Your browser doesn't support javascript.
Show: 20 | 50 | 100
Results 1 - 5 de 5
Add filters

Document Type
Year range
American Journal of Respiratory and Critical Care Medicine ; 203(9):1, 2021.
Article in English | Web of Science | ID: covidwho-1407488
American Journal of Respiratory and Critical Care Medicine ; 203(9), 2021.
Article in English | EMBASE | ID: covidwho-1277763


Rationale: Lesion segmentation is a critical step in medical image analysis, and methods to identify pathology without time-intensive manual labeling of data are of utmost importance during a pandemic and in resource-constrained healthcare settings. Here, we describe an unsupervised method of automatic lesion segmentation and quantification of COVID-19 lung tissue on chest Computed Tomography (CT) scans. Methods: Anonymized human COVID-19 (n=53), and non-pathologic control (n=87) inspiratory CT scans were used to train a publically available cycle-consistent generative adversarial network (CycleGAN), to convert the COVID-19 CT scans into generated "healthy" equivalents. Difference maps were created by subtracting the Hounsfield Units (HU) value for each voxel in the generated image from that of the original COVID-19 image. We then used these difference maps to construct 3D lesion segmentations to further quantitatively characterize COVID-19 lesions in an automated pipeline. Results: The CycleGAN produced lesion segmentations from COVID-19 CT scans of varying radiologic severity ranging from cases of patchy ground glass opacities to diffuse consolidative lesions. Images of COVID-19 patients showed higher HU intensity in original vs. generated images at sites of pulmonary lesions, while preserving normal parenchyma, fissures, vasculature, and airways (Figure 1, upper panels). The generated images showed larger lung gas volumes and lower tissue masses compared to their corresponding original COVID-19 images (p<0.001). Subtraction of the generated images from their corresponding original COVID-19 CT scans yielded difference maps showing the pathological tissue alone (Figure 1). Control, non-pathologic CT images were given as input to the CycleGAN, resulting in generated images nearly superimposable with the originals with no difference in gas volume or tissue mass (Figure 1, lower panel). Conclusions: To our knowledge, this is the first unsupervised COVID-19 lesion segmentation approach. Our automated lesion model performed well in mild and severe COVID-19 cases without the need for manually labelled lung segmentations as inputs. An automated lesion segmentation model can be used clinically to rapidly and objectively quantify pathologic pulmonary tissue to inform disease prognosis and treatment. Automated radiologic techniques, such as our model, circumvent the traditional bottle-neck of manually labeling data which has limited the scale and thus the impact of quantitative radiologic medical research.

American Journal of Respiratory and Critical Care Medicine ; 203(9), 2021.
Article in English | EMBASE | ID: covidwho-1277558


Rationale The pulmonary vasculature is critical for gas exchange, impacts both pulmonary and cardiac function, and has renewed importance due to COVID-19. Pulmonary blood volume is, however, technically difficult to assess, generally requiring invasive methodology for quantification. Prior studies are limited in size and participant enrollment was selective;therefore, variation in the general population is largely unknown. We performed contrast-enhanced dual-energy computed tomography (DECT) in a multicenter, community-based cohort to describe variation in pulmonary perfused blood volume (PBV) in the community. MethodsThe Multi-Ethnic Study of Atherosclerosis (MESA) recruited adults from six sites. The MESA Lung Study invited all MESA participants attending Exam 6 (2017-18), excluding those with kidney disease and contrast allergy, to undergo DECT at functional residual capacity via Siemens Flash or Force scanner: CareDose on, pitch 0.55, 0.25 sec exposure, 0.5mm slice thickness, iterative reconstruction (Admire) with Qr40 Kernel. Half concentration 370mg/ml Iopamidol was delivered at 4ml/s for the full scan, starting 17 seconds prior to scanning, including a ∼4 sec breath hold. PBV was calculated by material decomposition and normalized with iodine concentration in the pulmonary trunk. Generalized linear regression models included age, sex, race/ethnicity, height, weight, smoking status, site, and education.ResultsDECT scans were acquired for 714 participants, 36 of which were excluded due to image quality. Mean age of the remaining 678 participants was 71 years (range 63 - 79), 55% were male, 51% were ever smokers, and the race/ethnic distribution was 41% White, 29% Black, 17% Hispanic, and 13% Asian. Mean PBV was 468 + 151mL. The strongest demographic correlate was lower PBV with greater age (-30 mL per 10 years, 95% CI: -43, -18, p<0.001). Pulmonary PBV was positively associated with height, weight, and male sex (all P<0.001). PBV was lower in former compared to never smokers (p =0.04) and in Black than White participants (p=0.002), but not in Hispanic or Asian participants. There were no consistent differences across education or study site. Results were similar after adjustment for lung function and percent emphysema on CT.ConclusionsTo our knowledge, this is the first assessment of pulmonary PBV in a large, multiethnic, general community sample. Pulmonary PBV assessed by contrast-enhanced DECT was substantially reduced with advancing age and varied with body size, sex, former smoking, and, to a lesser extent, Black race. Understanding variation in pulmonary PBV in the general population may elucidate risk of cardiopulmonary disease and physical function.

American Journal of Respiratory and Critical Care Medicine ; 203(9), 2021.
Article in English | EMBASE | ID: covidwho-1277352


RATIONALE: Chest computed tomography (CT) has a potential role in the diagnosis, detection of complications, and prognosis of coronavirus disease 2019 (COVID-19). The value of chest CT can be further amplified when associated to physiological variables. Some studies have done efforts to correlate chest CT findings with overall oxygenation and respiratory mechanics, which although they are easily obtained may not be specifically related to COVID-19. Very few studies have tried to correlate chest CT findings with specific biomarkers related to COVID-19. For this purpose, temporal changes of chest CT were evaluated and then correlated with laboratory data in multicenter randomized clinical trial. METHODS: Adult patients who presented chest CT scan features compatible with viral pneumonia were admitted in the hospital and followed during 7 days (NCT: 04561219). CT scans and laboratory data [D-dimer, ferritin, and lactate dehydrogenase (LDH)] in blood were obtained at the moment of admission (Baseline) and on day 7 (Final). Qualitative and quantitative chest CT scan parameters were evaluated in ventral, middle and dorsal regions of interest (ROI) and classified as: hyper-, normal-, poor-, and non-aerated. RESULTS: In this study involving 45 COVID-19 patients no statistically significant differences in the overall Hounsfield Units (HU) ranges and percent of whole lung mass were found overtime. Normally aerated lung tissue reduced from Baseline to Final (p=0.004), mainly associated with a decrease in ventral (p=0.001) and middle (p=0.026) ROIs. At dorsal ROI, a reduction in CT lung mass in poorly aerated areas was observed from Baseline to Final. Poorly aerated and non-aerated lung areas were well correlated only with D-dimer blood levels (r=0.55, p<0.001;and r=0.52, p=0.001, respectively). CONCLUSION: In patients with COVID-19 pneumonia, changes in poor-and non-aerated were associated to changes in D-dimer blood levels, which may be a specific biomarker to be follow in facilities without CT as a way to infer radiologic changes.