Anavekar, Nandan S.

Abstract Artificial intelligence is spreading through public health agencies and healthcare systems faster than workforces are trained to govern it. We describe RHAMI (Responsible Health AI Readiness and Maturity Index), the first human-centered automated composite index for health organizations to score AI readiness and maturity as one measurable system of science, economics and ethics, and the physician certification-to-Master’s pathway built to teach what it measures—both designed jointly by frontline public health professionals, clinicians, ethicists, engineers, leaders, and policy-makers in low- and middle-income countries and high-income countries.

BACKGROUND: Technetium-based cardiac amyloid radionuclide imaging enables noninvasive diagnosis of transthyretin amyloid cardiomyopathy (ATTR-CM), but conventional planar and visual semiquantitative approaches have important limitations. We developed the quantitative 3-dimensional Mayo 3D Score, defined as the ratio of mean left ventricular myocardial radiotracer uptake to right atrial blood-pool uptake using computed tomography (CT)-fused single-photon emission CT (SPECT) volumes of interest, and sought to validate its diagnostic performance against myocardial pathology. METHODS: Consecutive patients who underwent 99m Tc-pyrophosphate SPECT/CT at Mayo Clinic for suspected ATTR-CM between June 2017 and January 2025 with archived myocardial tissue were included. Baseline characteristics, including demographics, echocardiogram, biomarkers, visual semiquantitative, and Mayo 3D Score metrics, were collected. Myocardial specimens underwent histopathologic confirmation of amyloid, with amyloid subtype determination and further semiquantitative grading of amyloid burden and deposition patterns. RESULTS: Among 230 patients (median age, 72 years; 78% male), 107 (47%) had ATTR-CM, of whom 9 had hereditary ATTR-CM. The Mayo 3D Score showed excellent discrimination against pathology (area under the receiver operating characteristic curve, 0.92 [0.88–0.96]), exceeding planar Perugini (0.87 [0.83–0.92]; P =0.006), and heart-to-contralateral lung ratio (0.86 [0.81–0.91]; P =0.004) and comparable to SPECT Perugini (0.90 [0.86–0.94]; P =0.12). A cutoff ≥1.0 yielded 86.0% sensitivity, 92.7% specificity, 91.1% positive predictive value, and 88.4% negative predictive value. Among ATTR-positive cases, the Mayo 3D Score correlated strongly with semiquantitative amyloid burden (Spearman ρ=0.69), was higher in pericellular than nodular deposition patterns ( P <0.001), and was not associated with vascular involvement ( P =0.42). CONCLUSIONS: In patients with suspected ATTR-CM undergoing 99m Tc-pyrophosphate SPECT/CT, the Mayo 3D Score demonstrated superior diagnostic performance compared with planar metrics and comparable to SPECT Perugini, while correlating with myocardial ATTR burden and deposition patterns on pathology. These findings support its integration as a quantitative adjunct within clinical diagnostic ATTR-CM workflows.

Poor diet is the leading preventable risk factor for death worldwide, associated with over 10 million premature deaths and USD 8 trillion related costs every year. Artificial intelligence or AI is rapidly emerging as the most historically disruptive, innovatively dynamic, rapidly scaled, cost-efficient, and economically productive technology (which is increasingly providing transformative countermeasures to these negative health trends, especially in low- and middle-income countries (LMICs) and underserved communities which bear the greatest burden from them). Yet widespread confusion persists among healthcare systems and policymakers on how to best identify, integrate, and evolve the safe, trusted, effective, affordable, and equitable AI solutions that are right for their communities, especially in public health nutrition. We therefore provide here the first known global, comprehensive, and actionable narrative review of the state of the art of AI-accelerated nutrition assessment and healthy eating for healthcare systems, generated by the first automated end-to-end empirical index for responsible health AI readiness and maturity: the Responsible Health AI readiness and Maturity Index (RHAMI). The index is built and the analysis and review conducted by a multi-national team spanning the Global North and South, consisting of front-line clinicians, ethicists, engineers, executives, administrators, public health practitioners, and policymakers. RHAMI analysis identified the top-performing healthcare systems and their nutrition AI, along with leading use cases including multimodal edge AI nutrition assessments as ambient intelligence, the strategic scaling of practical embedded precision nutrition platforms, and sovereign swarm agentic AI social networks for sustainable healthy diets. This index-based review is meant to facilitate standardized, continuous, automated, and real-time multi-disciplinary and multi-dimensional strategic planning, implementation, and optimization of AI capabilities and functionalities worldwide, aligned with healthcare systems’ strategic objectives, practical constraints, and local cultural values. The ultimate strategic objectives of the RHAMI’s application for AI-accelerated public health nutrition are to improve population health, financial efficiency, and societal equity through the global cooperation of the public and private sectors stretching across the Global North and South.

BACKGROUND: Guidelines recommend 3-hour cardiac amyloid radionuclide imaging (CARI) for transthyretin amyloid cardiomyopathy. Citing rapid blood clearance of 99m Tc-hydroxymethylene-diphosphonate (HMDP) and efficient laboratory throughput, 1-hour imaging is increasingly practiced despite limited supporting evidence. We sought to compare diagnostic performance and interpreter experience of 1-hour versus 3-hour HMDP-CARI. METHODS: Consecutive patients with suspected transthyretin amyloid cardiomyopathy (n=114) underwent both 1-hour and 3-hour HMDP single photon emission computed tomography (CT)/CT. Two cardiologist-radiologist reader teams, blinded to imaging timepoint (1 versus 3 hours), assessed overall interpretation, single photon emission CT–based Perugini grade, interpretation difficulty, interpreter confidence, and need for CT-fused images for anatomic localization. Discordant, equivocal, and difficult cases were arbitrated by a third tie-breaking team. The myocardial-to-blood-pool radiotracer uptake ratio (3-dimensional Score) was measured as a surrogate of contrast resolution. RESULTS: Interinterpreter agreement was high at both time points (κ≥0.81), with more cases requiring arbitration at 3 hours versus 1 hour (22% versus 13%; P =0.049). Overall interpretation and Perugini grades were concordant between time points in 111/114 (97%) and 106/114 (93%) patients, respectively. Three patients (3%) were negative at 1 hour but equivocal at 3 hours, all of which were clinically ruled out for transthyretin amyloid cardiomyopathy. Interpreter confidence was comparable at both timepoints (97% versus 95%; P =0.317). Compared with 3-hour imaging, contrast resolution was inferior (lower 3-dimensional score, P <0.001) and CT fusion was more frequently needed (57% versus 31%, P <0.001) at 1-hour imaging. CONCLUSIONS: In a prospective, blinded comparison of 1-hour versus 3-hour HMDP-CARI, diagnostic performance and interpreter experience were similar, with readers requesting CT fusion more frequently at 1 hour to optimize myocardial-to-blood pool discrimination.

Abstract Cardiac implantable electronic device infections (CIEDIs) present substantial challenges for infectious diseases specialists, encompassing diagnosis, management, and complex decision making involving patients, families, and multidisciplinary teams. This review, guided by a common clinical case presentation encountered in daily practice, navigates through the diagnostic process, management strategies in unique scenarios, long-term follow-up, and critical discussions required for CIEDIs.

Abstract Cardiac implantable electronic device infections (CIEDIs) present substantial challenges for infectious diseases specialists, encompassing diagnosis, management, and complex decision-making involving patients, families, and multidisciplinary teams. This review, guided by a common clinical case presentation encountered in daily practice, navigates through the diagnostic process, management strategies in unique scenarios, long-term follow-up, and critical discussions required for CIEDIs.

Introduction: Metabolic pharmacotherapies [Glucagon-like peptide-1 receptor agonists (GLP1RA) and sodium-glucose cotransporter type 2 inhibitors (SGLT2i)] have shown benefit in patients with heart failure and preserved ejection fraction (HFpEF). It remains unclear whether the two classes would have additive effects. Methods: We performed a retrospective study using a de-identified federated database with 89 contributing health care organizations (HCO), including over 126 million patients (TriNetX Research Network, Cambridge, MA; date of data access: May 23, 2024). TriNetX regularly aggregates clinical data directly from participating HCOs, and performs extensive data quality and accuracy assessment. While TriNetX obfuscates institutional information relating to participating HCOs, a typical HCO includes a large academic health center with inpatient, outpatient, and specialty care services. We identified all patients with a diagnosis of Heart failure with preserved ejection fraction (HFpEF, ICD-10: I50.3), and SGLT2i usage. We then categorized this patient population into two cohorts, including patients with concomitant GLP1RA use, and patients without GLP1RA use. For both cohorts, we collected data on patient demographics, and the outcomes of all-cause mortality, and all-cause hospitalization, occurring up to 10-years after the index diagnosis. We evaluated differences in outcome risk using hazards ratios and 95% confidence intervals. Results: Prior to propensity score matching, we identified 114,329 patients, with 7931 patients receiving SGLT2i +GLP1RA vs 106398 receiving SGLT2i-alone. Majority (55.7%) identified as white and 24.4% as African Americans and >90% of the patients had a diagnosis of type 2 diabetes mellitus. After propensity matching, the rate of all-cause hospitalization within the combination group was lower than SGLT2i group (29.6% vs 32.4%, HR 0.76 [0.73- 0.79], P<0.001). All-cause mortality for the combination group was 6.8% versus 10.6% within the SGLT2i-alone group (HR 0.52 [0.47-0.56], P<0.001)). Conclusion: In this propensity-matched cohort, patients with HFpEF receiving GLP1RA and SGLT2i had lower rates of hospitalizations and all-cause mortality compared with those receiving SGLT2i alone. Large randomized controlled trials are warranted to corroborate these findings.

Abstract Background Atrial high-rate episodes (AHRE) are commonly detected in individuals with cardiac implantable electronic devices. Outcomes of atrial high-rate episodes remain unclear in those without a prior history of atrial fibrillation. Purpose This study aims to evaluate clinical outcomes of atrial high-rate episodes in patients who have no prior history of atrial fibrillation. Methods PubMed, EMBASE, and Cochrane databases were searched through February 2024 for studies evaluating clinical outcomes in individuals with cardiac implantable electronic devices who experienced atrial high-rate episodes, compared to those who did not. Individuals who had a prior history of atrial fibrillation were excluded. Outcomes assessed were clinical atrial fibrillation, thromboembolic events, and all-cause mortality. A fixed-effects model was used to calculate risk ratios (RRs) with 95% confidence intervals (CI). A subgroup analysis of thromboembolic events was conducted based on the duration of atrial high-rate episodes. Results A total of 9 studies met inclusion criteria with a combined enrollment of 7440 patients. Compared to those who did not experience atrial high-rate episodes, those who did had significantly increased risks of clinical atrial fibrillation (RR 3.33; 95% CI 1.99 - 5.57; P &lt; 0.0001) and thromboembolic events (RR 2.21; 95% CI 1.72 -2.85; P &lt; 0.0001). The risk of all-cause mortality was similar between both groups (RR 1.19; 95% CI 0.95 – 1.49; P = 0.13). Subgroup analysis revealed an increased risk of thromboembolic events among atrial high-rate episodes ≥ 24 hours (RR 12.34; 95% CI 2.70 – 56.36) while atrial high-rate episodes lasting &lt; 24 hours were insignificantly associated with thromboembolic events (RR 0.83; 95% CI 0.12 – 5.58). Conclusions Our study indicates a heightened risk of clinical atrial fibrillation and thromboembolic events among patients with atrial high-rate episodes, with no observed increase in mortality risk. Furthermore, our findings suggest that the duration of atrial high-rate episodes influences the risk of thromboembolism. Further studies are warranted to confirm our findings.

Abstract Background The role of anticoagulation in patients with atrial high-rate episodes (AHRE), who have no prior history of atrial fibrillation or atrial flutter, is being extensively studied. Purpose Given the increased thrombo-embolic risk in patients who have atrial high-rate episodes, our aim was to evaluate whether anticoagulation leads to enhanced clinical outcomes compared to no anticoagulation (1). Methods We searched PubMed, EMBASE, and the Cochrane Library, through February 2024, for studies evaluating outcomes in patients with atrial high-rate episodes managed with oral anticoagulation (OAC) compared to those not managed with oral anticoagulation. Outcomes assessed were ischemic stroke, thromboembolic events, major bleeding, and all-cause mortality. A fixed-effects model was used to calculate pooled risk risk ratios (RR) and their 95% confidence intervals (CI). Results A total of 4 studies comprising 6788 patients were included. Anticoagulation was associated with a significant reduction in the risks of ischemic stroke (RR 0.66; 95% CI 0.49 - 0.90; P = 0.0084) and thromboembolic events (RR 0.70; 95% CI 0.55 - 0.89; P = 0.0037). Compared to atrial high-rate episodes not managed with anticoagulation, there was a significantly increased risk of major bleeding with anticoagulation (RR 1.57; 95% CI 1.23 - 2.00; P = 0.0002), while the risk of all-cause mortality was similar (RR 1.07; 95% CI 0.95 - 1.21; P = 0.2437). Conclusion Our study suggests that anticoagulation in patients with atrial high-rate episodes decreases the risks of ischemic stroke and thromboembolic events, while increasing the risk of major bleeding, compared patients not managed with anticoagulation. Our study does not show a mortality benefit with the use of anticoagulation in patients with atrial high-rate episodes. Further studies are needed to confirm the safety and efficacy of anticoagulation in atrial high-rate episodes.Forest Plots of Outcomes Assessed

Abstract Background Transcatheter aortic valve replacement (TAVR) continues to be increasingly performed. While machine learning has demonstrated utility in predicting adverse outcomes in various cardiovascular procedures, its effectiveness in predicting in-hospital mortality in TAVR remains uncertain. Purpose We aimed to examine the accuracy of machine learning in predicting in-hospital mortality in TAVR. Methods We searched PubMed and Cochrane databases up to February 2024 for observational studies assessing the diagnostic accuracy of machine learning for predicting in-hospital mortality in TAVR. Extracted data included study characteristics, patient demographics, type of predictive model used, and area under the receiver operating characteristic curves (AUC). Predictive accuracy was assessed by pooling AUC of machine-learning models across studies. AUC values ranging from 0.80 to 0.90, 0.70 to 0.80, 0.60 to 0.70, and 0.50 to 0.60 corresponded to excellent, good, fair, and poor discrimination ability, respectively (1). We used a random effects model to estimate the overall effect size with a 95% confidence interval (CI). Heterogeneity was evaluated with the I2 test.We conducted subgroup analyses based on the type of machine-learning model utilized to assess for any potential differences in predictive accuracy. Results A total of three studies were identified, comprising a study population of 70,714 patients. Across studies, eight machine learning models were utilized, consisting of three artificial neural network models, one logistic regression model, one naive Bayes model, two random forest models, and one support vector machine model. Pooled AUC showed excellent accuracy of machine learning in predicting in-hospital mortality in TAVR (0.92; 95% CI 0.90 - 0.95). Subgroup analyses revealed similar yet excellent accuracy across all types of machine learning models utilized: artificial neural network (0.92; 95% CI 0.86-0.99), logistic regression (0.92; 95% CI 0.89-0.95), naive Bayes (0.90; 95% CI 0.88-0.92), random forest (0.94; 95% CI 0.87-1.00), and support vector machine (0.94; 95% CI 0.91-0.96). Conclusion Our study suggests that utilizing machine learning can accurately predict in-hospital mortality in TAVR. Operators of TAVR procedures may benefit from incorporating machine learning prediction models in their clinical practice. Further exploration of machine learning on outcomes in TAVR is needed to verify our findings.