Accuracy of magnetic resonance imaging for diagnosing hallux valgus

Abstract Objective: To assess the accuracy of magnetic resonance imaging (MRI) for the diagnosis of hallux valgus using radiography during weight bearing as the gold standard. Materials and Methods: This was a retrospective analysis of all patients undergoing MRI of the foot and radiography of the foot during weight bearing at our institution between January and June of 2015. The hallux valgus angle (HVA) was measured on MRI and radiography. The Wilcoxon signed-rank test and simple linear regression were used in order to compare measurements. Patients were divided into two groups according to the HVA determined on radiography: > 15° (hallux valgus) and ≤ 15° (control). Qualitative and quantitative assessments of MRI scans were performed. For quantitative assessment, receiver operating characteristic curves were used in order to determine the HVA cutoff with the highest accuracy. Results: A total of 66 MRI scans were included, 22 in the hallux valgus group and 44 in the control group. Wilcoxon signed-rank tests indicated a significant difference between the radiography and MRI measurements. Simple linear regression showed a nonlinear relationship between the measurements and values did not present a strong correlation. In comparison with the radiography measurements, MRI with an HVA cutoff of 16.4° exhibited the highest accuracy (86%). The accuracy of the subjective (qualitative) assessment was inferior to the objective assessment (measurement of the HVA). Conclusion: Hallux valgus can be diagnosed by measuring the HVA on MRI, satisfactory accuracy being achieved with an HVA cutoff of 16.4°.

摘要
目的：以负重位足部X线摄影（radiography）作为金标准，评估磁共振成像（magnetic resonance imaging, MRI）用于拇外翻（hallux valgus）诊断的准确性。
材料与方法：本研究为回顾性分析，纳入2015年1月至6月于本机构接受足部MRI检查及负重位足部X线摄影的所有患者。分别于MRI及X线片上测量拇外翻角（hallux valgus angle, HVA）。采用威尔科克森符号秩检验（Wilcoxon signed-rank test）及简单线性回归（simple linear regression）对测量结果进行比较。根据X线片测得的HVA将患者分为两组：HVA＞15°者为拇外翻组，HVA≤15°者为对照组。对MRI影像进行定性及定量评估；定量评估方面，采用受试者工作特征曲线（receiver operating characteristic curves, ROC曲线）确定诊断准确性最高的HVA截断值。
结果：最终共纳入66例患者的MRI影像资料，其中拇外翻组22例，对照组44例。威尔科克森符号秩检验结果显示，X线与MRI的测量结果存在显著差异。简单线性回归分析显示，两类测量结果间呈非线性关系，未呈现较强相关性。以X线测量结果为金标准，当MRI的HVA截断值设为16.4°时，诊断准确性最高，达86%。主观（定性）评估的诊断准确性劣于客观（HVA测量）评估。
结论：通过测量MRI上的HVA可诊断拇外翻，当HVA截断值设为16.4°时，可获得令人满意的诊断准确性。