The present study investigated the OS of patients with clinical stage T1-T3, N0-N3 ESCC who were treated with definitive CRT or surgery with adjuvant CRT in Taiwan. Our results demonstrated that Charlson score, clinical T stage, tumor differentiation, tumor length and treatment modality were independent prognostic factors for OS in multivariable analysis after propensity score matching. The results also suggested surgery with adjuvant CRT had better survival than definitive CRT in clinical stage II and III ESCC patients. With regard to clinical stage I ESCC patients, no significant difference in OSl was found between the definitive CRT group and the surgery with adjuvant CRT group.
Definitive CRT is recommended as one of the definitive treatments for clinical stage T1b-T4a, N0/N + ESCC by NCCN guidelines[7]. In Taiwan, almost 46% of ESCC patients received definitive CRT as definitive treatment between 2008 and 2014[4, 9]. The treatment efficacy of definitive CRT was established by some clinical trials, such as the RTOG 85 − 01 and RTOG 94 − 05 trials[10, 11]. Mikhail et al. reviewed 12 trials and illustrated that 2-year OS of definitive chemotherapy ranged from 18–54% and 3-year OS ranged from 22.2–64.3%[10, 12–17]. The present study reported that 2-/3-year OS were 24.8%/17.8% before matching, and 31.9%/23.8% after matching. These survival outcomes were worse than the previous literature possibly because patients who received definitive CRT were diagnosed at advanced stages and were not suitable candidates for surgery. Furthermore, patient preference and hospital facilities may influence the definitive treatment modality. This might explain why definitive CRT could not provide a favorable survival outcome to patients with ESCC. As for clinical stage I ESCC patients, there were small sample size after propensity score matching.
Although surgery with adjuvant CRT was not one of the recommended treatments for locally advanced ESCC by NCCN guidelines, more than a few ESCC patients chose it as a definitive treatment in Taiwan[6, 9]. For adjuvant CRT, NCCN guidelines suggested only patients with positive surgical margins should receive adjuvant CRT regardless of their nodal status[7]. However, some studies reported that adjuvant CRT could increase OS, and decrease the rates of metastasis and overall recurrence in node-positive ESCC[18, 19]. Rice et al. demonstrated that adjuvant CRT significantly doubled survival time, time to recurrence and recurrence-free survival in patients with locoregionally advanced ESCC[20]. Hsu et al. showed that post-operative CRT was associated with longer OS, longer disease-free survival and lower locoregional recurrence[21]. However, Chen et al. illustrated that patients with resectable thoracic ESCC may not benefit from adjuvant CRT[22]. Chen et al. illustrated that 1-year and 2-year OS rates for patients with ESCC who received surgery with adjuvant CRT were 67.5% and 41.4%. In the present study, 1-year and 2-year OS rates for patients with ESCC who received surgery with adjuvant CRT were 74.35% and 47.81% before matching and 72.35% and 45.74% after matching. Our results demonstrated that surgery with adjuvant CRT provided acceptable survival outcome. Usually, the side effects and toxicity of adjuvant CRT are worries for patients who undergo esophagectomy. Before the development of minimally invasive esophagectomy, patients who underwent open esophagectomy were unable to cope with any adjuvant therapy. With the advancement of minimally invasive esophagectomy, the early restoration of physical function may make patients able to endure adjuvant treatment[23]. The early recovery after minimally invasive esophagectomy may improve the delivery of adjuvant CRT. Thus, the role of adjuvant CRT in ESCC should be reappraised in the era of minimally invasive esophagectomy.
Neoadjuvant CRT plus esophagectomy and esophagectomy alone were main treatment strategies for clinical stage T1b-T4a, N0/N + ESCC patients in the NCCN guidelines as well. Neoadjuvant CRT plus esophagectomy is a widely acceptable treatment strategy in the world. The CROSS trial and other systematic reviews reported the treatment efficacy of neoadjuvant CRT plus esophagectomy[24, 25]. However, some clinical trials were unable to find out the role of esophagectomy when added to CRT[26, 27]. As for esophagectomy alone, multiple studies were conducted to compare surgery alone with definitive CRT[28–30]. These studies concluded that no statistically significant difference in OS between surgery alone and definitive CRT was seen. Therefore, further head-to-head comparison studies are needed to conclude the advantages and disadvantages of each treatment strategy.
In Taiwan, 14,394 patients were newly diagnosed with ESCC between 2008 and 2014. Most of them received definitive CRT, CRT plus esophagectomy, esophagectomy alone, or esophagectomy plus adjuvant CRT[4, 9]. In order to find out which treatment strategy could provide the best treatment outcome for patients with locally advanced ESCC, head-to-head comparisons between each treatment strategy were necessary. Wang et al. had compared definitive CRT with neoadjuvant CRT plus esophagectomy, and definitive CRT with esophagectomy alone[4, 6]. Until today, there was no literature comparing definitive CRT and esophagectomy plus adjuvant CRT. This is the first study to directly compare definitive CRT and esophagectomy plus adjuvant CRT using a large number of cases and propensity score matching to reduce the selection bias.
The Taiwan Cancer Registry database is a multi-center and population-based database. It has limitations that are inherent to all cohort studies because it is a retrospectivelyprospectively maintained database. First of all, the definitive CRT group were tended to be at advanced stage in the present study. Therefore, propensity score matching was performed to minimize selection bias. There are still some confounding factors that we cannot overcome. First, there is no information about the performance status before surgery and CRT. Second, the regimen of chemotherapy and the dosage and field of radiotherapy were not disclosed in the present study. Third, details about surgical procedures, postoperative complications and recurrence were lacking in this database. Fourth, the TCR database didn’t include the information about how many of these ESCC patients were planned to treat with esophagectomy and adjuvant CRT from the initial diagnosis and staging. Due to the retrospective nature of our study, a randomized control trial to validate results is necessary.