Search

Background: This multicenter and retrospective study investigated the additive value of tumor morphologic features derived from the functional tumor volume (FTV) tumor mask at pre-treatment (T0) and the early treatment time point (T1) in the prediction of pathologic outcomes for breast cancer patients undergoing neoadjuvant chemotherapy., Methods: A total of 910 patients enrolled in the multicenter I-SPY 2 trial were included. FTV and tumor morphologic features were calculated from the dynamic contrast-enhanced (DCE) MRI. A poor response was defined as a residual cancer burden (RCB) class III (RCB-III) at surgical excision. The area under the receiver operating characteristic curve (AUC) was used to evaluate the predictive performance. The analysis was performed in the full cohort and in individual sub-cohorts stratified by hormone receptor (HR) and human epidermal growth factor receptor 2 (HER2) status., Results: In the full cohort, the AUCs for the use of the FTV ratio and clinicopathologic data were 0.64 ± 0.03 (mean ± SD [standard deviation]). With morphologic features, the AUC increased significantly to 0.76 ± 0.04 ( p < 0.001). The ratio of the surface area to volume ratio between T0 and T1 was found to be the most contributing feature. All top contributing features were from T1. An improvement was also observed in the HR+/HER2- and triple-negative sub-cohorts. The AUC increased significantly from 0.56 ± 0.05 to 0.70 ± 0.06 ( p < 0.001) and from 0.65 ± 0.06 to 0.73 ± 0.06 ( p < 0.001), respectively, when adding morphologic features., Conclusion: Tumor morphologic features can improve the prediction of RCB-III compared to using FTV only at the early treatment time point.

Purpose: This study aimed to predict the probability of grade ≥2 pneumonitis or dyspnea within 12 months of receiving conventionally fractionated or mildly hypofractionated proton beam therapy for locally advanced lung cancer using machine learning., Methods and Materials: Demographic and treatment characteristics were analyzed for 965 consecutive patients treated for lung cancer with conventionally fractionated or mildly hypofractionated (2.2-3 Gy/fraction) proton beam therapy across 12 institutions. Three machine learning models (gradient boosting, additive tree, and logistic regression with lasso regularization) were implemented to predict Common Terminology Criteria for Adverse Events version 4 grade ≥2 pulmonary toxicities using double 10-fold cross-validation for parameter hyper-tuning without leak of information. Balanced accuracy and area under the curve were calculated, and 95% confidence intervals were obtained using bootstrap sampling., Results: The median age of the patients was 70 years (range, 20-97), and they had predominantly stage IIIA or IIIB disease. They received a median dose of 60 Gy in 2 Gy/fraction, and 46.4% received concurrent chemotherapy. In total, 250 (25.9%) had grade ≥2 pulmonary toxicity. The probability of pulmonary toxicity was 0.08 for patients treated with pencil beam scanning and 0.34 for those treated with other techniques (P = 8.97e-13). Use of abdominal compression and breath hold were highly significant predictors of less toxicity (P = 2.88e-08). Higher total radiation delivered dose (P = .0182) and higher average dose to the ipsilateral lung (P = .0035) increased the likelihood of pulmonary toxicities. The gradient boosting model performed the best of the models tested, and when demographic and dosimetric features were combined, the area under the curve and balanced accuracy were 0.75 ± 0.02 and 0.67 ± 0.02, respectively. After analyzing performance versus the number of data points used for training, we observed that accuracy was limited by the number of observations., Conclusions: In the largest analysis of prospectively enrolled patients with lung cancer assessing pulmonary toxicities from proton therapy to date, advanced machine learning methods revealed that pencil beam scanning, abdominal compression, and lower normal lung doses can lead to significantly lower probability of developing grade ≥2 pneumonitis or dyspnea., (Published by Elsevier Inc.)

Liver transplantation (LT) is a treatment for acute-on-chronic liver failure (ACLF), but high post-LT mortality has been reported. Existing post-LT models in ACLF have been limited. We developed an Expert-Augmented Machine Learning (EAML) model to predict post-LT outcomes. We identified ACLF patients who underwent LT in the University of California Health Data Warehouse. We applied the RuleFit machine learning (ML) algorithm to extract rules from decision trees and create intermediate models. We asked human experts to rate the rules generated by RuleFit and incorporated these ratings to generate final EAML models. We identified 1384 ACLF patients. For death at 1 year, areas under the receiver-operating characteristic curve were 0.707 (confidence interval [CI] 0.625-0.793) for EAML and 0.719 (CI 0.640-0.800) for RuleFit. For death at 90 days, areas under the receiver-operating characteristic curve were 0.678 (CI 0.581-0.776) for EAML and 0.707 (CI 0.615-0.800) for RuleFit. In pairwise comparisons, both EAML and RuleFit models outperformed cross-sectional models. Significant discrepancies between experts and ML occurred in rankings of biomarkers used in clinical practice. EAML may serve as a method for ML-guided hypothesis generation in further ACLF research., Competing Interests: Disclosure The authors of this manuscript have conflicts of interest to disclose as described by the American Journal of Transplantation. J. Ge receives research support from Merck and Co, and consults for Astellas Pharmaceuticals., (Copyright © 2023 American Society of Transplantation & American Society of Transplant Surgeons. Published by Elsevier Inc. All rights reserved.)

Introduction: Heart failure (HF) is a frequent cause of readmissions. Despite caring for underresourced patients and dependence on government funding, safety net hospitals frequently incur penalties for failing to meet pay-for-performance readmission metrics. Limited research exists on the causes of HF readmissions in safety net hospitals. Therefore, we sought to investigate predictors of 30-day all-cause readmission in HF patients in the safety net setting., Methods: We performed a retrospective chart review of patients admitted for HF from October 2018 to April 2019. We extracted data on demographics and medical comorbidities and performed patient-specific review of social determinants and mental health in 4 domains: race/ethnicity, housing status, substance use, and mental illness. Multivariable Poisson regression modeling was employed to evaluate associations with 30-day all-cause readmission., Results: The study population included 290 patients, among whom the mean age was 59 years and 71% ( n  = 207) were male; 42% (120) were Black/African American (AA), 22% (64) were Hispanic/Latino, and 96% (278) had public insurance; 28% (79) were not housed, 19% (56) had a diagnosis of mental illness, and active substance use was common. The 30-day readmission rate was 25.5% ( n  = 88). Factors that were associated with increased risk of readmission included self-identifying as Black/AA (relative risk 2.28, 95% confidence interval 1.00-5.20) or Hispanic/Latino (2.53, 1.07-6.00), experiencing homelessness (2.07, 1.21-3.56), living in a shelter (3.20, 1.27-8.02), or intravenous drug use (IVDU) (2.00, 1.08-3.70)., Conclusion: Race/ethnicity, housing status, and substance use were associated with increased risk of 30-day all-cause readmission in HF patients in a safety net hospital. In contrast to prior studies, medical comorbidities were not associated with increased risk of readmission., Competing Interests: The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (© 2023 The Author(s).)

The estimation of nested functions (i.e., functions of functions) is one of the central reasons for the success and popularity of machine learning. Today, artificial neural networks are the predominant class of algorithms in this area, known as representational learning. Here, we introduce Representational Gradient Boosting (RGB), a nonparametric algorithm that estimates functions with multi-layer architectures obtained using backpropagation in the space of functions. RGB does not need to assume a functional form in the nodes or output (e.g., linear models or rectified linear units), but rather estimates these transformations. RGB can be seen as an optimized stacking procedure where a meta algorithm learns how to combine different classes of functions (e.g., Neural Networks (NN) and Gradient Boosting (GB)), while building and optimizing them jointly in an attempt to compensate each other's weaknesses. This highlights a stark difference with current approaches to meta-learning that combine models only after they have been built independently. We showed that providing optimized stacking is one of the main advantages of RGB over current approaches. Additionally, due to the nested nature of RGB we also showed how it improves over GB in problems that have several high-order interactions. Finally, we investigate both theoretically and in practice the problem of recovering nested functions and the value of prior knowledge.

Background: In comparison with the general population, adolescents with juvenile idiopathic arthritis (JIA) are at higher risk for morbidity and mortality. However, limited evidence is available about this condition's underlying metabolic profile in adolescents with JIA relative to healthy controls. In this untargeted, cross-sectional metabolomics study, we explore the plasma metabolites in this population., Methods: A sample of 20 adolescents with JIA and 20 controls aged 13-17 years were recruited to complete surveys, provide medical histories and biospecimens, and undergo assessments. Fasting morning plasma samples were processed with liquid chromatography-mass spectrometry. Data were centered, scaled, and analyzed using generalized linear models accounting for age, sex, and medications (p-values adjusted for multiple comparisons using the Holm method). Spearman's correlations were used to evaluate relationships among metabolites, time since diagnosis, and disease severity., Results: Of 72 metabolites identified in the samples, 55 were common to both groups. After adjustments, 6 metabolites remained significantly different between groups. Alpha-glucose, alpha-ketoglutarate, serine, and N-acetylaspartate were significantly lower in the JIA group than in controls; glycine and cystine were higher. Seven additional metabolites were detected only in the JIA group; 10 additional metabolites were detected only in the control group. Metabolites were unrelated to disease severity or time since diagnosis., Conclusions: The metabolic signature of adolescents with JIA relative to controls reflects a disruption in oxidative stress; neurological health; and amino acid, caffeine, and energy metabolism pathways. Serine and N-acetylaspartate were promising potential biomarkers, and their metabolic pathways are linked to both JIA and cardiovascular disease risk. The pathways may be a source of new diagnostic, treatment, or prevention options. This study's findings contribute new knowledge for systems biology and precision health approaches to JIA research. Further research is warranted to confirm these findings in a larger sample., (© 2022. The Author(s).)