Accuracy of Conventional and Machine Learning Enhanced Chest Radiography for the Assessment of COVID-19 Pneumonia: Intra-Individual Comparison with CT.

PURPOSE: To evaluate diagnostic accuracy of conventional radiography (CXR) and machine learning enhanced CXR (mlCXR) for the detection and quantification of disease-extent in COVID-19 patients compared to chest-CT. METHODS: Real-time polymerase chain reaction (rt-PCR)-confirmed COVID-19-patients undergoing CXR from March to April 2020 together with COVID-19 negative patients as control group were retrospectively included. Two independent readers assessed CXR and mlCXR images for presence, disease extent and type (consolidation vs. ground-glass opacities (GGOs) of COVID-19-pneumonia. Further, readers had to assign confidence levels to their diagnosis. CT obtained ≤ 36 h from acquisition of CXR served as standard of reference. Inter-reader agreement, sensitivity for detection and disease extent of COVID-19-pneumonia compared to CT was calculated. McNemar test was used to test for significant differences. RESULTS: Sixty patients (21 females; median age 61 years, range 38-81 years) were included. Inter-reader agreement improved from good to excellent when mlCXR instead of CXR was used (k = 0.831 vs. k = 0.742). Sensitivity for pneumonia detection improved from 79.5% to 92.3%, however, on the cost of specificity 100% vs. 71.4% (p = 0.031). Overall, sensitivity for the detection of consolidation was higher than for GGO (37.5% vs. 70.4%; respectively). No differences could be found in disease extent estimation between mlCXR and CXR, even though the detection of GGO could be improved. Diagnostic confidence was better on mlCXR compared to CXR (p = 0.013). CONCLUSION: In line with the current literature, the sensitivity for detection and quantification of COVID-19-pneumonia was moderate with CXR and could be improved when mlCXR was used for image interpretation.

Volumetric assessment of solid pulmonary nodules on ultralow-dose CT: a phantom study.

BACKGROUND: To reduce the radiation exposure from chest computed tomography (CT), ultralow-dose CT (ULDCT) protocols performed at sub-millisievert levels were previously tested for the evaluation of pulmonary nodules (PNs). The purpose of our study was to investigate the effect of ULDCT and iterative image reconstruction on volumetric measurements of solid PNs. METHODS: CT datasets of an anthropomorphic chest phantom containing solid microspheres were obtained with a third-generation dual-source CT at standard dose, 1/8th, 1/20th and 1/70th of standard dose [CT volume dose index (CTDIvol): 0.03-2.03 mGy]. Semi-automated volumetric measurements were performed on CT datasets reconstructed with filtered back projection (FBP) and advanced modelled iterative reconstruction (ADMIRE), at strength level 3 and 5. Absolute percentage error (APE) evaluated measurement accuracy related to the effective volume. Scan repetition differences were evaluated using Bland-Altman analysis. Two-way analysis of variance (ANOVA) assessed influence of different scan parameters on APE. Proportional differences (PDs) tested the effect of dose settings and reconstruction algorithms on volumetric measurements, as compared to the standard protocol (standard dose-FBP). RESULTS: Bland-Altman analysis revealed small mean interscan differences of APE with narrow limits of agreement (-0.1%±4.3% to -0.3%±3.8%). Dose settings (P<0.001), reconstruction algorithms (P<0.001), nodule diameters (P<0.001) and nodule density (P=0.011) had statistically significant influence on APE. Post-hoc Bonferroni tests showed slightly higher APE when scanning with 1/70th of standard dose [mean difference: 3.4%, 95% confidence interval (CI): 2.5-4.3%; P<0.001], and for image reconstruction with ADMIRE5 (mean difference: 1.8%, 95% CI: 1.0-2.5%; P<0.001). No significant differences for scanning with 1/20th of standard dose (P=0.42), and image reconstruction with ADMIRE3 (P=0.19) were found. Scanning with 1/70th of standard dose and image reconstruction with FBP showed the widest range of PDs (-16.8% to 23.4%) compared to standard dose-FBP. CONCLUSIONS: Our phantom study showed no significant difference between nodule volume measurements on standard dose CT (CTDIvol: 2 mGy) and ULDCT with 1/20th of standard dose (CTDIvol: 0.10 mGy).

Erratum to the impact of slice-reduced computed tomography on histogram based densitometry assessment of lung fibrosis in patients with systemic sclerosis.

[This corrects the article DOI: 10.21037/jtd.2018.04.39.].

The impact of slice-reduced computed tomography on histogram-based densitometry assessment of lung fibrosis in patients with systemic sclerosis.

Background: To evaluate usability of slice-reduced sequential computed tomography (CT) compared to standard high-resolution CT (HRCT) in patients with systemic sclerosis (SSc) for qualitative and quantitative assessment of interstitial lung disease (ILD) with respect to (I) detection of lung parenchymal abnormalities, (II) qualitative and semiquantitative visual assessment, (III) quantification of ILD by histograms and (IV) accuracy for the 20%-cut off discrimination. Methods: From standard chest HRCT of 60 SSc patients sequential 9-slice-computed tomography (reduced HRCT) was retrospectively reconstructed. ILD was assessed by visual scoring and quantitative histogram parameters. Results from standard and reduced HRCT were compared using non-parametric tests and analysed by univariate linear regression analyses. Results: With respect to the detection of parenchymal abnormalities, only the detection of intrapulmonary bronchiectasis was significantly lower in reduced HRCT compared to standard HRCT (P=0.039). No differences were found comparing visual scores for fibrosis severity and extension from standard and reduced HRCT (P=0.051-0.073). All scores correlated significantly (P<0.001) to histogram parameters derived from both, standard and reduced HRCT. Significant higher values of kurtosis and skewness for reduced HRCT were found (both P<0.001). In contrast to standard HRCT histogram parameters from reduced HRCT showed significant discrimination at cut-off 20% fibrosis (sensitivity 88% kurtosis and skewness; specificity 81% kurtosis and 86% skewness; cut-off kurtosis ≤26, cut-off skewness ≤4; both P<0.001). Conclusions: Reduced HRCT is a robust method to assess lung fibrosis in SSc with minimal radiation dose with no difference in scoring assessment of lung fibrosis severity and extension in comparison to standard HRCT. In contrast to standard HRCT histogram parameters derived from the approach of reduced HRCT could discriminate at a threshold of 20% lung fibrosis with high sensitivity and specificity. Hence it might be used to detect early disease progression of lung fibrosis in context of monitoring and treatment of SSc patients.

Impact of stroke volume assessment by integrating multi-detector computed tomography and Doppler data on the classification of aortic stenosis.

BACKGROUND: The prevalence of low flow low gradient (LFLG) severe aortic stenosis (AS) may be overrated due to underestimation of stroke volume in two-dimensional (2D) echocardiography. The implications of 3D imaging on stroke volume calculation for AS classification have not been elucidated. Integrating multi-detector computed tomography (MDCT) and Doppler data may improve diagnostic accuracy in patients with LFLG AS. METHODS: A total of 186 patients with severe AS evaluated for transcatheter aortic valve replacement were classified according to indexed stroke volume (SVI, cut-off 35mL/m2) and mean transaortic pressure gradient (cut-off 40mmHg). SVI was calculated using a) the biplane Simpson's method, b) left ventricular outflow tract (LVOT) velocity time integral (VTI) and LVOT diameter determined by 2D echocardiography, or c) LVOT VTI and LVOT area planimetered by MDCT. RESULTS: SVI assessed by the biplane Simpson's method was smaller than that obtained from 2D echocardiography LVOT diameter (29.5±0.6 vs 34.9±0.8mL/m2, p<0.001). The latter was smaller than SVI calculated by integrating MDCT and Doppler data (47.5±1.4mL/m2, p<0.001). LFLG and paradoxical LFLG severe AS were diagnosed in 42.5% and 27.4% of patients using the biplane Simpson's method, in 30.1% and 16.7% using 2D echocardiography LVOT diameter, and in 17.2% and 8.1% when integrating MDCT and Doppler data. CONCLUSIONS: The prevalence of LFLG and paradoxical LFLG severe AS was overestimated by 2.5- and 3.4-fold based on 2D echocardiography alone. Integration of MDCT and Doppler data should be considered for stroke volume assessment in the classification of severe AS.

Assessment of bronchial and pulmonary blood supply in non-small cell lung cancer subtypes using computed tomography perfusion.

OBJECTIVES: The aim of this study was to investigate the dual blood supply of non-small cell lung cancer (NSCLC) and its association with tumor subtype, size, and stage, using computed tomography perfusion (CTP). MATERIALS AND METHODS: A total of 54 patients (median age, 65 years; range, 42-79 years; 15 women, 39 men) with suspected lung cancer underwent a CTP scan of the lung tumor. Pulmonary and bronchial vasculature regions of interest were used to calculate independently CTP parameters (blood flow [BF], blood volume [BV], and mean transit time [MTT]) of the tumor tissue. The mean and maximum pulmonary and bronchial perfusion indexes (PImean and PImax) were calculated. The tumoral volume and the largest tumoral diameter were assessed. Differences in CTP parameters and indexes among NSCLC subtypes, tumor stages and tumor dimensions were analyzed using non-parametric tests. RESULTS: According to biopsy, 37 patients had NSCLC (22 adenocarcinomas [ACs], 8 squamous cell carcinomas [SCCs], 7 large-cell carcinomas [LCC]). The mean bronchial BF/pulmonary BF, bronchial BV/pulmonary BV, and bronchial MTT/pulmonary MTT was 41.2 ± 30.0/36.9 ± 24.2 mL/100 mL/min, 11.4 ± 9.7/10.4 ± 9.4 mL/100 mL, and 11.4 ± 4.3/14.9 ± 4.4 seconds, respectively. In general, higher bronchial BF than pulmonary BF was observed in NSCLC (P = 0.014). Using a tumoral volume cutoff of 3.5 cm, a significant difference in pulmonary PImax was found (P = 0.028). There was a significantly higher mean pulmonary BF in LCCs and SCCs compared with ACs (P = 0.018 and P = 0.044, respectively), whereas the mean bronchial BF was only significantly higher in LCCs compared with ACs (P = 0.024). Correspondingly, the PImax was significantly higher in LCCs and SCCs than in ACs (P = 0.001 for both). Differences between bronchial and pulmonary PImean and PImax among T stages and Union Internationale Contre le Cancer stages were not statistically significant (P values ranging from 0.691 to 0.753). CONCLUSIONS: The known dual blood supply of NSCLC, which depends on tumor size and histological subtype, is reflected in CTP parameters, with parameters depending both on tumor size and histological subtype. This has to be accounted for when analyzing NSCLC with CTP.

Cross-sectional area measurements versus volumetric assessment of the quadriceps femoris muscle in patients with anterior cruciate ligament reconstructions.

OBJECTIVE: Our aim was to validate the use of cross-sectional area (CSA) measurements at multiple quadriceps muscle levels for estimating the total muscle volume (TMV), and to define the best correlating measurement level. METHODS: Prospective institutional review board (IRB)-approved study with written informed patient consent. Thighs of thirty-four consecutive patients with ACL-reconstructions (men, 22; women, 12) were imaged at 1.5-T using three-dimensional (3D) spoiled dual gradient-echo sequences. CSA was measured at three levels: 15, 20, and 25 cm above the knee joint line. TMV was determined using dedicated volumetry software with semiautomatic segmentation. Pearson's correlation and regression analysis (including standard error of the estimate, SEE) was used to compare CSA and TMV. RESULTS: The mean ± standard deviation (SD) for the CSA was 60.6 ± 12.8 cm(2) (range, 35.6-93.4 cm(2)), 71.1 ± 15.1 cm(2) (range, 42.5-108.9 cm(2)) and 74.2 ± 17.1 cm(2) (range, 40.9-115.9 cm(2)) for CSA-15, CSA-20 and CSA-25, respectively. The mean ± SD quadriceps' TMV was 1949 ± 533.7 cm(3) (range, 964.0-3283.0 cm(3)). Pearson correlation coefficient was r = 0.835 (p < 0.01), r = 0.906 (p < 0.01), and r = 0.956 (p < 0.01) for CSA-15, CSA-20 and CSA-25, respectively. Corresponding SEE, expressed as percentage of the TMV, were 15.2%, 11.6% and 8.1%, respectively. CONCLUSION: The best correlation coefficient between quadriceps CSA and TMV was found for CSA-25, but its clinical application to estimate the TMV is limited by a relatively large SEE. KEY POINTS: • Cross-sectional area was used to estimate QFM size in patients with ACL-reconstruction • A high correlation coefficient exists between quadriceps CSA and volume • Best correlation was seen 25 cm above the knee joint line • A relatively large standard error of the estimate limits CSA application.