Scholarlink

Background <p>Contemporary cardiac surgical risk prediction relies on two established models — STS-PROM and EuroSCORE II — which share the same risk stratification framework: low risk below 4%, intermediate 4–8%, high above 8% predicted mortality. Both demonstrate satisfactory discrimination but systematic calibration failure across diverse populations. Independent studies have confirmed frailty, liver dysfunction, LV dimensions, SYNTAX score as the most important absent physiological domains with need for machine learning integration.</p> Objectives <p>To design the UCSRS v1.0: a unified physiology-informed risk architecture incorporating frailty (CFS), hepatorenal reserve (MELD), hemodynamic reserve (RHC), LV dimensions (LVEDD/LVESVI), and SYNTAX score as optional inputs. All domains except the core STS/EuroSCORE II baseline are optional — the score calculates regardless of which domains are available, with informative alerts for absent data rather than hard stops.</p> Methods <p>Published aggregate validation statistics from 14 international cardiac surgery registries from 2013 to 2024. No individual patient-level data accessed.</p> Results <p>The UCSRS provides a four-layer physiology-informed architecture: Layer 1 (baseline: STS-PROM 50% + EuroSCORE II 50%); Layer 2a (MELD correction, additive); Layer 2b (CFS frailty multiplier, multiplicative); Layer 2c (LV dimensions and SYNTAX score, additive, optional); Layer 3 (RHC hemodynamic corrections, additive, optional). Because all domain weights and correction terms are provisional and not regression-derived, the purpose of UCSRS v1.0 is to evaluate architecture rather than to establish final coefficients. The score calculates at any level of data completeness. No composite UCSRS performance metrics are reported.</p> Conclusions <p>The UCSRS is a physiology-informed risk prediction framework designed to incorporate frailty, hepatorenal reserve, ventricular remodeling, coronary anatomical complexity, and hemodynamic reserve into cardiac surgical risk assessment. Its tiered optional architecture calculates with any combination of available inputs, alerting clinicians to absent domains rather than requiring complete data. Prospective validation is required before clinical implementation.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Universal cardiac surgical risk score (UCSRS v1.0) a unified physiology-informed risk architecture for operative mortality prediction across all adult cardiac surgical procedures: algorithm design and prospective validation protocol

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年7月27日

查看全文
  • Atiq Rehman

摘要

Background

Contemporary cardiac surgical risk prediction relies on two established models — STS-PROM and EuroSCORE II — which share the same risk stratification framework: low risk below 4%, intermediate 4–8%, high above 8% predicted mortality. Both demonstrate satisfactory discrimination but systematic calibration failure across diverse populations. Independent studies have confirmed frailty, liver dysfunction, LV dimensions, SYNTAX score as the most important absent physiological domains with need for machine learning integration.

Objectives

To design the UCSRS v1.0: a unified physiology-informed risk architecture incorporating frailty (CFS), hepatorenal reserve (MELD), hemodynamic reserve (RHC), LV dimensions (LVEDD/LVESVI), and SYNTAX score as optional inputs. All domains except the core STS/EuroSCORE II baseline are optional — the score calculates regardless of which domains are available, with informative alerts for absent data rather than hard stops.

Methods

Published aggregate validation statistics from 14 international cardiac surgery registries from 2013 to 2024. No individual patient-level data accessed.

Results

The UCSRS provides a four-layer physiology-informed architecture: Layer 1 (baseline: STS-PROM 50% + EuroSCORE II 50%); Layer 2a (MELD correction, additive); Layer 2b (CFS frailty multiplier, multiplicative); Layer 2c (LV dimensions and SYNTAX score, additive, optional); Layer 3 (RHC hemodynamic corrections, additive, optional). Because all domain weights and correction terms are provisional and not regression-derived, the purpose of UCSRS v1.0 is to evaluate architecture rather than to establish final coefficients. The score calculates at any level of data completeness. No composite UCSRS performance metrics are reported.

Conclusions

The UCSRS is a physiology-informed risk prediction framework designed to incorporate frailty, hepatorenal reserve, ventricular remodeling, coronary anatomical complexity, and hemodynamic reserve into cardiac surgical risk assessment. Its tiered optional architecture calculates with any combination of available inputs, alerting clinicians to absent domains rather than requiring complete data. Prospective validation is required before clinical implementation.