Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2022 May 17;12(1):8143.
doi: 10.1038/s41598-022-12311-4.

Predefined and data driven CT densitometric features predict critical illness and hospital length of stay in COVID-19 patients

Tamar Shalmon  1   2 , Pascal Salazar  3 , Miho Horie  1   2 , Kate Hanneman  1   2 , Mini Pakkal  1   2 , Vahid Anwari  1   2 , Jennifer Fratesi  4   5
Affiliations

Predefined and data driven CT densitometric features predict critical illness and hospital length of stay in COVID-19 patients

Tamar Shalmon et al. Sci Rep. .

Abstract

The aim of this study was to compare whole lung CT density histograms to predict critical illness outcome and hospital length of stay in a cohort of 80 COVID-19 patients. CT chest images on segmented lungs were retrospectively analyzed. Functional Principal Component Analysis (FPCA) was used to find the main modes of variations on CT density histograms. CT density features, the CT severity score, the COVID-GRAM score and the patient clinical data were assessed for predicting the patient outcome using logistic regression models and survival analysis. ROC analysis predictors of critically ill status: 87.5th percentile CT density (Q875)-AUC 0.88 95% CI (0.79 0.94), F1-CT-AUC 0.87 (0.77 0.93) Standard Deviation (SD-CT)-AUC 0.86 (0.73, 0.93). Multivariate models combining CT-density predictors and Neutrophil-Lymphocyte Ratio showed the highest accuracy. SD-CT, Q875 and F1 score were significant predictors of hospital length of stay (LOS) while controlling for hospital death using competing risks models. Moreover, two multivariate Fine-Gray regression models combining the clinical variables: age, NLR, Contrast CT factor with either Q875 or F1 CT-density predictors revealed significant effects for the prediction of LOS incidence in presence of a competing risk (death) and acceptable predictive performances (Bootstrapped C-index 0.74 [0.70 0.78]).

PubMed Disclaimer

Conflict of interest statement

J.F., T.S., M.H., K.H., M.P., V.H. and P.S. declare no competing interests. P.S. is employee of Canon Medical—HIT division.

Figures

Figure 1
Figure 1
First four modes of variation of the lung CT density curves (smoothed CT histograms). (a) F1 explains 76.7% of the fraction of variance explained (FVE). The curves vary from homogeneous low CT density (10 percentile curve) to heterogeneous higher CT density (90 percentile curve). (b) F2 explains 13.5% of the FVE. The lung CT density vary from homogeneous low density (90 percentile curve) to include mid-low density in the – 700 HU to – 350 HU range (10 percentile curve). (c) F3 explains 3.8% of the FVE. This mode of variation represents a redistribution of CT lung densities from (− 700 HU to – 600 HU) to about (− 600 HU to – 300 HU). (d) F4 explains 2.6% of the FVE. F4 represents a small shift of about 140HU from low homogeneous CT densities (10 percentile curve) toward higher CT densities (90 percentile curve).
Figure 2
Figure 2
Odd ratios with 95% confidence intervals for critically ill outcome. The continuous variables are split into two groups below / above their median value. Significant P values (P < 0.05) are bold. Variables with P values > 0.5 are not shown. Notice that 6/10 of the highest ranked predictors are lung CT density quantitative features.
Figure 3
Figure 3
Cumulative incidence plot for hospital length-of-stay (LOS) and in-hospital deaths in Covid-19 patients stratified in Low Q875 (resp. F1) patients and high Q875 (resp. F1) patients using an optimal cutoff point for combined contrast and non-contrast studies found in the univariate analysis: Q875 > − 380 HU (resp. F1 > 0.099). Both F1 and Q875 features give similar predictions of cumulative incidences for LOS and Death during the follow-up period.
Figure 4
Figure 4
Non-contrast chest CT images in a 28-year-old man with mild COVID-19 pneumonia. The patient was discharged 3 days later without ventilation or ICU admission. (a) CT images show small, rounded ground-glass opacities in the upper lobes. (b) 3D view with volume. (c) Histogram of CT density values for both lungs shows most lung voxels in a normal lung density distribution and peak at about − 800 HU. (perc. = percentile). Mean lung CT density: − 786.1HU (21st perc.), SD-CT: 131.9HU (29th perc.), Skewness: 2.58 (97th perc.), Kurtosis: 13.40 (92nd perc.), Q875: − 664HU (19th perc.), F1: − 0.43 (19th perc.), F2: -0.05 (39th perc.).
Figure 5
Figure 5
Non-contrast low dose chest CT images in a 62-year-old woman with severe COVID-19 pneumonia. The patient was admitted to ICU, required ventilation, and died in hospital. (a) CT images show bilateral extensive ground glass opacities and septal thickening. (b) 3D view with volume. (c) Histogram of CT density values demonstrates 2 peaks, the first peak at − 800 HU representing normal well aerated lung and the second peak at 0 HU representing the lung consolidations and ground glass opacities. Mean lung density: − 375.4 HU (98th percent.), SD-CT: 348.9 HU (94th perc.), skewness: − 0.02 (7th perc.), kurtosis: 1.53 (0.6th perc.), Q875: 36HU (88th perc.), F1: 0.77 (98th perc.), F2: − 0.33 (9th perc.).
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

References

    1. Wong KT, et al. Thin-section CT of severe acute respiratory syndrome: Evaluation of 73 patients exposed to or with the disease. Radiology. 2003;228:395–400. doi: 10.1148/radiol.2283030541. - DOI - PubMed
    1. Kang H, et al. Computed tomography findings of influenza A (H1N1) pneumonia in adults: Pattern analysis and prognostic comparisons. J. Comput. Assist. Tomogr. 2012;36:285–290. doi: 10.1097/RCT.0b013e31825588e6. - DOI - PubMed
    1. Das KM, et al. CT correlation with outcomes in 15 patients with acute Middle East respiratory syndrome coronavirus. Am. J. Roentgenol. 2015;204:736–742. doi: 10.2214/AJR.14.13671. - DOI - PubMed
    1. Yang R, et al. Chest CT severity score: An imaging tool for assessing severe COVID-19. Radiol. Cardiothorac. Imaging. 2020;2:e200047. doi: 10.1148/ryct.2020200047. - DOI - PMC - PubMed
    1. Liang W, et al. Development and validation of a clinical risk score to predict the occurrence of critical illness in hospitalized patients with COVID-19. JAMA Intern Med. 2020;180:1081–1089. doi: 10.1001/jamainternmed.2020.2033. - DOI - PMC - PubMed
-