S156 ENDOBRONCHIAL ULTRASOUND: TRANSOESOPHAGEAL ULTRASOUND-GUIDED LYMPH NODE SAMPLING USING A LINEAR ULTRASONIC BRONCHOSCOPE IN THE STAGING OF LUNG CANCER

KM Skwarski, PS Hodkinson, H Huang, D Gilbert, JB McCafferty. Department of Respiratory Medicine, Royal Infirmary Edinburgh, Edinburgh, UK

Lung cancer treatment planning requires accurate staging of mediastinal lymph nodes. Current methods include mediastinoscopy, positron-emission tomography (PET-CT) and endobronchial (EBUS) or transoesophageal (EUS) ultrasound-guided lymph node aspiration. PET-CT is non-invasive but has only a moderate positive-predictive value in the assessment of lymph node metastases. The “gold standard” is mediastinoscopy, which allows surgical evaluation of lymph node stations 2, 3, 4 and 7. EBUS and EUS are less invasive than mediastinoscopy and are performed as outpatient procedures under conscious sedation. EBUS allows sampling from stations 2, 3, 4, 7, 10 and 11. EUS allows staging of stations L4, 7 and 8. We have performed over 500 EBUS procedures (2005–8) using an Olympus linear ultrasonic bronchoscope (BF-UM40). Lymph node samples were obtained using a 22G needle and tissue processed as a cell block before immunohistochemical analysis. Although EBUS is well tolerated, we have found that in a proportion of cases it was easier to aspirate lymph node stations L4, 7 and 8 via the transoesophageal route using the EBUS bronchoscope (EUS (EBUS)) and identical sample collection technique. We have analysed data from the first 40 EUS (EBUS) cases. This approach was employed to evaluate station L4 (different angle to transbronchial approach) in 15 cases, to provide additional sample from L7 in 11 cases, to access L8 in one case and to access a lung mass in four cases. Technical difficulties with the EBUS approach (extrinsic compression of the airway and inability to penetrate tracheal wall) prompted EUS (EBUS) in five cases. In addition, the EUS (EBUS) approach was used to minimise patient distress (cough, SpO₂ <90%, FEV₁ <1.0 l) in five cases. Tissue samples from lymph nodes or lung mass were obtained at 28/40 EUS (EBUS) procedures. 16 confirmed a diagnosis of malignancy (lung cancer in 12) and 12 were negative for cancer with no “false-negative” aspirates. In 12 cases no aspirates were taken either due to the absence or small (<5 mm) size of lymph nodes. There were no complications and the procedure was well tolerated. In conclusion, transoesophageal aspiration using the linear ultrasonic bronchoscope should be considered a safe and effective method of mediastinal staging as an alternative to mediastinoscopy.

S157 TEST PERFORMANCE OF ENDOBRONCHIAL ULTRASOUND AND TRANSBRONCHIAL NEEDLE ASPIRATION BIOPSY FOR MEDIASTINAL STAGING IN PATIENTS WITH LUNG CANCER: SYSTEMATIC REVIEW AND META-ANALYSIS

¹L Edmonds, ²KJ Adams, ²P Shah, ²E Lim. ¹Papworth Hospital, Cambridge, UK, ²The Royal Brompton Hospital, London, UK

Introduction: Endobronchial ultrasound (EBUS) and transbronchial needle aspiration (TBNA) is becoming widely used for mediastinal lymph node staging in patients with known or suspected lung cancer. Whereas there are a number of case series evaluating the sensitivity and specificity of this investigation, many have small numbers that limit the ability to assess the precision of EBUS TBNA as a staging modality. The aims of our study were to perform a systematic review of published studies evaluating EBUS TBNA for mediastinal lymph node staging and to ascertain the pooled sensitivity and specificity of this investigation in comparison with published results of computed tomography (CT) and positron emission tomography (PET).

Methods: A literature search was constructed and performed by a professional medical librarian from 1960 to February 2008. Pooled specificity and sensitivity was estimated from the extracted data with an exact binomial rendition of the bivariate mixed-effects regression model.

Results: Of 365 EBUS publications, we identified 24 that specifically focused on mediastinal lymph node staging, with only 10 reporting data possible for extraction. The overall test performance was excellent, with an area under the summary receiver operating characteristics curve of 0.99 (95% CI 0.96 to 1.00), similarly the EBUS TBNA has excellent pooled specificity of 1.00 (95% CI 0.92 to 1.00) and good pooled sensitivity of 0.88 (95% CI 0.79 to 0.94).

Conclusions: The results of our study indicate excellent overall test performance and specificity of EBUS TBNA when used for mediastinal lymph node staging in patients with lung cancer. In addition, the pooled test specificity and sensitivity compares favourably with published results of CT and PET. More work is required to compare directly the role of EBUS TBNA in relation to mediastinoscopy and define its clinical utility in the era of integrated PET-CT.

S158 ENDOSCOPIC ULTRASOUND-GUIDED TRUCUT BIOPSY FOR EVALUATION OF MEDIASTINAL LYMPHADENOPATHY AND LUNG CANCER STAGING

MP Sovani, M Jameson, T Thomas, K Ragunath, G Aital. Queen’s Medical Centre, Nottingham, UK

Introduction and Objectives: Transoesophageal endoscopic ultrasound (EUS) guided biopsy has been used to diagnose and stage mediastinal lymphadenopathy and can reduce the need for mediastinoscopy.1 EUS fine needle aspiration (FNA) has been performed at Queen’s Medical Centre, Nottingham since May 2002 and Trucut biopsy since December 2002. We evaluated the diagnostic yield from EUS-guided Trucut biopsies of mediastinal lymph nodes.

Methods: A retrospective analysis of all mediastinal lymph node sampling using EUS at the Queen’s Medical Centre, Nottingham.

Results: A total of 40 biopsies was performed on 37 patients during the period of the study. The mean age was 60 years (32–79), and 67.6% were men. Of the total referrals, 92% were for diagnosis of mediastinal lymphadenopathy. 5% patients had lung cancer and were referred for staging and 3% patients had a radiological diagnosis of metastatic malignancy referred for obtaining histology to find the source of malignancy. Lymph nodes sampled were as follows: subcarinal 59%; AP window 14%; paraoesophageal 8%; para-aortic 5%; perigastric 3%; adrenal 3%; not specified 8% Mean lymph node diameter was 2.3 cm and mean length of the tissue core obtained was 1.3 cm. Sensitivity 94% (95% CI 77 to 99), specificity 100% (95% CI 62 to 100). Assuming a prevalance of lung cancer of 65%, positive predictive value was 1 and negative predictive value was 0.66.

Complications: One patient each developed chest pain, minor haemoptysis and retro-oesophageal cold abscess (in a patient with tuberculosis).

Conclusion: Trucut biopsy allows larger core size thereby improving diagnostic ability as it allows immunohisto, special stains, etc, especially for lymphomas. Our diagnostic accuracy was comparable with most published studies.1 Importantly, we were able to diagnose lymphoma/granulomatous diseases as a result of Trucut biopsy cores. Despite its high diagnostic yield and its ability to access coeliac node, liver, adrenal EUS Trucut biopsy is underused in Nottingham.

S159 THE SENSITIVITY, NEGATIVE PREDICTIVE VALUE AND ACCURACY OF A NEW ENDOBRONCHIAL ULTRASOUND SERVICE

¹N Navani, ²H Makker, ¹G Kocjan, ¹SP Pereira, ²J Griffith, ¹SG Spiro, ¹SM Janes. ¹University College London, London, UK, ²North Middlesex University Hospital, London, UK

Background: Curvilinear endobronchial ultrasound with real-time transbronchial needle aspiration (EBUS–TBNA) is a promising, minimally invasive technique for the diagnosis of mediastinal lymphadenopathy. In patients with lung cancer, meta-analysis of cohort studies has demonstrated a sensitivity of 90% and negative predictive value (NPV) of 76% for mediastinal metastases. Emerging case series suggest EBUS–TBNA may play a role in the diagnosis of sarcoid and lymphoma. We wished to examine the sensitivity and NPV of a new EBUS service in all patients.

Methods: Prospective data were collected from the first 35 consecutive patients referred for EBUS–TBNA. All patients had mediastinal lymphadenopathy of 1 cm or greater in short axis in lymph node stations accessible to EBUS–TBNA. In each case, the procedure was performed on an outpatient basis and under conscious sedation. A sample was deemed adequate if lymphocytes and follicle centre fragments were obtained.

Results: 20 patients (out of 35) had a definitive diagnosis by EBUS–TBNA (nine non-small cell lung cancer, six tuberculosis, three sarcoid, two metastatic breast cancer). 15 patients have had benign lymphocytes aspirated with no specific pathology. These have been confirmed by mediastinoscopy or clinical follow-up in 11 cases. Two patients had inadequate samples. One patient with adenocarcinoma and one with tuberculosis were diagnosed on mediastinoscopy, following negative EBUS–TBNA. For all patients referred for EBUS, sensitivity, specificity, NPV and accuracy were 83%, 100%, 73% and 89%, respectively. For patients with known or suspected lung cancer (see table) similar results were obtained, with a prevalence of mediastinal metastases of 69%. No complications were observed and 16 patients had mediastinoscopies prevented.

Conclusion: EBUS–TBNA is an accurate and safe method for the diagnosis of mediastinal lymphadenopathy and may prevent mediastinoscopies. The learning curve is fast and our results are in line with published data. Due to the relatively low NPV, negative EBUS–TBNA results should be followed by further investigation. Randomised trials to investigate the effect of EBUS–TBNA on patient outcomes and healthcare costs are required.

S160 AN AUDIT TO ASSESS THE ACCURACY OF POSITRON EMISSION TOMOGRAPHY SCANNING IN LUNG CANCER

¹AL Rich, ¹N Nandi, ²GM Cox, ²KA Amsha, ¹DR Baldwin, ¹WS Lim. ¹Nottingham University Hospital, City Campus, Nottingham, UK, ²Kings Mill Hospital, Sutton-in-Ashfield, UK

Introduction and Objectives: Although positron emission tomography (PET) has been routinely used in the UK for lung cancer staging for over 5 years, its accuracy is still debated. The aim of this audit was to assess the accuracy of PET in lung cancer staging when applied in routine practice in a regional centre.

Method: All patients referred for PET scanning by the lung cancer multidisciplinary team (MDT) since March 2003 were identified retrospectively from databases and prospectively recorded since July 2007. Basic demographic information was obtained, along with details of co-morbidity, radiological and surgical staging and the treatment provided.

Results: Of 200 patients identified (134 retrospectively and 66 prospectively; mean age 67 years (range 34–87); 61% male; 91% WHO performance status 0 or 1; mean FEV₁ 1.92 litres (mean 72.6% predicted)), 178 (89%) had non-small cell carcinoma. 129 received radical treatment for lung cancer (85 surgery including six wedge resections, 44 radical radiotherapy). The accuracy of PET staging compared with surgical staging (n  =  79) was calculated for each tumour and nodal stage. Results are shown in the table. All those incorrectly staged as T1 at PET were T2 at pathology. The majority of those incorrectly staged by PET as T3, were downstaged at surgery to T2, illustrating the important role of the MDT. Similarly, of patients incorrectly staged by PET as N0 or N1, only 4/30 were surgically staged as N2. For the depiction of malignant nodes in the subgroup of patients with PET T1 disease, the sensitivity, specificity, positive predictive value and negative predictive value of PET were 87.5%, 100%, 100% and 50%, respectively. This compares favourably with published data.1 Of 16 patients staged Mx by PET, nine patients had 10 investigations, seven were downstaged to M0, of whom five received radical treatment.

Conclusion: The data suggest that whereas the positive predictive value of PET stage may not always be high, the MDT has an important role in interpreting PET results and recommending surgery appropriately. PET had a strong tendency to overstage metastatic spread, highlighting the need for prompt investigation to reduce delay from referral to treatment.