Significant Predictors Of Malignancy Research Articles

Prediction of Carcinoma After Resection in Subjects With Ampullary Adenomas on Endoscopic Biopsy

The endoscopic treatment of ampullary adenomas is established; however, the false-negative rate of endoscopic biopsy for carcinoma is 20% to 30%, and it remains uncertain whether identifiable features predict malignancy. Our aim in this study was to evaluate the predictable factors of malignancy in ampullary adenomas on endoscopic biopsy. Ninety-one subjects diagnosed with ampullary adenoma on endoscopic biopsy were confirmed after endoscopic or surgical resection of ampullary lesions between 1995 and 2011 respectively. Clinical, laboratory, radiologic, and endoscopic findings were compared between patients with adenoma and carcinoma after resection. We examined the predictors of malignancy in ampullary adenoma on endoscopic biopsy. The malignancy rate in ampullary adenomas on endoscopic biopsy was 26.4%. Univariate analysis revealed that presence of symptoms, villous components, high-grade dysplasia (HGD), papilla enlargement on computed tomography, duct dilatation on radiologic imaging, bilirubin>2 mg/dL, aspartate aminotransferase>40 IU/L, alanine aminotransferase>40 IU/L, and alkaline phosphatase>90 U/L were associated with malignancy in patients over 65 years of age. HGD [odds ratio, 6.86 (95% confidence interval, 1.58-29.79)] and ductal dilatation [odds ratio, 11.12 (95% confidence interval, 2.27-54.37)] were independently associated with malignancy in multivariate analysis. The sensitivity and negative predictive value for ≥1 risk factors were 95.83% and 96.77%, respectively. The presence of 2 risk factors resulted in a high specificity (96%) and positive predictive value (84%) for malignancy. HGD and ductal dilatation are significant predictors of malignancy in ampullary adenomas. When these risk factors are present, precautions should be taken in the consideration of malignancy in patients with ampullary adenoma.

Differentiation of transitional zone prostate cancer from benign hyperplasia nodules: Evaluation of discriminant criteria at multiparametric MRI

To assess topographical and magnetic resonance imaging (MRI) features in characterizing prostate transitional zone (TZ) nodules. Two radiologists evaluated all TZ nodules visible at multiparametric MRI in 52 consecutive patients who underwent radical prostatectomy. The radiologists assessed topographical (anteroposterior and superior-inferior location, crossing of the sagittal midline) and T2-weighted (shape, presence and distinctness of capsule, distinctness of contours, presence of cysts) features, the apparent diffusion coefficient (ADC), and eight semi-quantitative and quantitative enhancement parameters derived from dynamic contrast-enhanced (DCE) imaging. The nature of the nodules was assessed using prostatectomy specimens. Five statistical methods taking into account multiple testing were used. One hundred and thirty-seven nodules (117 benign, 20 malignant) were evaluated. Mean ADC, all topographical, and all T2-weighted features were significant predictors of malignancy according to at least four out of the five statistical methods. Particularly, 20/20 and 18/20 cancers involved the anterior and apical third of the TZ, respectively. None of the enhancement parameters was significantly different between cancers and benign nodules. By assessing the presence of cysts, the nodules' capsule, and their anteroposterior and superior-inferior location, 111/117 benign nodules were correctly diagnosed, without misclassifying any cancer. Topographical, T2-weighted, and diffusion-weighted features can be used to characterize TZ nodules. DCE imaging does not seem to provide additional information.

Isolated and unexplained dilation of the common bile duct on computed tomography scanscans

Isolated dilation of common bile duct (CBD) with normal sized pancreatic duct and without identifiable stones or mass lesion (unexplained) is frequently encountered by computed tomography/magnetic resonance imaging. We studied the final diagnoses in these patients and tried to elucidate factors that can predict a malignant etiology. This is a retrospective analysis of prospective database from a University based clinical practice (2002- 2008). We included 107 consecutive patients who underwent endoscopic ultrasound (EUS) for evaluation of isolated and unexplained CBD dilation noted on contrast computed tomography scans. EUS examination was performed using a radial echoendoscope followed by a linear echoechoendoscope, if a focal mass lesion was identified. Fine-needle aspirates were assessed immediately by an attending cytopathologist. Main outcome measurements included i) prevalence of neoplasms, CBD stones and chronic pancreatitis and ii) performance characteristics of EUS/EUS-fine needle aspiration (EUS-FNA). A malignant neoplasm was found in 16 patients (14.9%) of the study subjects, all with obstructive jaundice (ObJ). Six patients had CBD stones; three with ObJ and three with abnormal liver function tests. EUS findings suggestive of chronic pancreatitis were identified in 27 patients. EUSFNA had 97.3% accuracy (94.1% in subset with ObJ) with a sensitivity of 81.2% and specificity of 100% for diagnosing malignancy. Presence of ObJ and older patient age were only significant predictors of malignancy in our cohort. Amongst patients with isolated and unexplained dilation of CBD, the risk of malignancy is significantly higher in older patients presenting with ObJ. EUS-FNA can diagnose malignancy in these patients with high accuracy besides identifying other potential etiologies including missed CBD stones and chronic pancreatitis.

High-Spatial-Resolution 3-T Breast MRI of Nonmasslike Enhancement Lesions: An Analysis of Their Features as Significant Predictors of Malignancy

The purpose of this study was to analyze the features of nonmasslike enhancement detected on 3-T MRI and to determine which of these features are significant predictors of malignancy. Retrospective review was performed of 124 consecutive patients with nonmasslike enhancement detected on 3-T MRI after biopsy or surgery. We described nonmasslike enhancement using the descriptors in the BI-RADS MRI lexicon. In addition to the BI-RADS descriptors, whether clustered ring enhancement was present and whether surrounding high signal intensity (SI) was present on T2-weighted imaging were assessed. Cancer was identified in 85 lesions (69%). Of these lesions, ductal carcinoma in situ (DCIS) was found in 41 (48%) and invasive cancer in 44 (52%). The features found to be significant predictors of malignancy were segmental (p = 0.001), focal (p = 0.006), dendritic (p = 0.017), and clustered ring enhancement (p = 0.026) and surrounding high SI on T2-weighted imaging (p < 0.0001). The features found to be significant predictors of invasive cancer were dendritic enhancement (p < 0.0001) and surrounding high SI on T2-weighted imaging (p < 0.0001). There were no significant predictive features for DCIS. Homogeneous enhancement was found to be a significant predictor of benignancy (p = 0.001). Kinetic patterns were not significant predictors of malignancy. Nonmasslike enhancement of 1 cm or larger was more often malignant than lesions smaller than 1 cm (p < 0.0001). In multivariate analysis, a lesion size of 1 cm or larger was found to be the only significant predictor of malignancy for nonmasslike enhancement. Segmental, focal, dendritic, and clustered ring enhancement; surrounding high SI on T2-weighted imaging; and a lesion size of 1 cm or larger can act as predictors of malignancy for nonmasslike enhancement detected on 3-T MRI, but kinetic characteristics cannot.

Su1217 Histologic Differentiation of Hepatocellular Carcinoma and Response to Transcatheter Arterial Chemoembolization

malignancy. High grade dysplasia (OR 6.66 [95% CI, 1.52-29.08]) and ductal dilatation (OR 10.96 [95% CI, 2.25-53.42]) were independently associated with malignancy in multivariate analysis. Fourteen (56%) of 25 patients with high grade dysplasia on endoscopic biopsy were revealed as carcinoma after resection. Conclusions The malignancy rate in ampullary adenomas on endoscopic biopsy was 26.4%. Adenoma with high grade dysplasia and ductal dilatation are significant predictors of malignancy in the ampullary adenoma.

Su1218 Combination Therapy of Intraarterial 5-Fluorouracil and Pegylated Interferon-Alpha Subcutaneous Injection for Unresectable Biliary Tract Adenocarcinoma

malignancy. High grade dysplasia (OR 6.66 [95% CI, 1.52-29.08]) and ductal dilatation (OR 10.96 [95% CI, 2.25-53.42]) were independently associated with malignancy in multivariate analysis. Fourteen (56%) of 25 patients with high grade dysplasia on endoscopic biopsy were revealed as carcinoma after resection. Conclusions The malignancy rate in ampullary adenomas on endoscopic biopsy was 26.4%. Adenoma with high grade dysplasia and ductal dilatation are significant predictors of malignancy in the ampullary adenoma.

The Carcinoembryonic Antigen Level in Pancreatic Juice and Mural Nodule Size Are Predictors of Malignancy for Branch Duct Type Intraductal Papillary Mucinous Neoplasms of the Pancreas

Identification of predictors of malignancy for branch duct type intraductal papillary mucinous neoplasms (IPMN). Main duct type IPMN has been recommended for resection. However, the indications for resection of the branch duct type IPMN have been controversial. We retrospectively analyzed the clinicopathological factors of 134 patients undergoing resection for branch duct type IPMN, excluding main duct type IPMN, to identify predictors of the malignant behavior of this neoplasm. The cutoff values of tumor size, main pancreatic duct (MPD) size, mural nodule size, and carcinoembryonic antigen (CEA) level in the pancreatic juice obtained during preoperative endoscopic retrograde pancreatography (ERP) were analyzed using receiver-operator characteristic curves. We found 7 significant predictors for malignancy in the branch duct type IPMN in a univariate analysis; jaundice, tumor occupying the pancreatic head, MPD size >5 mm, mural nodule size >5 mm, serum carbohydrate antigen (CA)19-9 level, positive cytology in the pancreatic juice, and CEA level in the pancreatic juice >30 ng/mL. In a multivariate analysis, a mural nodule size >5 mm and a CEA level in the pancreatic juice >30 ng/mL were independent factors associated with malignancy. The positive predictive value of a mural nodule size >5 mm and a CEA level in the pancreatic juice >30 ng/mL was 100%, and the negative predictive value was 96.3%. We identified 2 useful predictive factors for malignancy in branch duct type IPMN; a mural nodule size >5 mm and a CEA level in the pancreatic juice obtained by preoperative ERP >30 ng/mL.

Gallbladder Lesions Identified on Ultrasound. Lessons from the Last 10 Years

BackgroundPossible mass lesions identified on ultrasound (US) of the gallbladder may prompt an aggressive surgical intervention due to the possibility of a malignant neoplasm. AimThis study aims to utilize a large modern series of patients with gallbladder lesions identified on US to evaluate imaging characteristics consistent with malignancy. MethodsA retrospective review was conducted of gallbladder ultrasound reports and clinicopathologic data of patients with a mass identified on US. ResultsApproximately 59,271 abdominal ultrasounds and 9,117 cholecystectomies were performed between February 2000 and February 2010. We identified 213 patients with a questionable gallbladder neoplasm on ultrasonography who underwent surgical exploration. Median age was 52 years (range = 11–87 years) and 147 (69%) were females. Final pathology demonstrated no neoplasm in 130 patients (61%), while 32 patients (15%) had a wall adenomyoma, 36 (17%) had a polyp (five of which were malignant), 14 (7%) had an adenocarcinoma not arising from a polyp, and one patient had a cystic papillary neoplasm. The smaller the lesion, the more likely it was to be a pseudo-mass. For lesions measuring <5 mm on US, 83% had no lesion found on final pathology. Significant predictors of malignancy were age >52 years (p < 0.001), presence of gallstones on US (p = 0.004), size >9 mm (p < 0.001), evidence of invasion at the liver interface (p < 0.001), and wall thickening >5 mm (p < 0.001). Shape (sessile or penduculated), echogenicity (echogenic or isoechoic), or presence of flow on Doppler were not predictors of malignancy. An US size of ≤9 mm had a negative predictive value of 100% for malignancy. ConclusionsDespite improvements in imaging, most apparent lesions measuring <5 mm on US are not identified in the surgical specimen. US size >9 mm, age >52 years, US suggestion of invasion at the liver interface, and wall thickening >5 mm, especially in the presence of gallstones, should raise the suspicion of malignancy.

Malignancy in Patients with Crohnʼs Disease: Data from the TREAT™ Registry with More Than 5 Years of Follow-up

Purpose: The association between malignancy and anti-tumor necrosis factor (TNF) therapy remains under investigation. Methods: We prospectively evaluated the incidence of malignancy in the large-scale, observational TREAT Registry, which examines the long-term outcomes of various treatments, including infliximab, used in the management of Crohn's disease (CD) in community-based and academic practice settings in North America. The influences of baseline patient (pt), disease characteristics and medication use were assessed via multivariate regression analysis and generation of hazard ratios (HRs). Standardized incidence ratios (SIRs) and exact 95% confidence intervals (CIs) were determined by dividing the number of patients with malignancies observed (in TREAT) with expected number (for the general U.S. population) using the Surveillance, Epidemiology End Results [SEER] 2009 database. Results: As of February 23, 2010, 6,273 pts were enrolled: 3,764 received infliximab and 2,509 received other-treatments-only. Incidences of malignancies were similar between infliximab-treated pts and pts receiving other-treatments-only within the categories of hematologic malignancies, lymphoma, nonmelanoma skin cancer, and malignant solid tumors (Table 1). Baseline age, disease duration, smoking, and immunomodulator therapy, but not infliximab therapy were significant predictors of malignancy (Table 2). Results of comparisons between the TREAT Registry and SEER database yielded 95% CIs containing 1 for all categories of malignancy assessed, indicating no significant difference, with the exception of breast cancer and lymphoma (Table 1). Breast cancer was less common in TREAT than SEER, and for lymphoma, the SIR was approximately twice that of the background population in both the infliximabtreated and other-treatments-only cohorts.Table: Table. Malignancies in TREAT: Comparison versus SEER databaseTable: Table. Malignancies in TREAT: effect of risk factorsConclusion: Infliximab did not significantly affect the risk of malignancies; immunomodulator use, age, duration of disease and smoking independently predicted time to first malignancy. Consistent with reports in the literature, CD patients, independent of infliximab treatment, appear to have a higher lymphoma risk than the general U.S. population. Disclosure: GR Lichtenstein: Investigator, Centocor Research and Development, Inc.; B Feagan: Investigator, Centocor Research and Development, Inc.; RD Cohen: Investigator, Centocor Research and Development, Inc.; BA Salzberg: Investigator, Centocor Research and Development, Inc.; RH Diamond: Employee, Centocor Ortho Biotech Services, LLC; W Langholff: Consultant, Centocor Research and Development, Inc.; A Londhe: Employee, Johnson & Johnson PRD.; WJ Sandborn: Investigator, Centocor Research and Development, Inc. This research was supported by an industry grant from This study was sponsored by Centocor Research and Development, Inc.

Lung nodules in pediatric oncology patients: a prediction rule for when to biopsy

PurposeThe purpose of the study was to develop a prediction rule regarding the factors that most accurately predict the diagnosis of a malignancy in a lung nodule in the pediatric oncology patient. MethodsA retrospective review of pediatric oncology patients that underwent lung nodule resection between 1998 and 2007 was performed. Multivariable logistic regression was used to create a prediction rule. ResultsFifty pediatric oncology patients underwent 21 thoracotomies and 48 thoracoscopies to resect discrete lung nodules seen on computed tomographic scans during workup for metastasis or routine surveillance. The mean nodule size was 10.43 ± 7.08 mm. The most significant predictors of malignancy in a nodule were peripheral location (odds ratio [OR], 9.1); size between 5 and 10 mm (OR, 2.78); location within the right lower lobe (OR, 2.43); and patients with osteosarcoma (OR, 10.8), Ewing sarcoma (OR, 3.05), or hepatocellular carcinoma (OR, 2.38). ConclusionsLesions that are between 5 and 10 mm in size and peripherally located in patients with osteosarcoma, Ewing sarcoma, or hepatocellular carcinoma are most likely to be malignant. Use of a prediction rule can help guide clinical practice by determining which patients should undergo surgical resection of lung nodules and which patients may be closely observed with continued radiologic studies.

Sa1431 Minimally Invasive Sampling of Abnormal PET Targets by EUS-Guided FNA Improves Clinical Staging of Patients With Suspected Non-Small Cell Lung Cancer (NSCLC)

Minimally Invasive Sampling of Abnormal PET Targets by EUSGuided FNA Improves Clinical Staging of Patients With Suspected Non-Small Cell Lung Cancer (NSCLC) James P. Callaway, Ashutosh Tamhane, Gurudatta Naik, Ayesha S. Bryant, Whitney Jennings, Robert J. Cerfolio, Mohamad A. Eloubeidi Internal Medicine, University of Alabama at Birmingham, Birmingham, AL; Gastroenterology, University of Alabama at Birmingham, Birmingham, AL; Biostatistics, University of Alabama at Birmingham, Birmingham, AL; Cardiothoracic Surgery, University of Alabama at Birmingham, Birmingham, AL Background: Current guidelines recommend that all abnormal lesions by PET scan should be sampled prior to thoracotomy in patients with lung cancer. EUSFNA has been demonstrated to be safe and minimally invasive in diagnosing N2/ N3 and metastatic disease to the adrenal glands. We currently report our updated large experience to determine the role of EUS-FNA in identifying abnormal PET scan lesions. The secondary objective was to determine predictors of lymph node malignancy. Methods: Consecutive patients (N 282) who underwent both PET and EUS-FNA between June 2003 and June 2009 at a tertiary care teaching hospital were evaluated. Targets sampled by EUS included: the adrenal glands and/or (N2/N3) stations 9, 8, 7, 5, 4L, or 2. All EUS examinations were performed by a single endosonographer. The reference standard included surgery or clinical follow-up ( 6 months). Abnormal PET targets were defined by a maximum standardized uptake value (SUV) of 2.5. Results: Mean age was 66 years and 66% were men. Mediastinal lymph nodes were solely sampled in 243 patients, while only Stage IV lesions in 18 and both in 21. In the 264 patients who underwent EUS-FNA of lymph nodes, sensitivity, specificity, positive predictive value, negative predictive value, accuracy, of EUS-FNA were: 76%, 100%, 100%, 81%, 88% respectively and for PET: 84%, 53%, 64%, 78%, 69%, respectively for confirming N2/N3 disease. EUS-FNA confirmed PET staging in 168 patients while 63 were “downstaged” (25.8%) and 11 “upstaged” (4.1%); 22 patients were falsely downstaged by EUS-FNA. In 11 patients with false negative FNA results, the node was not routinely accessible by EUS. In multivariable analysis, the short axis of the lymph node by EUS (ORadjusted1.25; 95% CI: 1.091.44; p 0.002) and SUV of the lymph node (ORadjusted1.38; 95% CI: 1.12-1.69; p 0.002) were significant predictors of malignancy when adjusted for age, race, sex, lymph node long axis, primary mass SUV and location, and prior chemotherapy. EUS-FNA confirmed Stage IV lung cancer in 19 patients (1 gastrohepatic node, 1 perigastric node, 1 pancreas, 3 right adrenals and 13 left adrenals) and in 1 patient who had left adrenal involvement without PET uptake. Conclusions: EUS-FNA improved clinical staging in patients with suspected NSCLC. Its ability to accurately confirm the nature of abnormal PET scan targets, along with the ability to upstage and downstage false positive and false negative PETs, makes it an ideal choice for initial minimally invasive sampling. Lymph node short axis and lymph node SUV were the significant predictors of malignancy and may help guide endosonographers to which lymph node needs to be sampled first during EUS.

Predictive Value for Malignancy of Suspicious Breast Masses of BI-RADS Categories 4 and 5 Using Ultrasound Elastography and MR Diffusion-Weighted Imaging

The aim of this study is to evaluate the ability of ultrasound elastography and MR diffusion-weighted imaging (DWI) to predict malignancy of breast masses, with subsequent recommendation for biopsy. For 115 breast masses classified as BI-RADS category 4 or 5, which were assessed according to combined findings of mammography, B-mode sonography, and dynamic contrast-enhanced MRI, two radiologists retrospectively evaluated the elasticity scores using ultrasound elastography and the apparent diffusion coefficient (ADC) values using MR DWI. The diagnostic abilities of these two techniques were analyzed by using univariate and multivariate logistic regression analysis. In the analysis of all 115 breast masses, the elasticity score was predictive of malignancy, whereas the ADC value was not independently predictive. In an analysis of the 52 masses assessed as BI-RADS category 4, the elasticity score was found to be a significant predictor of malignancy, compared with the ADC value, which was a nonsignificant predictor. In an analysis of the 63 masses assessed as BI-RADS category 5, neither the elasticity score nor the ADC value was a significant predictor of malignancy. Our results show that elasticity imaging provides relatively reliable predictions for malignancy, especially in BI-RADS category 4 masses, compared with MR DWI.

Endosonographic features predictive of malignancy in mediastinal lymph nodes in patients with lung cancer

EUS is useful in determining mediastinal lymph node (LN) metastases in patients undergoing staging for lung cancer. However, FNA of LNs is often performed only if suspicious features are present. The utility of individual LN features in predicting malignant cytology remains unclear. To evaluate the utility of EUS-determined LN features for predicting malignant cytology. Prospective observational study. Two U.S. tertiary-care centers. This study involved 425 patients with primary lung cancer who underwent EUS. All mediastinal LNs were described according to size, shape, echogenicity, and margin characteristics. FNA was performed on LNs with any features suggestive of malignancy. EUS-guided FNA cytology was classified as benign or abnormal (suspicious/malignant). The utility of LN features in predicting malignant cytology was determined and further analyzed by logistic regression, and a predictive model was established. Accuracy of individual LN features for predicting malignancy. EUS detected 836 LNs in 425 patients, and FNA was obtained in 698 patients. On multivariable analysis, only round shape, a short axis of >8.3 mm, and sharp margins were predictive of malignant cytology. According to the predictive model, the calculated probability of having malignancy is less than 4% (95% confidence interval [CI], 0.022-0.064) when none of the LN features are present and 63% (95% CI, 51%-72.2%) when all features were seen. No surgical histology as the criterion standard. Among patients with lung cancer, EUS features of round shape, sharp margins, and short axis of >8.3 mm are significant predictors of malignancy. The probability of malignancy is low when none of the features are present.

A large single-center experience of EUS-guided FNA of the left and right adrenal glands: diagnostic utility and impact on patient management

EUS-guided FNA of the left and right adrenals has been described, but data are very limited. Our primary objective was to determine the impact of the diagnostic utility of EUS-guided FNA of adrenal glands on patient management. Our secondary objective was to determine predictors of malignant adrenal involvement. Observational study. Tertiary referral center. Patients with enlarged adrenal(s) on abdominal imaging underwent EUS-guided FNA. The left adrenal (n = 54) was sampled via the transgastric approach and the right adrenal (n = 5) via a transduodenal approach. Fifty-nine patients (63% men, median age 65 years) were evaluated. The median adrenal gland size was 25 x 17 mm. Adrenal tissue adequate for interpretation was obtained in all of the patients. EUS-guided FNA confirmed malignancy in 22 (37%) patients. Based on size (> or =30 mm) alone, EUS had an accuracy of 68%. Patients with malignant cytology had higher standard uptake value scores on positron-emission tomography compared with patients with benign adrenal masses (P < .001). Malignant masses were more likely to have an altered adrenal gland shape compared with benign masses (crude odds ratio [OR] 12.0; P < .001). On multivariable analysis, altered adrenal gland shape was a significant predictor of malignancy (adjusted OR 7.94; P = .015), whereas a size of 30 mm or larger (adjusted OR 1.30; P = .774) and hypoechoic nature (adjusted OR 12.05; P = .148) were not. All patients except 2 with malignant cytology were treated with systemic therapy without the need for additional invasive biopsies or surgery. No immediate complications were encountered. Lack of surgical criterion standard; 1 experienced endosonographer. EUS-guided FNA of the adrenal glands is a minimally invasive and safe approach that documents or excludes malignant involvement. EUS-guided FNA should be the first next test to evaluate enlarged adrenal glands because it directs therapy and affects patient management.

Combined Use of EUS-guided FNA and Immunostains Discloses Metastatic and Unusual Diseases in the Evaluation of Mediastinal Lymphadenopathy of Unknown Etiology

Purpose: Mediastinal lymphadenopathy (ML) is a cause for concern, especially in patients with previous malignancy. EUS-guided FNA has significantly expanded the diagnostic capability of EUS in GI disorders including tissue evaluation of ML. Therefore, we sought to investigate the use of EUS-guided- FNA with immunostains in patients being evaluated for ML. Methods: We prospectively analyzed adult patients with ML of unknown origin who underwent EUS guided-FNA at a tertiary referral center between September 2000 and May 2007. EUS-guided FNA was performed with linear echoendoscopes and 22-gauge needles. On-site and final evaluation was performed by experienced cytologist, in addition to referral to special immunostains if necessary. Patients' demographics were documented, as well as lymph node (LN) characteristics. Results: One-hundred and nineteen patients were included, and a total of 138 mediastinal LN were sampled. Mean age of patients was 54.4 years (SD 16.6), 53.8% were men (64 patients) and 77.3% were white (92 patients). Of the 119 subjects, prior malignancy was present in 44.5% (61 patients). Prior chest CT was performed in 87% (100 patients); PET scan in 38% (44 patients, median SUV 6, IQR 3-10, range 2-52 n=22). Prior LN biopsy was performed in 24 patients (20.3%). Site of examined LN: 101 subcarinal (76.5%), 11 lower paraesophageal (8%), 8 aorto-pulmonary window (6%). The median ML diameter was 28 mm × 13mm. Sharp borders were seen in 28 (23.7%), round shape in 21 (15.9%), hypoechoic appearance in 53 (40.1%). Mean number of passes was 3.83, SD1.34. FNA was read on-site as malignant 21 (16.1%), benign 100 (76.9%), suspicious 6 (4.6%), atypical 3 (2.3%). Final FNA read was malignant in 28 (21.8%), benign 98 (76%), suspicious 3 (2.3%). Sixty-four LN were deferred for additional studies; 22 for immunostains, 26 for GMS stain and 22 for flow cytometry. The sensibility, specificity and overall accuracy of on-site read cytology was, respectively, 95%, 94% and 86%. Metastatic malignancies disclosed included Hodgkins lymphoma (2), NHL (5), lung (4), melanoma (1), breast (6), colon (3), renal (2), endometrial (1), Fallopian tube (1), hepatoma (1), unknown primary site(2).The significant predictors of malignancy in FNA were larger short axis(OR 1.08, 95%CI 0.99-1.18, P = 0.093), prior malignancy, (OR 12.90, 95%CI 2.90-57.62, P = 0.001) and presence of sharp borders (OR 5.44, 95%CI 1.43-20.70, P=0.013). Conclusion: EUS-guided FNA with additional ancillary studies and immunostains is useful in disclosing metastatic lymphadenopathy to the mediastinum from a variety of neoplasms. Predictors of malignancy include prior history of malignancy, larger short axis and presence of ML sharp borders.

Predictors of Malignancy in EUS-Guided FNA for Mediastinal Lymphadenopathy in Patients Without Known Lung Cancer

Background: Mediastinal lymphadenopathy (ML) poses a great challenge to patients and physicians. To date, no studies have evaluated the factors associated with risk of having a EUS-guided FNA positive for malignancy in patients without lung cancer. Therefore, we sought to investigate the predictors of malignancy in EUS-guided FNA for ML in a large, multi-ethnic cohort. Methods: We evaluated data from 113 patients without known lung cancer who underwent EUS guided-FNA for ML in a tertiary referral center between September 2000 and May 2007. Patients' demographics were included in our analyses, as well as lymph node characteristics such as location, largest and smallest diameters (mm), shape (round or not), appearance (hypoechoic or not) and presence of sharp borders. All subjects underwent EUS performed by a single endoscopist using linear endoscopes and 22-gauge needles for cytology. We used multivariable logistic regression to evaluate the predictors of cancer. Results: Of the 113 patients analyzed, 60 (53%) patients were men and 87 (77%) were Caucasian. The mean age was 53.8 years. History of prior malignancy was present in 48 (43%) patients. A total of 132 FNA samples from 132 lymph nodes were sent for cytology. Twenty-five (20%) lymph nodes were positive for malignancy. Mean smallest and largest lymph node diameters were, respectively, 15.0 and 28.9 mm. Round shape and sharp borders were found in 20 (16%) and 25 (22%) lymph nodes, correspondingly. The significant predictors of malignancy in FNA were: prior history of a cancer (OR 12.81; 95% CI 2.7-60.76; P = 0.001), smallest lymph node diameter (OR 1.11; 95% CI 1.00-1.22; P = 0.041) and sharp lymph node borders (OR 5.04; 95% CI 1.20-21.20; P = 0.028). Age, race, gender, largest lymph node diameter, round shape were not associated with cancer in our cohort. Conclusions: We demonstrated that, in patients evaluated by EUS-guided FNA for ML, increased risk of malignancy was associated with history of prior cancer, larger dimension of the smallest lymph node diameter and presence of sharp borders. The combination of the above predictors will help endosonographers perform FNA in lymph nodes that are more likely to reveal malignancy, resulting in less unnecessary biopsies and an increase in overall accuracy of EUS-guided FNA for ML.

Radiologic and Pathologic Findings of Nonpalpable Thyroid Carcinomas Detected by Ultrasonography in a Medical Screening Center

The purpose of this study was to investigate the ultrasonographic and pathologic findings of nonpalpable thyroid carcinomas and reliable guidelines for fine-needle aspiration (FNA). Our study was approved by our Institutional Review Board, and written informed consent was waived. Between April 2004 and June 2006, screening ultrasonography was performed for 16,352 self-referred patients in the health care center. Among 1325 nonpalpable thyroid nodules in 1009 patients, pathologic results of FNA revealed 823 benign, 154 indeterminate, 198 nondiagnostic, and 150 malignant nodules. Fifty-eight malignant thyroid nodules (39 microcarcionomas and 19 carcinomas >1 cm, confirmed by both FNA and thyroidectomy) in 55 patients and 82 benign nodules (confirmed by both FNA and follow-up over 2 years) in 75 patients were included for the analysis. Three radiologists retrospectively analyzed the ultrasonographic features of these nonpalpable thyroid nodules for echogenicity, shape, margin, calcification, degree of cystic changes, and size. We compared the radiologic and pathologic findings between microcarcinomas and carcinomas larger than 1 cm for extra-capsular invasion, lymph node metastasis, bilaterality, and multicentricity using univariate analysis. Marked hypoechogenicity, an irregular shape, a taller-than-wide shape, a well-defined spiculated margin, microcalcification, and an entirely solid nature were significant predictors for malignancy (P < .05), whereas a cutoff value of 1 cm in the longest diameter was not significant (P = .184). However, extracapsular invasion (P = .024) and lymph node metastasis (P = .019) were observed more frequently in carcinomas larger than 1 cm (73.7% and 42.1%, respectively) than in microcarcinomas (38.5% and 12.8%). Ultrasonographic findings suggesting malignancy should be preferentially considered as indicators for FNA, regardless of size, in nonpalpable thyroid nodules. However, extracapsular invasion and lymph node metastasis are closely related to the size of the thyroid nodule.

Analysis of cytological specimens from mediastinal lesions obtained by endoscopic ultrasound-guided fine-needle aspiration

Endoscopic ultrasound-guided fine-needle aspiration (EUS-FNA) seems to be a powerful tool to obtain cytologic specimens from mediastinal and celiac lymph nodes, enlarged left adrenal glands, and intrapulmonary tumors with mediastinal extension. The diagnostic yield of EUS-FNA and the accuracy of cytologic specimens was evaluated. Cytologic assessment of EUS-FNA specimens was performed and specimens were classified as positive, negative, suspicious for malignancy, or unsatisfactory for diagnosis. Cytology was compared with histologic and clinical (> or = 6 months) follow-up. Cytologic specimens were collected from 155 lymph nodes, 10 left adrenal glands, and 9 intrapulmonary tumor masses. For lymph nodes, the diagnostic yield was 0.65. After exclusion of unsatisfactory specimens, sensitivity, specificity, accuracy, and positive (PPV) and negative (NPV) predictive values of cytologic specimens were 0.92, 1.00, 0.93, 1.00, and 0.63, respectively. Subgroup analysis of lymph nodes with a dimension of > or = 10 mm showed similar results. With EUS imaging only, lymph node diameter and a round or irregular shape were significant predictors of malignancy at multiple logistic regression analysis, but their clinical usefulness is very limited (PPV = 0.78 and NPV = 0.45). For left adrenal gland specimens, sensitivity and specificity were 0.89 and 1.00, respectively. From intrapulmonary masses, 8 true-positive and 1 true-negative specimens were obtained. Cytologic specimens from mediastinal or celiac lymph nodes obtained with EUS-FNA were reliable and accurate. Specimens from left adrenal glands and intrapulmonary tumor masses showed promising results.

Comparison of diagnostic usefulness of predictive models in preliminary differentiation of adnexal masses

The purpose of this study was to compare prognostic models evaluating the probability of an ovarian cancer occurrence based on a number of clinical and ultrasonographic data in women with adnexal masses. A total of 686 women with adnexal masses underwent the examinations between 1994 and 2002. The recorded parameters included: age, menopausal status, body mass index, the grayscale and Doppler ultrasonographic examination, and selected markers concentration levels. In order to find the best combination of features, which significantly influences the probability of malignancy, stepwise logistic regression analysis, as well as artificial neural network, was used. The diagnostic efficiency of received models was estimated and compared using receiver-operating characteristics (ROC) curve. The results indicate that 431 and 255 patients had a benign and malignant ovarian tumor, respectively. Application of stepwise logistic regression analysis revealed statistically significant importance of eight features. The sensitivity and specificity for the received model were 65.71% and 77.59%, respectively. Three-layer perceptron network shows 13 features as significant predictors of malignancy. The network gave a sensitivity of 85.7% and specificity of 93.1%. Comparison of area under ROC curve for received models was 0.9679 vs 0.9716. Prognostic values of the analyzed neural model are not optimal but seem to surpass logistic regression model in terms of the predictive possibilities.

Diagnostic Value of CYFRA 21-1, CEA, CA 19-9, CA 15-3, and CA 125 Assays in Pleural Effusions: Analysis of 116 Cases and Review of the Literature

Levels of tumor markers in pleural effusions may help to establish the diagnosis of pleural malignancy, but the precise diagnostic value of each marker remains unclear. The aim of this study was to assess the diagnostic value of five common pleural fluid tumor markers, carcinoembryonic antigen (CEA), cytokeratin fragment (CYFRA) 21-1, cancer antigen (CA) 15-3, CA 19-9, and CA 125, and to review the literature from the past 15 years. Pleural fluid samples were collected prospectively from 116 patients and assayed for CEA, CYFRA 21-1, CA 15-3, CA 19-9, and CA 125 levels. A MEDLINE search of the English-language literature from the past 15 years was also done. Effusions were classified as benign or malignant on the basis of their definitive pathologic or cytologic diagnoses. The levels of all pleural tumor markers were statistically significantly higher in the malignant group than in the benign group. The marker with the highest accuracy was CEA (85.3%); CA 15-3, CYFRA 21-1, and CA 19-9 had similar accuracies (75.2%, 72.4%, and 71.5%, respectively), and CA 125 had the lowest accuracy (40.5%). On univariate analysis, tumor-marker combinations did not result in a greater accuracy than that of CEA alone. On multivariate logistic regression, CA 15-3 and CYFRA 21-1 were significant predictors of malignancy. Among the nine reports in the literature comparing 11 different tumor markers, CEA, CA 15-3, and CYFRA 21-1 yielded the best results. We conclude that pleural fluid analysis should include CEA for the diagnosis of malignancy. CA 15-3 and CYFRA 21-1 may serve as alternative options.