基于机器学习的血清学标志物预测粘连性小肠梗阻患者肠坏死的研究

打开文本图片集

【Abstract】Objective To explorethe value of machine learning-based serological markers in predicting irreversible transmuralintestialnecrosis ITN）insurgicalpatientswithadhesivesmallbowelobstruction（ASBO）.MethodsAtotalof133 ASBO patientswhounderwent surgical treatment at Xuzhou Central Hospital from February2023 to February2025 were prospectivelyerolled.According toitraoperativeexplorationand pathologicalresults，patients weredivided intonrois group （n=68） ） and non-necrosis group （n=65 ）. Fourteen indicators were assessed， including serum homocysteine （HCY）， endotoxin， procalcitonin （PCT）， C-reactive protein （CRP）， interleukin-6 （IL-6）， IL- 1β ，IL-5，neutrophil gelatinase-associated lipocalin （NGAL）， lactate dehydrogenase （LDH）， vitamin B12（VB12） ，folate， and age， gender， and body mass index （BMI）. Twenty machine learning models were constructed. The dataset was randomly divided into a training set （n=106） and atest set （n=27 ）at an 8：2 ratio.Model performance was evaluatedonthetestsetusingROCcurves，decisioncurveanalysis （DCA），calibrationcuves，nd SHAP feature importance analysis wasperformed.ResultsLevelsofHCY，endotoxin，PCTandCRPwere higherinteosis group than in the non-necrosis group （all P<0.05 ）. The Extra Trees model demonstrated optimal performance with an AUC of 0.977 （ 95% CI： 0.955-0.999）， sensitivity of 92.6% （ 95% CI： 83.9%96.8% ），andspecificityof 95.4% ？ 95% CI： 87.3%-98.4% ）

SHAP analysis identified HCY as the most important predictor （mean |SHAP value Λ=0.119 6 ）， followed by endotoxin （0.100 8） and CRP（0.0557）.Decisioncurveanalysisshowedthat withinathresholdprobabilityrangeof02-0.8，thenetbenefitoftheExtra Tres modelwassignificantlyhigherthanthatofthe“treat-all’or“treat-none”strategy.Thecalibrationcurvedemonstrated good agreement （Brier Score =0.098 ）.Conclusions Amachine learning-based multi-biomarker models can accurately predict the risk ofintestinal ncrosis insurgical ASBO patients， withthe Extra Trees model showing the best performance.HCYis the most importantpredictor，providinganobjectivebasisforpreoperativeclinicalriskassessment.Futuredevelopmentofacomprehesive prediction model applicable to conservatively treated ASBO patients is needed.

【 Key words 】Machine learning; Homocysteine; Adhesive smallbowel obstruction; Intestinal necrosis; Serum biomarkers;Extra trees;SHAP analysis

粘连性小肠梗阻（adhesivesmallbowelobstruction，ASBO）是主要外科急症之一，术后粘连占小肠梗阻的 65% [1]。（剩余16883字）