S26 Identification and prognostic importance of non-expansile lung following drainage of suspected malignant pleural effusion

Introduction Malignant Pleural Effusion (MPE) is common and often Results in disabling breathlessness. Non-expansile lung (NEL) frequently complicates pleural drainage, resulting in talc pleurodesis failure. Reliable detection of NEL would allow better clinical decision-making and more rational design of MPE trials. We developed 2 semi-objective definitions of NEL, which we hypothesised might prove more accurate and more consistent than the currently used subjective British Thoracic Society (BTS) method. Materials and Methods A retrospective cohort study was performed, involving 93 consecutive patients who underwent local anaesthetic thoracoscopy at our centre (July 2010- January 2015). NEL was defined prospectively at 3 month follow-up in all. Post-drainage chest radiographs were retrospectively classified as ‘NEL’ or ‘expansile’ by 2 independent assessors using the subjective BTS method and the 2 semi-objective methods (Re-expansion Proportion (REP) and Lateral Apposition Ratio (LAR), shown in figure 1). Sensitivity, Specificity and Inter-observer Agreement (Cohen’s Kappa, k) for NEL by each method were compared. Overall Survival (OS) based on expansion status by each method was compared using Kaplan-Meier methodology (MPE cases only). Results 65/93 patients had MPE. Sensitivity (0.81 (95%CI 0.71–0.89)) and specificity (0.87 (95%CI 0.81–1.00)) by the BTS method were highest. REP (sensitivity 0.61 (95%CI 0.49–0.72), specificity 0.94 (95%CI 0.73–1.00)) and LAR (sensitivity 0.56 (95%CI 0.44–0.67), specificity 0.94 (95%CI 0.73–1.00)) were less accurate. Inter-rater agreement (k) for BTS, REP, LAR were 0.68, 0.46 and 0.53, respectively. In MPE patients, NEL was consistently associated with a 2–4-fold lower median OS by all methods. Discussion The subjective BTS method appeared more accurate in predicting NEL than REP or LAR in this retrospective study, however all methods were subject to significant inter-observer variation. NEL is strongly associated with mortality. Our data highlight the clinical importance of NEL and its potential impact on MPE trial design, but do not strongly support any of these reported end-points as reliable clinical decision-making tools, trial end-points, or entry/stratification criteria. Further prospective research is needed to standardise the definition of NEL for these purposes, ideally prior to pleural drainage, and link this to patient-centred end-points.