Radiomics analysis of pulmonary nodules in low-dose CT for early detection of lung cancer

Abstract

Purpose: To develop a radiomics prediction model to improve pulmonary nodule (PN) classification in low-dose CT. To compare the model with the American College of Radiology (ACR) Lung CT Screening Reporting and Data System (Lung-RADS) for early detection of lung cancer.

Methods: We examined a set of 72 PNs (31 benign and 41 malignant) from the Lung Image Database Consortium image collection (LIDC-IDRI). One hundred three CT radiomic features were extracted from each PN. Before the model building process, distinctive features were identified using a hierarchical clustering method. We then constructed a prediction model by using a support vector machine (SVM) classifier coupled with a least absolute shrinkage and selection operator (LASSO). A tenfold cross-validation (CV) was repeated ten times (10 × 10-fold CV) to evaluate the accuracy of the SVM-LASSO model. Finally, the best model from the 10 × 10-fold CV was further evaluated using 20 × 5- and 50 × 2-fold CVs.

Results: The best SVM-LASSO model consisted of only two features: the bounding box anterior-posterior dimension (BB_AP) and the standard deviation of inverse difference moment (SD_IDM). The BB_AP measured the extension of a PN in the anterior-posterior direction and was highly correlated (r = 0.94) with the PN size. The SD_IDM was a texture feature that measured the directional variation of the local homogeneity feature IDM. Univariate analysis showed that both features were statistically significant and discriminative (P = 0.00013 and 0.000038, respectively). PNs with larger BB_AP or smaller SD_IDM were more likely malignant. The 10 × 10-fold CV of the best SVM model using the two features achieved an accuracy of 84.6% and 0.89 AUC. By comparison, Lung-RADS achieved an accuracy of 72.2% and 0.77 AUC using four features (size, type, calcification, and spiculation). The prediction improvement of SVM-LASSO comparing to Lung-RADS was statistically significant (McNemar's test P = 0.026). Lung-RADS misclassified 19 cases because it was mainly based on PN size, whereas the SVM-LASSO model correctly classified 10 of these cases by combining a size (BB_AP) feature and a texture (SD_IDM) feature. The performance of the SVM-LASSO model was stable when leaving more patients out with five- and twofold CVs (accuracy 84.1% and 81.6%, respectively).

Conclusion: We developed an SVM-LASSO model to predict malignancy of PNs with two CT radiomic features. We demonstrated that the model achieved an accuracy of 84.6%, which was 12.4% higher than Lung-RADS.

Keywords: CT; SVM; lung cancer; pulmonary nodule; radiomics.

Figure 1

Distributions of pulmonary nodule size, type, calcification, and spiculation in the dataset. [Color figure can be viewed at wileyonlinelibrary.com]

Figure 2

A flowchart of the extraction of radiomic features and the construction of a prediction model. [Color figure can be viewed at wileyonlinelibrary.com]

Figure 3

Performance of the prediction model with increasing number of features in the CV. [Color figure can be viewed at wileyonlinelibrary.com]

Figure 4

The box plots show the difference between benign and malignant for PN size and the selected features (BB_AP and SD_IDM). The Wilcoxon rank‐sum test obtained P‐values. [Color figure can be viewed at wileyonlinelibrary.com]

Figure 5

ROC curve analysis on the best model of SVM‐LASSO and Lung‐RADS for predicting malignant PNs. [Color figure can be viewed at wileyonlinelibrary.com]

Figure 6

Scatter plot of the two important features and the classification curve (dashed line) by the SVM‐LASSO model for all PNs. [Color figure can be viewed at wileyonlinelibrary.com]

Figure 7

Cases misclassified by Lung‐RADS but correctly classified by the SVM‐LASSO model (a–d). A case correctly classified by Lung‐RADS but misclassified by the SVM‐LASSO model (e). The scale bar indicates 10 mm, window/level: 1400/−500 HU. The value in the parenthesis for size is the diameter of the solid component of a part‐solid PN. Spiculation is on a 1(no) to 5(marked) scale. [Color figure can be viewed at wileyonlinelibrary.com]

Figure 8

Lung‐RADS categorization on scatter plots for solid PNs (a) and part‐solid PNs (b). The SVM‐LASSO classification curve is approximately mapped on the plots (green dashed line). Lung‐RADS categorization is shown on top with black vertical dashed lines (the bold line indicates classification between benign and suspicious); PNs with calcification (category 1) are filled with dark gray, and PNs with spiculation (category 4X) are filled with light gray (red color online version). [Color figure can be viewed at wileyonlinelibrary.com]

Figure 9

IDM and SD_IDM for small benign and malignant PNs (BB_AP = 10 mm for both). The length of each arrow indicates IDM value for left–right (LR), anterior–posterior (AP), and superior–inferior (SI) directions. [Color figure can be viewed at wileyonlinelibrary.com]

Figure 10

IDM and SD_IDM for large benign and malignant PNs (BB_AP = 17 mm for both). The length of each arrow indicates IDM value for the three directions. [Color figure can be viewed at wileyonlinelibrary.com]