Abstract

Purpose

The aim of this study was to propose and validate a new six-item topographical classification of knee bone marrow lesions in coronal MRI images, to provide an easy-to-use aid to describe their location in a more reproducible and accurate way.

Methods

This study was conducted in four phases. The first was to do a literature search for methods of describing bone marrow lesions in MRI of the knee. The second was creation of a six-area topographic classification of bone marrow lesions in coronal MRI of the knee. The third phase was selection of cases with bone marrow lesions on knee MRI performed in a single hospital between January of 2017 and December of 2018. The fourth phase was categorization of the bone marrow lesions’ location according to the new proposed classification by three independent examiners, two orthopedic surgeons and one radiologist. Patient’s demographic data and associated lesions were collected. The inter-observer and intra-observer reliability of the proposed classification was then calculated.

Results

MRI examination of 4000 patients were studied and in 520 patients a total of 666 bone marrow lesions were identified and their location classified using the new system. The inter-observer and intra-observer reliability analysis found a Fleiss’ Kappa value of 0.96 (0.95–0.97) and 0.97 (0.96–0.97), respectively, confirming the high reproducibility of the proposed classification.

Conclusions

The proposed six-location classification of bone marrow lesions is highly reproducible and can help researchers develop studies and share information in a more accurate and reliable way. The correct classification of bone marrow lesions can lead to a more accurate description of this pathology and help clinicians to propose appropriate therapies for this group of patients.

Level of evidence

V.

中文翻译：

---

验证膝关节骨髓病变的新地形分类：六字母系统

摘要

目的

这项研究的目的是提出并验证冠状MRI图像中膝部骨髓病变的新的六项地形学分类，以提供易于使用的辅助手段，以更可重复和准确的方式描述其位置。

方法

这项研究分四个阶段进行。首先是进行文献搜索，以寻找描述膝盖MRI中骨髓病变的方法。第二个是在膝部冠状MRI中创建了六个区域的骨髓病变的地形分类。第三阶段是在2017年1月至2018年12月期间在一家医院中通过膝部MRI进行骨髓病变病例的选择。第四阶段是由三位独立检查员根据新提议的分类对骨髓病变的位置进行分类，两名骨科医生和一名放射科医生。收集患者的人口统计学数据和相关病变。然后计算了建议分类的观察者间和观察者内可靠性。

结果

对4000例患者的MRI检查进行了研究，在520例患者中，共识别出666个骨髓病变，并使用新系统对它们的位置进行了分类。观察者之间和观察者内部的可靠性分析发现，弗莱斯的Kappa值分别为0.96（0.95-0.97）和0.97（0.96-0.97），证实了拟议分类的高度可重复性。

结论

拟议的骨髓损伤的六位分类具有高度可重复性，可以帮助研究人员以更准确和可靠的方式开展研究并共享信息。骨髓病变的正确分类可以导致这种病理的更准确描述，并帮助临床医生为这一组患者提出适当的治疗方法。

证据水平

V.