Lung hyperaeration assessment by computed tomography: correction of reconstruction-induced bias

Early versus late intubation in COVID-19 patients failing helmet CPAP: A quantitative computed tomography study

PURPOSE

To describe the effects of timing of intubation in COVID-19 patients that fail helmet continuous positive airway pressure (h-CPAP) on progression and severity of disease.

METHODS

COVID-19 patients that failed h-CPAP, required intubation, and underwent chest computed tomography (CT) at two levels of positive end-expiratory pressure (PEEP, 8 and 16 cmH₂O) were included in this retrospective study. Patients were divided in two groups (early versus late) based on the duration of h-CPAP before intubation. Endpoints included percentage of non-aerated lung tissue at PEEP of 8 cmH₂O, respiratory system compliance and oxygenation.

RESULTS

Fifty-two patients were included and classified in early (h-CPAP for ≤2 days, N = 26) and late groups (h-CPAP for >2 days, N = 26). Patients in the late compared to early intubation group presented: 1) lower respiratory system compliance (median difference, MD -7 mL/cmH₂O, p = 0.044) and PaO₂/FiO₂ (MD -29 mmHg, p = 0.047), 2) higher percentage of non-aerated lung tissue (MD 7.2%, p = 0.023) and 3) similar lung recruitment increasing PEEP from 8 to 16 cmH₂O (MD 0.1%, p = 0.964).

CONCLUSIONS

In COVID-19 patients receiving h-CPAP, late intubation was associated with worse clinical presentation at ICU admission and more advanced disease. The possible detrimental effects of delaying intubation should be carefully considered in these patients.

Automatic Lung Segmentation and Quantification of Aeration in Computed Tomography of the Chest Using 3D Transfer Learning

BACKGROUND

Identification of lung parenchyma on computer tomographic (CT) scans in the research setting is done semi-automatically and requires cumbersome manual correction. This is especially true in pathological conditions, hindering the clinical application of aeration compartment (AC) analysis. Deep learning based algorithms have lately been shown to be reliable and time-efficient in segmenting pathologic lungs. In this contribution, we thus propose a novel 3D transfer learning based approach to quantify lung volumes, aeration compartments and lung recruitability.

METHODS

Two convolutional neural networks developed for biomedical image segmentation (uNet), with different resolutions and fields of view, were implemented using Matlab. Training and evaluation was done on 180 scans of 18 pigs in experimental ARDS (u2Net Pig ) and on a clinical data set of 150 scans from 58 ICU patients with lung conditions varying from healthy, to COPD, to ARDS and COVID-19 (u2Net Human ). One manual segmentations (MS) was available for each scan, being a consensus by two experts. Transfer learning was then applied to train u2Net Pig on the clinical data set generating u2Net Transfer . General segmentation quality was quantified using the Jaccard index (JI) and the Boundary Function score (BF). The slope between JI or BF and relative volume of non-aerated compartment (S JI and S BF , respectively) was calculated over data sets to assess robustness toward non-aerated lung regions. Additionally, the relative volume of ACs and lung volumes (LV) were compared between automatic and MS.

RESULTS

On the experimental data set, u2Net Pig resulted in JI = 0.892 [0.88 : 091] (median [inter-quartile range]), BF = 0.995 [0.98 : 1.0] and slopes S JI = -0.2 {95% conf. int. -0.23 : -0.16} and S BF = -0.1 {-0.5 : -0.06}. u2Net Human showed similar performance compared to u2Net Pig in JI, BF but with reduced robustness S JI = -0.29 {-0.36 : -0.22} and S BF = -0.43 {-0.54 : -0.31}. Transfer learning improved overall JI = 0.92 [0.88 : 0.94], P < 0.001, but reduced robustness S JI = -0.46 {-0.52 : -0.40}, and affected neither BF = 0.96 [0.91 : 0.98] nor S BF = -0.48 {-0.59 : -0.36}. u2Net Transfer improved JI compared to u2Net Human in segmenting healthy (P = 0.008), ARDS (P < 0.001) and COPD (P = 0.004) patients but not in COVID-19 patients (P = 0.298). ACs and LV determined using u2Net Transfer segmentations exhibited < 5% volume difference compared to MS.

CONCLUSION

Compared to manual segmentations, automatic uNet based 3D lung segmentation provides acceptable quality for both clinical and scientific purposes in the quantification of lung volumes, aeration compartments, and recruitability.

Assessment of Lung Reaeration at 2 Levels of Positive End-expiratory Pressure in Patients With Early and Late COVID-19-related Acute Respiratory Distress Syndrome

PURPOSE

Patients with novel coronavirus disease (COVID-19) frequently develop acute respiratory distress syndrome (ARDS) and need invasive ventilation. The potential to reaerate consolidated lung tissue in COVID-19-related ARDS is heavily debated. This study assessed the potential to reaerate lung consolidations in patients with COVID-19-related ARDS under invasive ventilation.

MATERIALS AND METHODS

This was a retrospective analysis of patients with COVID-19-related ARDS who underwent chest computed tomography (CT) at low positive end-expiratory pressure (PEEP) and after a recruitment maneuver at high PEEP of 20 cm H2O. Lung reaeration, volume, and weight were calculated using both CT scans. CT scans were performed after intubation and start of ventilation (early CT), or after several days of intensive care unit admission (late CT).

RESULTS

Twenty-eight patients were analyzed. The median percentages of reaerated and nonaerated lung tissue were 19% [interquartile range, IQR: 10 to 33] and 11% [IQR: 4 to 15] for patients with early and late CT scans, respectively (P=0.049). End-expiratory lung volume showed a median increase of 663 mL [IQR: 483 to 865] and 574 mL [IQR: 292 to 670] after recruitment for patients with early and late CT scans, respectively (P=0.43). The median decrease in lung weight attributed to nonaerated lung tissue was 229 g [IQR: 165 to 376] and 171 g [IQR: 81 to 229] after recruitment for patients with early and late CT scans, respectively (P=0.16).

CONCLUSIONS

The majority of patients with COVID-19-related ARDS undergoing invasive ventilation had substantial reaeration of lung consolidations after recruitment and ventilation at high PEEP. Higher PEEP can be considered in patients with reaerated lung consolidations when accompanied by improvement in compliance and gas exchange.

Extension of Collagen Deposition in COVID-19 Post Mortem Lung Samples and Computed Tomography Analysis Findings

Lung fibrosis has specific computed tomography (CT) findings and represents a common finding in advanced COVID-19 pneumonia whose reversibility has been poorly investigated. The aim of this study was to quantify the extension of collagen deposition and aeration in postmortem cryobiopsies of critically ill COVID-19 patients and to describe the correlations with qualitative and quantitative analyses of lung CT. Postmortem transbronchial cryobiopsy samples were obtained, formalin fixed, paraffin embedded and stained with Sirius red to quantify collagen deposition, defining fibrotic samples as those with collagen deposition above 10%. Lung CT images were analyzed qualitatively with a radiographic score and quantitatively with computer-based analysis at the lobe level. Thirty samples from 10 patients with COVID-19 pneumonia deceased during invasive mechanical ventilation were included in this study. The median [interquartile range] percent collagen extension was 6.8% (4.6-16.2%). In fibrotic compared to nonfibrotic samples, the qualitative score was higher (260 (250-290) vs. 190 (120-270), p = 0.036) while the gas fraction was lower (0.46 (0.32-0.47) vs. 0.59 (0.37-0.68), p = 0.047). A radiographic score above 230 had 100% sensitivity (95% confidence interval, CI: 66.4% to 100%) and 66.7% specificity (95% CI: 41.0% to 92.3%) to detect fibrotic samples, while a gas fraction below 0.57 had 100% sensitivity (95% CI: 66.4% to 100%) and 57.1% specificity (95% CI: 26.3% to 88.0%). In COVID-19 pneumonia, qualitative and quantitative analyses of lung CT images have high sensitivity but moderate to low specificity to detect histopathological fibrosis. Pseudofibrotic CT findings do not always correspond to increased collagen deposition.

An Experimental Pre-Post Study on the Efficacy of Respiratory Physiotherapy in Severe Critically III COVID-19 Patients

Background: Respiratory physiotherapy (RPT) is considered essential in patients' management during intensive care unit (ICU) stay. The role of RPT in critically ill COVID-19 patients is poorly described. We aimed to investigate the effects of RPT on oxygenation and lung aeration in critically ill COVID-19 patients admitted to the ICU. Methods: Observational pre-post study. Patients with severe COVID-19 admitted to the ICU, who received a protocolized CPT session and for which a pre-and post-RPT lung ultrasound (LUS) was performed, were included. A subgroup of patients had an available quantitative computed tomography (CT) scan performed within 4 days from RPT. The primary aim was to evaluate whether RPT improved oxygenation; secondary aims included correlations between LUS, CT and response to RPT. Results: Twenty patients were included. The median (1st-3rd quartile) PaO2/FiO2 was 181 (105-456), 244 (137-497) and 246 (137-482) at baseline (T0), after RPT (T1), and after 6 h (T2), respectively. PaO2/FiO2 improved throughout the study (p = 0.042); particularly, PaO2/FiO2 improved at T1 in respect to T0 (p = 0.011), remaining higher at T2 (p = 0.007) compared to T0. Correlations between LUS, volume of gas (rho = 0.58, 95%CI 0.05-0.85, p = 0.033) and hyper-aerated mass at CT scan (rho = 0.54, 95% CI 0.00-0.84, p = 0.045) were detected. No significant changes in LUS score were observed before and after RPT. Conclusions: RPT improved oxygenation and the improvement persisted after 6 h. Oxygenation improvement was not reflected by aeration changes assessed with LUS. Further studies are warranted to assess the efficacy of RPT in COVID-19 ICU patients.

Magnetic Resonance Imaging for Quantitative Assessment of Lung Aeration: A Pilot Translational Study

Background: Computed tomography is the gold standard for lung aeration assessment, but exposure to ionizing radiation limits its application. We assessed the ability of magnetic resonance imaging (MRI) to detect changes in lung aeration in ex vivo isolated swine lung and the potential of translation of the findings to human MRI scans.

Methods: We performed MRI scans in 11 isolated non-injured and injured swine lungs, as well as 6 patients both pre- and post-operatively. Images were obtained using a 1.5 T MRI scanner, with T₁ – weighted volumetric interpolated breath-hold examination (VIBE) and T₂ – weighted half-Fourier acquisition single-shot turbo spin-echo (HASTE) sequences. We scanned swine lungs, with reference samples of water and muscle, at different airway pressure levels: 0, 40, 10, 2 cmH₂O. We investigated the relations between MRI signal intensity and both lung density and gas content fraction. We analyzed patients’ images according to the findings of the ex vivo model.

Results: In the ex vivo samples, the lung T₁ – VIBE signal intensity normalized to water or muscle reference signal correlated with lung density (r² = 0.98). Thresholds for poorly and non-aerated lung tissue, expressed as MRI intensity attenuation factor compared to the deflated lung, were estimated as 0.70 [95% CI: 0.65–0.74] and 0.28 [95% CI: 0.27–0.30], respectively. In patients, dorsal versus ventral regions had a higher MRI signal intensity both pre- and post-operatively (p = 0.031). Comparing post- versus pre-operative scans, lung volume decreased (p = 0.028), while the following increased: MRI signal intensity in ventral (p = 0.043) and dorsal (p < 0.0001) regions, and percentages of non-aerated (p = 0.028) and poorly aerated tissue volumes (p = 0.028).

Conclusion: Magnetic resonance imaging signal intensity is a function of lung density, decreasing linearly with increasing gas content. Lung MRI might be useful for estimating lung aeration. Compared to CT, this technique is radiation-free but requires a longer acquisition time and has a lower spatial resolution.

Lung imaging: how to get better look inside the lung

In the last years, imaging has played a key role in the diagnosis and monitoring and critical illness, including acute respiratory distress syndrome (ARDS). Chest X-ray (CXR) and computed tomography (CT) are the conventional techniques most performed in the critically ill patients, the latter being the gold standard to assess lung aeration in ARDS patients. In addition, two bedside techniques are now gaining popularity alongside the conventional ones: lung ultrasound (LUS) and electrical impedance tomography (EIT). These techniques do not involve the use of ionizing radiations, are non-invasive and relatively easy to use, and are under extensive investigation as a complement, and for some application a substitution of conventional techniques. At last, positron emission tomography (PET) and magnetic resonance imaging (MRI) can provide functional information on the lung and respiratory function, and are increasingly used in research to improve the understanding of the pathophysiological mechanisms underlying ARDS. The purpose of this review is to give an up-to-date overview of the conventional and emerging imaging techniques available the diagnosis and management of patients with ARDS.

Ultra-low-dose sequential computed tomography for quantitative lung aeration assessment-a translational study

Background

Quantitative lung computed tomography (CT) provides fundamental information about lung aeration in critically ill patients. We tested a scanning protocol combining reduced number of CT slices and tube current, comparing quantitative analysis and radiation exposure to conventional CT.

Methods

In pigs, CT scans were performed during breath hold in a model of lung injury with three different protocols: standard spiral with 180 mAs tube current-time product (Spiral180), sequential with 20-mm distance between slices and either 180 mAs (Sequential180) or 50 mAs (Sequential50). Spiral scans of critically ill patients were collected retrospectively, and subsets of equally spaced slices were extracted. The agreement between CT protocols was assessed with Bland–Altman analysis.

Results

In 12 pigs, there was good concordance between the sequential protocols and the spiral scan (all biases ≤1.9%, agreements ≤±6.5%). In Spiral180, Sequential180 and Sequential50, estimated dose exposure was 2.3 (2.1–2.8), 0.21 (0.19–0.26), and 0.09 (0.07–0.10) mSv, respectively (p < 0.001 compared to Spiral180); number of acquired slices was 244 (227–252), 12 (11–13) and 12 (11–13); acquisition time was 7 (6–7), 23 (21–25) and 24 (22–26) s. In 32 critically ill patients, quantitative analysis extrapolated from 1-mm slices interleaved by 20 mm had a good concordance with the analysis performed on the entire spiral scan (all biases <1%, agreements ≤2.2%).

Conclusions

In animal CT data, combining sequential scan and low tube current did not affect significantly the quantitative analysis, with a radiation exposure reduction of 97%, reaching a dose comparable to chest X-ray, but with longer acquisition time. In human CT data, lung aeration analysis could be extrapolated from a subset of thin equally spaced slices.

Electronic supplementary material

The online version of this article (doi: 10.1186/s40635-017-0133-6) contains supplementary material, which is available to authorized users.