Prediction of cervical cancer recurrence using textural features extracted from 18F-FDG PET images acquired with different scanners

Abstract

ObjectivesTo identify an imaging signature predicting local recurrence for locally advanced cervical cancer (LACC) treated by chemoradiation and brachytherapy from baseline 18F-FDG PET images, and to evaluate the possibility of gathering images from two different PET scanners in a radiomic study.Methods118 patients were included retrospectively. Two groups (G1, G2) were defined according to the PET scanner used for image acquisition. Eleven radiomic features were extracted from delineated cervical tumors to evaluate: (i) the predictive value of features for local recurrence of LACC, (ii) their reproducibility as a function of the scanner within a hepatic reference volume, (iii) the impact of voxel size on feature values.ResultsEight features were statistically significant predictors of local recurrence in G1 (p < 0.05). The multivariate signature trained in G2 was validated in G1 (AUC=0.76, p<0.001) and identified local recurrence more accurately than SUVmax (p=0.022). Four features were significantly different between G1 and G2 in the liver. Spatial resampling was not sufficient to explain the stratification effect.ConclusionThis study showed that radiomic features could predict local recurrence of LACC better than SUVmax. Further investigation is needed before applying a model designed using data from one PET scanner to another.

Prediction of outcome using pretreatment 18F-FDG PET/CT and MRI radiomics in locally advanced cervical cancer treated with chemoradiotherapy

“…A first study associated high pretreatment PET GLNU GLRLM with poorer prognosis [9]. Another showed that 18 F-FDG PET/CT features could predict local recurrence of LACC better than SUV max [13]. A third study showed that DCE-MRI 2 nd -order textures could predict treatment outcome [10].…”

“…Recently, in addition to conventional parameters on FDG PET/CT and MRI used to stage disease, plan treatment and assess response there has been a growing interest in the extraction of quantitative features from medical images, denoted radiomics [6]. Radiomics features are statistical or model-based metrics to quantify tumor intensity, shape and heterogeneity which have been shown to reflect intratumoral histopathological properties [7] and to provide prognostic information in several pathologies [8] including CC [9][10][11][12][13]. Radiomics on pretreatment PET/CT has been shown to predict response to therapy and risk of pelvic recurrence in CC [9,11,13] and used to characterize CC lesions from diffusion-weighted MRI (DW-MRI) [12] or dynamic contrast enhancement MRI (DCE-MRI) [10].…”

Radiomics for radiation oncologists: are we ready to go?

“…15 In cervical cancer for example, a small subset of PET/CT radiomic features has been found to be of greater value than the SUV max or standard clinical variables to predict local recurrence. 32,37 The case of neoadjuvant treatments (e.g. radiochemotherapy) represents an interesting challenge for prediction modelling, as the definitive answer is provided by histological analysis.…”

Big Imaging Data: Klinische Bildanalyse mit Radiomics und Deep Learning

“…Hierbei wurde auch gezeigt, dass die Wahl der Raster-bzw. der Voxelgröße, siehe [7,16,21] und [25], durchaus einen sehr großen Einfluss auf die Analyse hat.…”