Abstract
70
Objectives CRT is a mainstay of treatment for potentially operable EC, but not all patients benefit from post-CRT adjuvant surgery. There is a need for biomarkers to identify which patients would benefit from adjuvant surgery over CRT alone.
Methods We examined whether textural analysis (using proprietary 'TexRAD' software,www.texrad.org) of non-contrast CT(CTTA) and 18F-FDG PET images of baseline and post-treatment PET/CT scans, along with SUVmax and tumor metabolic area, could act as imaging biomarkers that identify patients benefiting from surgery after CRT. Imaging biomarkers were used to categorise 40 patients(adenocarcinoma 80%, SCC 20%) with potentially operable EC into subgroups based on good and poor imaging phenotypes, and results were correlated with mean overall survival(OS).
Results Stratification based on pre-CRT CTTA identified a sub-group of patients with significantly enhanced OS with CRT plus surgery vs CRT alone(46.0 vs 17.1 months respectively, p=0.028). Stratification by SUVmax post-CRT showed the low SUVmax group to have significantly longer OS with CRT plus surgery vs CRT alone(51.3 vs 30.3 months respectively,p=0.046). Stratifying by metabolic area showed the high metabolic area group survived significantly longer with CRT plus surgery vs CRT alone(41.2 vs 9.8 months, p=0.034). Textural analysis of PET images failed to show a significant relationship between texture and survival in baseline or post-treatment scans.
Conclusions CTTA features and 18F-FDG PET scan parameters have the potential to delineate imaging phenotypes which can predict CRT response and guide the use of adjuvant surgery in EC. This could enable more accurate stratification of patients into a subgroup who would clearly benefit from surgery.
Research Support CRT is a mainstay of therapy in esophageal cancer (EC), however there is uncertainty surrounding the role of post CRT surgery. There is a need to identify which EC patients benefit from post-CRT surgery given the significant associated risks. Textural analysis can be applied to already clinically obtained CT and FDG-PET images, without need for further image acquisition. CT texture analysis (CTTA) features and FDG uptake have been shown to be imaging biomarkers associated with a poor prognosis and lack of treatment response in several malignancies. This study assesses the potential for CTTA and PET textural analysis as well as PET metabolic parameters to define subgroups of patients in whom surgery significantly extends overall survival. Our results show CTTA, tumor metabolism and metabolic area, but not PET textural parameters, to correlate with OS. CTTA and tumor metabolic area identified subgroups of patients who clearly benefited from surgery.