BJS.03Markov Decision Analysis of Treatment Pathways for Resectable Malignancy of the Oesophagus or Gastrooesophageal Junction

Aims To perform decision-analysis of treatment options for resectable malignancy of the oesophagus or gastroesophageal junction including: surgery alone, neoadjuvant chemoradiotherapy, neoadjuvant chemotherapy and surgery followed by adjuvant therapy based on current highest-level evidence. Methods A Markov decision analysis model in an advanced decision-tree format was constructedand populated with data from existing randomised controlled trials. Markov model transition probabilities were based on weighted pooled estimates of proportions from included studies, calculated using Freeman-Tukey arcsine square root transformation under random effects model to account for heterogeneity.Each Markov cycle equated with one month and Markov states within the model included: alive without disease, alive with disease and dead. Extensive deterministic and Monte Carlo probabilistic sensitivity analysis was performed to test all parameters contained within the model. Results 23 randomised controlled trials were included. Intention-to-treat analysis of the treatment pathways showed that neoadjuvant chemoradiotherapy was the superior pathway with an overall survival time of 50.52 months (42.26 QALMs). Monte Carlo sensitivity analysis run over 10000 iterations showed that neoadjuvant chemoradiotherapy was selected as the superior treatment pathway at a frequency of 93.21% followed by surgery followed by adjuvant therapy with a frequency of 6.79%. Subgroup analysis of only squamous cell carcinomas demonstrated that neoadjuvant chemoradiotherapy was the superior pathway with an overall survival time of 62.67 months (55.22 QALMs). Conclusions Based on current best available evidence this decision analysis supports neoadjuvant chemoradiotherapy as the treatment strategy of choice for resectable malignancy of the oesophagus or gastroesophageal junction.