Validation of a Deep Learning-Based Model to Predict Lung Cancer Risk Using Chest Radiographs and Electronic Medical Record Data.

Importance: Lung cancer screening with chest computed tomography (CT) prevents lung cancer death; however, fewer than 5% of eligible Americans are screened. CXR-LC, an open-source deep learning tool that estimates lung cancer risk from existing chest radiograph images and commonly available electronic medical record (EMR) data, may enable automated identification of high-risk patients as a step toward improving lung cancer screening participation. Objective: To validate CXR-LC using EMR data to identify individuals at high-risk for lung cancer to complement 2022 US Centers for Medicare & Medicaid Services (CMS) lung cancer screening eligibility guidelines. Design, Setting, and Participants: This prognostic study compared CXR-LC estimates with CMS screening guidelines using patient data from a large US hospital system. Included participants were persons who currently or formerly smoked cigarettes with an outpatient posterior-anterior chest radiograph between January 1, 2013, and December 31, 2014, with no history of lung cancer or screening CT. Data analysis was performed between May 2021 and June 2022. Exposures: CXR-LC lung cancer screening eligibility (previously defined as having a 3.297% or greater 12-year risk) based on inputs (chest radiograph image, age, sex, and whether currently smoking) extracted from the EMR. Main Outcomes and Measures: 6-year incident lung cancer. Results: A total of 14 737 persons were included in the study population (mean [SD] age, 62.6 [6.8] years; 7154 [48.5%] male; 204 [1.4%] Asian, 1051 [7.3%] Black, 432 [2.9%] Hispanic, 12 330 [85.2%] White) with a 2.4% rate of incident lung cancer over 6 years (361 patients with cancer). CMS eligibility could be determined in 6277 patients (42.6%) using smoking pack-year and quit-date from the EMR. Patients eligible by both CXR-LC and 2022 CMS criteria had a high rate of lung cancer (83 of 974 patients [8.5%]), higher than those eligible by 2022 CMS criteria alone (5 of 177 patients [2.8%]; P < .001). Patients eligible by CXR-LC but not 2022 CMS criteria also had a high 6-year incidence of lung cancer (121 of 3703 [3.3%]). In the 8460 cases (57.4%) where CMS eligibility was unknown, CXR-LC eligible patients had a 5-fold higher rate of lung cancer than ineligible (127 of 5177 [2.5%] vs 18 of 2283 [0.5%]; P < .001). Similar results were found in subgroups, including female patients and Black persons. Conclusions and Relevance: Using routine chest radiographs and other data automatically extracted from the EMR, CXR-LC identified high-risk individuals who may benefit from lung cancer screening CT.

Are risk factors necessary for pretest probability assessment of coronary artery disease? A patient similarity network analysis of the PROMISE trial.

BACKGROUND: Pretest probability (PTP) calculators utilize epidemiological-level findings to provide patient-level risk assessment of obstructive coronary artery disease (CAD). However, their limited accuracies question whether dissimilarities in risk factors necessarily result in differences in CAD. Using patient similarity network (PSN) analyses, we wished to assess the accuracy of risk factors and imaging markers to identify ≥50% luminal narrowing on coronary CT angiography (CCTA) in stable chest-pain patients. METHODS: We created four PSNs representing: patient characteristics, risk factors, non-coronary imaging markers and calcium score. We used spectral clustering to group individuals with similar risk profiles. We compared PSNs to a contemporary PTP score incorporating calcium score and risk factors to identify ≥50% luminal narrowing on CCTA in the CT-arm of the PROMISE trial. We also conducted subanalyses in different age and sex groups. RESULTS: In 3556 individuals, the calcium score PSN significantly outperformed patient characteristic, risk factor, and non-coronary imaging marker PSNs (AUC: 0.81 vs. 0.57, 0.55, 0.54; respectively, p â€‹< â€‹0.001 for all). The calcium score PSN significantly outperformed the contemporary PTP score (AUC: 0.81 vs. 0.78, p â€‹< â€‹0.001), and using 0, 1-100 and â€‹> â€‹100 cut-offs provided comparable results (AUC: 0.81 vs. 0.81, p â€‹= â€‹0.06). Similar results were found in all subanalyses. CONCLUSION: Calcium score on its own provides better individualized obstructive CAD prediction than contemporary PTP scores incorporating calcium score and risk factors. Risk factors may not be able to improve the diagnostic accuracy of calcium score to predict ≥50% luminal narrowing on CCTA.

Assessment of Coronary Artery Disease With Computed Tomography Angiography and Inflammatory and Immune Activation Biomarkers Among Adults With HIV Eligible for Primary Cardiovascular Prevention.

Importance: Cardiovascular disease (CVD) is increased among people with HIV (PWH), but little is known regarding the prevalence and extent of coronary artery disease (CAD) and associated biological factors in PWH with low to moderate traditional CVD risk. Objectives: To determine unique factors associated with CVD in PWH and to assess CAD by coronary computed tomography angiography (CTA) and critical pathways of arterial inflammation and immune activation. Design, Setting, and Participants: This cohort study among male and female PWH, aged 40 to 75 years, without known CVD, receiving stable antiretroviral therapy, and with low to moderate atherosclerotic cardiovascular disease (ASCVD) risk according to the 2013 American College of Cardiology/American Heart Association pooled cohort equation, was part of the Randomized Trial to Prevent Vascular Events in HIV (REPRIEVE), a large, ongoing primary prevention trial of statin therapy among PWH conducted at 31 US sites. Participants were enrolled from May 2015 to February 2018. Data analysis was conducted from May to December 2020. Exposure: HIV disease. Main Outcomes and Measures: The primary outcome was the prevalence and composition of CAD assessed by coronary CTA and, secondarily, the association of CAD with traditional risk indices and circulating biomarkers, including insulin, monocyte chemoattractant protein 1 (MCP-1), interleukin (IL) 6, soluble CD14 (sCD14), sCD163, lipoprotein-associated phospholipase A2 (LpPLA2), oxidized low-density lipoprotein (oxLDL), and high-sensitivity C-reactive protein (hsCRP). Results: The sample included 755 participants, with a mean (SD) age of 51 (6) years, 124 (16%) female participants, 267 (35%) Black or African American participants, 182 (24%) Latinx participants, a low median (interquartile range) ASCVD risk (4.5% [2.6%-6.8%]), and well-controlled viremia. Overall, plaque was seen in 368 participants (49%), including among 52 of 175 participants (30%) with atherosclerotic CVD (ASCVD) risk of less than 2.5%. Luminal obstruction of at least 50% was rare (25 [3%]), but vulnerable plaque and high Leaman score (ie, >5) were more frequently observed (172 of 755 [23%] and 118 of 743 [16%], respectively). Overall, 251 of 718 participants (35%) demonstrated coronary artery calcium score scores greater than 0. IL-6, LpPLA2, oxLDL, and MCP-1 levels were higher in those with plaque compared with those without (eg, median [IQR] IL-6 level, 1.71 [1.05-3.04] pg/mL vs 1.45 [0.96-2.60] pg/mL; P = .008). LpPLA2 and IL-6 levels were associated with plaque in adjusted modeling, independent of traditional risk indices and HIV parameters (eg, IL-6: adjusted odds ratio, 1.07; 95% CI, 1.02-1.12; P = .01). Conclusions and Relevance: In this study of a large primary prevention cohort of individuals with well-controlled HIV and low to moderate ASCVD risk, CAD, including noncalcified, nonobstructive, and vulnerable plaque, was highly prevalent. Participants with plaque demonstrated higher levels of immune activation and arterial inflammation, independent of traditional ASCVD risk and HIV parameters.

Validation of Wall Shear Stress Assessment in Non-invasive Coronary CTA versus Invasive Imaging: A Patient-Specific Computational Study.

Endothelial shear stress (ESS) identifies coronary plaques at high risk for progression and/or rupture leading to a future acute coronary syndrome. In this study an optimized methodology was developed to derive ESS, pressure drop and oscillatory shear index using computational fluid dynamics (CFD) in 3D models of coronary arteries derived from non-invasive coronary computed tomography angiography (CTA). These CTA-based ESS calculations were compared to the ESS calculations using the gold standard with fusion of invasive imaging and CTA. In 14 patients paired patient-specific CFD models based on invasive and non-invasive imaging of the left anterior descending (LAD) coronary arteries were created. Ten patients were used to optimize the methodology, and four patients to test this methodology. Time-averaged ESS (TAESS) was calculated for both coronary models applying patient-specific physiological data available at the time of imaging. For data analysis, each 3D reconstructed coronary artery was divided into 2 mm segments and each segment was subdivided into 8 arcs (45°).TAESS and other hemodynamic parameters were averaged per segment as well as per arc. Furthermore, the paired segment- and arc-averaged TAESS were categorized into patient-specific tertiles (low, medium and high). In the ten LADs, used for optimization of the methodology, we found high correlations between invasively-derived and non-invasively-derived TAESS averaged over segments (n = 263, r = 0.86) as well as arcs (n = 2104, r = 0.85, p < 0.001). The correlation was also strong in the four testing-patients with r = 0.95 (n = 117 segments, p = 0.001) and r = 0.93 (n = 936 arcs, p = 0.001).There was an overall high concordance of 78% of the three TAESS categories comparing both methodologies using the segment- and 76% for the arc-averages in the first ten patients. This concordance was lower in the four testing patients (64 and 64% in segment- and arc-averaged TAESS). Although the correlation and concordance were high for both patient groups, the absolute TAESS values averaged per segment and arc were overestimated using non-invasive vs. invasive imaging [testing patients: TAESS segment: 30.1(17.1-83.8) vs. 15.8(8.8-63.4) and TAESS arc: 29.4(16.2-74.7) vs 15.0(8.9-57.4) p < 0.001]. We showed that our methodology can accurately assess the TAESS distribution non-invasively from CTA and demonstrated a good correlation with TAESS calculated using IVUS/OCT 3D reconstructed models.

Cost-effectiveness Analysis of Anatomic vs Functional Index Testing in Patients With Low-Risk Stable Chest Pain.

Importance: Both noninvasive anatomic and functional testing strategies are now routinely used as initial workup in patients with low-risk stable chest pain (SCP). Objective: To determine whether anatomic approaches (ie, coronary computed tomography angiography [CTA] and coronary CTA supplemented with noninvasive fractional flow reserve [FFRCT], performed in patients with 30% to 69% stenosis) are cost-effective compared with functional testing for the assessment of low-risk SCP. Design, Setting, and Participants: This cost-effectiveness analysis used an individual-based Markov microsimulation model for low-risk SCP. The model was developed using patient data from the Prospective Multicenter Imaging Study for Evaluation of Chest Pain (PROMISE) trial. The model was validated by comparing model outcomes with outcomes observed in the PROMISE trial for anatomic (coronary CTA) and functional (stress testing) strategies, including diagnostic test results, referral to invasive coronary angiography (ICA), coronary revascularization, incident major adverse cardiovascular event (MACE), and costs during 60 days and 2 years. The validated model was used to determine whether anatomic approaches are cost-effective over a lifetime compared with functional testing. Exposure: Choice of index test for evaluation of low-risk SCP. Main Outcomes and Measures: Downstream ICA and coronary revascularization, MACE (death, nonfatal myocardial infarction), cost, quality-adjusted life-years (QALYs), and incremental cost-effectiveness ratio (ICER) of competing strategies. Results: The model cohort included 10 003 individual patients (median [interquartile range] age, 60.0 [54.4-65.9] years; 5270 [52.7%] women; 7693 [77.4%] White individuals), who entered the model 100 times. The Markov model accurately estimated the test assignment, results of anatomic and functional index testing, referral to ICA, revascularization, MACE, and costs at 60 days and 2 years compared with observed data in PROMISE (eg, coronary CTA: ICA, 12.2% [95% CI, 10.9%-13.5%] vs 12.3% [95% CI, 12.2%-12.4%]; revascularization, 6.2% [95% CI, 5.5%-6.9%] vs 6.4% [95% CI, 6.3%-6.5%]; functional strategy: ICA, 8.1% [95% CI, 7.4%-8.9%] vs 8.2% [95% CI, 8.1%-8.3%]; revascularization, 3.2% [95% CI, 2.7%-3.7%] vs 3.3% [95% CI, 3.2%-3.4%]; 2-year MACE rates: coronary CTA, 2.1% [95% CI, 1.7%-2.5%] vs 2.3% [95% CI, 2.2%-2.4%]; functional strategy, 2.2% [95% CI, 1.8%-2.6%] vs 2.4% [95% CI, 2.3%-2.4%]). Anatomic approaches led to higher ICA and revascularization rates at 60 days, 2 years, and 5 years compared with functional testing but were more effective in patient selection for ICA (eg, 60-day revascularization-to-ICA ratio, CTA: 53.7% [95% CI, 53.3%-54.0%]; CTA with FFRCT: 59.5% [95% CI, 59.2%-59.8%]; functional testing: 40.7% [95% CI, 40.4%-50.0%]). Over a lifetime, anatomic approaches gained an additional 6 months in perfect health compared with functional testing (CTA, 25.16 [95% CI, 25.14-25.19] QALYs; CTA with FFRCT, 25.14 [95% CI, 25.12-25.17] QALYs; functional testing, 24.68 [95% CI, 24.66-24.70] QALYs). Anatomic strategies were less costly and more effective; thus, CTA with FFRCT dominated and CTA alone was cost-effective (ICERs ranged from $1912/QALY for women and $3,559/QALY for men) compared with functional testing. In probabilistic sensitivity analyses, anatomic approaches were cost-effective in more than 65% of scenarios, assuming a willingness-to-pay threshold of $100 000/QALY. Conclusions and Relevance: The results of this study suggest that anatomic strategies may present a more favorable initial diagnostic option in the evaluation of low-risk SCP compared with functional testing.

Radiomics versus Visual and Histogram-based Assessment to Identify Atheromatous Lesions at Coronary CT Angiography: An ex Vivo Study.

Background Visual and histogram-based assessments of coronary CT angiography have limited accuracy in the identification of advanced lesions. Radiomics-based machine learning (ML) could provide a more accurate tool. Purpose To compare the diagnostic performance of radiomics-based ML with that of visual and histogram-based assessment of ex vivo coronary CT angiography cross sections to identify advanced atherosclerotic lesions defined with histologic examination. Materials and Methods In this prospective study, 21 coronary arteries from seven hearts obtained from male donors (mean age, 52.3 years ± 5.3) were imaged ex vivo with coronary CT angiography between February 23, 2009, and July 31, 2010. From 95 coronary plaques, 611 histologic cross sections were coregistered with coronary CT cross sections. Lesions were considered advanced if early fibroatheroma, late fibroatheroma, or thin-cap atheroma was present. CT cross sections were classified as showing homogeneous, heterogeneous, or napkin-ring sign plaques on the basis of visual assessment. The area of low attenuation (<30 HU) and the average Hounsfield unit were quantified. Radiomic parameters were extracted and used as inputs to ML algorithms. Eight radiomics-based ML models were trained on randomly selected cross sections (training set, 75% of the cross sections) to identify advanced lesions. Visual assessment, histogram-based assessment, and the best ML model were compared on the remaining 25% of the data (validation set) by using the area under the receiver operating characteristic curve (AUC) to identify advanced lesions. Results After excluding sections with no visible plaque (n = 134) and with heavy calcium (n = 32), 445 cross sections were analyzed. Of those 445 cross sections, 134 (30.1%) were advanced lesions. Visual assessment of the 445 cross sections indicated that 207 (46.5%) were homogeneous, 200 (44.9%) were heterogeneous, and 38 (8.5%) demonstrated the napkin-ring sign. A radiomics-based ML model incorporating 13 parameters outperformed visual assessment (AUC = 0.73 with 95% confidence interval [CI] of 0.63, 0.84 vs 0.65 with 95% CI of 0.56, 0.73, respectively; P = .04), area of low attenuation (AUC = 0.55 with 95% CI of 0.42, 0.68; P = .01), and average Hounsfield unit (AUC = 0.53 with 95% CI of 0.42, 0.65; P = .004) in the identification of advanced atheromatous lesions. Conclusion Radiomics-based machine learning analysis improves the discriminatory power of coronary CT angiography in the identification of advanced atherosclerotic lesions. Published under a CC BY 4.0 license.

Secondary Interpretation of CT Examinations: Frequency and Payment in the Medicare Fee-for-Service Population.

PURPOSE: Secondary interpretation of diagnostic imaging examinations (providing a second formal interpretation for imaging performed at another institution) may reduce repeat imaging after transfer of care. Recently, CMS requested information to guide payment policy. We aimed to study historic trends in submitted claims and payments for secondary interpretation services in the Medicare fee-for-service population. METHODS: Applying current procedural terminology codes by body part to Medicare Part B aggregate claims files, we identified all CT interpretation services rendered between 1999 and 2012. Secondary interpretation services were identified using combined code modifiers 26 and 77, in accordance with CMS billing guidelines. The frequencies of billed and denied services were extracted for primary and secondary CT interpretation services. Primary versus secondary interpretation denial rates were calculated and compared. RESULTS: Of all 227 million Medicare Part B claims for CT services, 299,468 (0.13%) were for secondary interpretation services. From 1999 to 2012, growth in secondary interpretation claims outpaced that in primary interpretation claims (+811% versus +56%; compound annual growth rate 17% versus 3.2%). As a percentage of all services, secondary interpretations increased from 0.05% in 1999 to 0.30% in 2012. Denial rates for second interpretations decreased from 1999 to 2012 (12.7% to 7.0%), and now approach those for primary interpretations (5.4% in 2012). CONCLUSIONS: Medicare claims for secondary interpretation of CT examinations are growing but account for less than 1% of all billed CT interpretation services. Denial rates are similar to those of primary interpretation services.

Quantitative assessment of asymmetric aortic dilation with valve-related aortic disease.

RATIONALE AND OBJECTIVES: Previous work suggests that ascending aortic (AsAo) dilation can be asymmetric and is potentially related to valve-related blood flow abnormalities. The aim of this study was to investigate the relationship between the aortic valve and AsAo dilation using a quantitative, three-dimensional assessment of aortic shapes. MATERIALS AND METHODS: Computed tomographic and magnetic resonance images of the thorax were retrospectively reviewed. Four groups with aortic dilation were studied: those with tricuspid aortic valves (TAVs) with and without stenosis and those with bicuspid aortic valves (BAVs) with and without stenosis. Controls had either TAVs or BAVs but no aortic stenosis or dilation. In additional to standard orthogonal diameters, a unique measurement of AsAo asymmetry was used: the ratio of the greater to lesser curvatures measured using three-dimensional reformats in a "candy-cane" orientation. RESULTS: A total of 105 patients were identified. Ratios of greater to lesser curvature in patients with aortic dilation and nonstenotic TAVs were not significantly different from those in controls (1.69 vs 1.55, P > .20), but the asymmetry reflected by this ratio was markedly increased in patients with aortic dilation and stenotic TAVs (1.94, P < .001). Patients with aortic dilation and BAVs had significantly elevated ratios regardless of the status of the aortic valve (1.96 for nonstenotic and 2.05 for stenotic vs 1.53 for controls, P < .001). CONCLUSIONS: Asymmetric AsAo dilation with relative bulging of the greater curvature is linked to aortic stenosis, but it is also seen with nonstenotic BAVs. This suggests that the hemodynamic forces that contribute to aortic dilation are not fully revealed by conventional assessment of the aortic valve.

Subjective assessment of right ventricle enlargement from computed tomography pulmonary angiography images.

To retrospectively evaluate prognostic accuracy of subjective assessment of right ventricle (RV) enlargement on CT pulmonary angiography (CTPA) images in comparison with objective measures of RV enlargement in patients with acute pulmonary embolism (PE). For 200 consecutive patients with acute PE, two readers blinded to patient outcomes subjectively determined whether the maximum RV diameter was greater than that of the left ventricle (LV) using axial CTPA images. For the objective measurements, RV/LV diameter ratios were calculated using axial images and 4-chamber reformatted images. For all assessments, sensitivities and specificities for predicting PE-related death within 30-days and a composite outcome including PE-related death or the need for intensive therapies were compared. The agreement between two readers was 91.5% (kappa = 0.83) and all other assessments had pair-wise agreement over 75% (kappa = 0.53-0.72). There was no significant difference in sensitivity between the subjective and objective methods for predicting both outcomes. The specificity for subjective RV enlargement (55.4-67.7%) was significantly higher than objective measures (45.8-53.1%), except for the 4-chamber views where, for one reader, the specificity of the subjective evaluation was higher but did not reach statistical significance. Complex measurements of RV/LV diameter ratios may not be needed to maximize the prognostic value from CTPA. The radiologist who interprets the CTPA images should report RV enlargement when the RV diameter subjectively appears larger than the LV.