Scholarlink

<p>Anastomotic leak is a life-threatening complication following colorectal surgery. This study benchmarks Quantum Neural Networks (QNNs) against hyperparameter-tuned classical models (logistic regression, multi-layer perceptrons, boosting algorithms) for anastomotic leak prediction. Using a 200-patient clinical dataset strictly bounded by a priori medical constraints, we simulated QNNs with ZZFeatureMap encoding and EfficientSU2/RealAmplitudes ansatze under realistic hardware noise. To ensure statistical reliability, performance metrics were averaged across 10 independent optimization runs. The EfficientSU2-BFGS configuration achieved the highest mean AUC of <InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(0.797 \pm 0.024\)</EquationSource> </InlineEquation>, while RealAmplitudes with CMA-ES maximized Average Precision (<InlineEquation ID="IEq2"> <EquationSource Format="TEX">\(0.504 \pm 0.121\)</EquationSource> </InlineEquation>). Crucially, at a fixed, clinically necessary sensitivity of <InlineEquation ID="IEq3"> <EquationSource Format="TEX">\(83\%\)</EquationSource> </InlineEquation>, specific QNN configurations achieved significantly higher specificity (up to <InlineEquation ID="IEq4"> <EquationSource Format="TEX">\(66\%\)</EquationSource> </InlineEquation>) and Negative Predictive Value (up to <InlineEquation ID="IEq5"> <EquationSource Format="TEX">\(96\%\)</EquationSource> </InlineEquation>) compared to classical models (maximum <InlineEquation ID="IEq6"> <EquationSource Format="TEX">\(44\%\)</EquationSource> </InlineEquation> and <InlineEquation ID="IEq7"> <EquationSource Format="TEX">\(94\%\)</EquationSource> </InlineEquation>, respectively), effectively minimizing false positives. However, classical models maintained superior probability calibration for continuous risk stratification. We conclude that QNNs offer robust discriminative performance for clinical screening, warranting further validation on larger, independent cohorts.</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

Quantum machine learning for predicting anastomotic leak: a clinical study

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年5月19日

查看全文
  • Vojtěch Novák,
  • Ivan Zelinka,
  • Lenka Přibylová,
  • Lubomír Martínek,
  • Vladimír Benčurik

摘要

Anastomotic leak is a life-threatening complication following colorectal surgery. This study benchmarks Quantum Neural Networks (QNNs) against hyperparameter-tuned classical models (logistic regression, multi-layer perceptrons, boosting algorithms) for anastomotic leak prediction. Using a 200-patient clinical dataset strictly bounded by a priori medical constraints, we simulated QNNs with ZZFeatureMap encoding and EfficientSU2/RealAmplitudes ansatze under realistic hardware noise. To ensure statistical reliability, performance metrics were averaged across 10 independent optimization runs. The EfficientSU2-BFGS configuration achieved the highest mean AUC of \(0.797 \pm 0.024\) , while RealAmplitudes with CMA-ES maximized Average Precision ( \(0.504 \pm 0.121\) ). Crucially, at a fixed, clinically necessary sensitivity of \(83\%\) , specific QNN configurations achieved significantly higher specificity (up to \(66\%\) ) and Negative Predictive Value (up to \(96\%\) ) compared to classical models (maximum \(44\%\) and \(94\%\) , respectively), effectively minimizing false positives. However, classical models maintained superior probability calibration for continuous risk stratification. We conclude that QNNs offer robust discriminative performance for clinical screening, warranting further validation on larger, independent cohorts.