RESUMEN
OBJECTIVE: This study aims to assess the efficacy of polyenergetic reconstruction methods in reducing streak artifacts caused by dual source imaging in Photon Counting Detector Computed Tomography (PCD-CT) imaging, thereby improving image quality and diagnostic accuracy. METHODS: A retrospective cohort study was conducted, involving 50 patients who underwent chest Computed Tomography Angiography with PCD-CT, focusing on those with streak artifacts. Quantitative and qualitative analyses were performed on images reconstructed using monoenergetic and polyenergetic techniques. Quantitative evaluations measured the attenuation of tracheal air density in regions affected by streak artifacts, while qualitative assessments employed a modified Likert scale to rate image quality. Statistical analyses included Wilcoxon's signed-rank tests and Spearman's correlation, alongside assessments of inter-rater reliability. RESULTS: There was significantly lower attenuation of tracheal air density on the polyenergetic reconstructions (Median - 1010 ± 62 HU vs -930 ± 110 HU; P < 0.001), and significantly decreased variation on the polyenergetic reconstructions (Median 65.2 ± 79.5 HU vs 38.8 ± 33.9 HU; P < 0.001). The median modified-Likert scale were significantly better for the polyenergetic reconstructions (median modified-Likert 4 ± 0.5 vs 2.5 ± 1; P < 0.001). The inter-rater agreement was substantial and not significantly different between reconstructions (Gwet's ACPolyenergetic = 0.78 vs Gwet's ACVMI = 0.775). CONCLUSION: Polyenergetic reconstruction significantly mitigates streak artifacts in PCD-CT imaging, enhancing quantitative and qualitative image quality. This advancement addresses a known limitation of current PCD-CT reconstruction techniques, offering a promising approach to improving diagnostic reliability and accuracy in clinical practice. We demonstrate that future software implementations can resolve this artifact.
RESUMEN
Purpose To investigate the impact of plaque size and density on virtual noncontrast (VNC)-based coronary artery calcium scoring (CACS) using photon-counting detector CT and to provide safety net reconstructions for improved detection of subtle plaques in patients whose VNC-based CACS would otherwise be erroneously zero when compared with true noncontrast (TNC)-based CACS. Materials and Methods In this prospective study, CACS was evaluated in a phantom containing calcifications with different diameters (5, 3, and 1 mm) and densities (800, 400, and 200 mg/cm3) and in participants who underwent TNC and contrast-enhanced cardiac photon-counting detector CT (July 2021-March 2022). VNC images were reconstructed at different virtual monoenergetic imaging (55-80 keV) and quantum iterative reconstruction (QIR) levels (QIR,1-4). TNC scans at 70 keV with QIR off served as the reference standard. In vitro CACS was analyzed using standard settings (3.0-mm sections, kernel Qr36, 130-HU threshold). Calcification detectability and CACS of small and low-density plaques were also evaluated using 1.0-mm sections, kernel Qr44, and 120- or 110-HU thresholds. Safety net reconstructions were defined based on background Agatston scores and evaluated in vivo in TNC plaques initially nondetectable using standard VNC reconstructions. Results The in vivo cohort included 63 participants (57.8 years ± 15.5 [SD]; 37 [59%] male, 26 [41%] female). Correlation and agreement between standard CACSVNC and CACSTNC were higher in large- and medium-sized and high- and medium-density than in low-density plaques (in vitro: intraclass correlation coefficient [ICC] ≥ 0.90; r > 0.9 vs ICC = 0.20-0.48; r = 0.5-0.6). Small plaques were not detectable using standard VNC reconstructions. Calcification detectability was highest using 1.0-mm sections, kernel Qr44, 120- and 110-HU thresholds, and QIR level of 2 or less VNC reconstructions. Compared with standard VNC, using safety net reconstructions (55 keV, QIR 2, 110-HU threshold) for in vivo subtle plaque detection led to higher detection (increased by 89% [50 of 56]) and improved correlation and agreement of CACSVNC with CACSTNC (in vivo: ICC = 0.51-0.61; r = 0.6). Conclusion Compared with TNC-based calcium scoring, VNC-based calcium scoring was limited for small and low-density plaques but improved using safety net reconstructions, which may be particularly useful in patients with low calcium scores who would otherwise be treated based on potentially false-negative results. Keywords: Coronary Artery Calcium CT, Photon-Counting Detector CT, Virtual Noncontrast, Plaque Size, Plaque Density Supplemental material is available for this article. © RSNA, 2024.
Asunto(s)
Enfermedad de la Arteria Coronaria , Fantasmas de Imagen , Placa Aterosclerótica , Humanos , Masculino , Femenino , Estudios Prospectivos , Placa Aterosclerótica/diagnóstico por imagen , Placa Aterosclerótica/patología , Persona de Mediana Edad , Enfermedad de la Arteria Coronaria/diagnóstico por imagen , Enfermedad de la Arteria Coronaria/patología , Anciano , Fotones , Vasos Coronarios/diagnóstico por imagen , Vasos Coronarios/patología , Calcificación Vascular/diagnóstico por imagen , Calcificación Vascular/patología , Tomografía Computarizada por Rayos X/métodos , Interpretación de Imagen Radiográfica Asistida por Computador/métodos , Angiografía Coronaria/métodos , Medios de ContrasteRESUMEN
Purpose: Computed tomography (CT) pulmonary angiography is considered the gold standard for pulmonary embolism (PE) diagnosis, relying on the discrimination between contrast and embolus. Photon-counting detector CT (PCD-CT) generates monoenergetic reconstructions through energy-resolved detection. Virtual monoenergetic images (VMI) at low keV can be used to improve pulmonary artery opacification. While studies have assessed VMI for PE diagnosis on dual-energy CT (DECT), there is a lack of literature on optimal settings for PCD-CT-PE reconstructions, warranting further investigation. Material and methods: Twenty-five sequential patients who underwent PCD-CT pulmonary angiography for suspicion of acute PE were retrospectively included in this study. Quantitative metrics including signal-to-noise ratio (SNR) and contrast-to-noise (CNR) ratio were calculated for 4 VMI values (40, 60, 80, and 100 keV). Qualitative measures of diagnostic quality were obtained for proximal to distal pulmonary artery branches by 2 cardiothoracic radiologists using a 5-point modified Likert scale. Results: SNR and CNR were highest for the 40 keV VMI (49.3 ± 22.2 and 48.2 ± 22.1, respectively) and were inversely related to monoenergetic keV. Qualitatively, 40 and 60 keV both exhibited excellent diagnostic quality (mean main pulmonary artery: 5.0 ± 0 and 5.0 ± 0; subsegmental pulmonary arteries 4.9 ± 0.1 and 4.9 ± 0.1, respectively) while distal segments at high (80-100) keVs had worse quality. Conclusions: 40 keV was the best individual VMI for the detection of pulmonary embolism by quantitative metrics. Qualitatively, 40-60 keV reconstructions may be used without a significant decrease in subjective quality. VMIs at higher keV lead to reduced opacification of the distal pulmonary arteries, resulting in decreased image quality.
RESUMEN
PURPOSE: Acute chest syndrome (ACS) is secondary to occlusion of the pulmonary vasculature and a potentially life-threatening complication of sickle cell disease (SCD). Dual-energy CT (DECT) iodine perfusion map reconstructions can provide a method to visualize and quantify the extent of pulmonary microthrombi. METHODS: A total of 102 patients with sickle cell disease who underwent DECT CTPA with perfusion were retrospectively identified. The presence or absence of airspace opacities, segmental perfusion defects, and acute or chronic pulmonary emboli was noted. The number of segmental perfusion defects between patients with and without acute chest syndrome was compared. Sub-analyses were performed to investigate robustness. RESULTS: Of the 102 patients, 68 were clinically determined to not have ACS and 34 were determined to have ACS by clinical criteria. Of the patients with ACS, 82.4% were found to have perfusion defects with a median of 2 perfusion defects per patient. The presence of any or new perfusion defects was significantly associated with the diagnosis of ACS (P = 0.005 and < 0.001, respectively). Excluding patients with pulmonary embolism, 79% of patients with ACS had old or new perfusion defects, and the specificity for new perfusion defects was 87%, higher than consolidation/ground glass opacities (80%). CONCLUSION: DECT iodine map has the capability to depict microthrombi as perfusion defects. The presence of segmental perfusion defects on dual-energy CT maps was found to be associated with ACS with potential for improved specificity and reclassification.
Asunto(s)
Síndrome Torácico Agudo , Anemia de Células Falciformes , Yodo , Embolia Pulmonar , Humanos , Síndrome Torácico Agudo/diagnóstico por imagen , Estudios Retrospectivos , Angiografía/métodos , Reproducibilidad de los Resultados , Tomografía Computarizada por Rayos X/métodos , Pulmón , Embolia Pulmonar/diagnóstico por imagen , Anemia de Células Falciformes/complicaciones , Anemia de Células Falciformes/diagnóstico por imagen , PerfusiónRESUMEN
Purpose: Left atrial calcification (LAC), a primarily radiologic diagnosis, has been associated with rheumatic heart disease (RHD) and rheumatic fever (RF). However, left atrial calcification continues to be observed despite a significant decrease in the prevalence of rheumatic heart disease. The purpose of this study was to investigate other possible etiologies of left atrial calcification. Material and methods: This retrospective, observational single-center study included patients from 2017 to 2022 identified as having left atrial calcification as well as age- and sex-matched controls. The prevalence of rheumatic heart disease, atrial ablation, and mitral valve disease was compared, and odds ratios were calculated for each independent variable. Results: Sixty-two patients with left atrial calcifications were included and compared with 62 controls. 87.1% of patients in the left atrial calcifications cohort had a history of atrial fibrillation compared with 21% in the control cohort (p < 0.001). 16.1% of patients in the calcifications cohort presented a history of rheumatic fever compared with zero in the control cohort (p = 0.004). 66.1% of the left atrial calcifications cohort had a history of atrial ablation compared with 6.5% of the control group (p < 0.001). The odds ratio for left atrial calcification was 19.0 vs. 4.8 for rheumatic fever (comparative odds = 4.0 for atrial ablation vs. rheumatic fever). Multivariable log model found atrial ablation to explain 79.8% of left atrial calcifications identified. Conclusions: Our study found a 4-fold higher association between history of atrial ablation and left atrial calcification compared with rheumatic heart disease, suggesting a potential shift in etiology.
RESUMEN
PURPOSE: Photon-counting-detector computed tomography (PCD-CT) offers enhanced noise reduction, spatial resolution, and image quality in comparison to energy-integrated-detectors CT (EID-CT). These hypothesized improvements were compared using PCD-CT ultra-high (UHR) and standard-resolution (SR) scan-modes. METHODS: Phantom scans were obtained with both EID-CT and PCD-CT (UHR, SR) on an adult body-phantom. Radiation dose was measured and noise levels were compared at a minimum achievable slice thickness of 0.5 mm for EID-CT, 0.2 mm for PCD-CT-UHR and 0.4 mm for PCD-CT-SR. Signal-to-noise ratios (SNR) and contrast-to-noise ratios (CNR) were calculated for five tissue densities. Additionally, data from 25 patients who had PCD-CT of chest were reconstructed at 1 mm and 0.2 mm (UHR) slice-thickness and compared quantitatively (SNR) and qualitatively (noise, quality, sharpness, bone details). RESULTS: Phantom PCD-CT-UHR and PCD-CT-SR scans had similar measured radiation dose (16.0mGy vs 15.8 mGy). Phantom PCD-CT-SR (0.4 mm) had lower noise level in comparison to EID-CT (0.5 mm) (9.0HU vs 9.6HU). PCD-CT-UHR (0.2 mm) had slightly higher noise level (11.1HU). Phantom PCD-CT-SR (0.4 mm) had higher SNR in comparison to EID-CT (0.5 mm) while achieving higher resolution (Bone 115 vs 96, Acrylic 14 vs 14, Polyethylene 11 vs 10). SNR was slightly lower across all densities for PCD-CT UHR (0.2 mm). Interestingly, CNR was highest in the 0.2 mm PCD-CT group; PCD-CT CNR was 2.45 and 2.88 times the CNR for 0.5 mm EID-CT for acrylic and poly densities. Clinical comparison of SNR showed predictably higher SNR for 1 mm (30.3 ± 10.7 vs 14.2 ± 7, p = 0.02). Median subjective ratings were higher for 0.2 mm UHR vs 1 mm PCD-CT for nodule contour (4.6 ± 0.3 vs 3.6 ± 0.1, p = 0.02), bone detail (5 ± 0 vs 4 ± 0.1, p = 0.001), image quality (5 ± 0.1 vs 4.6 ± 0.4, p = 0.001), and sharpness (5 ± 0.1 vs 4 ± 0.2). CONCLUSION: Both UHR and SR PCD-CT result in similar radiation dose levels. PCD-CT can achieve higher resolution with lower noise level in comparison to EID-CT.
Asunto(s)
Fotones , Tomografía Computarizada por Rayos X , Adulto , Humanos , Tomografía Computarizada por Rayos X/métodos , Pulmón , Dosis de Radiación , Relación Señal-Ruido , Fantasmas de ImagenRESUMEN
BACKGROUND: Detecting occlusions of coronary artery bypass grafts using non-contrast computed tomography (CT) series is understudied and underestimated. PURPOSE: To evaluate morphological findings for the diagnosis of chronic coronary artery bypass graft occlusion on non-contrast CT and investigate performance statistics for potential use cases. MATERIAL AND METHODS: Seventy-three patients with coronary artery bypass grafts who had CT angiography of the chest (non-contrast and arterial phases) were retrospectively included. Two readers applied pre-set morphologic findings to assess the patency of a bypass graft on non-contrast series. These findings included vessel shape (linear-band like), collapsed lumen and surgical graft marker without a visible vessel. Performance was tested using the simultaneously acquired arterial phase series as the ground truth. RESULTS: The per-patient diagnostic accuracy for occlusion was 0.890 (95% confidence interval = 0.795-0.951). Venous grafts overall had an 88% accuracy. None of the left internal mammary artery to left anterior descending artery arterial graft occlusions were detected. The negative likelihood ratio for an occluded graft that is truly patent was 0.121, demonstrating a true post-test probability of 97% for identifying a patent graft as truly patent given a prevalence of 20% occlusion at a median 8.4 years post-surgery. Neither years post-surgery, nor number of vessels was associated with a significant decrease in reader accuracy. CONCLUSION: Evaluation of coronary bypass grafts for chronic occlusion on non-contrast CT based off vessel morphology is feasible and accurate for venous grafts. Potential use cases include low-intermediate risk patients with chest pain or shortness of breath for whom non-contrast CT was ordered, or administration of iodine-based contrast is contraindicated.
Asunto(s)
Puente de Arteria Coronaria , Tomografía Computarizada por Rayos X , Humanos , Estudios Retrospectivos , Angiografía Coronaria/métodos , Grado de Desobstrucción Vascular , Sensibilidad y Especificidad , Puente de Arteria Coronaria/efectos adversos , Puente de Arteria Coronaria/métodos , Tomografía Computarizada por Rayos X/métodos , Oclusión de Injerto Vascular/diagnóstico por imagenRESUMEN
RATIONALE: Single-photon-emission-computerized-tomography/computed-tomography(SPECT/CT) is commonly used for pulmonary disease. Scant work has been done to determine ability of AI for secondary findings using low-dose-CT(LDCT) attenuation correction series of SPECT/CT. METHODS: 120 patients with ventilation-perfusion-SPECT/CT from 9/1/21-5/1/22 were included in this retrospective study. AI-RAD companion(VA10A,Siemens-Healthineers, Erlangen, Germany), an ensemble of deep-convolutional-neural-networks was evaluated for the detection of pulmonary nodules, coronary artery calcium, aortic ectasia/aneurysm, and vertebral height loss. Accuracy, sensitivity, specificity was measured for the outcomes. Inter-rater reliability were measured. Inter-rater reliability was measured using the intraclass correlation coefficient (ICC) by comparing the number of nodules identified by the AI to radiologist. RESULTS: Overall per-nodule accuracy, sensitivity, and specificity for detection of lung nodules were 0.678(95%CI 0.615-0.732), 0.956(95%CI 0.900-0.985), and 0.456(95%CI 0.376-0.543), respectively, with an intraclass correlation coefficient (ICC) between AI and radiologist of 0.78(95%CI 0.71-0.83). Overall per-patient accuracy for AI detection of coronary artery calcium, aortic ectasia/aneurysm, and vertebral height loss was 0.939(95%CI 0.878-0.975), 0.974(95%CI 0.925-0.995), and 0.857(95%CI 0.781-0.915), respectively. Sensitivity for coronary artery calcium, aortic ectasia/aneurysm, and vertebral height loss was 0.898(95%CI 0.778-0.966), 1 (95%CI 0.958-1), and 1 (95%CI 0.961-1), respectively. Specificity for coronary artery calcium, aortic ectasia/aneurysm, and vertebral height loss was 0.969(95% CI 0.893-0.996), 0.897 (95% CI 0.726-0.978), and 0.346 (95% CI 0.172-0.557), respectively. CONCLUSION: AI ensemble was accurate for coronary artery calcium and aortic ectasia/aneurysm, while sensitive for aortic ectasia/aneurysm, lung nodules and vertebral height loss on LDCT attenuation correction series of SPECT/CT.
Asunto(s)
Inteligencia Artificial , Calcio , Humanos , Estudios Retrospectivos , Dilatación Patológica , Reproducibilidad de los Resultados , Tomografía Computarizada de Emisión de Fotón Único , Tomografía Computarizada por Rayos X , Pulmón , PerfusiónRESUMEN
Noninvasive identification of active myocardial inflammation in patients with cardiac sarcoidosis plays a key role in management but remains elusive. T2 mapping is a proposed solution, but the added value of quantitative myocardial T2 mapping for active cardiac sarcoidosis is unknown. Retrospective cohort analysis of 56 sequential patients with biopsy-confirmed extracardiac sarcoidosis who underwent cardiac MRI for myocardial T2 mapping. The presence or absence of active myocardial inflammation in patients with CS was defined using a modified Japanese circulation society criteria within one month of MRI. Myocardial T2 values were obtained for the 16 standard American Heart Association left ventricular segments. The best model was selected using logistic regression. Receiver operating characteristic curves and dominance analysis were used to evaluate the diagnostic performance and variable importance. Of the 56 sarcoidosis patients included, 14 met criteria for active myocardial inflammation. Mean basal T2 value was the best performing model for the diagnosis of active myocardial inflammation in CS patients (pR2 = 0.493, AUC = 0.918, 95% CI 0.835-1). Mean basal T2 value > 50.8 ms was the most accurate threshold (accuracy = 0.911). Mean basal T2 value + JCS criteria was significantly more accurate than JCS criteria alone (AUC = 0.981 vs. 0.887, p = 0.017). Quantitative regional T2 values are independent predictors of active myocardial inflammation in CS and may add additional discriminatory capability to JCS criteria for active disease.
Asunto(s)
Cardiomiopatías , Miocarditis , Sarcoidosis , Humanos , Estudios Retrospectivos , Pueblos del Este de Asia , Valor Predictivo de las Pruebas , Imagen por Resonancia Magnética , InflamaciónRESUMEN
OBJECTIVES: The aim of this study was to evaluate the impact of virtual monoenergetic imaging (VMI) and quantum iterative reconstruction (QIR) on the accuracy of coronary artery calcium scoring (CACS) using a virtual noniodine (VNI) reconstruction algorithm on a first-generation, clinical, photon counting detector computed tomography system. MATERIALS AND METHODS: Coronary artery calcium scoring was evaluated in an anthropomorphic chest phantom simulating 3 different patient sizes by using 2 extension rings (small: 300 × 200 mm, medium: 350 × 250 mm, large: 400 × 300 mm) and in patients (n = 61; final analyses only in patients with coronary calcifications [n = 34; 65.4 ± 10.0 years; 73.5% male]), who underwent nonenhanced and contrast-enhanced, electrocardiogram-gated, cardiac computed tomography on a photon counting detector system. Phantom and patient data were reconstructed using a VNI reconstruction algorithm at different VMI (55-80 keV) and QIR (strength 1-4) levels (CACS VNI ). True noncontrast (TNC) scans at 70 keV and QIR "off" were used as reference for phantom and patient studies (CACS TNC ). RESULTS: In vitro and in vivo CACS VNI showed strong correlation ( r > 0.9, P < 0.001 for all) and excellent agreement (intraclass correlation coefficient > 0.9 for all) with CACS TNC at all investigated VMI and QIR levels. Phantom and patient CACS VNI significantly increased with decreasing keV levels (in vitro: from 475.2 ± 26.3 at 80 keV up to 652.5 ± 42.2 at 55 keV; in vivo: from 142.5 [7.4/737.7] at 80 keV up to 248.1 [31.2/1144] at 55 keV; P < 0.001 for all), resulting in an overestimation of CACS VNI at 55 keV compared with CACS TNC at 70 keV in some cases (in vitro: 625.8 ± 24.4; in vivo: 225.4 [35.1/959.7]). In vitro CACS increased with rising QIR at low keV. In vivo scores were significantly higher at QIR 1 compared with QIR 4 only at 60 and 80 keV (60 keV: 220.3 [29.6-1060] vs 219.5 [23.7/1048]; 80 keV: 152.0 [12.0/735.6] vs 142.5 [7.4/737.7]; P < 0.001). CACS VNI was closest to CACS TNC at 60 keV, QIR 2 (+0.1%) in the small; 55 keV, QIR 1 (±0%) in the medium; 55 keV, QIR 4 (-0.1%) in the large phantom; and at 60 keV, QIR 1 (-2.3%) in patients. CONCLUSIONS: Virtual monoenergetic imaging reconstructions have a significant impact on CACS VNI . The effects of different QIR levels are less consistent and seem to depend on several individual conditions, which should be further investigated.
Asunto(s)
Calcio , Vasos Coronarios , Humanos , Masculino , Femenino , Relación Señal-Ruido , Vasos Coronarios/diagnóstico por imagen , Tomografía Computarizada por Rayos X/métodos , AlgoritmosRESUMEN
BACKGROUND: The coronavirus disease 2019 (COVID-19) pandemic has disrupted the world since 2019, causing significant morbidity and mortality in developed and developing countries alike. Although substantial resources have been diverted to developing diagnostic, preventative, and treatment measures, disparities in the availability and efficacy of these tools vary across countries. We seek to assess the ability of commercial artificial intelligence (AI) technology to diagnose COVID-19 by analyzing chest radiographs. MATERIALS AND METHODS: Chest radiographs taken from symptomatic patients within two days of polymerase chain reaction (PCR) tests were assessed for COVID-19 infection by board-certified radiologists and commercially available AI software. Sixty patients with negative and 60 with positive COVID reverse transcription-polymerase chain reaction (RT-PCR) tests were chosen. Results were compared against results of the PCR test for accuracy and statistically analyzed by receiver operating characteristic (ROC) curves along with area under the curve (AUC) values. RESULTS: A total of 120 chest radiographs (60 positive and 60 negative RT-PCR tests) radiographs were analyzed. The AI software performed significantly better than chance (p = 0.001) and did not differ significantly from the radiologist ROC curve (p = 0.78). CONCLUSION: Commercially available AI software was not inferior compared with trained radiologists in accurately identifying COVID-19 cases by analyzing radiographs. While RT-PCR testing remains the standard, current advances in AI help correctly analyze chest radiographs to diagnose COVID-19 infection.
RESUMEN
Vascular spasm is well known and studied in the arterial system. There are only a few cases reported related to central venous spasms. We present the case of a 63-year-old male with an extensive medical history, including deep vein thrombosis (DVT), who underwent peripheral insertion of a central catheter in his left upper extremity with subsequent development of left upper extremity edema. The central catheter was removed before the patient underwent a contrast-enhanced computed tomography of the chest which revealed severe narrowing of the left brachiocephalic vein, consistent with venospasm in the clinical setting. Nitroglycerin might be useful to prevent vasospasm, or it might also be used for treatment. In our case, the catheter was removed, and no subsequent treatment was necessary.
RESUMEN
BACKGROUND: Airspace disease as seen on chest X-rays is an important point in triage for patients initially presenting to the emergency department with suspected COVID-19 infection. The purpose of this study is to evaluate a previously trained interpretable deep learning algorithm for the diagnosis and prognosis of COVID-19 pneumonia from chest X-rays obtained in the ED. METHODS: This retrospective study included 2456 (50% RT-PCR positive for COVID-19) adult patients who received both a chest X-ray and SARS-CoV-2 RT-PCR test from January 2020 to March of 2021 in the emergency department at a single U.S. INSTITUTION: A total of 2000 patients were included as an additional training cohort and 456 patients in the randomized internal holdout testing cohort for a previously trained Siemens AI-Radiology Companion deep learning convolutional neural network algorithm. Three cardiothoracic fellowship-trained radiologists systematically evaluated each chest X-ray and generated an airspace disease area-based severity score which was compared against the same score produced by artificial intelligence. The interobserver agreement, diagnostic accuracy, and predictive capability for inpatient outcomes were assessed. Principal statistical tests used in this study include both univariate and multivariate logistic regression. RESULTS: Overall ICC was 0.820 (95% CI 0.790-0.840). The diagnostic AUC for SARS-CoV-2 RT-PCR positivity was 0.890 (95% CI 0.861-0.920) for the neural network and 0.936 (95% CI 0.918-0.960) for radiologists. Airspace opacities score by AI alone predicted ICU admission (AUC = 0.870) and mortality (0.829) in all patients. Addition of age and BMI into a multivariate log model improved mortality prediction (AUC = 0.906). CONCLUSION: The deep learning algorithm provides an accurate and interpretable assessment of the disease burden in COVID-19 pneumonia on chest radiographs. The reported severity scores correlate with expert assessment and accurately predicts important clinical outcomes. The algorithm contributes additional prognostic information not currently incorporated into patient management.
Asunto(s)
COVID-19 , Aprendizaje Profundo , Adulto , Inteligencia Artificial , COVID-19/diagnóstico por imagen , Humanos , Pronóstico , Radiografía Torácica , Estudios Retrospectivos , SARS-CoV-2 , Tomografía Computarizada por Rayos X , Rayos XRESUMEN
BACKGROUND. Deep learning-based convolutional neural networks have enabled major advances in development of artificial intelligence (AI) software applications. Modern AI applications offer comprehensive multiorgan evaluation. OBJECTIVE. The purpose of this article was to evaluate the impact of an automated AI platform integrated into clinical workflow for chest CT interpretation on radiologists' interpretation times when evaluated in a real-world clinical setting. METHODS. In this prospective single-center study, a commercial AI software solution was integrated into clinical workflow for chest CT interpretation. The software provided automated analysis of cardiac, pulmonary, and musculoskeletal findings, including labeling, segmenting, and measuring normal structures as well as detecting, labeling, and measuring abnormalities. AI-annotated images and autogenerated summary results were stored in the PACS and available to interpreting radiologists. A total of 390 patients (204 women, 186 men; mean age, 62.8 ± 13.3 [SD] years) who underwent out-patient chest CT between January 19, 2021, and January 28, 2021, were included. Scans were randomized using 1:1 allocation between AI-assisted and non-AI-assisted arms and were clinically interpreted by one of three cardiothoracic radiologists (65 scans per arm per radiologist; total of 195 scans per arm) who recorded interpretation times using a stopwatch. Findings were categorized according to review of report impressions. Interpretation times were compared between arms. RESULTS. Mean interpretation times were significantly shorter in the AI-assisted than in the non-AI-assisted arm for all three readers (289 ± 89 vs 344 ± 129 seconds, p < .001; 449 ± 110 vs 649 ± 82 seconds, p < .001; 281 ± 114 vs 348 ± 93 seconds, p = .01) and for readers combined (328 ± 122 vs 421 ± 175 seconds, p < .001). For readers combined, the mean difference was 93 seconds (95% CI, 63-123 seconds), corresponding with a 22.1% reduction in the AI-assisted arm. Mean interpretation time was also shorter in the AI-assisted arm compared with the non-AI-assisted arm for contrast-enhanced scans (83 seconds), noncontrast scans (104 seconds), negative scans (84 seconds), positive scans without significant new findings (117 seconds), and positive scans with significant new findings (92 seconds). CONCLUSION. Cardiothoracic radiologists exhibited a 22.1% reduction in chest CT interpretations times when they had access to results from an automated AI support platform during real-world clinical practice. CLINICAL IMPACT. Integration of the AI support platform into clinical workflow improved radiologist efficiency.
Asunto(s)
Inteligencia Artificial , Tomografía Computarizada por Rayos X , Masculino , Humanos , Femenino , Persona de Mediana Edad , Anciano , Estudios Prospectivos , Tomografía Computarizada por Rayos X/métodos , Radiólogos , Redes Neurales de la Computación , Estudios RetrospectivosRESUMEN
RATIONALE AND OBJECTIVES: The burden of coronavirus disease 2019 (COVID-19) airspace opacities is time consuming and challenging to quantify on computed tomography. The purpose of this study was to evaluate the ability of a deep convolutional neural network (dCNN) to predict inpatient outcomes associated with COVID-19 pneumonia. MATERIALS AND METHODS: A previously trained dCNN was tested on an external validation cohort of 241 patients who presented to the emergency department and received a chest computed tomography scan, 93 with COVID-19 and 168 without. Airspace opacity scoring systems were defined by the extent of airspace opacity in each lobe, totaled across the entire lungs. Expert and dCNN scores were concurrently evaluated for interobserver agreement, while both dCNN identified airspace opacity scoring and raw opacity values were used in the prediction of COVID-19 diagnosis and inpatient outcomes. RESULTS: Interobserver agreement for airspace opacity scoring was 0.892 (95% CI 0.834-0.930). Probability of each outcome behaved as a logistic function of the opacity scoring (25% intensive care unit admission at score of 13/25, 25% intubation at 17/25, and 25% mortality at 20/25). Length of hospitalization, intensive care unit stay, and intubation were associated with larger airspace opacity score (p = 0.032, 0.039, 0.036, respectively). CONCLUSION: The tested dCNN was highly predictive of inpatient outcomes, performs at a near expert level, and provides added value for clinicians in terms of prognostication and disease severity.
Asunto(s)
COVID-19 , Aprendizaje Profundo , Algoritmos , COVID-19/diagnóstico por imagen , Prueba de COVID-19 , Humanos , Pacientes Internos , Pulmón/diagnóstico por imagen , Morbilidad , Estudios Retrospectivos , SARS-CoV-2 , Tomografía Computarizada por Rayos X/métodosRESUMEN
BACKGROUND: Determination of the total number and size of all pulmonary metastases on chest CT is time-consuming and as such has been understudied as an independent metric for disease assessment. A novel artificial intelligence (AI) model may allow for automated detection, size determination, and quantification of the number of pulmonary metastases on chest CT. OBJECTIVE: To investigate the utility of a novel AI program applied to initial staging chest CT in breast cancer patients in risk assessment of mortality and survival. METHODS: Retrospective imaging data from a cohort of 226 subjects with breast cancer was assessed by the novel AI program and the results validated by blinded readers. Mean clinical follow-up was 2.5 years for outcomes including cancer-related death and development of extrapulmonary metastatic disease. AI measurements including total number of pulmonary metastases and maximum nodule size were assessed by Cox-proportional hazard modeling and adjusted survival. RESULTS: 752 lung nodules were identified by the AI program, 689 of which were identified in 168 subjects having confirmed lung metastases (Lmet+) and 63 were identified in 58 subjects without confirmed lung metastases (Lmet-). When compared to the reader assessment, AI had a per-patient sensitivity, specificity, PPV and NPV of 0.952, 0.639, 0.878, and 0.830. Mortality in the Lmet + group was four times greater compared to the Lmet-group (p = 0.002). In a multivariate analysis, total lung nodule count by AI had a high correlation with overall mortality (OR 1.11 (range 1.07-1.15), p < 0.001) with an AUC of 0.811 (R2 = 0.226, p < 0.0001). When total lung nodule count and maximum nodule diameter were combined there was an AUC of 0.826 (R2 = 0.243, p < 0.001). CONCLUSION: Automated AI-based detection of lung metastases in breast cancer patients at initial staging chest CT performed well at identifying pulmonary metastases and demonstrated strong correlation between the total number and maximum size of lung metastases with future mortality. CLINICAL IMPACT: As a component of precision medicine, AI-based measurements at the time of initial staging may improve prediction of which breast cancer patients will have negative future outcomes.