Sensitivity analysis of FDG PET tumor voxel cluster radiomics and dosimetry for predicting mid-chemoradiation regional response of locally advanced lung cancer

Abstract

We investigated the sensitivity of regional tumor response prediction to variability in voxel clustering techniques, imaging features, and machine learning algorithms in 25 patients with locally advanced non-small cell lung cancer (LA-NSCLC) enrolled on the FLARE-RT clinical trial. Metabolic tumor volumes (MTV) from pre-chemoradiation (PETpre) and mid-chemoradiation fluorodeoxyglucose-positron emission tomography (FDG PET) images (PETmid) were subdivided into K-means or hierarchical voxel clusters by standardized uptake values (SUV) and 3D-positions. MTV cluster separability was evaluated by CH index, and morphologic changes were captured by Dice similarity and centroid Euclidean distance. PETpre conventional features included SUVmean, MTV/MTV cluster size, and mean radiation dose. PETpre radiomics consisted of 41 intensity histogram and 3D texture features (PET Oncology Radiomics Test Suite) extracted from MTV or MTV clusters. Machine learning models (multiple linear regression, support vector regression, logistic regression, support vector machines) of conventional features or radiomic features were constructed to predict PETmid response. Leave-one-out-cross-validated root-mean-squared-error (RMSE) for continuous response regression (ΔSUVmean) and area-under-receiver-operating-characteristic-curve (AUC) for binary response classification were calculated. K-means MTV 2-clusters (MTVhi, MTVlo) achieved maximum CH index separability (Friedman p < 0.001). Between PETpre and PETmid, MTV cluster pairs overlapped (Dice 0.70-0.87) and migrated 0.6-1.1 cm. PETmid ΔSUVmean response prediction was superior in MTV and MTVlo (RMSE = 0.17-0.21) compared to MTVhi (RMSE = 0.42-0.52, Friedman p < 0.001). PETmid ΔSUVmean response class prediction performance trended higher in MTVlo (AUC = 0.83-0.88) compared to MTVhi (AUC = 0.44-0.58, Friedman p = 0.052). Models were more sensitive to MTV/MTV cluster regions (Friedman p = 0.026) than feature sets/algorithms (Wilcoxon signed-rank p = 0.36). Top-ranked radiomic features included GLZSM-LZHGE (large-zone-high-SUV), GTSDM-CP (cluster-prominence), GTSDM-CS (cluster-shade) and NGTDM-CNT (contrast). Top-ranked features were consistent between MTVhi and MTVlo cluster pairs but varied between MTVhi-MTVlo clusters, reflecting distinct regional radiomic phenotypes. Variability in tumor voxel cluster response prediction can inform robust radiomic target definition for risk-adaptive chemoradiation in patients with LA-NSCLC. FLARE-RT trial: NCT02773238.

Figures

Figure 1.. Prediction modeling sensitivity study network.

Regions of interest define the primary metabolic tumor volume, 2 intra-tumor voxel clusters, or 2* intra-tumor voxel clusters. Within each ROI, conventional PET SUV + RT features or PET radiomic features are extracted. Linear and support vector machine regression models were built under leave-1-out cross-validation to predict Week 3 mid-treatment change in SUVmean as a continuous variable. Logistic and support vector machine classifiers are built under leave-1-out cross-validation to predict Week 3 SUVmean response class as a dichotomized variable.

Figure 2.

Tumor voxel spatial clustering on baseline FDG PET imaging (PET preTx) and 3-week response (PET midTx) to define 2 clusters during chemoradiotherapy in two example patients (A, B) with locally advanced non-small cell lung cancer (LA-NSCLC).

Figure 3.

Tumor voxel spatial clustering on baseline FDG PET imaging (PET preTx) and 3-week response (PET midTx) to define 3 clusters during chemoradiotherapy in two example patients (A, B) with locally advanced non-small cell lung cancer (LA-NSCLC).

Figure 4.

Tumor voxel spatial clustering on baseline FDG PET imaging (PET preTx) and 3-week response (PET midTx) to define 2* clusters during chemoradiotherapy in two example patients (A, B) with locally advanced non-small cell lung cancer (LA-NSCLC).