Development of End-to-End Artificial Intelligence Models for Surgical Planning in Transforaminal Lumbar Interbody Fusion

Abstract

Transforaminal lumbar interbody fusion (TLIF) is a commonly used technique for treating lumbar degenerative diseases. In this study, we developed a fully computer-supported pipeline to predict both the cage height and the degree of lumbar lordosis subtraction from the pelvic incidence (PI-LL) after TLIF surgery, utilizing preoperative X-ray images. The automated pipeline comprised two primary stages. First, the pretrained BiLuNet deep learning model was employed to extract essential features from X-ray images. Subsequently, five machine learning algorithms were trained using a five-fold cross-validation technique on a dataset of 311 patients to identify the optimal models to predict interbody cage height and postoperative PI-LL. LASSO regression and support vector regression demonstrated superior performance in predicting interbody cage height and postoperative PI-LL, respectively. For cage height prediction, the root mean square error (RMSE) was calculated as 1.01, and the model achieved the highest accuracy at a height of 12 mm, with exact prediction achieved in 54.43% (43/79) of cases. In most of the remaining cases, the prediction error of the model was within 1 mm. Additionally, the model demonstrated satisfactory performance in predicting PI-LL, with an RMSE of 5.19 and an accuracy of 0.81 for PI-LL stratification. In conclusion, our results indicate that machine learning models can reliably predict interbody cage height and postoperative PI-LL.

Keywords: artificial intelligence; interbody cage; machine learning; sagittal balance; spinal fusion; spinal parameters.

Figure 1

Study flowchart depicting four subprocesses: data cohort collection, feature extraction, feature validation, and ML model construction and validation. ML: machine learning; SVR: support vector regression; LR: LASSO regression; DT: decision tree; KNN: K-nearest neighbor; MLP: multilayer perceptron; RFE: recursive feature elimination; RMSE: root mean square error; MAE: mean absolute error.

Figure 2

Distribution of actual interbody cage heights and postoperative PI-LL values.

Figure 3

RFECV curves of two baseline models with negative MAEs for different numbers of features: (A) an LR model for interbody cage height prediction and (B) an SVR model for postoperative PI-LL prediction.

Figure 4

Confusion matrix for final model performance in the prediction of interbody cage height.

Figure 5

Performance of SVR. (A) Calibration plot (actual and predicted values) for predicting postoperative PI-LL on both training and testing data. (B) Confusion matrix for stratifying postoperative PI-LL on the testing set into three groups: 0 (<10), 1 (10–20), and 2 (>20).

Figure 6

(A) Most crucial features for the model of interbody cage height prediction. (B) Most crucial features for the model of postoperative PI-LL prediction. Note: The explanations of feature abbreviations are provided in Supplementary Table S1.