Comparison of PET and CT radiomics for prediction of local tumor control in head and neck squamous cell carcinoma

Abstract

(2017). Material and methodsData from HNSCC patients (n=121) treated with definitive radiochemotherapy were used for model training. In total, 569 radiomic features were extracted from both contrast enhanced CT and 18F-FDG PET images in the primary tumor region.CT, PET and combined PET/CT radiomic models to assess local tumor control were trained separately. Five feature selection and three classification methods were implemented. The performance of the models was quantified using concordance index (CI) in 5-fold cross-validation in the training cohort. The best models, per image modality, were compared and verified in the independent validation cohort (n=51). The difference in CI was investigated using bootstrapping. Additionally, the observed and radiomics-based estimated probabilities of local tumor control were compared between two risk groups. ResultsThe feature selection using principal component analysis and the classification based on the multivariabale Cox regression with backward selection of the variables resulted in the best models for all image modalities (CICT=0.72, CIPET=0.74, CIPET/CT=0.77). Tumors more homogenous in CT density (decreased GLSZMsize_zone_entropy) and with a focused region of high FDG uptake (higher GLSZMSZLGE) indicated better prognosis. No significant difference in the performance of the models in the validation cohort was observed (CICT=0.73, CIPET=0.71, CIPET/CT=0.73). However, the CT radiomics based model overestimated the probability of tumor control in the poor prognostic group (predicted=68%, observed=56%). ConclusionBoth CT and PET radiomics showed equally good discriminative power for local tumor control modeling in HNSCC. However, CT-based predictions overestimated the local control rate in the poor prognostic validation cohort and thus we recommend to base the local control modeling on the 18F-FDG PET.

Influence of inter-observer delineation variability on radiomics stability in different tumor sites

“…In total, 1404 features were extracted per patient/image (see Supplementary Tables 1-4). The definitions of analyzed intensity and texture features are described by Zwanenburg et al [16], whereas the details on implementation of the shape features and the wavelet transform can be found in work by Bogowicz et al [17,18]. For intensity, texture and wavelet analysis images were resized to cubic voxels of 3 mm for NSCLC, 2 mm for HNSCC and 3.3 mm for MPM using linear interpolation.…”

Personalizing Medicine Through Hybrid Imaging and Medical Big Data Analysis

“…To date, in vivo disease characterization with hybrid imaging data-especially in the light of oncological applications-is performed mainly by analyzing engineered features [63,129]. This process is widely referred to as "Radiomics, " even though, this kind of approach was originally applied to morphological images only [8].…”

“…There is increasing evidence for in vivo tissue characterization with both PET/CT and PET/MRI hybrid imaging [11,146,174,175]. In one study, PET, CT, and PET/CT features were used to predicting local tumor control in head and neck cancer [129] by multivariate cox regression with a confidence interval (CI) CI CT and CI PET/CT of 0.73, however, CT-based radiomics overestimated the probability of tumor control in the poor prognostic groups. Another study found a correlation of PET and CT features using lymph node density [176] and concluded that CT density measurements together with PET uptake analysis increases the differentiation between malignant and benign LN.…”

Prediction of prostate cancer aggressiveness with a combination of radiomics and machine learning-based analysis of dynamic contrast-enhanced MRI

“…This result was strongly consistent with those of previous studies. 22,26,27 Among all the 10 features, most of them were higherorder statistics features and utilised wavelet transform. Higher-order statistics features could display subtle alterations in tissue morphology more explicitly.…”