Loco-regional staging accuracy in oesophageal cancer—How good are we in the modern era?

Abstract Introduction Accuracy of locoregional staging in patients with oesophageal cancer is critical in determining operability and the need for neoadjuvant treatment. Imaging technology has advanced significantly in recent years but it is not known whether this translates to improved staging accuracy. This study investigates staging accuracy in relation to CT, EUS, PET-CT and final pre-operative stage. It specifically addresses the accuracy of staging with respect to the threshold for administering neoadjuvant therapies. Materials and methods Pre-operative staging according to CT, EUS, PET-CT and final pre-operative stage were compared to the postoperative histological staging in 133 patients undergoing potentially curative surgery (without neoadjuvant therapy) for oesophageal cancer between January 2010 and January 2015. T and N stage accuracies were reported separately for each imaging modality. Patients were also divided into two groups depending on whether the final pre-operative stage was below (≤T2, N0, early tumours) or above (≥T3 and/or ≥N1, locally advanced tumours) the threshold for offering neoadjuvant therapy. Accuracy of pre-operative staging was then analysed with respect to identification of patients below/above this threshold. The additional benefit offered by EUS for this purpose was investigated. Results T stage accuracies were 72.6%, 76.7% and 79.3% for CT, EUS and final pre-operative stage respectively. N stage accuracies were 75.6%, 77.2%, 74.5% and 78.6% for CT, EUS, PET-CT and final pre-operative stage respectively. Staging accuracy with respect to threshold for neoadjuvant treatment showed 62.0% early tumours were correctly staged and 80.5% advanced tumours were correctly staged. Whether or not patients underwent EUS did not affect the staging accuracy with respect to neoadjuvant treatment threshold. Conclusions Staging accuracy with respect to the threshold for treatment with neoadjuvant therapy is poor, leading to potential over/under treatment. Predicting individual response to neoadjuvant therapy would provide a better way to determine which patients should receive this additional treatment.