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

Abstract

ObjectivesThis retrospective cohort study developed a prognostic model incorporating PET texture analysis in patients with oesophageal cancer (OC). Internal validation of the model was performed.MethodsConsecutive OC patients (n = 403) were chronologically separated into development (n = 302, September 2010-September 2014, median age = 67.0, males = 227, adenocarcinomas = 237) and validation cohorts (n = 101, September 2014-July 2015, median age = 69.0, males = 78, adenocarcinomas = 79). Texture metrics were obtained using a machine-learning algorithm for automatic PET segmentation. A Cox regression model including age, radiological stage, treatment and 16 texture metrics was developed. Patients were stratified into quartiles according to a prognostic score derived from the model. A p-value < 0.05 was considered statistically significant. Primary outcome was overall survival (OS).ResultsSix variables were significantly and independently associated with OS: age [HR =1.02 (95% CI 1.01-1.04), p < 0.001], radiological stage [1.49 (1.20-1.84), p < 0.001], treatment [0.34 (0.24–0.47), p < 0.001], log(TLG) [5.74 (1.44–22.83), p = 0.013], log(Histogram Energy) [0.27 (0.10–0.74), p = 0.011] and Histogram Kurtosis [1.22 (1.04–1.44), p = 0.017]. The prognostic score demonstrated significant differences in OS between quartiles in both the development (X2 143.14, df 3, p < 0.001) and validation cohorts (X2 20.621, df 3, p < 0.001).ConclusionsThis prognostic model can risk stratify patients and demonstrates the additional benefit of PET texture analysis in OC staging.Key points• PET texture analysis adds prognostic value to oesophageal cancer staging.• Texture metrics are independently and significantly associated with overall survival.• A prognostic model including texture analysis can help risk stratify patients.Electronic supplementary materialThe online version of this article (doi:10.1007/s00330-017-4973-y) contains supplementary material, which is available to authorized users.

Radiomics in esophageal and gastric cancer

“…Similarly, in a study of 65 patients Paul et al found that a model incorporating GLCM homogeneity was a predictor of response (with an AUROC value of 0.823) but not of survival [29]. However, in a larger study with 403 patients, Foley et al found that total lesion glycolysis, histogram energy and kurtosis were independently associated with overall-survival [30].…”

Novel imaging techniques in staging oesophageal cancer

“…[88] One large, single centre study developed and internally validated a prognostic model including PET radiomics and found that model performance for overall survival improved compared to traditional clinical factors. [89] One critical aspect and potential limitation of radiomics is the reliability and reproducibility of these features. Much radiomics research is now concentrating on defining stable features for future application.…”

External validation of a prognostic model incorporating quantitative PET image features in oesophageal cancer

“…[14] A previously published prognostic model had been developed and internally validated in patients with esophageal cancer. Details of model development have been provided in Foley et al [11] Briefly, the prognostic model had only been evaluated by same-centre internal validation in patients managed by the South-East Wales Regional Upper Gastrointestinal (GI) Cancer Multi-Disciplinary Team (MDT), United Kingdom. A suitable independent cohort was not accessible at the time of publication.…”

“…[9] Post-reconstruction harmonisation methods have been proposed to adjust for these differences, thus promoting standardised research between centres. [10] The primary aim of this study was to test the generalizability of a UK single-centre esophageal cancer prognostic model incorporating radiomic features [11] firstly preharmonisation, then post-harmonisation, against a cohort of esophageal cancer patients treated exclusively with neo-adjuvant chemoradiotherapy (NACRT) according to the Dutch NACRT plus surgery versus surgery alone for oesophageal/junctional cancer (CROSS) trial regimen. [12] A widely generalizable prognostic model incorporating radiomic features of primary tumours might offer clinicians complimentary data beyond traditional prognostic factors that will assist treatment decision making and risk stratification.…”

“…[12] A widely generalizable prognostic model incorporating radiomic features of primary tumours might offer clinicians complimentary data beyond traditional prognostic factors that will assist treatment decision making and risk stratification. [11,13] The secondary aim was to compare prognostic models with and without PET image features between cohorts to provide further validation.…”