Reduction in Radiologist Interpretation Time of Serial CT and MR Imaging Findings with Deep Learning Identification of Relevant Priors, Series and Finding Locations
RATIONALE AND OBJECTIVES/OBJECTIVE:Finding comparison to relevant prior studies is a requisite component of the radiology workflow. The purpose of this study was to evaluate the impact of a deep learning tool simplifying this time-consuming task by automatically identifying and displaying the finding in relevant prior studies. MATERIALS AND METHODS/METHODS:The algorithm pipeline used in this retrospective study, TimeLens (TL), is based on natural language processing and descriptor-based image-matching algorithms. The dataset used for testing comprised 3872 series of 246 radiology examinations from 75 patients (189 CTs, 95 MRIs). To ensure a comprehensive testing, five finding types frequently encountered in radiology practice were included: aortic aneurysm, intracranial aneurysm, kidney lesion, meningioma, and pulmonary nodule. After a standardized training session, nine radiologists from three university hospitals performed two reading sessions on a cloud-based evaluation platform resembling a standard RIS/PACS. The task was to measure the diameter of the finding-of-interest on two or more exams (a most recent and at least one prior exam): first without use of TL, and a second session at an interval of at least 21 days with the use of TL. All user actions were logged for each round, including time needed to measure the finding at all timepoints, number of mouse clicks, and mouse distance traveled. The effect of TL was evaluated in total, per finding type, per reader, per experience (resident vs. board-certified radiologist), and per modality. Mouse movement patterns were analyzed with heatmaps. To assess the effect of habituation to the cases, a third round of readings was performed without TL. RESULTS:Across scenarios, TL reduced the average time needed to assess a finding at all timepoints by 40.1% (107 vs. 65 seconds; p < 0.001). Largest accelerations were demonstrated for assessment of pulmonary nodules (-47.0%; p < 0.001). Less mouse clicks (-17.2%) were needed for finding evaluation with TL, and mouse distance traveled was reduced by 38.0%. Time needed to assess the findings increased from round 2 to round 3 (+27.6%; p < 0.001). Readers were able to measure a given finding in 94.4% of cases on the series initially proposed by TL as most relevant series for comparison. The heatmaps showed consistently simplified mouse movement patterns with TL. CONCLUSION/CONCLUSIONS:A deep learning tool significantly reduced both the amount of user interactions with the radiology image viewer and the time needed to assess findings of interest on cross-sectional imaging with relevant prior exams.
Evaluation of Socioeconomic Disparities in Follow-up Completion for Incidental Pulmonary Nodules
OBJECTIVE:To evaluate the association between census-tract level measures of social vulnerability and residential segregation and IPN follow up. METHODS:This retrospective cohort study included patients with IPN ≥6 mm in size or multiple subsolid/ground-glass IPNs <6 mm (with non-optional follow-up recommendations) diagnosed between January 1, 2018 and December 30, 2019 at a large urban tertiary center and followed ≥two years. Geographic sociodemographic context was characterized by 2018 U.S. Centers for Disease Control and Prevention Social Vulnerability Index (SVI) and the Index of Concentration at the Extreme (ICE), categorized in quartiles. Multivariable binomial regression models were utilized with a primary outcome of inappropriate IPN follow up (late or no follow up). Models were also stratified by nodule risk. RESULTS:The study consisted of 2,492 patients (mean age 65.6 years +/- 12.6 years; 1,361 women). Top-quartile SVI patients were more likely to have inappropriate follow up (Risk Ratio [RR]: 1.24, 95% Confidence Interval [95% CI], 1.12-1.36]), compared with the bottom quartile; risk was also elevated in top-quartile SVI subcategories of Socioeconomic Status (RR: 1.23, 95% CI, 1.13-1.34), Minority Status and Language (RR: 1.24, 95% CI, 1.03-1.48), Housing and Transportation (RR: 1.13, 95% CI, 1.02-1.26), and ICE (RR: 1.20, 95% CI, 1.11-1.30). Further, top-quartile ICE was associated with greater risk of inappropriate follow up among high-risk vs. lower-risk IPN (1.33 [1.18-1.50] vs. 1.13 [1.02-1.25]), respectively, P for interaction= 0.017). DISCUSSION/CONCLUSIONS:Local social vulnerability and residential segregation are associated with inappropriate IPN follow up and may inform policy or interventions tailored for neighborhoods.
ACR Appropriateness Criteria® Routine Chest Imaging
Routine chest imaging has been used to identify unknown or subclinical cardiothoracic abnormalities in the absence of symptoms. Various imaging modalities have been suggested for routine chest imaging. We review the evidence for or against the use of routine chest imaging in different clinical scenarios. This document aims to determine guidelines for the use of routine chest imaging as initial imaging for hospital admission, initial imaging prior to noncardiothoracic surgery, and surveillance imaging for chronic cardiopulmonary disease. The American College of Radiology Appropriateness Criteria are evidence-based guidelines for specific clinical conditions that are reviewed annually by a multidisciplinary expert panel. The guideline development and revision process support the systematic analysis of the medical literature from peer reviewed journals. Established methodology principles such as Grading of Recommendations Assessment, Development, and Evaluation or GRADE are adapted to evaluate the evidence. The RAND/UCLA Appropriateness Method User Manual provides the methodology to determine the appropriateness of imaging and treatment procedures for specific clinical scenarios. In those instances where peer reviewed literature is lacking or equivocal, experts may be the primary evidentiary source available to formulate a recommendation.
Deep Learning Denoising of Low-Dose Computed Tomography Chest Images: A Quantitative and Qualitative Image Analysis
PURPOSE/OBJECTIVE:To assess deep learning denoised (DLD) computed tomography (CT) chest images at various low doses by both quantitative and qualitative perceptual image analysis. METHODS:Simulated noise was inserted into sinogram data from 32 chest CTs acquired at 100 mAs, generating anatomically registered images at 40, 20, 10, and 5 mAs. A DLD model was developed, with 23 scans selected for training, 5 for validation, and 4 for test.Quantitative analysis of perceptual image quality was assessed with Structural SIMilarity Index (SSIM) and Fréchet Inception Distance (FID). Four thoracic radiologists graded overall diagnostic image quality, image artifact, visibility of small structures, and lesion conspicuity. Noise-simulated and denoised image series were evaluated in comparison with one another, and in comparison with standard 100 mAs acquisition at the 4 mAs levels. Statistical tests were conducted at the 2-sided 5% significance level, with multiple comparison correction. RESULTS:At the same mAs levels, SSIM and FID between noise-simulated and reconstructed DLD images indicated that images were closer to a perfect match with increasing mAs (closer to 1 for SSIM, and 0 for FID).In comparing noise-simulated and DLD images to standard-dose 100-mAs images, DLD improved SSIM and FID. Deep learning denoising improved SSIM of 40-, 20-, 10-, and 5-mAs simulations in comparison with standard-dose 100-mAs images, with change in SSIM from 0.91 to 0.94, 0.87 to 0.93, 0.67 to 0.87, and 0.54 to 0.84, respectively. Deep learning denoising improved FID of 40-, 20-, 10-, and 5-mAs simulations in comparison with standard-dose 100-mAs images, with change in FID from 20 to 13, 46 to 21, 104 to 41, and 148 to 69, respectively.Qualitative image analysis showed no significant difference in lesion conspicuity between DLD images at any mAs in comparison with 100-mAs images. Deep learning denoising images at 10 and 5 mAs were rated lower for overall diagnostic image quality (P < 0.001), and at 5 mAs lower for overall image artifact and visibility of small structures (P = 0.002), in comparison with 100 mAs. CONCLUSIONS:Deep learning denoising resulted in quantitative improvements in image quality. Qualitative assessment demonstrated DLD images at or less than 10 mAs to be rated inferior to standard-dose images.
Prevalence of Adenopathy at Chest Computed Tomography After Vaccination for Severe Acute Respiratory Syndrome Coronavirus 2
OBJECTIVE:This study aimed to determine the prevalence of axillary and subpectoral (SP) lymph nodes after ipsilateral COVID-19 vaccine administration on chest computed tomography (CT). METHODS:Subjects with chest CTs between 2 and 25 days after a first or second vaccine dose, December 15, 2020, to February 12, 2021, were included. Orthogonal measures of the largest axillary and SP nodes were recorded by 2 readers blinded to vaccine administration and clinical details. A mean nodal diameter discrepancy of ≥6 mm between contralateral stations was considered positive for asymmetry. Correlation with the side of vaccination, using a Spearman rank correlation, was performed on the full cohort and after excluding patients with diseases associated with adenopathy. RESULTS:Of the 138 subjects (81 women, 57 men; mean [SD] age, 74.4 ± 11.7 years), 48 (35%) had asymmetrically enlarged axillary and/or SP lymph nodes, 42 (30%) had ipsilateral, and 6 (4%) had contralateral to vaccination ( P = 0.003). Exclusion of 29 subjects with conditions associated with adenopathy showed almost identical correlation, with asymmetric nodes in 32 of 109 (29%) ipsilateral and in 5 of 109 (5%) contralateral to vaccination ( P = 0.002). CONCLUSIONS:Axillary and/or SP lymph nodes ipsilateral to vaccine administration represents a clinical conundrum. Asymmetric nodes were detected at CT in 30% of subjects overall and 29% of subjects without conditions associated with adenopathy, approximately double the prevalence rate reported to the Centers for Disease Control and Prevention by vaccine manufacturers. When interpreting examinations correlation with vaccine administration timing and site is important for pragmatic management.
OXYGENATOR IMPACT ON REMDESIVIR IN EXTRACORPOREAL MEMBRANE OXYGENATION CIRCUITS [Meeting Abstract]
INTRODUCTION: Extra-corporeal membrane oxygenation (ECMO) is a treatment modality known to alter drug pharmacokinetics (PK). The PK changes can result from drug binding to the oxygenator, alterations in clearance, and drug adsorption or sequestration. Levels of drug absorption by polymers, silicone rubber and other materials have been linked to the drugs' lipophilicity and the published literature is mostly outdated. Additionally, there is limited data regarding the impact of the oxygenator on drug changes in ECMO circuits in comparison to the other components of the ECMO circuit. The purpose of this study was to determine the impact of the Quadrox-i pediatric and adult oxygenators on the PK of remdesivir (RDV) in contemporary ECMO circuits.
METHOD(S): One 1/4-in. and one 3/8-in. closed loop ECMO circuits were prepared using custom tubing with polyvinylchloride and superTygon (Medtronic Inc., Minneapolis, MN) and a Quadrox-i adult or pediatric oxygenator (Maquet). Additionally, one 1/4-in. and one 3/8- in. closed loop ECMO circuits were assembled without an oxygenator in series. RDV was added to the circuit and levels were obtained pre-and post-oxygenator at the following time intervals; 5 mins, 1, 2, 3, 4, 5, 6, 8, 12, and 24 hrs. RDV was also maintained in a glass vial and samples obtained at the same time periods for control purposes. RDV samples were analyzed by liquid chromatography tandem mass spectrometry.
RESULT(S): For the 3/8-in. circuit with and without an oxygenator, there was a 60-70% RDV loss during the study period. For the 1/4-in. circuits with an oxygenator, there was a 35-60% RDV loss during the study period. For the 1/4-in. circuits without an oxygenator, there was a 5-20% RDV loss during the study period.
CONCLUSION(S): There was RDV loss within the circuit during the study period and the RDV loss was more pronounced with the larger 3/8-in circuit when compared with the 1/4-in. circuit. This preliminary data suggests RDV dosing may need to be adjusted for concern of drug loss via the ECMO circuit. Additional single and multiple dose studies are needed to validate these findings
Reduced CT iodine perfusion score is associated with adverse clinical outcomes in acute pulmonary embolism [Letter]
Computed Tomography Evaluation of In Vivo Pulmonary Cryoablation Zone Sizes
PURPOSE/OBJECTIVE:To evaluate ablation zone sizes in patients undergoing pulmonary tumor cryoablation with 14-gauge cryoablation probes. MATERIALS AND METHODS/METHODS:Single-center retrospective analysis of all consecutive patients who underwent cryoablation of pulmonary tumors with one or more 14-gauge probes (August 2017-June 2020) was performed. Intraprocedural and 1-2 month post-procedural chest CTs were evaluated to characterize pulmonary lesions, ice balls, and ablation zones. Single-probe 14-gauge ablation zone volumes were compared to manufacturer reference isotherms, and to single- and two-probe ablation zones from a prior investigation of 17-gauge probes. Overall and local recurrence-free survival were calculated to 3 years. RESULTS:) (all p<0.001). Three-year overall and local recurrence-free survival were 69% (95% CI: 53-89%) and 87% (95% CI: 74-100%), respectively. CONCLUSIONS:14-gauge probes generate larger ablation volumes than 17-gauge probes. Manufacturer-reported isotherms are significantly larger than actual cryoablation zones. Cryoablation can attain low rates of local recurrence.
Low-field 0.55Â T MRI for assessment of pulmonary groundglass and fibrosis-like opacities: Inter-reader and inter-modality concordance
PURPOSE/OBJECTIVE:To evaluate detection and characterization of groundglass and fibrosis-like opacities imaged by non-contrast 0.55 Tesla MRI, and versus clinically-acquired chest CT images, in a cohort of post-Covid patients. MATERIALS AND METHODS/METHODS:64 individuals (26 women, mean age 53Â Â±Â 14Â years, range 19-85) with history of Covid-19 pneumonia were recruited through a survivorship registry, with 106 non-contrast low-field 0.55Â T cardiopulmonary MRI exams acquired from 9/8/2020-9/28/2021. MRI exams were obtained at an average interval of 9.5Â Â±Â 4.5Â months from initial symptom report (range 1-18Â months). Of these, 20 participants with 22 MRI exams had corresponding clinically-acquired CT chest imaging obtained within 30Â days of MRI (average interval 18Â Â±Â 9Â days, range 0-30). MR and CT images were reviewed and scored by two thoracic radiologists, for presence and extent of lung opacity by quadrant, opacity distribution, and presence versus absence of fibrosis-like subpleural reticulation and subpleural lines. Scoring was performed for each of four lung quadrants: right upper and middle lobe, right lower lobe, left upper lobe and lingula, and left lower lobe. Agreement between readers and modalities was assessed with simple and linear weighted Cohen's kappa (k) coefficients. RESULTS:Inter-reader concordance on CT for opacity presence, opacity extent, opacity distribution, and presence of subpleural lines and reticulation was 99%, 78%, 97%, 99%, and 94% (k 0.96, 0.86, 0.94, 0.97, 0.89), respectively. Inter-reader concordance on MR, among all 106 exams, for opacity presence, opacity extent, opacity distribution, and presence of subpleural lines and reticulation was 85%, 48%, 70%, 86%, and 76% (k 0.57, 0.32, 0.46, 0.47, 0.37), respectively. Inter-modality agreement between CT and MRI for opacity presence, opacity extent, opacity distribution, and presence subpleural lines and reticulation was 86%, 52%, 79%, 93%, and 76% (k 0.43, 0.63, 0.65, 0.80, 0.52). CONCLUSION/CONCLUSIONS:Low-field 0.55Â T non-contrast MRI demonstrates fair to moderate inter-reader concordance, and moderate to substantial inter-modality agreement with CT, for detection and characterization of groundglass and fibrosis-like opacities.
Current imaging of PE and emerging techniques: is there a role for artificial intelligence?
Acute pulmonary embolism (PE) is a critical, potentially life-threatening finding on contrast-enhanced cross-sectional chest imaging. Timely and accurate diagnosis of thrombus acuity and extent directly influences patient management, and outcomes. Technical and interpretive pitfalls may present challenges to the radiologist, and by extension, pose nuance in the development and integration of artificial intelligence support tools. This review delineates imaging considerations for diagnosis of acute PE, and rationale, hurdles and applications of artificial intelligence for the PE task.