Salvage esophagectomy was performed infrequently. The average study reported on only 27 patients (range 12–65). Ten articles reported a denominator that included the total number of esophagectomies performed during the study period (mean 558, range 268–780) and of all esophagectomies, salvage cases accounted for on average only 5.3 %. Only three articles reported the total number of patients undergoing dCRT during the study period [11, 14, 15]. Chao et al. reported salvage esophagectomy having been performed in 27 of 47 (57 %) patients with locally recurrent tumor out of a group of 84 patients who had undergone dCRT. Smithers et al. reported salvage esophagectomy in 11 of 235 (4 %) patients undergoing dCRT at their institution and Miyata and colleagues performed salvage surgery in 33 of 219 (15 %) patients treated with dCRT. Eight studies, all from Asia, reported exclusively on patients with recurrent or persistent squamous cell cancer (SCC) [7–10, 14–16, 19]. A single study from North America reported outcomes of patients undergoing salvage esophagectomy for adenocarcinoma (ACA) only [18]. Radiation doses used for dCRT were generally between 50 and 60 Gy, with two exceptions; Schieman et al. reported a dosage range between 30 and 72 Gy in patients with recurrent SCC of the proximal esophagus (defined as tumor arising less than 20 cm from the incisors), and in the study by Chao and colleagues the total dosage of radiation was 30 Gy in 25 fractions [15, 20]. All 16 studies reviewed described salvage esophagectomy for persistent or recurrent tumor, however, D’Journo et al. included in their analysis six patients who underwent esophagectomy for benign etiologies (intractable stenosis in three, perforation in two and radiation induced esophagitis in one) [12]. Eleven studies provided details regarding whether salvage esophagectomy was performed for persistent or recurrent disease.

Combined analysis of 251 patients included in these studies shows even distribution between persistent (50.2 %) and recurrent tumors (49.8 %), however studies by Swisher et al and Yoo et al consisted of a very high proportion of recurrent tumors (100 and 92 %, respectively), whereas the study by Nakamura et al had a predominance of persistent tumors following dCRT (89 %) [6, 19, 22]. This may be relevant with respect to perioperative outcomes because the deleterious fibrotic effects of radiation are more likely to be encountered when performing esophagectomy for recurrent tumors which typically occur many months or even years after the completion of dCRT. The mean time from completion of dCRT until esophagectomy ranged from 1.7 to 18 months, and averaged 7.1 months among 11 studies where data were available (Table 28.2).

Surgical approaches usually included either a transthoracic or a three-hole technique. Reports from Asia tended to include a higher utilization of three-hole (McKeown) esophagectomy and cervical anastomoses compared to North American, Australian and European series. Three-field lymphadenectomy was reported exclusively in reports from Asian centers. The transhiatal approach was used infrequently (mean 5.5 %, range 0–15 % of cases, 11 studies). Complete resection with negative margins averaged 79 % (range 50–100 %, 15 studies) and the mean complete pathologic response rate (ypT0N0M0) was 4.1 % (range 0–13 %, 12 studies, benign cases excluded). Regarding perioperative morbidity, complications were reported in 35–79 % cases (mean 53 %, 8 studies). The most common postoperative events were respiratory complications, which were reported in 9–62 % of cases (mean 28 %, 15 studies [6–16, 18, 20]).

Seven comparative studies evaluated incidence of postoperative complications and all showed higher rates in patients undergoing salvage procedures compared to controls undergoing surgery alone or planned surgery after nCRT (Table 28.3). In only two studies was the difference significant, however. In 27 patients undergoing salvage surgery, Chao et al. reported respiratory complications in 33 % compared to 12 % in non-matched patients undergoing planned surgery after nCRT (p = 0.006) [15]. Similarly, Smithers et al reported higher rates of respiratory events in 14 patients undergoing salvage surgery (57 % vs. 30 %, p < 0.05) [11]. However analysis of a larger group of patients (n = 65) undergoing salvage esophagectomy by Marks and colleagues showed no significant differences in rates postoperative respiratory events between salvage patients and propensity score matched controls (23 % vs. 19 %, p = 0.664) [18]. Anastomotic leak was reported to occur in between 8 and 38 % of cases (mean 23 %, 15 studies). The highest leak rate (38 %) was reported by Swisher and colleagues, however, this small series of 13 patients consisted exclusively of patients who had recurrent tumors that presented at an average of 18 months after dCRT (mean dose 56 Gy) which is the longest disease free interval reported in any of the studies reviewed [6].