Research

BACKGROUND: Prenatal per- and polyfluoroalkyl substance (PFAS) exposures are associated with adverse offspring health outcomes, yet the underlying pathological mechanisms are unclear. Cord blood metabolomics can identify potentially important pathways associated with prenatal PFAS exposures, providing mechanistic insights that may help explain PFAS' long-term health effects.

METHODS: The study included 590 mother-infant dyads from the Boston Birth Cohort. We measured PFAS in maternal plasma samples collected 24-72 h after delivery and metabolites in cord plasma samples. We used metabolome-wide association studies and pathway enrichment analyses to identify metabolites and pathways associated with individual PFAS, and quantile-based g-computation models to examine associations of metabolites with the PFAS mixture. We used False Discovery Rate to account for multiple comparisons.

RESULTS: We found that 331 metabolites and 18 pathways were associated with ≥ 1 PFAS, and 38 metabolites were associated with the PFAS mixture, predominantly amino acids and lipids. Amino acids such as alanine and lysine and their pathways, crucial to energy generation, biosynthesis, and bone health, were associated with PFAS and may explain PFAS' effects on fetal growth restriction. Carnitines and carnitine shuttle pathway, associated with 7 PFAS and the PFAS mixture, are involved in mitochondrial fatty acid β-oxidation, which may predispose higher risks of fetal and child growth restriction and cardiovascular diseases. Lipids, such as glycerophospholipids and their related pathway, can contribute to insulin resistance and diabetes by modulating transporters on cell membranes, participating in β-cell signaling pathways, and inducing oxidative damage. Neurotransmission-related metabolites and pathways associated with PFAS, including cofactors, precursors, and neurotransmitters, may explain the PFAS' effects on child neurodevelopment. We observed stronger associations between prenatal PFAS exposures and metabolites in males.

CONCLUSIONS: This prospective birth cohort study contributes to the limited literature on potential metabolomic perturbations for prenatal PFAS exposures. Future studies are needed to replicate our findings and link prenatal PFAS associated metabolomic perturbations to long-term child health outcomes.

BACKGROUND: Late gadolinium enhancement (LGE) cardiovascular magnetic resonance (CMR) imaging enables imaging of scar/fibrosis and is a cornerstone of most CMR imaging protocols. CMR imaging can benefit from image acceleration; however, image acceleration in LGE remains challenging due to its limited signal-to-noise ratio. In this study, we sought to evaluate a rapid two-dimensional (2D) LGE imaging protocol using a generative artificial intelligence (AI) algorithm with inline reconstruction.

METHODS: A generative AI-based image enhancement was used to improve the sharpness of 2D LGE images acquired with low spatial resolution in the phase-encode direction. The generative AI model is an image enhancement technique built on the enhanced super-resolution generative adversarial network. The model was trained using balanced steady-state free-precession cine images, readily used for LGE without additional training. The model was implemented inline, allowing the reconstruction of images on the scanner console. We prospectively enrolled 100 patients (55 ± 14 years, 72 males) referred for clinical CMR at 3T. We collected three sets of LGE images in each subject, with in-plane spatial resolutions of 1.5 × 1.5-3-6 mm2. The generative AI model enhanced in-plane resolution to 1.5 × 1.5 mm2 from the low-resolution counterparts. Images were compared using a blur metric, quantifying the perceived image sharpness (0 = sharpest, 1 = blurriest). LGE image sharpness (using a 5-point scale) was assessed by three independent readers.

RESULTS: The scan times for the three imaging sets were 15 ± 3, 9 ± 2, and 6 ± 1 s, with inline generative AI-based images reconstructed time of ∼37 ms. The generative AI-based model improved visual image sharpness, resulting in lower blur metric compared to low-resolution counterparts (AI-enhanced from 1.5 × 3 mm2 resolution: 0.3 ± 0.03 vs 0.35 ± 0.03, P < 0.01). Meanwhile, AI-enhanced images from 1.5 × 3 mm2 resolution and original LGE images showed similar blur metric (0.30 ± 0.03 vs 0.31 ± 0.03, P = 1.0) Additionally, there was an overall 18% improvement in image sharpness between AI-enhanced images from 1.5 × 3 mm2 resolution and original LGE images in the subjective blurriness score (P < 0.01).

CONCLUSION: The generative AI-based model enhances the image quality of 2D LGE images while reducing the scan time and preserving imaging sharpness. Further evaluation in a large cohort is needed to assess the clinical utility of AI-enhanced LGE images for scar evaluation, as this proof-of-concept study does not provide evidence of an impact on diagnosis.

BACKGROUND: Reliable, noninvasive tools to diagnose at-risk metabolic dysfunction-associated steatohepatitis (MASH) are urgently needed to improve management. We developed a risk stratification score incorporating proteomics-derived serum markers with clinical variables to identify high-risk patients with MASH (NAFLD activity score >4 and fibrosis score >2).

METHODS: In this 3-phase proteomic study of biopsy-proven metabolic dysfunction-associated steatotic fatty liver disease, we first developed a multi-protein predictor for discriminating NAFLD activity score >4 based on SOMAscan proteomics quantifying 1305 serum proteins from 57 US patients. Four key predictor proteins were verified by ELISA in the expanded US cohort (N = 168) and enhanced by adding clinical variables to create the 9-feature MASH Dx score, which predicted MASH and also high-risk MASH (F2+). The MASH Dx score was validated in 2 independent, external cohorts from Germany (N = 139) and Brazil (N = 177).

RESULTS: The discovery phase identified a 6-protein classifier that achieved an AUC of 0.93 for identifying MASH. Significant elevation of 4 proteins (THBS2, GDF15, SELE, and IGFBP7) was verified by ELISA in the expanded discovery and independently in the 2 external cohorts. MASH Dx score incorporated these proteins with established MASH risk factors (age, body mass index, ALT, diabetes, and hypertension) to achieve good discrimination between MASH and metabolic dysfunction-associated steatotic fatty liver disease without MASH (AUC: 0.87-discovery; 0.83-pooled external validation cohorts), with similar performance when evaluating high-risk MASH F2-4 (vs. MASH F0-1 and metabolic dysfunction-associated steatotic fatty liver disease without MASH).

CONCLUSIONS: The MASH Dx score offers the first reliable noninvasive approach combining novel, biologically plausible ELISA-based fibrosis markers and clinical parameters to detect high-risk MASH in patient cohorts from the United States, Brazil, and Europe.

BACKGROUND: Patients undergoing colectomy are at risk of numerous major complications. However, existing binary risk stratification models do not predict when a patient may be at highest risks of each complication. Accurate prediction of the timing of complications facilitates targeted, resource-efficient monitoring. We sought to develop and internally validate Cox proportional hazards models to predict time-to-complication of major complications within 30 days after elective colectomy.

METHODS: We studied a retrospective cohort from the multicentered American College of Surgeons National Surgical Quality Improvement Program procedure-targeted colectomy dataset. Patients aged 18 years or above, who underwent elective colectomy between January 1, 2014 and December 31, 2019 were included. A priori candidate predictors were selected based on variable availability, literature review, and multidisciplinary team consensus. Outcomes were mortality, hospital readmission, myocardial infarction, cerebral vascular events, pneumonia, venous thromboembolism, acute renal failure, and sepsis or septic shock within 30 days after surgery.

RESULTS: The cohort consisted of 132145 patients (mean ± SD age, 61 ± 15 years; 52% females). Complication rates ranged between 0.3% (n = 383) for cardiac arrest and acute renal failure to 5.3% (n = 6986) for bleeding requiring transfusion, with readmission rate of 8.6% (n = 11415). We observed distinct temporal patterns for each complication: the median [quartiles] postoperative day of complication diagnosis ranged from 1 [0, 2] days for bleeding requiring transfusion to 12 [6, 18] days for venous thromboembolism. Models for mortality, myocardial infarction, pneumonia, and renal failure showed good discrimination with a concordance > 0.8, while models for readmission, venous thromboembolism, and sepsis performed poorly with a concordance of 0.6 to 0.7. Models exhibited good calibration but ranges were limited to low probability areas.

CONCLUSIONS: We developed and internally validated time-to-event prediction models for complications after elective colectomy. Once further validated, the models can facilitate tailored monitoring of high risk patients during high risk periods.

TRIAL REGISTRATION: Clinicaltrials.gov (NCT05150548; Principal Investigator: Janny Xue Chen Ke, M.D., M.Sc., F.R.C.P.C.; initial posting: November 25, 2021).

BACKGROUND: Diagnostic scope is the range of diagnoses found in a clinical setting. Although the diagnostic scope is an essential feature of training and evaluating artificial intelligence (AI) systems to promote diagnostic excellence, its impact on AI systems and the diagnostic process remains under-explored.

CONTENT: We define the concept of diagnostic scope, discuss its nuanced role in building safe and effective AI-based diagnostic decision support systems, review current challenges to measurement and use, and highlight knowledge gaps for future research.

SUMMARY: The diagnostic scope parallels the differential diagnosis although the latter is at the level of an encounter and the former is at the level of a clinical setting. Therefore, diagnostic scope will vary by local characteristics including geography, population, and resources. The true, observed, and considered scope in each setting may also diverge, both posing challenges for clinicians, patients, and AI developers, while also highlighting opportunities to improve safety. Further work is needed to systematically define and measure diagnostic scope in terms that are accurate, equitable, and meaningful at the bedside. AI tools tailored to a particular setting, such as a primary care clinic or intensive care unit, will each require specifying and measuring the appropriate diagnostic scope.

OUTLOOK: AI tools will promote diagnostic excellence if they are aligned with patient and clinician needs and trained on an accurately measured diagnostic scope. A careful understanding and rigorous evaluation of the diagnostic scope in each clinical setting will promote optimal care through human-AI collaborations in the diagnostic process.

BACKGROUND: Dietary guidelines recommend substituting animal protein with plant protein, however, the ideal ratio of plant-to-animal protein (P:A) remains unknown.

OBJECTIVES: We aimed to evaluate associations between the P:A ratio and incident cardiovascular disease (CVD), coronary artery disease (CAD), and stroke in 3 cohorts.

METHODS: Multivariable-adjusted Cox proportional hazard models were used to estimate hazard ratios (HRs) for CVD outcomes among 70,918 females in the Nurses' Health Study (NHS) (1984-2016), 89,205 females in the NHSII (1991-2017) and 42,740 males from the Health Professionals Follow-up Study (1986-2016). The P:A ratio was based on percent energy from plant and animal protein and assessed using food frequency questionnaires every 4 y.

RESULTS: During 30 y of follow-up, 16,118 incident CVD cases occurred. In the pooled multivariable-adjusted models, participants had a lower risk of total CVD [HR: 0.81; 95% confidence interval (CI): 0.76, 0.87; P trend < 0.001], CAD (HR: 0.73; 95% CI: 0.67, 0.79; P trend < 0.001), but not stroke (HR: 0.98; 95% CI: 0.88, 1.09; P trend = 0.71), when comparing highest to lowest deciles of the P:A ratio (ratio: ∼0.76 compared with ∼0.24). Dose-response analyses showed evidence of linear and nonlinear relationships for CVD and CAD, with more marked risk reductions early in the dose-response curve. Lower risk of CVD (HR: 0.72; 95% CI: 0.64, 0.82) and CAD (HR: 0.64; 95% CI: 0.55, 0.75) were also observed with higher ratios and protein density (20.8% energy) combined. Substitution analyses indicated that replacing red and processed meat with several plant protein sources showed the greatest cardiovascular benefit.

CONCLUSIONS: In cohort studies of United States adults, a higher plant-to-animal protein ratio is associated with lower risks of CVD and CAD, but not stroke. Furthermore, a higher ratio combined with higher protein density showed the most cardiovascular benefit.