Development and validation of a prognostic model incorporating texture analysis derived from standardised segmentation of PET in patients with oesophageal cancer

• DOI: 10.1007/s00330-017-4973-y
• Publication Date: 2017-08-02T11:02:19Z
• Authors: Kieran G Foley; Robert K. Hills; Beatrice Berthon; Christopher Marshall; Craig Parkinson; Wyn G. Lewis; Tom D. L. Crosby; Emiliano Spezi; Stuart Ashley Rob...
• Keywords: Quartile; Neuroradiology

This retrospective cohort study developed a prognostic model incorporating PET texture analysis in patients with oesophageal cancer (OC). Internal validation of the was performed. Consecutive OC (n = 403) were chronologically separated into development 302, September 2010-September 2014, median age 67.0, males 227, adenocarcinomas 237) and cohorts 101, 2014-July 2015, 69.0, 78, 79). Texture metrics obtained using machine-learning algorithm for automatic segmentation. A Cox regression including age, radiological stage, treatment 16 developed. Patients stratified quartiles according to score derived from model. p-value < 0.05 considered statistically significant. Primary outcome overall survival (OS). Six variables significantly independently associated OS: [HR =1.02 (95% CI 1.01-1.04), p 0.001], stage [1.49 (1.20-1.84), [0.34 (0.24–0.47), log(TLG) [5.74 (1.44–22.83), 0.013], log(Histogram Energy) [0.27 (0.10–0.74), 0.011] Histogram Kurtosis [1.22 (1.04–1.44), 0.017]. The demonstrated significant differences OS between both (X2 143.14, df 3, 0.001) 20.621, 0.001). can risk stratify demonstrates additional benefit staging. • adds value are survival. help patients.