Application of Deauville or Hopkins visual 5-point scale in response assessment of esophageal cancer receiving neoadjuvant chemotherapy and surgery

Objectives: Deauville visual 5-point scale (DS) on FDG PET is adopted as a new response assessment method of lymphoma in the Lugano classification published in 2014. Similar visual scale was reported to be useful in head and neck cancer as Hopkins criteria (HC). The purpose of this study was to evaluate if DS or HC was applicable in response assessment of esophageal cancer (EC) receiving neoadjuvant chemotherapy (NAC) and surgery.

Methods: FDG PET-CT examinations were performed before and after 2 cycles of NAC in 41 EC patients before surgery. SUVmax of EC lesion was obtained in each exam and reduction of SUVmax after NAC was calculated. Patients with reduction of SUVmax 蠅30% were defined as responders. PET images after NAC were evaluated regarding residual disease using DS and HC. In the evaluation with DS, two patterns were tested as a definition of non-residual disease: scores 1 and 2 (uptake &#8804; mediastinum) [D1] or scores 1, 2, and 3 (uptake > mediastinum but &#8804; liver) [D2]. HC scores are as follows: 1- FDG uptake < blood pool(bp); 2- focal uptake > bp but < liver; 3- diffuse uptake > bp or liver; 4- focal uptake > liver; 5- focal and intense uptake. HC scores 1, 2, and 3 were considered to represent non-residual disease although scores 3 and 4 both showed uptake > liver (diffuse vs. focal). Results provided with these scores were compared to the reduction of SUVmax and pathologic results after NAC as well as recurrence within 1 year after subsequent surgery. Statistical analysis was performed with a McNemar test.

Results: D1, D2, and HC provided results of residual disease respectively in 31, 25, and 23 pts after NAC. 33 (80%) of 41 pts were considered as responders with reduction of SUVmax after NAC (蠅70% reduction in 16 of them). Although exhibiting mixed results even in responders with 蠅70% SUVmax reduction, D1, D2, and HC provided results of residual disease in all 8 non-responders. Good pathologic response (viable cancer cells: < 1/3 of specimens) was achieved in 5 (12%) pts and recurrence was observed in 19 (46%) of the 41 pts. D1 exhibited significantly higher accuracy (73%, p <0.05) than D2 (59%) or HC (59%) in evaluating pathologic response, if non-good pathologic response was regarded as residual disease. HS exhibited significantly higher accuracy (66%, p <0.01) than D1 (46%) in evaluating recurrence (D2: 61%).

Conclusion: This study demonstrated both DS and HC were applicable in response assessment of EC after NAC. DS had the promise in evaluating pathologic response, when uptake > mediastinum was treated as residual disease. HC was useful in predicting recurrence in this study population. Research Support: