Resumen es: Introduccion La hipertrofia del ventriculo izquierdo (HVI) incluye diferentes etiologias, estados evolutivos y pronostico. El strain rate sistolico (S...
Resumen es: Introduccion En la estenosis aortica, el mecanismo de adaptacion miocardica a la sobrecarga de presion es la hipertrofia ventricular. Diferentes trabajos...
BACKGROUND & AIMS:Acute-on-chronic liver failure (ACLF) is characterised by multiorgan failure and high short-term mortality in hospitalised patients with acute decompensation of cirrhosis. Although the EASL-CLIF criteria are widely used for diagnosis and prognostication, evolving definitions of organ dysfunction and emerging therapies require updated, tailored criteria to improve diagnostic accuracy, treatment assessment, and applicability in clinical trials. We aimed to develop and validate the A-TANGO organ failure (OF) score to refine ACLF diagnosis and enhance its utility for treatment response evaluation and risk stratification. METHODS:We performed a retrospective analysis of prospective observational cohorts. The derivation cohort comprised three EF-CLIF consortium studies conducted in Europe and Latin America (CANONIC, PREDICT, ACLARA; n = 3,896). Validation cohorts included one study from India (Ambi-spective study n = 2,055) and one from China (CATCH-LIFE; n = 2,568). Patients were enrolled between 2011 and 2023, with follow-up completed in 2023. The primary objective was to redefine thresholds for organ dysfunction and failure using three subscores per organ, with subscore 3 corresponding to ≥15% 28-day mortality and defining organ failure. RESULTS:Compared with the CLIF-C OF score, the A-TANGO OF score introduced revised thresholds for organ failure and added an ACLF grade 4 to address the wide mortality variation within CLIF-C OF grade 3. A-TANGO identified more organ failures, increasing ACLF diagnosis from 24% to 36% and improving the net reclassification index by 16%, while maintaining similar predictive accuracy for 28- and 90-day mortality. Two additional prognostic models (A-TANGO ACLF-WBC and A-TANGO ACLF-CRP) demonstrated strong associations with 28- and 90-day mortality and improved prognostic performance. Findings were confirmed in external validation cohorts. CONCLUSIONS:The A-TANGO OF score is a reproducible and comprehensive tool for ACLF diagnosis with preserved prognostic performance, validated across large international cohorts. It provides a robust framework for clinical trials by enabling more accurate diagnosis, reducing required sample sizes, and offering clinically meaningful endpoints such as ACLF resolution for treatment response assessment. IMPACT AND IMPLICATIONS:The A-TANGO organ failure (OF) score provides a scientifically justified advancement in ACLF research by refining organ-specific dysfunction thresholds and introducing a new grade 4, thereby addressing limitations in current EASL-CLIF criteria and improving identification of high-risk patients. These findings are important for clinicians, researchers, and healthcare systems globally, as they increase detection of organ failure, enhance risk stratification, and enable more accurate prediction of short-term mortality in hospitalized patients with cirrhosis. The A-TANGO OF score and its associated prognostic scores (ACLF-WBC and ACLF-CRP) can be applied in clinical practice to guide treatment decisions, and serve as reliable, measurable endpoints in clinical trials evaluating emerging therapies. Although limitations such as missing data, cohort-specific recruitment differences, and historical classification criteria exist, the consistent and robust performance of the A-TANGO scores across large, multinational cohorts highlights their potential applicability and utility on a global scale.
Code generation is important in software engineering, and Reinforcement Learning with Verifiable Rewards (RLVR) is a powerful paradigm to improve it through execution-based feedback. However, most RLVR pipelines rely on human-curated tests, making progress bottlenecked by scarce and costly supervision. Existing work tried to use self-generated tests to ground rewards, but the lack of discriminative tests constrains the effect due to the sub-optimal performance of the model on test generation. We aim to improve code generation without ground-truth supervision by co-evolving code and test generation, so that their interactions yield progressively more informative supervision. To this end, we present ZeroCoder, a fully label-free co-evolutionary framework that jointly trains a Coder and a Tester using execution feedback from self-generated code-test interactions. For each problem, ZeroCoder executes sampled solutions against sampled tests to form a passing matrix, identifies a consensus subset of likely-correct solutions and consistent tests via a pluggable selection algorithm, and derives role-specific rewards. To ensure reward quality, ZeroCoder filters low-information instances via rank-based pre-filtering and trains the Tester with a curriculum balancing validity and mutation-driven discriminativeness. We further identify selector drift, the progressive miscalibration of fixed selection rules during co-evolution, and introduce DyB4, a Bayesian selector that uses as few as 10 labeled instances to recalibrate its priors dynamically. Across three models and six benchmarks, ZeroCoder consistently improves code generation and test generation. In the fully label-free setting, it improves code generation by up to 14.5
