Buwaider Ali, El-Hajj Victor Gabriel, MacDowall Anna, Gerdhem Paul, Staartjes Victor E, Edström Erik, Elmi-Terander Adrian
Department of Clinical Neuroscience, Karolinska Institutet, Stockholm, Sweden.
Department of Surgical Sciences, Uppsala University, Uppsala, Sweden; Department of Orthopaedics and Hand surgery, Uppsala University Hospital, Uppsala, Sweden.
Spine J. 2025 Mar;25(3):419-428. doi: 10.1016/j.spinee.2024.10.010. Epub 2024 Nov 4.
Dysphonia is one of the more common complications following anterior cervical discectomy and fusion (ACDF). ACDF is the gold standard for treating degenerative cervical spine disorders, and identifying high-risk patients is therefore crucial.
This study aimed to evaluate different machine learning models to predict persistent dysphonia after ACDF.
A retrospective review of the nationwide Swedish spine registry (Swespine).
All adults in the Swespine registry who underwent elective ACDF between 2006 and 2020.
The primary outcome was self-reported dysphonia lasting at least 1 month after surgery. Predictive performance was assessed using discrimination and calibration metrics.
Patients with missing dysphonia data at the 1-year follow-up were excluded. Data preprocessing involved one-hot encoding categorical variables, scaling continuous variables, and imputing missing values. Four machine learning models (logistic regression, random forest (RF), gradient boosting, K-nearest neighbor) were employed. The models were trained and tested using an 80:20 data split and 5-fold cross-validation, with performance metrics guiding the selection of the best model for predicting persistent dysphonia.
In total, 2,708 were included in the study. Twelve key predictors were identified. Four machine learning models were tested, with the RF model achieving the best performance (AUC=0.794). The most significant predictors across models included preoperative NDI, EQ5D, preoperative neurology, number of operated levels, and use of a fusion cage. The RF model, chosen for its superior performance, showed high sensitivity and consistent accuracy, but a low specificity and positive predictive value.
In this study, machine learning models were employed to identify predictors of persistent dysphonia following ACDF. Among the models tested, the RF classifier demonstrated superior performance, with an AUC value of 0.790. The RF model identified NDI, EQ5D, and number of fused vertebrae as key variables. These findings underscore the potential of machine learning models in identifying patients at increased risk for dysphonia persisting for more than 1 month after surgery.
发音障碍是颈椎前路椎间盘切除融合术（ACDF）后较常见的并发症之一。ACDF是治疗退行性颈椎疾病的金标准，因此识别高危患者至关重要。
本研究旨在评估不同的机器学习模型，以预测ACDF术后持续性发音障碍。
对瑞典全国脊柱登记处（Swespine）进行回顾性研究。
Swespine登记处中2006年至2020年间接受择期ACDF的所有成年人。
主要结局是术后自我报告的发音障碍持续至少1个月。使用区分度和校准指标评估预测性能。
排除在1年随访时发音障碍数据缺失的患者。数据预处理包括对分类变量进行独热编码、对连续变量进行缩放以及插补缺失值。采用了四种机器学习模型（逻辑回归、随机森林（RF）、梯度提升、K近邻）。使用80:20的数据分割和5折交叉验证对模型进行训练和测试，性能指标指导选择预测持续性发音障碍的最佳模型。
本研究共纳入2708例患者。确定了12个关键预测因素。测试了四种机器学习模型，其中RF模型表现最佳（AUC = 0.794）。各模型中最显著的预测因素包括术前NDI、EQ5D、术前神经学情况、手术节段数量以及融合器的使用。因性能优越而被选中的RF模型显示出高敏感性和一致的准确性，但特异性和阳性预测值较低。
在本研究中，使用机器学习模型来识别ACDF术后持续性发音障碍的预测因素。在所测试的模型中，RF分类器表现出优越性能，AUC值为0.790。RF模型将NDI、EQ5D和融合椎体数量确定为关键变量。这些发现强调了机器学习模型在识别术后发音障碍持续超过1个月风险增加的患者方面的潜力。
