A subgroup analysis of patients with later-stage vs earlier-stage disease showed no difference in overall disease-specific or disease-free survival for earlier-stage patients with CR vs no CR to neoadjuvant therapy. In the later-stage patients, however, there appeared to be "a trend toward improved survival with a complete response . . . to neoadjuvant therapy, although this did not reach statistical significance," he said.
In assessing the entire cohort, using a Cox proportional hazards model to determine prognostic factors for survival, again, late clinical stage at presentation and transthoracic approach predicted a worse overall survival in all patients, while use of neoadjuvant therapy was not a significant predictor for survival.
Morbidity and mortality in the two treatment groups were comparable: "Our overall perioperative mortality was six patients, or 3.9%," Dr. Schmidt said. "We saw no differences in mortality, overall complications, or major complications between those who had surgical resection alone vs those who underwent preoperative chemoradiation with surgical resection. An exception to that was a surprising higher rate of leak in our patients who underwent surgical resection alone vs the neoadjuvant group, but our overall leak rate was fairly low at 7%."
Dr. Schmidt concluded: "We believe continued assessment of the value of neoadjuvant therapy in esophageal cancer is necessary, including prospective randomized trials, molecularly targeted therapies, and other avenues of therapy."
Limitations to the study, Dr. Schmidt noted, include the fact that patients received a variety of chemoradiation regimens but separate assessments were not made in evaluating responses. "If there was one regimen that was more effective than the others, its effect would have been diluted by our decision to include all patients," he said.
Other possible limitations included the fact that the investigators could not know whether preoperative T stage was well-matched between the groups that did or did not receive neoadjuvant chemoradiation, because more than half of the patients did not have an endoscopic ultrasound. The procedure was not widely used at VUMC in the earlier years for which data were collected, so some patients may have been understaged.
"Given that we saw a trend toward improved survival in late-stage patients," Dr. Schmidt said, "perhaps if the staging of our patients had been more late-stage rather than early-stage patients, we might have seen an overall survival benefit."
Another limiting factor may have been lack of standardization over the 10-year period in the pathologic assessment of esophageal cancer specimens and the criteria for designating "complete response." At VUMC, Dr. Schmidt noted, beginning in 1996, a rigorous protocol for assessment of response was instituted that included much more careful sectioning of esophageal tumors and very close inspection of the gastroesophageal junction. It is possible, he said, that prior to 1996, some patients may have been inappropriately assessed as having had a complete response.