Rischmüller Karen, Caton Vanessa, Wolfien Markus, Ehlers Luise, van Welzen Matti, Brauer David, Sautter Lea F, Meyer Fatuma, Valentini Luzia, Wiese Mats L, Aghdassi Ali A, Jaster Robert, Wolkenhauer Olaf, Lamprecht Georg, Bej Saptarshi
Division of Gastroenterology and Endocrinology, Department of Internal Medicine II, Rostock University Medical Center, Rostock, Germany.
Department of Systems Biology and Bioinformatics, Institute of Computer Science, University of Rostock, Rostock, Germany.
Front Nutr. 2024 Dec 12;11:1479501. doi: 10.3389/fnut.2024.1479501. eCollection 2024.
Disease-related malnutrition is common but often underdiagnosed in patients with chronic gastrointestinal diseases, such as liver cirrhosis, short bowel and intestinal insufficiency, and chronic pancreatitis. To improve malnutrition diagnosis in these patients, an evaluation of the current Global Leadership Initiative on Malnutrition (GLIM) diagnostic criteria, and possibly the implementation of additional criteria, is needed.
This study aimed to identify previously unknown and potentially specific features of malnutrition in patients with different chronic gastrointestinal diseases and to validate the relevance of the GLIM criteria for clinical practice using machine learning (ML).
Between 10/2018 and 09/2021, = 314 patients and controls were prospectively enrolled in a cross-sectional study. A total of = 230 features (anthropometric data, body composition, handgrip strength, gait speed, laboratory values, dietary habits, physical activity, mental health) were recorded. After data preprocessing (cleaning, feature exploration, imputation of missing data), = 135 features were included in the ML analyses. Supervised ML models were used to classify malnutrition, and key features were identified using SHapley Additive exPlanations (SHAP).
Supervised ML effectively classified malnourished versus non-malnourished patients and controls. Excluding the existing GLIM criteria and malnutrition risk reduced model performance (sensitivity -19%, specificity -8%, F1-score -10%), highlighting their significance. Besides some GLIM criteria (weight loss, reduced food intake, disease/inflammation), additional anthropometric (hip and upper arm circumference), body composition (phase angle, SMMI), and laboratory markers (albumin, pseudocholinesterase, prealbumin) were key features for malnutrition classification.
ML analysis confirmed the clinical applicability of the current GLIM criteria and identified additional features that may improve malnutrition diagnosis and understanding of the pathophysiology of malnutrition in chronic gastrointestinal diseases.
疾病相关的营养不良很常见，但在患有慢性胃肠疾病的患者中常常未被诊断出来，这些疾病包括肝硬化、短肠综合征和肠道功能不全以及慢性胰腺炎。为了改善这些患者的营养不良诊断，需要对当前的全球营养不良领导倡议（GLIM）诊断标准进行评估，并可能实施额外的标准。
本研究旨在识别不同慢性胃肠疾病患者中先前未知且可能具有特异性的营养不良特征，并使用机器学习（ML）验证GLIM标准在临床实践中的相关性。
在2018年10月至2021年9月期间，前瞻性地招募了314名患者和对照参与一项横断面研究。总共记录了230个特征（人体测量数据、身体成分、握力、步速、实验室值、饮食习惯、身体活动、心理健康）。经过数据预处理（清理、特征探索、缺失数据插补）后，135个特征被纳入ML分析。使用监督式ML模型对营养不良进行分类，并使用夏普利值附加解释（SHAP）识别关键特征。
监督式ML有效地将营养不良患者与非营养不良患者及对照进行了分类。排除现有的GLIM标准和营养不良风险会降低模型性能（敏感性降低19%，特异性降低8%，F1分数降低10%），突出了它们的重要性。除了一些GLIM标准（体重减轻、食物摄入量减少、疾病/炎症）外，其他人体测量指标（臀围和上臂围）、身体成分（相位角、体细胞质量指数）和实验室标志物（白蛋白、假性胆碱酯酶、前白蛋白）是营养不良分类的关键特征。
ML分析证实了当前GLIM标准的临床适用性，并识别出了可能改善慢性胃肠疾病中营养不良诊断及对营养不良病理生理学理解的其他特征。
