Machine learning-based CT radiomics features for the prediction of pulmonary metastasis in osteosarcoma

Abstract

Objective: This study aims to build machine learning-based CT radiomic features to predict patients developing metastasis after osteosarcoma diagnosis.

Methods and materials: This retrospective study has included 81 patients with a histopathological diagnosis of osteosarcoma. The entire dataset was divided randomly into training (60%) and test sets (40%). A data augmentation technique for the minority class was performed in the training set, along with feature's selection and model's training. The radiomic features were extracted from CT's image of the local osteosarcoma. Three frequently used machine learning models tried to predict patients with lung metastases (MT) and those without lung metastases (non-MT). According to the higher area under the curve (AUC), the best classifier was chosen and applied in the testing set with unseen data to provide an unbiased evaluation of the final model.

Results: The best classifier for predicting MT and non-MT groups used a Random Forest algorithm. The AUC and accuracy results of the test set were bulky (accuracy of 73% [ 95% coefficient interval (CI): 54%; 87%] and AUC of 0.79 [95% CI: 0.62; 0.96]). Features that fitted the model (radiomics signature) derived from Laplacian of Gaussian and wavelet filters.

Conclusions: Machine learning-based CT radiomics approach can provide a non-invasive method with a fair predictive accuracy of the risk of developing pulmonary metastasis in osteosarcoma patients.

Advances in knowledge: Models based on CT radiomic analysis help assess the risk of developing pulmonary metastases in patients with osteosarcoma, allowing further studies for those with a worse prognosis.

Figure 1.

Twenty-two years old, a young male with axial NECT image segmentation of a osteosarcoma of distal femur from the most cranial to caudal images (A,B,C) and the segment volume provided (D).

Figure 2.

Features selections steps used to build a radiomic signature and the best classifier algorithm to test the radiomics model.

Figure 3.

Flowchart for the selection of eligible patients.

Figure 4.

Kaplan–Meier curves for 5 years survival (n = 62) of patients without pulmonary metastases (non-MT) and with pulmonary metastases (MT) at presentation and pulmonary metastases arising during follow-up. In the Kaplan–Meier analysis, significantly shorter survival was observed in the MT group, mainly those with pulmonary metastases at presentation (log-rank p < 0.0001).

Figure 5.

Boxplots show difference between patients without lung metastasis (0) and with lung metastasis in radiomics signature based on NECT images.