Back to Annual Meeting Program

Background: Attempts to define the optimal treatment strategy for resectable but locoregionally advanced esophagogastric (EG) adenocarcinoma have yielded conflicting results. As a result of high local failure rates, most thoracic oncologists favor induction therapy followed by surgery. In our center, a selective approach has been used based on the surgeon’s assessment of resectability at laparoscopic staging, (e.g. mobile tumors and low lymph node burden). Our study aim was to determine whether survival is impacted by a selective approach to treatment that includes minimally invasive esophagectomy (MIE) alone.
Methods: Patients with stage II or higher EG adenocarcinoma treated with MIE (n=375; 1997-2009) were reviewed. Demographics, comorbidities, tumor and treatment variables were abstracted. To determine the probability of treatment assignment to either induction therapy followed by MIE (E-) or to MIE alone (E+), propensity scores were calculated. Variables are listed in Table 1. Complete data for the propensity variables were available in 280 patients and 80 closely matched pairs (n=160) were generated. Data missingness was random; survival relevant variables in excluded patients did not differ significantly from the included patients. Hazard ratios for death were calculated by stratified Cox proportional-hazards regression model after controlling for age, gender, BMI, smoking history and age adjusted CCI.
Results: MIE was the primary therapy in 47% (n=178) of patients and induction therapy in 53% (n=197; 51% chemotherapy alone and 49% chemoradiation). Chemotherapy included cisplatin (61%), 5-FU (76%), and paclitaxel (42%). Complete response following induction therapy was confirmed at MIE in 13%; 44% were node-negative at resection compared to 20% for MIE alone. Median number of lymph nodes examined was 21 (IQR 15, 29). Median time to follow-up was 23 months (IQR 11, 38). Adjuvant therapy was given to 53% of patients following MIE. In the 80 matched pairs, there were 117 deaths (73%; median 18 months, IQR 9, 29). Surgery as primary mode of treatment was not associated with significantly different hazard for death after adjusting for age, sex, BMI, smoking history and age-adjusted Charlson Comorbidity index (0.96; 95% CI 0.58-1.6). Complete response was not associated with a significant improvement in median survival (19 versus 17 months; p=0.24).
Conclusion: We found that a selective approach to treatment of EG adenocarcinoma does not negatively impact patient survival. Complete response rates to induction therapy in this series were not associated with improved survival. Our data suggest that surgeon-driven treatment allocation for locoregionally advanced esophagogastric adenocarcinoma, including aggressive lymph node dissection, is a reasonable alternative to treating all patients with induction therapy.

Variables included in propensity matching

Age	Body mass indes	smoking history
Pack-years smoked	Sex	Surgeon
Alcohol use	Gerd history	Histologically confirmed Barrett's
History of MI	History of CHF	History of CABG or coronary stent
History of vascular disease	History of renal failure	History of diabetes requiring treatment
History of pulmonary disease	History of stroke or TIA	History of malignancy other than current
History of metastaatic cancer	History of liver disease	Final pretreatment clinical stage
Operation status	Prior esophageal surgery	Esophageal cancer location