br ACCEPTED MANUSCRIPT br compared with surgery alone
0.80-0.99) compared with surgery alone. In agreement with this 4u8C meta-analysis, we found that patients receiving neoadjuvant therapy followed by esophagectomy had significantly improved long-term survival compared to patients undergoing esophagectomy alone. However, there was no significant survival advantage in the neoadjuvant therapy group compared to esophagectomy followed by adjuvant therapy. It also appears that many patients in the surgery only and surgery plus adjuvant therapy groups would have been candidates for neoadjuvant CR by today’s standards (stage 2 and 3). (Table
1) However, candidacy for neoadjuvant therapy cannot be decided based on cancer stage alone. Surgical resection is the only potential curative option in patients with locoregional esophageal cancer. However, less than 50% of patients with locoregional disease may undergo surgical resection (15, 16). In the present study, only 21% of patients underwent surgical resection. The underutilization of esophagectomy in the treatment of EC may be related to the potential morbidity and mortality associated with this operation (17). However, there have been significant improvements in surgical technique and perioperative care and a recent meta-analysis suggests that esophagectomy can be performed safely at high-volume centers with an in-hospital mortality less than 3% (6). Additionally, many patients may not be operative candidates despite having resectable disease due to significant comorbidities or malnutrition. Other patients may simply refuse surgical therapy. Nevertheless, we suspect, the low number of esophageal resections could also at least in part be secondary to therapeutic nihilism.
Patients in the neoadjuvant therapy group had a significantly shorter median length of hospital stay and were less likely to be readmitted compared to patients in the surgery alone group in our study. We believe in the surgery only group patients had worse postoperative performance status likely because of post-surgical complications and were not found to be candidates for adjuvant therapy. This assumption is also reinforced by the significantly elevated 30 (0.7% vs 8.4%) and 90 (3.7% and 16.6%) day mortality between the adjuvant therapy and surgery only groups.
Over the last several decades, there has been an increasing number of studies demonstrating an inverse relationship between hospital volume and perioperative mortality in multiple high risk
surgical procedures including esophageal surgery (18, 19), cardiac surgery (20, 21), lung surgery (22), pancreatic surgery (23), complex gastrointestinal surgery (24), and hepatic resections (25, 26). In a propensity matched analysis, Speicher et al. demonstrated that in patients with EC that traveled to higher-volume centers were much more likely to undergo surgical resection and 5- year survival was 50% higher than patients who did not travel and were treated at low-volume centers (27). In the present study, there was a significant survival advantage in patients undergoing treatment at facilities performing 20 or more esophagectomies per year adjusting for patient and tumor related characteristics. Importantly, this volume relationship remained statistically significant in patients treated with definitive CR. This survival advantage may be related to increased surgical experience resulting in fewer perioperative complications and improved long-term oncologic outcomes. This was also true for non-operatively managed patients suggesting that high-volume facilities are likely to have multidisciplinary teams (MCTs) consisting of specialized surgical oncologists, medical oncologists, gastroenterologists, diagnostic and interventional radiologist, pathologists, and dieticians that may lead to earlier diagnosis and treatment, improved patient follow-up, and better management of malnutrition that is often associated with ESCC.
There are several limitations to the present study. It suffers from all the usual drawbacks of any publications using major national datasets like the NCDB. This study was limited by the variables collected in the NCDB. The type of chemotherapy agents and radiation utilized, post-operative complications, and recurrence data are not recorded in the NCDB, thus the impact of these variables on overall survival cannot be evaluated. In non-randomized samples, where clinical judgement about physiologic fitness plans a major role in treatment planning, less fit patients will be less likely to be offered tri-modality therapy. Those less fit patients probably have poorer survival. Pathologic staging has been used for patients undergoing ER, while clinical staging for patients in the CR only group. Tumor location within the esophagus was missing for most patients. Lastly, we were unable to evaluate other factors that may influence quality of care such as multidisciplinary clinics, advanced training,