Objective

To develop a machine learning-based prediction model for incident radiographic osteoarthritis (OA) of the knee over 8 years using MRI-based cartilage biochemical composition and knee joint structure, demographics, and clinical predictors including muscle strength and symptoms.

Design

Individuals (n = 1,044) with baseline Kellgren Lawrence (KL) grade 0–1 in the right knee from the Osteoarthritis Initiative database were analyzed. 3T MRI at baseline was used to quantify knee cartilage T₂, and Whole-Organ Magnetic Resonance Imaging Scores (WORMS) were obtained for cartilage, meniscus, and bone marrow. The outcome was set as true if a subject developed KL grade 2–4 OA in the right knee over 8 years (n = 183) and false if the subject remained at KL 0–1 over 8 years (n = 861). We developed and compared three models: Model 1: 112 predictors based on OA risk factors; Model 2: top ten predictors based on feature importance score from Model 1 and clinical relevance; Model 3: Model 2 without the imaging predictors. We compared the models using the area under the ROC curve derived from hold-out data.

Results

The 10-predictor model (Model 2, that includes cartilage and meniscus WORMS scores and cartilage T₂) had a slightly lower AUC (0.772) compared to the model with 112 predictors (Model 1: AUC = 0.792, p = 0.739); and had a significantly higher AUC compared to the model without MR imaging predictors (Model 3, AUC = 0.669, p = 0.011).

Conclusions

A 10-predictor model including MRI parameters coupled with demographics, symptoms, muscle, and physical activity scores provides good prediction of incident radiographic OA over 8 years.

中文翻译：

---

机器学习使用组合 MR 成像特征、人口统计和临床因素预测 8 年内发生的放射性膝关节骨关节炎：来自骨关节炎倡议的数据

客观的

使用基于 MRI 的软骨生化成分和膝关节结构、人口统计学和包括肌肉力量和症状在内的临床预测指标，开发基于机器学习的膝关节影像学骨关节炎 (OA) 预测模型超过 8 年。

设计

 分析了来自骨关节炎倡议数据库中右膝基线 Kellgren Lawrence (KL) 0-1 级的个体 ( n = 1,044)。基线时的 3T MRI 用于量化膝关节软骨 T ₂，并获得软骨、半月板和骨髓的全器官磁共振成像评分 (WORMS)。如果受试者在 8 年内右膝出现 KL 2-4 级 OA，则结果为真（n  = 183），如果受试者在 8 年内保持 KL 0-1 级（n  = 861），则结果为假。我们开发并比较了三个模型：模型 1：基于 OA 风险因素的 112 个预测变量；模型 2：基于模型 1 的特征重要性得分和临床相关性的前十个预测因子；模型 3：没有成像预测因子的模型 2。我们使用来自保留数据的 ROC 曲线下面积来比较模型。

结果

与具有 112 个预测因子的模型（模型 1：AUC = 0.792，p  = 0.739）相比， 10 个预测因子模型（模型 2，包括软骨和半月板 WORMS 评分以及软骨 T ₂）的 AUC（0.772）略低；并且与没有 MR 成像预测因子的模型相比，AUC 显着更高（模型 3， AUC = 0.669，p  = 0.011）。

结论

包括 MRI 参数以及人口统计、症状、肌肉和身体活动评分的 10 预测模型可以很好地预测 8 年内发生的放射学 OA。