A Machine Learning Model Based on PET/CT Radiomics and Clinical Characteristics Predicts ALK Rearrangement Status in Lung Adenocarcinoma

Abstract

Objectives: Anaplastic lymphoma kinase (ALK) rearrangement status examination has been widely used in clinic for non-small cell lung cancer (NSCLC) patients in order to find patients that can be treated with targeted ALK inhibitors. This study intended to non-invasively predict the ALK rearrangement status in lung adenocarcinomas by developing a machine learning model that combines PET/CT radiomic features and clinical characteristics.

Methods: Five hundred twenty-six patients of lung adenocarcinoma with PET/CT scan examination were enrolled, including 109 positive and 417 negative patients for ALK rearrangements from February 2016 to March 2019. The Artificial Intelligence Kit software was used to extract radiomic features of PET/CT images. The maximum relevance minimum redundancy (mRMR) and least absolute shrinkage and selection operator (LASSO) logistic regression were further employed to select the most distinguishable radiomic features to construct predictive models. The mRMR is a feature selection method, which selects the features with high correlation to the pathological results (maximum correlation), meanwhile retain the features with minimum correlation between them (minimum redundancy). LASSO is a statistical formula whose main purpose is the feature selection and regularization of data model. LASSO method regularizes model parameters by shrinking the regression coefficients, reducing some of them to zero. The feature selection phase occurs after the shrinkage, where every non-zero value is selected to be used in the model. Receiver operating characteristic (ROC) analysis was used to evaluate the performance of the models, and the performance of different models was compared by the DeLong test.

Results: A total of 22 radiomic features were extracted from PET/CT images for constructing the PET/CT radiomic model, and majority of these features used were based on CT features (20 out of 22), only 2 PET features were included (PET percentile 10 and PET difference entropy). Moreover, three clinical features associated with ALK mutation (age, burr and pleural effusion) were also employed to construct a combined model of PET/CT and clinical model. We found that this combined model PET/CT-clinical model has a significant advantage to predict the ALK mutation status in the training group (AUC = 0.87) and the testing group (AUC = 0.88) compared with the clinical model alone in the training group (AUC = 0.76) and the testing group (AUC = 0.74) respectively. However, there is no significant difference between the combined model and PET/CT radiomic model.

Conclusions: This study demonstrated that PET/CT radiomics-based machine learning model has potential to be used as a non-invasive diagnostic method to help diagnose ALK mutation status for lung adenocarcinoma patients in the clinic.

Keywords: anaplastic lymphoma kinase (ALK) rearrangement; lung adenocarcinoma; machine learning; positron emission tomography/computed tomography (PET/CT); radiomics.

Figure 1

Flowchart of lung adenocarcinoma patient selection.

Figure 2

The workflow of radiomic analysis. Feature extraction: AK software (402), 402 means the total number of extracted features from AK software. ROI, region of interest; GLCM, gray level co-occurrence matrix; GLSZM, grey level size zone matrix; RLM, run length matrix; mRMR, minimum redundancy and maximum correlation; LASSO, least absolute shrinkage and selection operator; ROC, receiver operating characteristic.

Figure 3

Construction of a PET/CT radiomic model based on PET/CT images. (A) the Selection of the tuning parameter (λ) in the LASSO model via 10-fold cross-validation based on minimum criteria. Binomial deviances of the LASSO regression cross-validation model are plotted as a function of ln (λ). The y-axis shows binomial deviances and the lower x-axis the ln (λ). Feature numbers along the upper x-axis indicate the number of features via the change of λ. (B) The final retained features selected by mRMR, y axis was the retained features and x axis shows the corresponding LASSO regression coefficients of them. The fitted coefficients of the features plotted vs. ln (λ). (C) Representative results of PET/CT radiomic model for predicting ALK rearrangement in training (left) and testing (right) group of lung adenocarcinoma patients. 0, negative ALK rearrangement; 1, positive ALK rearrangement. (D) Cross-validation analysis showed that PET/CT radiomic model has good reliability to predict ALK rearrangement in training (left) and testing (right) group of lung adenocarcinoma patients.

Figure 4

ROC curve analysis of three radiomics models, PET/CT, CT, and PET in training group (A) and testing group (B), respectively.

Figure 5

Evaluates the performances of integrated PET/CT radiomics-clinical model. (A) Receiver operating characteristic (ROC) curves of predictive performances of different methods in the training cohort (left) and test cohort (right). The curves of 3 colors represent different models: red, PET/CT radiomics + clinical model; blue, PET/CT radiomics model; green, clinical model. AUC, area under the curve. (B) Nomogram for ALK mutation prediction. The nomogram was developed by integrating radiomic score with 3 significant clinical features (age, burr and pleural effusion). The probability of each predictor can be converted into the “points” scale at the top of the nomogram. By sum up the points for each predictor and locate in the “Total points” scale, we can predict the probability of ALK mutation in the “Risk” scale. (C) Calibration curve with Hosmer-Lemeshow test of the nomogram in the training cohort (left panel) and test cohort (right panel). Calibration curve shows the calibration of the model in terms of consistence between predicated risk of ALK mutation and real observed ALK mutation status. The x-axis represents the predicted risk of ALK mutation and y-axis represents the real ALK mutation status. (D) Decision curve analysis for the nomograms. The y-axis measures the standardized net benefit. The net benefit is calculated by adding up the true positive results and subtracting the false positive results, weighting the latter by a factor relevant to the relative harm of an undetected cancer compared with the harm of unnecessary treatment. The red line represents the PET/CT radiomics and clinical features model, the green line represents the PET/CT clinical features model, the gray line represents the assumption than all patients are negative for ALK mutation and the blue line represents the assumption that all patients are positive for ALK mutation.