Machine learning-based computer-aided simple triage (CAST) for COVID-19 pneumonia as compared with triage by board-certified chest radiologists.

PURPOSE: Several reporting systems have been proposed for providing standardized language and diagnostic categories aiming for expressing the likelihood that lung abnormalities on CT images represent COVID-19. We developed a machine learning (ML)-based CT texture analysis software for simple triage based on the RSNA Expert Consensus Statement system. The purpose of this study was to conduct a multi-center and multi-reader study to determine the capability of ML-based computer-aided simple triage (CAST) software based on RSNA expert consensus statements for diagnosis of COVID-19 pneumonia. METHODS: For this multi-center study, 174 cases who had undergone CT and polymerase chain reaction (PCR) tests for COVID-19 were retrospectively included. Their CT data were then assessed by CAST and consensus from three board-certified chest radiologists, after which all cases were classified as either positive or negative. Diagnostic performance was then compared by McNemar's test. To determine radiological finding evaluation capability of CAST, three other board-certified chest radiologists assessed CAST results for radiological findings into five criteria. Finally, accuracies of all radiological evaluations were compared by McNemar's test. RESULTS: A comparison of diagnosis for COVID-19 pneumonia based on RT-PCR results for cases with COVID-19 pneumonia findings on CT showed no significant difference of diagnostic performance between ML-based CAST software and consensus evaluation (p > 0.05). Comparison of agreement on accuracy for all radiological finding evaluations showed that emphysema evaluation accuracy for investigator A (AC = 91.7%) was significantly lower than that for investigators B (100%, p = 0.0009) and C (100%, p = 0.0009). CONCLUSION: This multi-center study shows COVID-19 pneumonia triage by CAST can be considered at least as valid as that by chest expert radiologists and may be capable for playing as useful a complementary role for management of suspected COVID-19 pneumonia patients as well as the RT-PCR test in routine clinical practice.

Deep learning reconstruction allows low-dose imaging while maintaining image quality: comparison of deep learning reconstruction and hybrid iterative reconstruction in contrast-enhanced abdominal CT.

We aimed to compare the radiation dose and image quality of a low-dose abdominal computed tomography (CT) protocol reconstructed with deep learning reconstruction (DLR) with those of a routine-dose protocol reconstructed with hybrid-iterative reconstruction. This retrospective study enrolled 71 patients [61 men; average age, 71.9 years; mean body mass index (BMI), 24.3 kg/m2] who underwent both low-dose abdominal CT with DLR [advanced intelligent clear-IQ engine (AiCE)] and routine-dose abdominal CT with hybrid-iterative reconstruction [adaptive iterative dose reduction 3D (AIDR 3D)]. Radiation dose parameters included volume CT dose index (CTDIvol), effective dose (ED), and size-specific dose estimate (SSDE). Mean image noise and contrast-to-noise ratio (CNR) were calculated. Image noise was measured in the hepatic parenchyma and bilateral erector spinae muscles. Moreover, subjective assessment of perceived image quality and diagnostic acceptability was performed. The low-dose protocol helped reduce the CTDIvol by 44.3%, ED by 43.7%, and SSDE by 44.9%. Moreover, the noise was significantly lower and CNR significantly higher with the low-dose protocol than with the normal-dose protocol (P<0.001). In the subjective assessment of image quality, there was no significant difference between the protocols with regard to image noise. Overall, AiCE was superior to AIDR 3D in terms of diagnostic acceptability (P=0.001). The use of AiCE can reduce overall radiation dose by more than 40% without loss of image quality compared to routine-dose abdominal CT with AIDR 3D.

Newly developed artificial intelligence algorithm for COVID-19 pneumonia: utility of quantitative CT texture analysis for prediction of favipiravir treatment effect.

PURPOSE: Using CT findings from a prospective, randomized, open-label multicenter trial of favipiravir treatment of COVID-19 patients, the purpose of this study was to compare the utility of machine learning (ML)-based algorithm with that of CT-determined disease severity score and time from disease onset to CT (i.e., time until CT) in this setting. MATERIALS AND METHODS: From March to May 2020, 32 COVID-19 patients underwent initial chest CT before enrollment were evaluated in this study. Eighteen patients were randomized to start favipiravir on day 1 (early treatment group), and 14 patients on day 6 of study participation (late treatment group). In this study, percentages of ground-glass opacity (GGO), reticulation, consolidation, emphysema, honeycomb, and nodular lesion volumes were calculated as quantitative indexes by means of the software, while CT-determined disease severity was also visually scored. Next, univariate and stepwise regression analyses were performed to determine relationships between quantitative indexes and time until CT. Moreover, patient outcomes determined as viral clearance in the first 6 days and duration of fever were compared for those who started therapy within 4, 5, or 6 days as time until CT and those who started later by means of the Kaplan-Meier method followed by Wilcoxon's signed-rank test. RESULTS: % GGO and % consolidation showed significant correlations with time until CT (p < 0.05), and stepwise regression analyses identified both indexes as significant descriptors for time until CT (p < 0.05). When divided all patients between time until CT of 4 days and that of more than 4 days, accuracy of the combined quantitative method (87.5%) was significantly higher than that of the CT disease severity score (62.5%, p = 0.008). CONCLUSION: ML-based CT texture analysis is equally or more useful for predicting time until CT for favipiravir treatment on COVID-19 patients than CT disease severity score.

Diagnostic performance of coronary CT angiography with ultra-high-resolution CT: Comparison with invasive coronary angiography.

PURPOSE: Recently, ultra-high-resolution computed tomography (U-HRCT) with a 0.25 mm × 128-row detector was introduced. The purpose of this study was to evaluate the diagnostic performance of coronary CT angiography (CCTA) using U-HRCT. METHODS: This retrospective study included 38 consecutive patients with suspected coronary artery disease (CAD) who underwent CCTA with U-HRCT followed by invasive coronary angiography (ICA). Per-segment diameter stenosis was calculated. Diagnostic performance of CCTA relative to ICA as the reference standard was determined. For segments with >30% diameter stenosis, the correlation and agreement of percent diameter stenosis between CCTA and ICA were calculated. RESULTS: Obstructive CAD was observed in 65 segments (12%) of 51 vessels (45%) in 32 patients (84%) during ICA. The per-patient, vessel, and segment analyses showed a sensitivity of 100% (95% confidence interval [CI], 95%-100%), 96% (95% CI: 89%-99%) and 95% (95% CI: 89%-98%), respectively, and a specificity of 67% (95% CI: 38%-67%), 81% (95% CI: 75%-83%) and 96% (95% CI: 96%-97%), respectively. The percentage of diameter stenosis, as determined by CCTA, demonstrated an excellent correlation with ICA (R = 0.90; 95% CI: 0.83-0.95) and a slight significant overestimation (mean: 4% ±â€¯7%, p < .01), with the agreed range of limits being ±â€¯16%. The median effective radiation dose for CCTA was 5.4 mSv (range: 2.9-18.0 mSv). CONCLUSIONS: CCTA with U-HRCT demonstrated an excellent correlation and agreement with ICA in the quantification of coronary artery stenosis.

Systematic evaluation of collateral pathways to the artery of Adamkiewicz using computed tomography.

OBJECTIVES: Preoperative identification of the artery of Adamkiewicz can help prevent postoperative spinal cord injury in patients undergoing thoracic and thoraco-abdominal aortic aneurysm repair. Although several studies have shown the feasibility of evaluating the artery of Adamkiewicz using multidetector row computed tomography (MDCT), no detailed investigations regarding the collateral circulation to the artery of Adamkiewicz have been performed. The purpose of this study was to investigate the collateral circulation to the artery of Adamkiewicz using MDCT in patients with thoracic and thoraco-abdominal aortic aneurysms. METHODS: Our institutional review board approved this study. Sixty-four patients with descending thoracic and thoraco-abdominal aortic aneurysms associated with the occlusion of the segmental artery from which the artery of Adamkiewicz originated were scanned using 64- or 320-detector row computed tomography. Two independent observers evaluated the MDCT images based on the degree of visualization of the artery of Adamkiewicz and its collateral circulation using a 4-point scale. RESULTS: The average visualization score was 2.8 ± 0.6. In 53 of the 64 (83%) patients, image quality was judged to be diagnostic. MDCT demonstrated 75 collateral pathways to the artery of Adamkiewicz in these 53 patients. Sixty-four of the 75 (85%) pathways were collaterals around the spinal column, and the remaining 11 (15%) pathways were collateral arteries in the thoracic wall. CONCLUSIONS: MDCT revealed the collateral pathways to the artery of Adamkiewicz around the spinal column and in the thoracic wall in 83% of our patients with thoracic and thoraco-abdominal aortic aneurysms.

Ultra-high-resolution CT angiography of the artery of Adamkiewicz: a feasibility study.

PURPOSE: Preoperative identification of the artery of Adamkiewicz can help prevent postoperative spinal cord injury following thoracic and thoracoabdominal aortic repair. Several studies have demonstrated the feasibility of evaluating the artery of Adamkiewicz using multi-detector row computed tomography (CT), but precise visualization remains a challenge. The present study was conducted to evaluate the usefulness of ultra-high-resolution CT for visualizing the artery of Adamkiewicz with a slice thickness of 0.25 versus 0.5 mm in patients with aortic aneurysms. METHODS: Our institutional review board approved this study. Twenty-four patients with thoracic and thoracoabdominal aneurysms were scanned with beam collimation of 0.25 mm × 128. Images were reconstructed with slice thicknesses of 0.25 and 0.5 mm. The signal-to-noise ratio (SNR) of the aorta and contrast-to-noise ratio (CNR) between the anterior spinal artery and spinal cord were measured. Two independent observers evaluated visualization of the artery of Adamkiewicz and its continuity between the anterior spinal artery and the aorta using a four-point scale. RESULTS: No significant differences in the SNR of the aorta or CNR of the anterior spinal artery were observed between 0.25- and 0.5-mm slices. The average visualization score was significantly higher for 0.25-mm slices (3.58 ± 0.78) than for 0.5-mm slices (3.13 ± 0.99) (p = 0.01). The percentage of patients with nondiagnostic image quality was significantly lower for 0.25-mm slices (8.3%) than for 0.5-mm slices (33.3%) (p = 0.03). CONCLUSION: In patients with aortic aneurysms, ultra-high-resolution CT with 0.25-mm slices significantly improves visualization of the artery of Adamkiewicz compared to 0.5-mm slices.

Assessment of coronary in-stent restenosis: value of subtraction coronary computed tomography angiography.

In conventional coronary computed tomography angiography (CCTA), metal artifacts are frequently observed where stents are located, making it difficult to evaluate in-stent restenosis. This study was conducted to investigate whether subtraction CCTA can improve diagnostic accuracy in the evaluation of in-stent restenosis. Subtraction CCTA was performed using 320-row CT in 398 patients with previously placed stents who were able to hold their breath for 25 s and in whom mid-diastolic prospective one-beat scanning was possible. Among these patients, 126 patients (94 men and 32 women, age 74 ± 8 years) with 370 stents who also underwent invasive coronary angiography (ICA) were selected as the subjects of this study. With ICA findings considered the gold standard, conventional CCTA was compared against subtraction CCTA to determine whether subtraction can improve diagnostic accuracy in the evaluation of in-stent restenosis. When non-assessable stents were considered to be stenotic, the diagnostic accuracy in the evaluation of in-stent restenosis was 62.7 % for conventional CCTA and 89.5 % for subtraction CCTA. When the non-assessable stents were considered to be non-stenotic the diagnostic accuracy was 90.3 % for conventional CCTA and 94.31 % for subtraction CCTA. When subtraction CCTA was used to evaluate only the 138 stents that were judged to be non-assessable by conventional CCTA, 116 of these stents were judged to be assessable, and the findings for 109 of them agreed with those obtained by ICA. Even for stents with an internal diameter of 2.5-3 mm, the lumen can be evaluated in more than 80 % of patients. Subtraction CCTA provides significantly higher diagnostic accuracy than conventional CCTA in the evaluation of in-stent restenosis.

Subtraction coronary computed tomography in patients with severe calcification.

To investigate the clinical usefulness of subtraction coronary computed tomographic angiography (CCTA) in patients with severe calcification. A 320-row area detector CT system was used in this study. The subjects were 78 patients (47 men and 31 women, 739 years of age) with an Agatston score of >300 who were able to undergo prospective one-beat scanning during a single breath-hold. The CCTA findings were compared against invasive coronary angiography. The diagnostic capabilities of CCTA for the severely calcified segments with and without the additional information provided by subtraction CCTA were compared. Severe calcification was observed in 174 (31.9%) of the 546 segments, and non-assessable regions were observed in 74 (13.6%) of the segments. The addition of subtraction CCTA information improved the diagnostic accuracy for segments with severe calcification from 67.8 to 82.8% on a per-segment basis and from 70.1 to 82.1% on a per-patient basis, with non-assessable segments considered to be stenotic. When non-assessable segments were considered to be an incorrect diagnosis, the diagnostic accuracy was improved from 48.3 to 75.9% on a per-segment basis and from 43.3 to 79.1% on a per-patient basis. In addition, when evaluation was limited to non-assessable segments, subtraction CCTA provided a diagnostic accuracy of 81.1% when non-assessable segments were considered to be stenotic or 66.2% when non-assessable segments were considered to be an incorrect diagnosis. Subtraction CCTA improves the diagnostic capabilities of CCTA in patients with severe calcification.

Feasibility of coronary calcium and stent image subtraction using 320-detector row CT angiography.

BACKGROUND: The reader confidence and diagnostic accuracy of coronary CT angiography (CCTA) can be compromised by the presence of calcified plaques and stents causing blooming artifacts. Compared to conventional invasive coronary angiography (ICA), this may cause an overestimation of stenosis severity leading to false-positive results. In a pilot study, we tested the feasibility of a new coronary calcium image subtraction algorithm in relation to reader confidence and diagnostic accuracy. METHODS: Forty-three patients underwent clinically indicated ICA and CCTA using a 320-detector row CT. Median Agatston score was 510. Two data sets were reconstructed: a conventional CCTA (CCTAconv) and a subtracted CCTA (CCTAsub), where calcifications detected on noncontrast images were subtracted from the CCTA. Reader confidence and concordance with ICA for identification of >50% stenosis were recorded. We defined target segments on CCTAconv as motion-free coronary segments with calcification or stent and low reader confidence. The effect of CCTAsub was assessed. No approval from the ethics committee was required according to Danish law. RESULTS: A total of 76 target segments were identified. The use of coronary calcium image subtraction improved the reader confidence in 66% of these segments. In target segments, specificity (86% vs 65%; P < .01) and positive predictive value (71% vs 51%, P = .03) were improved using CCTAsub compared to CCTAconv without loss in negative predictive value. CONCLUSIONS: Our initial experience with coronary calcium image subtraction suggests that it is feasible and could lead to an improvement in reader confidence and diagnostic accuracy for identification of significant coronary artery disease.