P199 Assessment of pneumothorax treatment response on chest radiograph: a comparison of methods of size measurement

Background There is recent evidence that the size of pneumothorax on chest radiograph (CXR) in patients with Primary Spontaneous Pneumothorax (PSP) can predict failure of needle aspiration or persistent air leak. However, there are a number of differing methods of measuring pneumothorax size on CXR: Collins and Rhea methods use a combination of 3 intrapleural distances (in centimetres), whereas Light’s index uses the cube of the ratio of lung/hemithorax measurement (figure 1). Developing a robust method of measuring pneumothorax size is useful for research and clinical care. The aim of this study is to compare multiple measurement techniques to a novel digital measurement method, and to describe its utility in assessing treatment response. Methods 56 CXRs from 28 patients recruited to the RAMPP pneumothorax trial,1 in our centre, were reviewed, comparing three existing methods of size measurement with our method. This involves measuring the absolute percentage area of pneumothorax (in pixels) compared to the hemithorax size (as described previously[2]). The treatment response was assessed by absolute change in size (% of hemithorax) on CXR from pre- to post-procedure (1 hour after intervention). Results There was good correlation between the calculated percentage area and Collins/Rhea methods (r2 0.845). However, for large pneumothoraces, the Collins method overestimates the size: 5 of 56 CXR (8.9%) had a calculated size >100%. 14 (50%) patients were managed with a Pleural Vent (ambulatory device) and 14/28 (50%) by standard care. 10/14 (71.4%) in standard care group had initial management with needle aspiration (as per BTS guidelines); whilst 4 had chest drains inserted. The treatment response was lower in the needle aspiration group (22.6% improvement in pneumothorax size) and chest drain group (28.5%), than the Pleural vent group (44.2%). Conclusion This new digital method of assessing pneumothorax has good correlation with published methods, but does not rely on measurement in centimetres, so is useful in multicentre research (when DICOM formats may not be available). On-going work will assess whether pneumothorax size can predict treatment outcomes and long-term recurrence rates. Reference Hallifax RJ, Lancet 2020; 396: 39–49; 2. Rahman NM, N Engl J Med2011; 365:518–526