Inspiratory Lung Expansion in Patients with Interstitial Lung Disease: CT Histogram Analyses

Volume changes of diseased and normal areas in progressive fibrosing interstitial lung disease on inspiratory and expiratory computed tomography

PURPOSE

The diagnosis of progressive fibrosing interstitial lung disease (PF-ILD) using computed tomography (CT) is an important medical practice in respiratory care, and most imaging findings for this disease have been obtained with inspiratory CT. It is possible that some characteristic changes in respiration may be seen in normal and diseased lung in PF-ILD, which may lead to a new understanding of the pathogenesis of interstitial pneumonia, but it has never been examined. In this study, we collected and selected inspiratory and expiratory CT scans performed in pure PF-ILD cases, and evaluated the volumes of diseased and normal lung separately by manual detection and 3-dimensional volumetry to characterize the dynamic features of PF-ILD.

MATERIALS AND METHODS

Cases were collected retrospectively from a total of 753 inspiratory and expiratory CT scans performed at our hospital over a 3-year period. Sixteen cases of pure PF-ILD, excluding almost all other diseases, were included. We measured their diseased, normal, and the whole lung volumes manually and evaluated the correlation of their values and their relationship with respiratory function tests (FVC, FVC%-predicted, and DLCO%-predicted).

RESULTS

The relative expansion rate of the diseased lung is no less than that of the normal lung. The "Expansion volume of total lung" divided by the "Expansion volume of normal lung" was found to be significantly associated with DLCO%-predicted abnormalities (p = 0.0073).

CONCLUSION

The diseased lung in PF-ILD retained expansion capacity comparable to the normal lung, suggesting a negative impact on respiratory function.

Heterogeneity Analysis of Chest CT Predict Individual Prognosis of COVID-19 Patients

BACKGROUND

Ground-glass opacity (GGO) and consolidation opacity (CLO) are the common CT lung opacities, and their heterogeneity may have potential for prognosis ofcoronavirus disease-19 (COVID-19) patients.

OBJECTIVE

This study aimed to estimate clinical outcomes in individual COVID-19 patients using histogram heterogeneity analysis based on CT opacities.

METHODS

71 COVID-19 cases' medical records were retrospectively reviewed from a designated hospital in Wuhan, China, from January 24th to February 28th at the early stage of the pandemic. Two characteristic lung abnormity opacities, GGO and CLO, were drawn on CT images to identify the heterogeneity using quantitative histogram analysis. The parameters (mean, mode, kurtosis, and skewness) were derived from histograms to evaluate the accuracy of clinical classification and outcome prediction. Nomograms were built to predict the risk of death and median length of hospital stays (LOS), respectively.

RESULTS

A total of 57 COVID-19 cases were eligible for the study cohort after excluding 14 cases. The highest lung abnormalities were GGO mixed with CLO in both the survival populations (26 in 42, 61.9%) and died population (10 in 15, 66.7%). The best performance heterogeneity parameters to discriminate severe type from mild/moderate counterparts were as follows: GGO_skewness: specificity=66.67%, sensitivity=78.12%, AUC=0.706; CLO_mean: specificity=70.00%, sensitivity=76.92%, and AUC=0.746. Nomogram based on histogram parameters can predict the individual risk of death and the prolonged median LOS of COVID-19 patients. C-indexes were 0.763 and 0.888 for risk of death and prolonged median LOS, respectively.

CONCLUSION

Histogram analysis method based on GGO and CLO has the ability for individual risk prediction in COVID-19 patients.

Right-Angled Traction Bronchiectasis in Differentiating Idiopathic Pulmonary Fibrosis Without Honeycombing From Idiopathic Nonspecific Interstitial Pneumonia

OBJECTIVES

The aim of this study was to conduct a radiopathologic evaluation of right-angled traction bronchiectasis to differentiate idiopathic pulmonary fibrosis (IPF) without honeycombing from idiopathic nonspecific interstitial pneumonia (NSIP).

MATERIALS AND METHODS

The derivation cohort included 78 consecutive patients with idiopathic NSIP (n = 39) or IPF (n = 39) without honeycombing who underwent preoperative thin-section computed tomography scans at a single tertiary hospital. The validation cohort comprised 22 patients (14 IPF and 8 NSIP) from another institution. We assessed conventional computed tomography findings, right-angled traction bronchiectasis on minimum intensity projection (MinIP) images, and pathologic features associated with right-angled bronchiectasis. Right-angled traction bronchiectasis was defined as abrupt kinking of a single bronchus by over 90 degrees or an abrupt angle close to 180 degrees of branching bronchi in the background of fibrosis. In the validation cohort, we evaluated the proportion of correct IPF diagnoses and interobserver agreement of 4 radiologists before and after reviewing MinIP images.

RESULTS

A probable usual interstitial pneumonia (UIP) pattern (odds ratio [OR], 6.948; 95% confidence interval [CI], 1.525-31.654; P = 0.012) and right-angled traction bronchiectasis (OR, 6.004; 95% CI, 1.980-18.209; P = 0.002) were independently associated with IPF. Patients with right-angled traction bronchiectasis were more likely to have extensive reticular opacity (OR, 1.149; 95% CI, 1.077-1.225; P < 0.001) and pathologically were more likely to have a broad extent of subpleural fibrosis (OR, 4.000; 95% CI, 1.457-10.987; P = 0.007) and relatively thick fibrosis (OR, 7.750; 95% CI, 2.504-23.991; P < 0.001). After reviewing MinIP images, the proportion of correct diagnoses increased from 40.9% to 54.5% to 50.0% to 77.3%. The mean kappa value for right-angled traction bronchiectasis was 0.489 ± 0.192.

CONCLUSIONS

Right-angled traction bronchiectasis pathologically reflected a subpleural predominance of fibrosis and partly supported the radiologic differentiation of IPF without honeycombing from idiopathic NSIP.

CT quantification of the heterogeneity of fibrosis boundaries in idiopathic pulmonary fibrosis

OBJECTIVES

To quantify the heterogeneity of fibrosis boundaries in idiopathic pulmonary fibrosis (IPF) using the Gaussian curvature analysis for evaluating disease severity and predicting survival.

METHODS

We retrospectively included 104 IPF patients and 52 controls who underwent baseline chest CT scans. Normal lungs below - 500 HU were segmented, and the boundary was three-dimensionally reconstructed using in-house software. Gaussian curvature analysis provided histogram features on the heterogeneity of the fibrosis boundary. We analyzed the correlations between histogram features and the gender-age-physiology (GAP) and CT fibrosis scores. We built a regression model to predict diffusing capacity of carbon monoxide (DLCO) using the histogram features and calculated the modified GAP (mGAP) score by replacing DLCO with the predicted DLCO. The performances of the GAP, CT-GAP, and mGAP scores were compared using 100 repeated random-split sets.

RESULTS

Patients with moderate-to-severe IPF had more numerous Gaussian curvatures at the fibrosis boundary, lower uniformity, and lower 10th to 30th percentiles of Gaussian curvature than controls or patients with mild IPF (all p < 0.0033). The 20th percentile was most significantly correlated with the GAP score (r = - 0.357; p < 0.001) and the CT fibrosis score (r = - 0.343; p = 0.001). More numerous Gaussian curvatures, higher entropy, lower uniformity, and 10th to 30th percentiles (p < 0.001-0.041) were associated with mortality. The mGAP score was comparable to the GAP and CT-GAP scores for survival prediction (mean C-indices, 0.76 vs. 0.79 vs. 0.77, respectively).

CONCLUSIONS

Gaussian curvatures of fibrosis boundaries became more heterogeneous as the disease progressed, and heterogeneity was negatively associated with survival in IPF.

KEY POINTS

• Gaussian curvature of the fibrotic lung boundary was more heterogeneous in patients with moderate-to-severe IPF than those with mild IPF or normal controls. • The 20th percentile of the Gaussian curvature of the fibrosis boundary was linearly correlated with the GAP score and the CT fibrosis score. • A modified GAP score that replaced the diffusing capacity of carbon monoxide with a composite measure using histogram features of the Gaussian curvature of the fibrosis boundary showed a comparable ability to predict survival to both the GAP and the CT-GAP score.

High throughput image labeling on chest computed tomography by deep learning

When mining image data from PACs or clinical trials or processing large volumes of data without curation, the relevant scans must be identified among irrelevant or redundant data. Only images acquired with appropriate technical factors, patient positioning, and physiological conditions may be applicable to a particular image processing or machine learning task. Automatic labeling is important to make big data mining practical by replacing conventional manual review of every single-image series. Digital imaging and communications in medicine headers usually do not provide all the necessary labels and are sometimes incorrect. We propose an image-based high throughput labeling pipeline using deep learning, aimed at identifying scan direction, scan posture, lung coverage, contrast usage, and breath-hold types. They were posed as different classification problems and some of them involved further segmentation and identification of anatomic landmarks. Images of different view planes were used depending on the specific classification problem. All of our models achieved accuracy > 99 % on test set across different tasks using a research database from multicenter clinical trials.