenopathy, and it changed the management in 12 of the 17 patients (71 %) with suspected esophageal invasion and in eight of the 17 patients (47 %) with suspected mediastinal disease.CONCLUSIONS:EUS with FNA provides a viable approach to the diagnosis and staging of tumors in the head and neck region when there is a suggestion of esophageal invasion on CT or MRI, or enlarged mediastinal lymph nodes. EUS with FNA may avoid the need for mediastinoscopy or other more invasive techniques for staging of these neoplasms.

BACKGROUND:The accuracy, the safety, and the cost-effectiveness of EUS-guided FNA for screening patients with lung cancer for mediastinal metastasis are well established, but the utility of EUS-guided FNA in evaluating lung mass per se has not been investigated. This study retrospectively evaluated experience with EUS-guided FNA of lung mass lesions after unsuccessful attempts by CT-guided or bronchoscopic tissue sampling to establish a tissue diagnosis.METHODS:A database was searched for all patients who had EUS-guided FNA of lung mass lesions over a 3-year period. The diagnostic yield and safety of EUS-guided FNA were evaluated.OBSERVATIONS:Eighteen patients (11 men, 7 women) underwent EUS-guided FNA of lung mass lesions adjacent to or abutting the esophagus. The indication for EUS-guided FNA was evaluation of the mediastinum of patients with lung mass of unclear etiology. EUS-guided FNA yielded tissue for diagnosis in 100% of patients: 15 non-small-cell lung cancer, one small-cell lung cancer, two metastatic lung disease. Ten patients had unresectable disease; in 8, the mass was confined to the lung parenchyma. The mean number of needle passes required to establish a diagnosis was two (range 1-6). No complication was encountered (mean follow-up 141 days; range 72-396 days). Five patients underwent curative surgery, and 13 had palliative chemoradiation.CONCLUSIONS:In this study, EUS-guided FNA of lung mass was safe, and it established a diagnosis in all patients with accessible lesions. Given these preliminary data, a prospective evaluation of this new indication for EUS-guided FNA is justified.

BACKGROUND:Preliminary data show that endosonography guided fine needle aspiration (EUS-FNA) may be an accurate method for diagnosing sarcoidosis. However, these data were obtained in a small selected group of patients with a very high pretest probability of sarcoidosis. This retrospective study reports on the use of EUS-FNA in an unselected group of patients with mediastinal lymphadenopathy of unknown origin.METHODS:The EUS database of a single tertiary referral centre was reviewed for patients who underwent EUS-FNA for mediastinal lymphadenopathy of unknown origin. Clinical presentation and imaging studies of each case were carefully reviewed and the diagnosis "sarcoidosis" or "no sarcoidosis" attributed if possible. The diagnoses were compared with the result of EUS-FNA.RESULTS:One hundred and twenty four patients were investigated. In 35 cases EUS-FNA identified granulomas (group 1); in the other 89 cases (group 2) no granulomas were detected. The definite diagnoses in group 1 were sarcoidosis (n = 25), indefinite (n = 7), no sarcoidosis (n = 3). The definite diagnoses in group 2 were sarcoidosis (n = 3), indefinite (n = 9), no sarcoidosis (n = 77). Of the 77 cases with no sarcoidosis, 44 were diagnosed with other diseases. The other 33 showed non-specific changes in the FNA and sarcoidosis was excluded by negative non-EUS pathology (n = 17) and clinical presentation. The sensitivity and specificity for EUS-FNA were 89% (95% CI 82 to 94) and 96% (95% CI 91 to 98), respectively, after exclusion of the indefinite cases in both groups.CONCLUSIONS:EUS-FNA is an accurate method for diagnosing sarcoidosis in an unselected group of patients with mediastinal lymphadenopathy. The reported sensitivity and specificity must be appreciated in the context of the difficult and often incomplete clinical diagnosis of sarcoidosis.

BACKGROUND:EUS with FNA is useful for staging non-small-cell lung cancer. However, benign mediastinal adenopathy is common. The aims of this study were to identify clinical factors, especially primary tumor location, and EUS lymph nodal characteristics predictive of aortopulmonary window and subcarinal lymph node metastases of non-small-cell lung cancer.METHODS:Patients with known or suspected non-small-cell lung cancer underwent EUS staging at which EUS-FNA was performed for all identified mediastinal lymph nodes. Clinical characteristics, primary tumor data, EUS findings, and histopathology were reviewed. Exact tests were performed for both aortopulmonary window and subcarinal lymph nodes to identify factors predictive of malignant cytology.RESULTS:Ninety-two patients with non-small-cell lung cancer were included. Fifty-one had aortopulmonary window, and 73 had subcarinal lymph nodes on EUS. The EUS with FNA specimens were interpreted as suspicious or diagnostic for malignancy for 9 aortopulmonary window and 9 subcarinal lymph nodes. When comparing benign vs. malignant EUS with FNA findings for aortopulmonary window and subcarinal lymph nodes, only lymph node size of 1 cm or greater and sharp lymph nodal edges were associated with malignancy in lymph nodes at both sites, whereas primary tumor site, lymph node shape, and echogenicity were associated with malignant subcarinal nodes. When 4 classic lymph nodal features of malignancy were evaluated, the presence of 3 or more typical features had positive and negative predictive values of, respectively, 41% and 96%.CONCLUSIONS:Although tumor location and EUS lymph nodal characteristics are associated with malignant involvement of lymph nodes, the accuracy of these predictors does not obviate the need for cytologic evaluation. EUS with FNA should be performed for all lymph nodes when an abnormal finding will alter management.

Background Increasingly, EUS is being used to stage lung cancer. Direct mediastinal invasion (T4) by lung cancer is stage IIIb disease. Patients in this stage have a 5-year survival of less than 5% and generally are offered chemotherapy without surgery. This study evaluated the accuracy of EUS in detecting T4 lung cancer. Methods The study included all patients with lung cancer who had EUS staging and subsequent staging at surgery, or for whom there was unequivocal confirmation of unresectability (T4) by thoracoscopy, thoracotomy or presence of malignant pleural effusion, or definite invasion of great vessels/adjacent organs on CT. Results A total of 175 of 308 patients with lung cancer who underwent EUS over a 5-year period (1997-2002) had subsequent confirmatory tumor staging. Ten patients were found by EUS to have stage T4 tumors; 7 were confirmed to be T4 by either surgical exploration (2), CT demonstration of aortic invasion (3), or documentation of malignant pleural effusion (2). Three of the 10 (30%) patients found to have stage T4 tumors by EUS had T2 disease at surgery and underwent curative resection. Of the remaining 165 patients without evidence of T4 disease at EUS, only one was found to have aortic invasion (T4) at surgery. EUS had a sensitivity of 87.5%, specificity of 98%, positive predictive value of 70%, and a negative predictive value of 99% for detecting T4 disease. Conclusions Caution is warranted when unresectability of lung cancer is based solely on tumor invasion into mediastinal soft tissue at EUS. Overstaging occurs when a tumor appears to invade the pleural layer without mediastinal organ invasion. Confirmation of unresectability by other diagnostic modalities is warranted in such instances.