Tidal changes on CT and progression of ARDS

Cereda, Maurizio; Xin, Yi; Hamedani, Hooman; Bellani, Giacomo; Kadlecek, Stephen; Clapp, Justin; Guerra, Luca; Meeder, Natalie; Rajaei, Jennia; Tustison, Nicholas J; Gee, James C; Kavanagh, Brian P; Rizi, Rahim R

The combination of reduced gas content and of locally unstable inflation was closely associated with progression of experimental injury. This unstable inflation partially overlapsbut does not coincide-with tidal recruitment of atelectasis and elevated local strain. There are multiple mechanisms that can explain why this is closely associated with disease progression. First, voxels with unstable inflation may reflect the presence of reversible atelectasis, possibly atelectrauma. 22 Second, the high-risk regions contained voxels with intermediate gas content, mixed aeration that represents pathological alveolar inflation, 23 airspace flooding by oedema 24 and sparse microscopic atelectasis. Third, in areas where ventilated, collapsed and/or flooded airspaces coexist, CT yields an averaged greyscale 7 and cannot resolve a 'single' inflation state. Moreover, alveolar interdependence results in focal augmentation of stretch, 25 a concept supported by studies using in vivo MRI with inhaled hyperpolarised gas. 23 Finally, flooding potentiates ventilator-induced injury 26 and studies in humans 7 and animals 27 indicate that voxels with heterogeneous inflation are clustered among regions of fixed airspace filling and aerated lung. In summary, areas of unstable inflation concentrate local lung stress and potentiate topographic propagation of ventilator-induced injury. 12 Lung micromechanics can be directly assessed using techniques such as intravital microscopy, 28 optical coherence tomography 29 and real-time fluorescence microscopy. 26 However, these approaches do not permit whole-lung (or patient) assessment. While MRI provides more accurate functional information, 23 30 CT is more accessible and easier to use. Furthermore, ARDS is clinically under-recognised and undertreated 31 and further refinement of conventional criteria (ie, plain radiography and oxygenation) is not likely to provide substantial improvement. However, the detection of high-risk tissue with unstable inflation using CT might, if validated, offer additional future options in properly screened patients. 32 For example, image analysis could help identify subjects at high risk of deterioration, perhaps where unstable inflation is not reversible by ventilator adjustment. In such a setting, this information could inform treatment decisions, especially where costly or high-risk management (eg, extracorporeal therapy) 33 is being considered. In addition, the information could be used to enrich clinical trials of interventions in ARDS, perhaps ensuring that only patients with a high (but not overwhelming) probability of poor outcome are recruited. 2 34 In the future, measurement of high-risk tissue might enable determination of choices of (or responses to) specific therapies such as open lung ventilation, neuromuscular blockade 35 or prone positioning. 5 There are important limitations. The current results are preliminary, and translation is limited by interspecies anatomical differences. 36 Furthermore, the human data are retrospective with a small sample size; thus, prospective human studies will be required to better calibrate the methodology for human microanatomy and to verify the prognostic value. Because our images were static, we likely underdetected viscoelastic and recruitment behaviours that affect dynamic inflation. 37 The study design was intended to correlate baseline imaging characteristics with variable injury progression using an observational design similar to that