Recurrent (after a disease-free period) or persistent esophageal cancer following definitive therapy, particularly for locally advanced cancer, is common. Treatment failure following definitive chemoradiotherapy in RTOG-8501 was 46% at a minimum of 5 years follow-up.¹ Twenty percent of patients had recurrence after clinical complete response and 26% persistent cancer. In RTOG-8911 following R0 esophagectomy, at a median follow-up of 8.8 years, cancer recurrence was reported in 52% of patients receiving induction chemotherapy and 56% of patients with esophagectomy alone.²

Treatment of recurrent or persistent esophageal cancer after primary therapy is typically palliative and most commonly nonsurgical. However, in rare situations, salvage esophagectomy has been reported to be useful in highly selected patients in a final attempt to cure. To determine indications for salvage esophagectomy, the literature must be carefully assessed.

Failed Definitive Chemoradiotherapy

In 2007, Gardner-Thorpe et al.³ reviewed nine single institution series of salvage esophagectomy following definitive chemoradiotherapy. A total of 105 patients, predominantly with squamous cell carcinoma, were included in this simple evaluation. Salvage esophagectomy was an uncommon operation. With centers performing one to two salvage esophagectomies a year, this represented 1.7% to 4.2% of their esophagectomy volume. Morbidity was significant and three centers reported it was more than following neoadjuvant chemoradiotherapy. Anastomotic leak was reported in 18% of patients and conduit complications (necrosis and fistula) in 5%. Complications accounted for 42% of the 11% in-hospital mortality. Median survival ranged from 7 to 32 months. Almost all deaths after discharge were due to cancer. R0 resection was associated with improved survival in four series. In three series, no patient with R1 or R2 resection survived more than 13 months. There was some evidence that patients with recurrent cancer after chemoradiotherapy did better than those with persistent cancer.

Nishimura et al.⁴ reported results of salvage esophagectomy in 46 patients treated 1 to 7 months following definitive chemoradiotherapy. This group comprised 16% of patients with persistent or recurrent cancer after chemoradiotherapy. Operative mortality was 15%. Complications were frequent; the most common was anastomotic leak in 22% of patients. Median survival was 22 months, and 3-year survival was 17%. There were no clear predictors of survival; however, the three long-term survivors were ypN0 at salvage esophagectomy.

Piessen et al.⁵ reported salvage esophagectomy in 98 (20%) of 472 patients treated with definitive chemoradiotherapy. Sixty-two percent had R0 resection which was associated with improved median survival, 19 months in R0 patients versus 9 months in R1 and R2 patients (p < 0.001). In-hospital mortality was 3.1% and morbidity 33%. There were no R1 or R2 survivors after 26 months. Predictors of R0 resection were cancer length ≤5 cm on barium esophagram (p = 0.05) and contact between the primary cancer and aorta ≤90° on CT (p = 0.04).

D’Journo et al.⁶ reported their 10-year salvage esophagectomy experience in 24 patients (9% of esophagectomies). In this series, 88% had R0 resection and 30- and 90-day mortality was high at 21% and 25%, respectively. Forty-five percent of patients had a complication. Radiation dose greater than 55 Gray (Gy, unit of absorbed radiation dose) was associated with increased mortality, morbidity, length of stay, and blood transfusions. Five-year disease-free survival was 21%. R0 resection was associated with best 5-year survival (36% for R0 vs. 0% for R1 and R2), but this small experience did not reach statistical significance (p = 0.7).

Four studies have compared salvage esophagectomy to esophagectomy alone or esophagectomy as a component of neoadjuvant chemoradiotherapy.^7–10 Smithers et al.⁷ reported the results of salvage esophagectomy in 14 patients (4% of patients treated with definitive chemoradiotherapy) and compared them to 53 patients treated with neoadjuvant chemoradiotherapy. No statistical assessments of the comparisons were made. Median time from chemoradiotherapy to esophagectomy was 28 weeks in salvage patients versus 4 weeks in neoadjuvant patients. Salvage esophagectomy patients had more respiratory complications and longer ICU and hospital stays. Anastomotic leaks, transfusion requirements, and operative mortality were greater in salvage patients. Survival at 3 years was 24% for salvage patients and 53% for neoadjuvant patients. Survival following salvage esophagectomy was better for patients with recurrent cancer (25 months) than for those with persistent cancer (13 months).

Chao et al. reported 84 patients treated definitively with chemoradiotherapy.⁸ Forty-seven had persistent or recurrent cancer, of which 27 had salvage esophagectomy. Compared to patients receiving definitive chemoradiotherapy alone, salvage esophagectomy patients had a significantly better survival, 0% 5-year survival for definitive chemoradiotherapy versus 25% for salvage esophagectomy (p = 0.003). Compared to patients receiving neoadjuvant therapy, salvage esophagectomy patients were significantly older (p = 0.04), had more cervical and upper third thoracic cancers (p = 0.05), had fewer R0 resections (p = 0.001), more pulmonary complications (p = 0.006), more anastomotic leaks (p = 0.002), and higher operative mortality. However, survival was similar. In salvage esophagectomy patients, multivariable analysis identified low preoperative albumin level (p = 0.07) and anastomotic leak (p = 0.02) as predictors of hospital mortality and R1 and R2 resections (p = 0.04) as the predictor of disease-specific mortality.

Morita et al.⁹ reported 27 salvage esophagectomy patients and compared them to both patients who received neoadjuvant therapy and patients who received esophagectomy alone. Complications were associated with the use of chemoradiotherapy and were most common in salvage patients (25% esophagectomy alone, 40% neoadjuvant therapy, and 59% salvage esophagectomy, p < 0.001). Salvage patients were four times more likely to experience a pulmonary complication than those receiving esophagectomy alone. Preoperative therapy and retrosternal or subcutaneous routes of reconstruction were associated with increased risk of anastomotic leak. Operative mortality was increased in salvage patients but not statistically significant (2.4% vs. 2.0% vs. 7.4%, respectively). Compared to patients who had salvage esophagectomy for recurrent cancer, patients who had salvage esophagectomy for persistent cancer were less likely to have an R0 resection (p = 0.07) and had a worse 5-year survival (15% vs. 70%, respectively, p = 0.008).

Tachimori et al.¹⁰ treated 59 patients who had failed definitive chemoradiotherapy (radiation dose >60 Gy) with salvage esophagectomy and compared them to 553 patients who had esophagectomy alone. Salvage esophagectomy patients had significantly more respiratory complications (32% vs. 20%, respectively, p = 0.05), anastomotic leaks (31% vs. 25%, p = 0.003), wound infections (27% vs. 15%, p = 0.02), and hospital mortality (8% vs. 2%, p = 0.01), compared to patients who had esophagectomy alone. Tracheobronchial and gastric conduit necroses were highly lethal complications of salvage esophagectomy. Three-year survival following salvage esophagectomy was 38% compared to 58% following esophagectomy alone, T4 cancer and R1 and R2 resections were confounded in the analysis, thus leading the authors to conclude that only R0 resection was associated with improved survival in salvage esophagectomy patients.

Failed Esophagectomy

The literature is scarce for salvage esophagectomy following esophagectomy and typically anecdotal. In this setting, salvage esophagectomy is primarily used to treat anastomotic recurrence. Schipper et al.¹¹ reported a 30-year experience in which 27 patients were considered for salvage esophagectomy, 23 (85%) for anastomotic recurrence, and 4 for cancer recurrence (3 esophageal remnant and 1 gastric conduit). Salvage surgery was possible in 19 (70%), 10 completion esophagogastrectomies and 9 anastomotic resections. Eight patients had exploration only. Operative mortality was 7% and 59% of patients had postoperative complications. Five-year survival was 35% in salvage patients, no patient undergoing exploration only survived 3 years. Predictors of survival were disease-free interval between esophagectomy and salvage surgery greater than 2 years (p < 0.05) and R0 resection (p < 0.02).

Although the quality of evidence is poor, the overwhelming message from the literature is R0 resection is a prerequisite if salvage esophagectomy is to offer a survival benefit.