• Amandine Crombé, Cynthia Périer, Michèle Kind, Baudouin Denis De Senneville, François Le Loarer, Antoine Italiano, Xavier Buy, and Olivier Saut.
• Department of Radiology, Institut Bergonie, Regional Comprehensive Cancer Center, Bordeaux, France.
• J Magn Reson Imaging. 2019 Aug 1; 50 (2): 497-510.

BackgroundStandard of care for patients with high-grade soft-tissue sarcoma (STS) are being redefined since neoadjuvant chemotherapy (NAC) has demonstrated a positive effect on patients' outcome. Yet response evaluation in clinical trials still relies on RECIST criteria.PurposeTo investigate the added value of a Delta-radiomics approach for early response prediction in patients with STS undergoing NAC.Study TypeRetrospective.PopulationSixty-five adult patients with newly-diagnosed, locally-advanced, histologically proven high-grade STS of trunk and extremities. All were treated by anthracycline-based NAC followed by surgery and had available MRI at baseline and after two chemotherapy cycles.Field Strength/SequencePre- and postcontrast enhanced T1 -weighted imaging (T1 -WI), turbo spin echo T2 -WI at 1.5 T.AssessmentA threshold of <10% viable cells on surgical specimens defined good response (Good-HR). Two senior radiologists performed a semantic analysis of the MRI. After 3D manual segmentation of tumors at baseline and early evaluation, and standardization of voxel-sizes and intensities, absolute changes in 33 texture and shape features were calculated.Statistical TestsClassification models based on logistic regression, support vector machine, k-nearest neighbors, and random forests were elaborated using crossvalidation (training and validation) on 50 patients ("training cohort") and was validated on 15 other patients ("test cohort").ResultsSixteen patients were good-HR. Neither RECIST status (P = 0.112) nor semantic radiological variables were associated with response (range of P-values: 0.134-0.490) except an edema decrease (P = 0.003), although 14 shape and texture features were (range of P-values: 0.002-0.037). On the training cohort, the highest diagnostic performances were obtained with random forests built on three features: Δ_Histogram_Entropy, Δ_Elongation, Δ_Surrounding_Edema, which provided: area under the curve the receiver operating characteristic = 0.86, accuracy = 88.1%, sensitivity = 94.1%, and specificity = 66.3%. On the test cohort, this model provided an accuracy of 74.6% but 3/5 good-HR were systematically ill-classified.Data ConclusionA T2 -based Delta-radiomics approach might improve early response assessment in STS patients with a limited number of features.Level Of Evidence3 Technical Efficacy: Stage 2 J. Magn. Reson. Imaging 2019;50:497-510.© 2018 International Society for Magnetic Resonance in Medicine.

23 keywords...

hide…