Prolonged interval between neoadjuvant chemoradiotherapy and esophagectomy does not benefit the outcome in esophageal cancer: a systematic review and meta-analysis

Prolonged interval between neoadjuvant chemoradiotherapy and esophagectomy does not benefit the... Summary Whether a prolonged interval between neoadjuvant chemoradiotherapy (nCRT) and esophagectomy could benefits conditions such as rectal cancer, still remains unknown. We therefore performed the current study to evaluate the influence of the interval between nCRT and esophagectomy on the clinical outcomes in patients with esophageal cancer. PubMed and Embase were searched to identify eligible cohort studies. The primary outcome was five-year overall survival (OS), and secondary outcomes included the incidence of anastomotic complications, perioperative mortality, pathologic complete response (pCR) rate, positive circumferential resection margin (CRM) rate, and R0 resection rate. A random-effects model was used for all meta-analyses irrespective of heterogeneity. Ten cohort studies with 2383 patients were included. Overall, the pooled estimate revealed that the prolonged interval has no impact on five-year OS (odds ratio (OR) 0.87, 95% CI 0.66 to 1.14, P = 0.30), with low heterogeneity (PH = 0.78, I2 = 0%). However, it was associated with an increased risk of anastomotic complication (OR 1.71, 95% CI 1.15 to 2.54, P = 0.008), with no effect on perioperative mortality (OR 1.20, 95% CI 0.79 to 1.83, P = 0.40). Additionally, the prolonged interval failed to increase the pCR rate (OR 1.02, 95% CI 0.78 to 1.33, P = 0.89). Even worse, it was correlated with a decreased R0 resection rate (OR 0.60, 95% CI 0.41 to 0.88, P = 0.009) and increased positive CRM rate (OR 2.20, 95% CI 1.44 to 3.36, P < 0.001). This study suggests that the prolonged interval between nCRT and esophagectomy fails to result in better outcomes, and in fact, could worsen clinical outcomes, with increasing anastomotic complications, and undermine resection completeness. However, this conclusion should be treated with caution because of the limitations of retrospective cohort study and substantial clinical heterogeneity. (The study was registered at PRESPERO as CRD42016048210). INTRODUCTION Esophageal cancer is one of the most aggressive malignancies, and surgical resection remains the essential cornerstone for patients with resectable disease. However, surgery alone is unsatisfactory and results in a poor prognosis in patients with locally advanced disease. With progresses in technology and multimodality therapy, the five-year overall survival (OS) rate has increased from 5% to 20% in the past three decades.1 Randomized controlled trials2–4 and meta-analyses5,6 suggest that preoperation chemoradiotherapy could induce a significant improvement in progression-free survival (PFS) and OS compared to surgery alone in patients with locally advanced esophageal cancer. Therefore, neoadjuvant chemoradiotherapy (nCRT) followed by radical resection has been recommended as the standard of care in the management of esophageal cancer.7,8 The interval between nCRT and esophagectomy allows for the resolution of acute inflammation, for patients to recover, and for such patients to be fit for surgery.9 To our knowledge, the optimal length of the interval between nCRT and esophagectomy remains unknown. Current evidence from several prospective randomized trials3,4,10 suggests that esophagectomy should be performed within 3–8 weeks after completion of nCRT. However, surgery is sometimes performed beyond this normal time interval due to the toxicity of nCRT, as well as personal and logistical reasons. Delaying surgery has raised important concerns about tumor growth, metastasis, and difficult dissection induced by increased radiation fibrosis.11 Nevertheless, the benefits of downstaging, increased pathologic complete response (pCR), and favorable long-term outcomes can be obtained by delaying surgery after nCRT in rectal cancer.12,13 Several retrospective studies have found that prolonging the interval between nCRT and esophagectomy resulted in improved pCR,14–18 reduced recurrence rate,19 and tumor downstaging,16 but paradoxically unchanged15–18 and even worse OS.14 Moreover, poor outcomes of increased mortality,20 decreased OS,20,21 and an increased incidence of anastomotic leak22–24 were seen in other studies. Additionally, most of the studies9,11,25–30 found no difference in clinical outcomes regarding the interval between nCRT and esophagectomy. Given the discrepancies of current evidence, we therefore decided to perform a systematic and meta-analysis to evaluate the influence of the prolonged interval between nCRT and esophagectomy on clinical outcomes in patients with esophageal cancer. METHODS AND MATERIALS The current systematic review and meta-analysis of cohort studies were conducted and reported following the guideline of MOOSE (meta-analysis of observational studies in epidemiology),31 and the protocol of the study was registered at PROSPERO as CRD42016048210 (https://www.crd.york.ac.uk/PROSPERO/). Literature search and selection criteria PubMed and Embase databases, from the inception to August 2016, were systematically searched to include all potentially eligible studies assessing the effect of the interval between nCRT and esophagectomy on clinical outcomes in patients with esophageal cancer. Free-text terms and subject terms of esophageal cancer, neoadjuvant chemoradiotherapy, and esophagectomy were used in combination, and the detailed search strategy is shown in Appendix 1 (Search strategy). Two investigators independently performed the literature search, removed duplicate records, and conducted study selection via screening the titles, abstracts, and fulltext of the retrieved studies. Additionally, references of included studies and relevant reviews were manually screened to identify additional eligible studies. Any discrepancies were resolved by consensus. Inclusion criteria were as follows: (a) Population: patients received esophagectomy after nCRT; (b) Exposure (Intervention): prolonged interval between nCRT and esophagectomy; (c) Control: short interval between nCRT and esophagectomy; (d) Outcome: no limitation; (e) Study design: cohort studies. Data extraction and outcome measure Two investigators independently extracted data from the included studies by using a predesigned standard form. The extracted information included the first author, year of publication, country, study periods, sample size, patients' baseline characteristics (patients age, tumor type, and tumor stage), scheme of nCRT, interval grouping, and outcomes (five-year OS, incidence of anastomotic complications, perioperative mortality, pCR rate, R0 resection rate, positive circumferential resection margin [CRM] rate). If essential data were not available, the corresponding author would be contacted repeatedly. In studies presenting the results as Kaplan–Meier curves, Engauge Digitizer software (http://digitizer.sourceforge.net/) was used to estimate the data. The extracted data were entered into a standardized Excel file and rechecked by two investigators. Discrepancies were eliminated by discussion. The primary outcomes were five-year OS. Secondary ones were the incidence of the anastomotic complications, perioperative mortality, pCR rate, R0 resection rate, and positive CRM rate. Quality assessment Newcastle–Ottawa scale (NOS),32 a scale that assesses the quality of nonrandomized studies, was used to assess risk of bias for the included cohorts. The scale consists of the selection of participants, comparability of study groups, and the outcomes, with a maximum of nine stars indicating the lowest risk of bias. Risks of bias were divided into three levels: high (0–3 stars), moderate (4–6 stars), and low (7–9 stars), respectively. Disagreements in quality assessment were settled by discussion. Statistical analysis Since all outcomes of the included cohorts were dichotomous variable, odds ratios (ORs) with 95% confidence intervals (CIs) were used to express the effect of the different intervals on clinical outcomes. Heterogeneity across the included studies was assessed by using the I2 statistic and Q test. The I2 statistic was perceived to be low (25% ≤ I2<50%), moderate (50% ≤ I2<75%), and high (I2 ≥ 75%), while P < 0.10 of Q test also suggested significant heterogeneity.33 The random-effects model was predesigned to be used for all meta-analyses regardless of heterogeneity, and sensitivity analyses were also performed. Publication bias was not evaluated because fewer than 10 studies were included for each outcome. A two-sided P < 0.05 was considered statistically significant, while P < 0.1 of the Q test for significant heterogeneity. All statistical analyses were carried out with Stata 12.0 software (StataCorp, College Station, TX, USA). RESULTS Study selection and identification The initial database search yielded a total of 2216 unique relevant publications, of which 2191 were excluded by screening the titles and abstracts. The remaining 25 articles were further assessed; seven of them were excluded since they were conference abstracts of the included ones, and two of them were excluded because they were performed to compare salvage and planned surgery in esophageal cancer. Additionally, two eligible studies were retrieved by hand screening. In the 18 eligible studies, four conference abstracts were excluded because relevant data were unavailable and could not be obtained from the authors despite repeated contact, and four were excluded because they were database research. Finally, ten cohort studies9,11,17,19,21,23–26,28 were included for meta-analysis. The study selection procedure is shown in Figure 1. Fig. 1 View largeDownload slide Flow chart for study selection. Fig. 1 View largeDownload slide Flow chart for study selection. Characteristics of eligible studies Baseline characteristics of ten included cohort studies are shown in Table 1, and the outcome data are presented in Appendix 2 (Table S1). All studies were retrospective cohort and published between 2008 and 2016, with the sample size varying from 69 to 665 and a total of 2383. Seven studies reported the nCRT, the chemotherapy regimens included platinum, paclitaxel, and fluorouracil, and total doses of radiotherapy ranged from 30 to 50.4 Gy. Cutoff of the prolonged interval between nCRT and esophagectomy varied from 30 to 60 days. The intervals were eight weeks in three studies,9,21,28 50 days in one study,26 seven weeks in one studies,11 30 days in two studies,23,24 and more than two groups in three studies.17,19,25 For the primary outcomes, six studies9,11,17,19,21,28 reported the five-year OS. While for secondary outcomes, six9,11,21,23,24,28 reported the incidence of anastomotic complications, six9,11,19,21,25,28 reported the perioperative mortality, eight,9,11,17,19,21,25,26,28 reported the pCR rate, four9,11,19,21 reported the R0 resection rate, and two21,25 reported the positive CRM rate. In one study, the positive CRM were determined according to the College of American Pathologists, which defined a positive CRM as tumor found at the surgical margin. While the definition of positive CRM was not mentioned in the other study. The quality of the six included cohorts was moderate with NOS scores of six stars because they were retrospective cohorts and the primary outcome was extracted from crude associations. The NOS scores for the remaining three cohorts were three stars in two studies and four in one since detailed items required for evaluation were missing. The detailed information of quality assessment is shown in Appendix 3 (Table S2). Table 1 Characteristics of the ten cohorts included in meta-analysis Study ID  Publishing type  Country; Study period  Neoadjuvant chemoradiotherapy  Sample  Sample (F/M)  Tumor type  cTNM stage  Interval grouping  Chiu et al.21  Full-text  China; 2002–2008  Chem: 5-fluorouracil + cisplatin; Radi: a total dose of 30 Gy  276  138 (6/132)138 (2/136)  SCC  II–IV  ≤8 weeks>8 weeks  Kathiravetpillai et al.9  Full-text  Netherlands; 2001–2014  Chem: Paclitaxel + carboplatin; Radi: a total dose of 41.4 Gy  190  65 (11/54)125 (10/115)  ADC  c(T1–T3)N0M0  ≤8 weeks>8 weeks  Kim et al.28  Full-text  America; 2002–2008  Chem: cisplatin-based in 69% of patients; Radi: a total dose of 45 Gy or a higher in 83% of patients  266  150 (15/135)116 (19/97)  SCC/ADC  II–IVa  ≤8 weeks>8 weeks  Parekh et al.24  Abstract  America; 2000–2007  Chem: NA; Radi: NA  69  35 (NA)34 (NA)  NA  NA  ≤30 days>30 days  Roh et al.23  Abstract  America; 2000–2013  Chem: NA; Radi: NA  216  13185  NA  NA  ≤30 days>30 days  Ruol et al.19  Full-text  Italy; 1998–2007  Chem: cisplatin + 5-fluorouracil; cisplatin + paclitaxel; oxaliplatin + 5-fluorouracil; Radi: a total dose of 45 to 50.4 Gy  129  66 (16/50);17 (6/11)63 (14/49);83 (17/66)NA;29 (7/22)  SCC  II–IV  ≤46 days>46 days  ≤30 days30–60 days61–90 days  Shaikh et al.17  Full-text  America; 2000–2011  Chem: 5-fluorouracil-based (71); paclitaxel -based (17); Radi: a total dose of 50.4 Gy  88  24 (3/21)20 (3/17)22 (2/20)22 (5/17)  SCC/ADC  I–IV  ≤45 days46–50 days51–63 days≥64 days  Singla et al.26  Abstract  America; 2004–2014  Chem: NA; Radi: NA  227  52 (NA)175 (NA)  NA  NA  <50 days>50 days  Tessier et al.11  Full-text  France; 1997–2011  Chem: 5-fluorouracil + cisplatin; Radi: a total dose of 45 Gy  257  122 (13/109)135 (17/118)  SCC/ADC  I–III  <7 weeks≥7 weeks  Wang et al.25  Full-text  China; 2008–2011  Chem: NA; Radi: an average total dose of 45.8 Gy  665  90 (5/85)385 (18/367)141 (3/138)49 (3/46)  SCC  I–III  <30 days30–59 days60–89 days≥90 days  Study ID  Publishing type  Country; Study period  Neoadjuvant chemoradiotherapy  Sample  Sample (F/M)  Tumor type  cTNM stage  Interval grouping  Chiu et al.21  Full-text  China; 2002–2008  Chem: 5-fluorouracil + cisplatin; Radi: a total dose of 30 Gy  276  138 (6/132)138 (2/136)  SCC  II–IV  ≤8 weeks>8 weeks  Kathiravetpillai et al.9  Full-text  Netherlands; 2001–2014  Chem: Paclitaxel + carboplatin; Radi: a total dose of 41.4 Gy  190  65 (11/54)125 (10/115)  ADC  c(T1–T3)N0M0  ≤8 weeks>8 weeks  Kim et al.28  Full-text  America; 2002–2008  Chem: cisplatin-based in 69% of patients; Radi: a total dose of 45 Gy or a higher in 83% of patients  266  150 (15/135)116 (19/97)  SCC/ADC  II–IVa  ≤8 weeks>8 weeks  Parekh et al.24  Abstract  America; 2000–2007  Chem: NA; Radi: NA  69  35 (NA)34 (NA)  NA  NA  ≤30 days>30 days  Roh et al.23  Abstract  America; 2000–2013  Chem: NA; Radi: NA  216  13185  NA  NA  ≤30 days>30 days  Ruol et al.19  Full-text  Italy; 1998–2007  Chem: cisplatin + 5-fluorouracil; cisplatin + paclitaxel; oxaliplatin + 5-fluorouracil; Radi: a total dose of 45 to 50.4 Gy  129  66 (16/50);17 (6/11)63 (14/49);83 (17/66)NA;29 (7/22)  SCC  II–IV  ≤46 days>46 days  ≤30 days30–60 days61–90 days  Shaikh et al.17  Full-text  America; 2000–2011  Chem: 5-fluorouracil-based (71); paclitaxel -based (17); Radi: a total dose of 50.4 Gy  88  24 (3/21)20 (3/17)22 (2/20)22 (5/17)  SCC/ADC  I–IV  ≤45 days46–50 days51–63 days≥64 days  Singla et al.26  Abstract  America; 2004–2014  Chem: NA; Radi: NA  227  52 (NA)175 (NA)  NA  NA  <50 days>50 days  Tessier et al.11  Full-text  France; 1997–2011  Chem: 5-fluorouracil + cisplatin; Radi: a total dose of 45 Gy  257  122 (13/109)135 (17/118)  SCC/ADC  I–III  <7 weeks≥7 weeks  Wang et al.25  Full-text  China; 2008–2011  Chem: NA; Radi: an average total dose of 45.8 Gy  665  90 (5/85)385 (18/367)141 (3/138)49 (3/46)  SCC  I–III  <30 days30–59 days60–89 days≥90 days  ADC, adenocarcinoma; Chem, chemotherapy; F, female; M, male; NA, not available; SCC, squamous cell carcinoma; Radi, radiotherapy. View Large Primary outcomes of five-year OS Six studies involving 1206 patients (621 in the prolonged interval, 585 in the short interval) contributed the outcome of five-year OS, and the pooled estimate of OR showed that the prolonged interval between nCRT and esophagectomy was not associated with five-year OS (OR 0.87, 95% CI 0.66 to 1.14, P = 0.30; Fig. 2), with low heterogeneity (P for heterogeneity = 0.78, I2 = 0%). The results remained consistent in sensitivity analysis by defining the cutoff as a different interval, as shown in Table 2. Fig. 2 View largeDownload slide Forest plot for 5-year overall survival. Fig. 2 View largeDownload slide Forest plot for 5-year overall survival. Table 2 Sensitivity analysis for primary outcomes according to different factors Sensitivity analysis  No. of study  No. of patients  I2  PH  OR (95%CI)  P value    5-year OS  All studies  6  1206  0%  0.78  0.87 (0.66–1.14)  0.30  Cutoff of 7–8 weeks  5  1077  0%  0.97  0.81 (0.61–1.08)  0.16  Cutoff of 8 weeks–60 days  4  861  0%  0.52  0.85 (0.62–1.17)  0.31  Cutoff of 30 days  1  129  –  –  10.7 (0.62–184.37)  0.10    Incidence of anastomotic complication  All studies  6  1274  11%  0.35  1.71 (1.15–2.54)  0.008  Anastomotic leak  5  1084  29%  0.23  1.73 (1.02– 2.94)  0.04  Cutoff of 7–8 weeks  4  989  0%  0.71  1.37 (0.89–2.09)  0.15  Cutoff of 30 days  2  285  0%  0.81  3.32 (1.59 to 6.96)  0.001  Sensitivity analysis  No. of study  No. of patients  I2  PH  OR (95%CI)  P value    5-year OS  All studies  6  1206  0%  0.78  0.87 (0.66–1.14)  0.30  Cutoff of 7–8 weeks  5  1077  0%  0.97  0.81 (0.61–1.08)  0.16  Cutoff of 8 weeks–60 days  4  861  0%  0.52  0.85 (0.62–1.17)  0.31  Cutoff of 30 days  1  129  –  –  10.7 (0.62–184.37)  0.10    Incidence of anastomotic complication  All studies  6  1274  11%  0.35  1.71 (1.15–2.54)  0.008  Anastomotic leak  5  1084  29%  0.23  1.73 (1.02– 2.94)  0.04  Cutoff of 7–8 weeks  4  989  0%  0.71  1.37 (0.89–2.09)  0.15  Cutoff of 30 days  2  285  0%  0.81  3.32 (1.59 to 6.96)  0.001  PH, P for heterogeneity; OR, odds ratio; OS, overall survival. View Large Secondary outcomes of the incidence of anastomotic complication Six studies involving 1274 patients reported the outcome of anastomotic complications. The combined estimate showed that the prolonged interval was significantly associated with increased incidence of anastomotic complications (OR 1.71, 95% CI 1.15 to 2.54, P = 0.008; Fig. 3), with low heterogeneity (P for heterogeneity = 0.35, I2 = 11%). Sensitivity analysis found that a significantly increased risk of the anastomotic leak was associated with the prolonged interval (OR 1.73, 95% CI 1.02 to 2.94, P = 0.004). The results of the sensitivity analysis are shown in Table 2. Fig. 3 View largeDownload slide Forest plot for anastomotic complication. Fig. 3 View largeDownload slide Forest plot for anastomotic complication. Secondary outcome of perioperative mortality In-hospital mortality, 30-day mortality, and postoperative mortality were treated as perioperative mortality. The pooled estimate of six studies with 1783 patients suggested that prolonged interval was not correlated with perioperative mortality (OR 1.20, 95% CI 0.79 to 1.83, P = 0.40; Fig. 4), with low heterogeneity (P for heterogeneity = 0.88, I2 = 0%). Fig. 4 View largeDownload slide Forest plot for perioperative mortality. Fig. 4 View largeDownload slide Forest plot for perioperative mortality. Secondary outcomes Prolonged interval was associated with decreased R0 resection rate (OR 0.60, 95% CI 0.41–0.88, P = 0.009; Appendix 4 (Fig. S1)), and increased positive CRM rate (OR 2.20, 95% CI 1.44–3.36, P < 0.001; Appendix 5 (Fig. S2)). However, no difference was observed in pCR rate (OR 1.02, 95% CI 0.78 to 1.33, P = 0.89; Appendix 6 (Fig. S3)). DISCUSSION Current meta-analysis suggests that the prolonged interval between nCRT and esophagectomy has no impact on five-year OS and perioperative mortality. Moreover, it has been associated with increased risk of anastomotic complications. The results are consistent in sensitivity analysis. Additionally, prolonged interval fails to increase the PCR rate, and, even worse it was correlated with decreased R0 resection rate and increased positive CRM rate. Four eligible studies with unavailable data are described in Table 3. Three of them found that the prolonged interval could increase the probability of pCR,18,27,30 and one found that it was associated with better relief in dysphagia and weight gain without an increase in postoperative complications.30 However, no effects on OS and DFS were found,18,30 and even worse it could increase the risk of postoperative complications18,22 and anastomotic leaks.22 In other studies, prolonged interval was associated with an increased trend of mortality. 22,27 The results of these studies are similar, and further consolidate our findings. They also show that the inconsistency in the increase of pCR was due to different statistical power caused by various sample sizes and other baseline characteristics. Table 3 Baseline characteristics and results of the four studies without available data Study ID  Article type  Patient  Grouping  Results  Gupta et al.30  Meeting article  117; mean interval 44.36 days  Group 1: <30 days; Group 2: 30–60 days; Group 3: 60–90 days  No effect on OS (P = 0.6) and DFS (P = 0.2); Better relief in dysphagia, better weight gain, and higher pCR rate without increase of postoperative complication  Muller et al.27  Meeting article  106; median interval 40 days  Group 1: ≤40 days; Group 2: >40 days  Trend toward a higher mortality risk (P = 0.064); Trend towards a higher pCR rate (P = 0.097).  Shapiro et al.18  Full text  325; median interval 48 days  Group 1: ≤45 days; Group 2: >45 days  Increased probability of pCR (OR 1.35 per additional week, P = 0.0004); Increased risk of postoperative complications (OR 1.20, per additional week, P < 0.001); No effect on DFS (HR 1.00, per additional week, P = 0.976) and OS (HR per additional week 1.06 P = 0.139)  Teman et al.22  Meeting article  457; Mean interval 49.99 days  Continuous/quintiles  Continuous: no impact on any of the outcome measures. Quintiles: increased postoperative pulmonary complications (P = 0.05), anastomotic leaks (P = 0.02) and a trend toward mortality (P = 0.09)  Study ID  Article type  Patient  Grouping  Results  Gupta et al.30  Meeting article  117; mean interval 44.36 days  Group 1: <30 days; Group 2: 30–60 days; Group 3: 60–90 days  No effect on OS (P = 0.6) and DFS (P = 0.2); Better relief in dysphagia, better weight gain, and higher pCR rate without increase of postoperative complication  Muller et al.27  Meeting article  106; median interval 40 days  Group 1: ≤40 days; Group 2: >40 days  Trend toward a higher mortality risk (P = 0.064); Trend towards a higher pCR rate (P = 0.097).  Shapiro et al.18  Full text  325; median interval 48 days  Group 1: ≤45 days; Group 2: >45 days  Increased probability of pCR (OR 1.35 per additional week, P = 0.0004); Increased risk of postoperative complications (OR 1.20, per additional week, P < 0.001); No effect on DFS (HR 1.00, per additional week, P = 0.976) and OS (HR per additional week 1.06 P = 0.139)  Teman et al.22  Meeting article  457; Mean interval 49.99 days  Continuous/quintiles  Continuous: no impact on any of the outcome measures. Quintiles: increased postoperative pulmonary complications (P = 0.05), anastomotic leaks (P = 0.02) and a trend toward mortality (P = 0.09)  DFS, disease-free survival; HR, hazard ratio; OR, odd ratios; OS, overall survival; pCR, pathologic complete response. View Large Table 4 summarizes four large-sample, database-based studies14–16,20 evaluating the effect of prolonged intervals between nCRT and esophagectomy on clinical outcomes. These studies were all performed by analyzing the data from the National Cancer Data Base, and sample sizes ranged from 4564 to 7900. In these four studies, the benefit of higher pCR14–16 and higher odds of tumor downstaging16 in the prolonged interval group were found, but one study also suggested that it had no effect on long-term OS, 30-day mortality and 30-day readmission.16 Moreover, prolonged interval was associated with a significantly increased risk of mortality and worse OS in three studies.14,15,20 Although, the studies reported a significant detrimental effect of a prolonged interval on mortality and OS with an advantage of large sample sizes (more than 4200 vs. 2383), important outcomes of anastomotic complications, R0 resection rate, and positive CRM rate were not assessed in any of them. Moreover, potential selection biases could not be accounted for because of the inherent nature of database-based study: their results could only indicate the associations and trends rather than practice-altering conclusions.20 Additionally, two of them14,16 were conference abstracts without detailed information. Combining the results of these studies with our own suggests that a prolonged interval between nCRT and esophagectomy might be associated with increased risk of mortality and worse OS, and the null association might be attributed to chance error induced by the limited sample size. Table 4 Baseline characteristics and results of the four database researches Study ID  Article type/Data source  Patient  Grouping  Results  Franko et al.20  Full-text; 2003–2011; NCDB  4284; median interval 7.1 weeks  <5 weeks; ≥5 and <7 weeks; ≥7 and <9 weeks; ≥9 weeks  Higher odds of 30-day mortality (aOR, 2.10; 95%CI, 1.10–4.22; P = 0.025); Higher odds of 90-day mortality (aOR, 1.91; 95%CI, 1.29–2.84; P < 0.001); Increased mortality risk (aHR, 1.19; 95%CI, 1.03–1.38; P = 0.017); Shorter OS in adjusted analyses  Hanna et al.14  Abstract; NR; NCDB  7900; NR  short vs. long  Higher rate of pCR; Worse OS  Lee et al.15  Full-text; 2003–2012; NCDB  5393; median interval 50 day  ≤40 days; 41–50 days; 51–63 days; ≥64 days  Increased pCR rate (12.3% for ≤40 days, 15.7% for 41–50 days, 16.8% for 51–63 days, and 18.3% for ≥64 days, P < 0.001) Worse OS in ≥64 days (HR, 1.16 95%CI: 1.01–1.33; P = 0.03)  Probst et al.16  Abstract; 2006–2013; NCDB  4564; median interval 51 day  <4 weeks; 4–6 weeks; 7–10 weeks; ≥11 weeks  Higher odds of tumor downstaging; Higher odds of pCR; No effect on long-term OS, 30-days mortality and 30-days readmission  Study ID  Article type/Data source  Patient  Grouping  Results  Franko et al.20  Full-text; 2003–2011; NCDB  4284; median interval 7.1 weeks  <5 weeks; ≥5 and <7 weeks; ≥7 and <9 weeks; ≥9 weeks  Higher odds of 30-day mortality (aOR, 2.10; 95%CI, 1.10–4.22; P = 0.025); Higher odds of 90-day mortality (aOR, 1.91; 95%CI, 1.29–2.84; P < 0.001); Increased mortality risk (aHR, 1.19; 95%CI, 1.03–1.38; P = 0.017); Shorter OS in adjusted analyses  Hanna et al.14  Abstract; NR; NCDB  7900; NR  short vs. long  Higher rate of pCR; Worse OS  Lee et al.15  Full-text; 2003–2012; NCDB  5393; median interval 50 day  ≤40 days; 41–50 days; 51–63 days; ≥64 days  Increased pCR rate (12.3% for ≤40 days, 15.7% for 41–50 days, 16.8% for 51–63 days, and 18.3% for ≥64 days, P < 0.001) Worse OS in ≥64 days (HR, 1.16 95%CI: 1.01–1.33; P = 0.03)  Probst et al.16  Abstract; 2006–2013; NCDB  4564; median interval 51 day  <4 weeks; 4–6 weeks; 7–10 weeks; ≥11 weeks  Higher odds of tumor downstaging; Higher odds of pCR; No effect on long-term OS, 30-days mortality and 30-days readmission  aHR, adjusted hazard ratio; aOR, adjusted odd ratios; OS, overall survival; pCR, pathologic complete response. View Large A previous meta-analysis found that the prolonged interval could decrease the R0 resection rate, but failed to impact the five-year OS, pCR rates, postoperative mortality, and anastomotic leak.34 In general, our study agreed with and further extended the partial results with six additional cohorts and 1,367 cases involved. Nevertheless, the risk of anastomotic complications, including anastomotic leak was significantly higher in the prolonged group, the inconsistency of anastomotic leak risk was mainly because of only three studies and 799 patients included in the previous meta-analysis. Additionally, a decreased R0 resection rate and concertedly increased positive CRM rate were found in our study. NCRT-induced tumor necrosis and regression may be strengthened by increasing the interval, and thus improved resectability and maximal effect of nCRT could be achieved. However, a long time delay may increase the risk of tumor repopulation, recurrence, and lose the benefit of tumor staging.21 Additionally, waiting longer could theoretically increase radiation-induced fibrosis,35 and subsequently make dissection more difficult. Difficult dissection in esophagectomy might be associated with an increased positive CRM rate, decreased the R0 resection rate, and the anastomosis complications. These detrimental effects of a longer interval were all observed in our meta-analysis. In rectal cancer, evidence12,36,37 has demonstrated that a longer interval could result in higher pCR rate, decreased postoperative morbidity and mortality, improved tumor down-staging, improved disease-free survival at a critical point, and decreased recurrence. Consistently, database research14–16 in our research, enrolling more than 4000 cases, supported a longer interval in increasing pCR. However, this point was not verified by the estimate of meta-analysis. We ascribed this controversy to the relatively small total sample size and various cutoff points for intervals in the included studies. In our meta-analysis, the cutoff points for intervals included 30 days, 50 days, 60 days, seven weeks, and eight weeks. Chiu et al.21 suggested that an eight-week interval was enough to get a maximal radiotherapy response and recovery from nCRT, since there was no increased rate of pCR, frequent closer CRM and non-R0 resection, increased amount of residual cancer, and decreased five-year OS in patients with complete pathological remission observed while the interval was more than eight weeks. Unlike rectal cancer, although an increased pCR rate was found in several reports; a longer interval not only failed to generate a favorable clinical outcome in patients with esophageal cancer,16,17 but also it could also lead to increased mortality and worsened OS.20 This suggests that delayed surgery might not be beneficial for other disease sites, and this organ-dependent response was also supported by another study, which concluded that lung surgery should be performed within six weeks after NCRT and that the longer interval resulted in worse clinical outcomes among patients with IIIA N2 non-small cell lung cancer.38 Several other reasons should be considered for potential failure of adopting longer interval for esophageal cancer. First, delaying surgery after nCRT is inevitable for various reasons, including reversible factors, such as surgery capacity and patient indecision, and medical objectives reasons (e.g., complications of nCRT and patient's condition).9,20,21 These medical factors could worsen clinical outcomes themselves and may mask the benefit of a longer interval. Second, significantly more advanced stage of esophageal cancers were noted in the delayed surgery group,11,25 and advanced stage cancer itself was associated with poor outcomes, which could confound the effect of the longer interval. Third, evidence indicated that esophageal squamous-cell carcinoma (SCC) was more chemoradiosensitive than adenocarcinoma (ADC).4 However, ADC and SCC were not analyzed separately. Owing to various confounding factors about the issue, a prospective randomized study may be necessary. Some limitations should be addressed. First, current meta-analysis was based on retrospective cohorts, and potential bias was always involved in the observation study. However, no substantial heterogeneity existed and our results remain similar to the large-sample database researches and conference abstracts included in the systematic review. Second, the cut-off point for the interval was diverse, which could confound the pooled estimate. Sensitivity analysis according to various intervals was performed, but the results of sensitivity analysis should be treated with caution because of decreasing the sample size. While in our study, all the results of sensitivity analysis, except the cut-offs of 7–8 weeks for anastomotic complications, remained still with the pooled estimate. Third, although the effect of nRCT was correlated to the type of esophageal cancer, both adenocarcinoma and squamous-cell cancers were enrolled in several included studies, and subgroup analysis could not be performed because of the sparse number of studies included. Considering the above limitations, the results should be interpreted with caution. In summary, current systematic review and meta-analysis suggested a prolonged interval between esophagectomy, and nCRT failed to result in better outcomes. Even worse, it could increase the risk of anastomotic complications and positive CRM rate and decrease the R0 resection rate. However, this conclusion should be interpreted with caution due to the limitations of retrospective observational study and substantial clinical heterogeneity. Accordingly, prospective randomized studies are urgently needed before a definitive conclusion can be reached. Notes Specific author contributions: Hongtao Tie, Feng He, and Jianfei Shen contributed equally to this work. Financial support: None. Conflicts of interest: The authors declare that they have no conflict of interest. Ethical approval: This article does not contain any studies with human participants performed by any of the authors. All analyses are based on previous published studies, thus ethical approval and patient consent are waived. SUPPLEMENTARY DATA Supplementary data are available at DOTESO online. Appendix 1 Search strategy Appendix 2 Table S1 Outcome data of the ten cohorts included in meta-analysis Appendix 3 Table S2 Quality assessment of the ten cohort included in meta-analysis Appendix 4 Fig S1 Forest plot for R0 resection rate Appendix 5 Fig S2 Forest plot for positive CRM rate Appendix 6 Fig S3 Forest plot for pathologic complete response References 1 Siegel R L, Miller K D, Jemal A. Cancer statistics, 2016. CA Cancer J Clin 2016 ; 66: 7– 30. CrossRef Search ADS   2 Bosset J F, Gignoux M, Triboulet J P et al.   Chemoradiotherapy followed by surgery compared with surgery alone in squamous-cell cancer of the esophagus. N Engl J Med  1997; 337: 161– 7. Google Scholar CrossRef Search ADS PubMed  3 Tepper J, Krasna M J, Niedzwiecki D et al.   Phase III trial of trimodality therapy with cisplatin, fluorouracil, radiotherapy, and surgery compared with surgery alone for esophageal cancer: CALGB 9781. J Clin Oncol  2008; 26: 1086– 92. Google Scholar CrossRef Search ADS PubMed  4 van Hagen P, Hulshof M C, van Lanschot J J et al.   Preoperative chemoradiotherapy for esophageal or junctional cancer. N Engl J Med  2012; 366: 2074– 84. Google Scholar CrossRef Search ADS PubMed  5 Gebski V, Burmeister B, Smithers B M et al.   Survival benefits from neoadjuvant chemoradiotherapy or chemotherapy in oesophageal carcinoma: a meta-analysis. Lancet Oncol  2007; 8: 226– 34. Google Scholar CrossRef Search ADS PubMed  6 Sjoquist K M, Burmeister B H, Smithers B M et al.   Survival after neoadjuvant chemotherapy or chemoradiotherapy for resectable oesophageal carcinoma: an updated meta-analysis. Lancet Oncol  2011; 12: 681– 92. Google Scholar CrossRef Search ADS PubMed  7 National Comprehensive Cancer Network. 2015 NCCN guidelines: Esophageal and esophagogastric junction cancers. Available at: https://www.nccn.org/professionals/physician_gls/f_guidelines.asp. 8 Little A G, Lerut A E, Harpole D H et al.   The Society of Thoracic Surgeons practice guidelines on the role of multimodality treatment for cancer of the esophagus and gastroesophageal junction. Ann Thorac Surg.  2014; 98: 1880– 5. Google Scholar CrossRef Search ADS PubMed  9 Kathiravetpillai N, Koeter M, van der Sangen M J et al.   Delaying surgery after neoadjuvant chemoradiotherapy does not significantly influence postoperative morbidity or oncological outcome in patients with oesophageal adenocarcinoma. Eur J Surg Oncol.  2016; 42: 1183– 90. Google Scholar CrossRef Search ADS PubMed  10 Burmeister B H, Smithers B M, Gebski V et al.   Surgery alone versus chemoradiotherapy followed by surgery for resectable cancer of the oesophagus: a randomised controlled phase III trial. Lancet Oncol  2005; 6: 659– 68. Google Scholar CrossRef Search ADS PubMed  11 Tessier W, Gronnier C, Messager M et al.   Does timing of surgical procedure after neoadjuvant chemoradiation affect outcomes in esophageal cancer? Ann Thorac Surg  2014; 97: 1181– 9. Google Scholar CrossRef Search ADS PubMed  12 Tulchinsky H, Shmueli E, Figer A et al.   An interval >7 weeks between neoadjuvant therapy and surgery improves pathologic complete response and disease-free survival in patients with locally advanced rectal cancer. Ann Surg Oncol  2008; 15: 2661– 7. Google Scholar CrossRef Search ADS PubMed  13 Kalady M F, de Campos-Lobato L F, Stocchi L et al.   Predictive factors of pathologic complete response after neoadjuvant chemoradiation for rectal cancer. Ann Surg  2009; 250: 582– 9. Google Scholar PubMed  14 Hanna A, Chuong M D, Søren B M et al.   An analysis of treatment timing parameters for patients receiving neoadjuvant chemoradiation followed by surgery for esophageal adenocarcinoma: a national cancer data base study. J Am Coll Surg  2015; 221: e126. Google Scholar CrossRef Search ADS   15 Lee A, Wong A T, Schwartz D et al.   Is there a benefit to prolonging the interval between neoadjuvant chemoradiation and esophagectomy in esophageal cancer? Ann Thorac Surg  2016; 102: 433– 8. Google Scholar CrossRef Search ADS PubMed  16 Probst C P, Aquina C T, Hensley B J et al.   Is more patience required between time from neoadjuvant therapy to esophagectomy? J Am Coll Surg  2015; 221: S150. Google Scholar CrossRef Search ADS   17 Shaikh T, Ruth K, Scott W J et al.   Increased time from neoadjuvant chemoradiation to surgery is associated with higher pathologic complete response rates in esophageal cancer. Ann Thorac Surg  2015; 99: 270– 6. Google Scholar CrossRef Search ADS PubMed  18 Shapiro J, van Hagen P, Lingsma H F et al.   Prolonged time to surgery after neoadjuvant chemoradiotherapy increases histopathological response without affecting survival in patients with esophageal or junctional cancer. Ann Surg  2014; 260: 807– 13. Google Scholar CrossRef Search ADS PubMed  19 Ruol A, Rizzetto C, Castoro C et al.   Interval between neoadjuvant chemoradiotherapy and surgery for squamous cell carcinoma of the thoracic esophagus: does delayed surgery have an impact on outcome? Ann Surg  2010; 252: 788– 95. Google Scholar CrossRef Search ADS PubMed  20 Franko J, Voynov G, Goldman C D. Esophagectomy timing after neoadjuvant therapy for distal esophageal adenocarcinoma. Ann Thorac Surg  2016; 101: 1123– 30. Google Scholar CrossRef Search ADS PubMed  21 Chiu C H, Chao Y K, Chang H K et al.   Interval between neoadjuvant chemoradiotherapy and surgery for esophageal squamous cell carcinoma: does delayed surgery impact outcome? Ann Surg Oncol  2013; 20: 4245– 51. Google Scholar CrossRef Search ADS PubMed  22 Teman N R, Silski L, Zhao L et al.   Delaying surgery for esophageal cancer increases postoperative complications. J Am Coll Surg  2013; 217: S35– 6. Google Scholar CrossRef Search ADS   23 Roh S, Iannettoni M, Keech J et al.   Timing of esophagectomy after neoadjuvant chemoradiation therapy affects clinically significant anastomotic leak rates. Ann Surg Oncol  2016; 23: S150. 24 Parekh K, F.T. V N T, Lynch W, Lannettoni M, Timing of esophagectomy after completion of neoadjuvant therapy affects anastomotic leak rates. Interactive Cardiovasc Thorac Surg  2008; S185. 25 Wang B Y, Chen H S, Hsu P K et al.   Clinical impact of the interval between chemoradiotherapy and esophagectomy in esophageal squamous cell carcinoma patients. Ann Thorac Surg  2015; 99: 947– 55. Google Scholar CrossRef Search ADS PubMed  26 Singla S, Kukar M, Alnaji R M et al.   Complete pathologic response is independent of the timing of esophagectomy and is predictive of improved survival following neoadjuvant chemoradiation for esophageal cancer. Ann Surg Oncol  2015; 22: S14. 27 Muller A K, Lenschow C, Palmes D et al.   Timing of esophagectomy in multimodal therapy of esophageal cancer: impact of time interval between neoadjuvant therapy and surgery on outcome and response. Chirurg  2015; 86: 874– 80. Google Scholar CrossRef Search ADS PubMed  28 Kim J Y, Correa A M, Vaporciyan A A et al.   Does the timing of esophagectomy after chemoradiation affect outcome? Ann Thorac Surg  2012; 93: 207– 13. Google Scholar CrossRef Search ADS PubMed  29 Howard T, Hsu H, Goldman C D et al.   Timing of esophagectomy after neoadjuvant treatment in squamous cell carcinoma. J Am Coll Surg  2015; 221: S17. Google Scholar CrossRef Search ADS   30 Gupta R, Shenvi S, Babu Y et al.   Impact of delayed surgery on outcome after neoadjuvant chemoradiotherapy in patients with carcinoma esophagus. J Gastroenterol Hepatol  2013; 28: 527. Google Scholar CrossRef Search ADS   31 Stroup D F, Berlin J A, Morton S C et al.   Meta-analysis of observational studies in epidemiology: a proposal for reporting. Meta-analysis of observational studies in epidemiology (MOOSE) group. JAMA  2000; 283: 2008– 12. Google Scholar CrossRef Search ADS PubMed  32 Wells G, Shea B, O’connell D et al.   The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses. in 3rd Symposium on Systematic Reviews: Beyond the Basics . 2000. 33 Higgins J P, Thompson S G, Deeks J J et al.   Measuring inconsistency in meta-analyses. BMJ  2003; 327: 557– 60. Google Scholar CrossRef Search ADS PubMed  34 Lin G, Han S Y, Xu Y P et al.   Increasing the interval between neoadjuvant chemoradiotherapy and surgery in esophageal cancer: a meta-analysis of published studies. Dis Esophagus  2015; 96: 75– 7. 35 Delanian S,, Lefaix J L. Current management for late normal tissue injury: radiation-induced fibrosis and necrosis. Semin Radiat Oncol  2007; 17: 99– 107. Google Scholar CrossRef Search ADS PubMed  36 Foster J D, Jones E L, Falk S et al.   Timing of surgery after long-course neoadjuvant chemoradiotherapy for rectal cancer: a systematic review of the literature. Dis Colon Rectum  2013; 56: 921– 30. Google Scholar CrossRef Search ADS PubMed  37 Kerr S F, Norton S, Glynne-Jones R. Delaying surgery after neoadjuvant chemoradiotherapy for rectal cancer may reduce postoperative morbidity without compromising prognosis. Br J Surg  2008; 95: 1534– 40. Google Scholar CrossRef Search ADS PubMed  38 Gao S J, Corso C D, Wang E H et al.   Timing of surgery after neoadjuvant chemoradiation in locally advanced non-small cell lung cancer. J Thorac Oncol.  2017; 12: 314– 32. Google Scholar CrossRef Search ADS PubMed  © The Authors 2017. Published by Oxford University Press on behalf of International Society for Diseases of the Esophagus. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Diseases of the Esophagus Oxford University Press

Prolonged interval between neoadjuvant chemoradiotherapy and esophagectomy does not benefit the outcome in esophageal cancer: a systematic review and meta-analysis

Loading next page...
 
/lp/ou_press/prolonged-interval-between-neoadjuvant-chemoradiotherapy-and-G3cW4IQlxo
Publisher
The International Society for Diseases of the Esophagus
Copyright
© The Authors 2017. Published by Oxford University Press on behalf of International Society for Diseases of the Esophagus. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com
ISSN
1120-8694
eISSN
1442-2050
D.O.I.
10.1093/dote/dox116
Publisher site
See Article on Publisher Site

Abstract

Summary Whether a prolonged interval between neoadjuvant chemoradiotherapy (nCRT) and esophagectomy could benefits conditions such as rectal cancer, still remains unknown. We therefore performed the current study to evaluate the influence of the interval between nCRT and esophagectomy on the clinical outcomes in patients with esophageal cancer. PubMed and Embase were searched to identify eligible cohort studies. The primary outcome was five-year overall survival (OS), and secondary outcomes included the incidence of anastomotic complications, perioperative mortality, pathologic complete response (pCR) rate, positive circumferential resection margin (CRM) rate, and R0 resection rate. A random-effects model was used for all meta-analyses irrespective of heterogeneity. Ten cohort studies with 2383 patients were included. Overall, the pooled estimate revealed that the prolonged interval has no impact on five-year OS (odds ratio (OR) 0.87, 95% CI 0.66 to 1.14, P = 0.30), with low heterogeneity (PH = 0.78, I2 = 0%). However, it was associated with an increased risk of anastomotic complication (OR 1.71, 95% CI 1.15 to 2.54, P = 0.008), with no effect on perioperative mortality (OR 1.20, 95% CI 0.79 to 1.83, P = 0.40). Additionally, the prolonged interval failed to increase the pCR rate (OR 1.02, 95% CI 0.78 to 1.33, P = 0.89). Even worse, it was correlated with a decreased R0 resection rate (OR 0.60, 95% CI 0.41 to 0.88, P = 0.009) and increased positive CRM rate (OR 2.20, 95% CI 1.44 to 3.36, P < 0.001). This study suggests that the prolonged interval between nCRT and esophagectomy fails to result in better outcomes, and in fact, could worsen clinical outcomes, with increasing anastomotic complications, and undermine resection completeness. However, this conclusion should be treated with caution because of the limitations of retrospective cohort study and substantial clinical heterogeneity. (The study was registered at PRESPERO as CRD42016048210). INTRODUCTION Esophageal cancer is one of the most aggressive malignancies, and surgical resection remains the essential cornerstone for patients with resectable disease. However, surgery alone is unsatisfactory and results in a poor prognosis in patients with locally advanced disease. With progresses in technology and multimodality therapy, the five-year overall survival (OS) rate has increased from 5% to 20% in the past three decades.1 Randomized controlled trials2–4 and meta-analyses5,6 suggest that preoperation chemoradiotherapy could induce a significant improvement in progression-free survival (PFS) and OS compared to surgery alone in patients with locally advanced esophageal cancer. Therefore, neoadjuvant chemoradiotherapy (nCRT) followed by radical resection has been recommended as the standard of care in the management of esophageal cancer.7,8 The interval between nCRT and esophagectomy allows for the resolution of acute inflammation, for patients to recover, and for such patients to be fit for surgery.9 To our knowledge, the optimal length of the interval between nCRT and esophagectomy remains unknown. Current evidence from several prospective randomized trials3,4,10 suggests that esophagectomy should be performed within 3–8 weeks after completion of nCRT. However, surgery is sometimes performed beyond this normal time interval due to the toxicity of nCRT, as well as personal and logistical reasons. Delaying surgery has raised important concerns about tumor growth, metastasis, and difficult dissection induced by increased radiation fibrosis.11 Nevertheless, the benefits of downstaging, increased pathologic complete response (pCR), and favorable long-term outcomes can be obtained by delaying surgery after nCRT in rectal cancer.12,13 Several retrospective studies have found that prolonging the interval between nCRT and esophagectomy resulted in improved pCR,14–18 reduced recurrence rate,19 and tumor downstaging,16 but paradoxically unchanged15–18 and even worse OS.14 Moreover, poor outcomes of increased mortality,20 decreased OS,20,21 and an increased incidence of anastomotic leak22–24 were seen in other studies. Additionally, most of the studies9,11,25–30 found no difference in clinical outcomes regarding the interval between nCRT and esophagectomy. Given the discrepancies of current evidence, we therefore decided to perform a systematic and meta-analysis to evaluate the influence of the prolonged interval between nCRT and esophagectomy on clinical outcomes in patients with esophageal cancer. METHODS AND MATERIALS The current systematic review and meta-analysis of cohort studies were conducted and reported following the guideline of MOOSE (meta-analysis of observational studies in epidemiology),31 and the protocol of the study was registered at PROSPERO as CRD42016048210 (https://www.crd.york.ac.uk/PROSPERO/). Literature search and selection criteria PubMed and Embase databases, from the inception to August 2016, were systematically searched to include all potentially eligible studies assessing the effect of the interval between nCRT and esophagectomy on clinical outcomes in patients with esophageal cancer. Free-text terms and subject terms of esophageal cancer, neoadjuvant chemoradiotherapy, and esophagectomy were used in combination, and the detailed search strategy is shown in Appendix 1 (Search strategy). Two investigators independently performed the literature search, removed duplicate records, and conducted study selection via screening the titles, abstracts, and fulltext of the retrieved studies. Additionally, references of included studies and relevant reviews were manually screened to identify additional eligible studies. Any discrepancies were resolved by consensus. Inclusion criteria were as follows: (a) Population: patients received esophagectomy after nCRT; (b) Exposure (Intervention): prolonged interval between nCRT and esophagectomy; (c) Control: short interval between nCRT and esophagectomy; (d) Outcome: no limitation; (e) Study design: cohort studies. Data extraction and outcome measure Two investigators independently extracted data from the included studies by using a predesigned standard form. The extracted information included the first author, year of publication, country, study periods, sample size, patients' baseline characteristics (patients age, tumor type, and tumor stage), scheme of nCRT, interval grouping, and outcomes (five-year OS, incidence of anastomotic complications, perioperative mortality, pCR rate, R0 resection rate, positive circumferential resection margin [CRM] rate). If essential data were not available, the corresponding author would be contacted repeatedly. In studies presenting the results as Kaplan–Meier curves, Engauge Digitizer software (http://digitizer.sourceforge.net/) was used to estimate the data. The extracted data were entered into a standardized Excel file and rechecked by two investigators. Discrepancies were eliminated by discussion. The primary outcomes were five-year OS. Secondary ones were the incidence of the anastomotic complications, perioperative mortality, pCR rate, R0 resection rate, and positive CRM rate. Quality assessment Newcastle–Ottawa scale (NOS),32 a scale that assesses the quality of nonrandomized studies, was used to assess risk of bias for the included cohorts. The scale consists of the selection of participants, comparability of study groups, and the outcomes, with a maximum of nine stars indicating the lowest risk of bias. Risks of bias were divided into three levels: high (0–3 stars), moderate (4–6 stars), and low (7–9 stars), respectively. Disagreements in quality assessment were settled by discussion. Statistical analysis Since all outcomes of the included cohorts were dichotomous variable, odds ratios (ORs) with 95% confidence intervals (CIs) were used to express the effect of the different intervals on clinical outcomes. Heterogeneity across the included studies was assessed by using the I2 statistic and Q test. The I2 statistic was perceived to be low (25% ≤ I2<50%), moderate (50% ≤ I2<75%), and high (I2 ≥ 75%), while P < 0.10 of Q test also suggested significant heterogeneity.33 The random-effects model was predesigned to be used for all meta-analyses regardless of heterogeneity, and sensitivity analyses were also performed. Publication bias was not evaluated because fewer than 10 studies were included for each outcome. A two-sided P < 0.05 was considered statistically significant, while P < 0.1 of the Q test for significant heterogeneity. All statistical analyses were carried out with Stata 12.0 software (StataCorp, College Station, TX, USA). RESULTS Study selection and identification The initial database search yielded a total of 2216 unique relevant publications, of which 2191 were excluded by screening the titles and abstracts. The remaining 25 articles were further assessed; seven of them were excluded since they were conference abstracts of the included ones, and two of them were excluded because they were performed to compare salvage and planned surgery in esophageal cancer. Additionally, two eligible studies were retrieved by hand screening. In the 18 eligible studies, four conference abstracts were excluded because relevant data were unavailable and could not be obtained from the authors despite repeated contact, and four were excluded because they were database research. Finally, ten cohort studies9,11,17,19,21,23–26,28 were included for meta-analysis. The study selection procedure is shown in Figure 1. Fig. 1 View largeDownload slide Flow chart for study selection. Fig. 1 View largeDownload slide Flow chart for study selection. Characteristics of eligible studies Baseline characteristics of ten included cohort studies are shown in Table 1, and the outcome data are presented in Appendix 2 (Table S1). All studies were retrospective cohort and published between 2008 and 2016, with the sample size varying from 69 to 665 and a total of 2383. Seven studies reported the nCRT, the chemotherapy regimens included platinum, paclitaxel, and fluorouracil, and total doses of radiotherapy ranged from 30 to 50.4 Gy. Cutoff of the prolonged interval between nCRT and esophagectomy varied from 30 to 60 days. The intervals were eight weeks in three studies,9,21,28 50 days in one study,26 seven weeks in one studies,11 30 days in two studies,23,24 and more than two groups in three studies.17,19,25 For the primary outcomes, six studies9,11,17,19,21,28 reported the five-year OS. While for secondary outcomes, six9,11,21,23,24,28 reported the incidence of anastomotic complications, six9,11,19,21,25,28 reported the perioperative mortality, eight,9,11,17,19,21,25,26,28 reported the pCR rate, four9,11,19,21 reported the R0 resection rate, and two21,25 reported the positive CRM rate. In one study, the positive CRM were determined according to the College of American Pathologists, which defined a positive CRM as tumor found at the surgical margin. While the definition of positive CRM was not mentioned in the other study. The quality of the six included cohorts was moderate with NOS scores of six stars because they were retrospective cohorts and the primary outcome was extracted from crude associations. The NOS scores for the remaining three cohorts were three stars in two studies and four in one since detailed items required for evaluation were missing. The detailed information of quality assessment is shown in Appendix 3 (Table S2). Table 1 Characteristics of the ten cohorts included in meta-analysis Study ID  Publishing type  Country; Study period  Neoadjuvant chemoradiotherapy  Sample  Sample (F/M)  Tumor type  cTNM stage  Interval grouping  Chiu et al.21  Full-text  China; 2002–2008  Chem: 5-fluorouracil + cisplatin; Radi: a total dose of 30 Gy  276  138 (6/132)138 (2/136)  SCC  II–IV  ≤8 weeks>8 weeks  Kathiravetpillai et al.9  Full-text  Netherlands; 2001–2014  Chem: Paclitaxel + carboplatin; Radi: a total dose of 41.4 Gy  190  65 (11/54)125 (10/115)  ADC  c(T1–T3)N0M0  ≤8 weeks>8 weeks  Kim et al.28  Full-text  America; 2002–2008  Chem: cisplatin-based in 69% of patients; Radi: a total dose of 45 Gy or a higher in 83% of patients  266  150 (15/135)116 (19/97)  SCC/ADC  II–IVa  ≤8 weeks>8 weeks  Parekh et al.24  Abstract  America; 2000–2007  Chem: NA; Radi: NA  69  35 (NA)34 (NA)  NA  NA  ≤30 days>30 days  Roh et al.23  Abstract  America; 2000–2013  Chem: NA; Radi: NA  216  13185  NA  NA  ≤30 days>30 days  Ruol et al.19  Full-text  Italy; 1998–2007  Chem: cisplatin + 5-fluorouracil; cisplatin + paclitaxel; oxaliplatin + 5-fluorouracil; Radi: a total dose of 45 to 50.4 Gy  129  66 (16/50);17 (6/11)63 (14/49);83 (17/66)NA;29 (7/22)  SCC  II–IV  ≤46 days>46 days  ≤30 days30–60 days61–90 days  Shaikh et al.17  Full-text  America; 2000–2011  Chem: 5-fluorouracil-based (71); paclitaxel -based (17); Radi: a total dose of 50.4 Gy  88  24 (3/21)20 (3/17)22 (2/20)22 (5/17)  SCC/ADC  I–IV  ≤45 days46–50 days51–63 days≥64 days  Singla et al.26  Abstract  America; 2004–2014  Chem: NA; Radi: NA  227  52 (NA)175 (NA)  NA  NA  <50 days>50 days  Tessier et al.11  Full-text  France; 1997–2011  Chem: 5-fluorouracil + cisplatin; Radi: a total dose of 45 Gy  257  122 (13/109)135 (17/118)  SCC/ADC  I–III  <7 weeks≥7 weeks  Wang et al.25  Full-text  China; 2008–2011  Chem: NA; Radi: an average total dose of 45.8 Gy  665  90 (5/85)385 (18/367)141 (3/138)49 (3/46)  SCC  I–III  <30 days30–59 days60–89 days≥90 days  Study ID  Publishing type  Country; Study period  Neoadjuvant chemoradiotherapy  Sample  Sample (F/M)  Tumor type  cTNM stage  Interval grouping  Chiu et al.21  Full-text  China; 2002–2008  Chem: 5-fluorouracil + cisplatin; Radi: a total dose of 30 Gy  276  138 (6/132)138 (2/136)  SCC  II–IV  ≤8 weeks>8 weeks  Kathiravetpillai et al.9  Full-text  Netherlands; 2001–2014  Chem: Paclitaxel + carboplatin; Radi: a total dose of 41.4 Gy  190  65 (11/54)125 (10/115)  ADC  c(T1–T3)N0M0  ≤8 weeks>8 weeks  Kim et al.28  Full-text  America; 2002–2008  Chem: cisplatin-based in 69% of patients; Radi: a total dose of 45 Gy or a higher in 83% of patients  266  150 (15/135)116 (19/97)  SCC/ADC  II–IVa  ≤8 weeks>8 weeks  Parekh et al.24  Abstract  America; 2000–2007  Chem: NA; Radi: NA  69  35 (NA)34 (NA)  NA  NA  ≤30 days>30 days  Roh et al.23  Abstract  America; 2000–2013  Chem: NA; Radi: NA  216  13185  NA  NA  ≤30 days>30 days  Ruol et al.19  Full-text  Italy; 1998–2007  Chem: cisplatin + 5-fluorouracil; cisplatin + paclitaxel; oxaliplatin + 5-fluorouracil; Radi: a total dose of 45 to 50.4 Gy  129  66 (16/50);17 (6/11)63 (14/49);83 (17/66)NA;29 (7/22)  SCC  II–IV  ≤46 days>46 days  ≤30 days30–60 days61–90 days  Shaikh et al.17  Full-text  America; 2000–2011  Chem: 5-fluorouracil-based (71); paclitaxel -based (17); Radi: a total dose of 50.4 Gy  88  24 (3/21)20 (3/17)22 (2/20)22 (5/17)  SCC/ADC  I–IV  ≤45 days46–50 days51–63 days≥64 days  Singla et al.26  Abstract  America; 2004–2014  Chem: NA; Radi: NA  227  52 (NA)175 (NA)  NA  NA  <50 days>50 days  Tessier et al.11  Full-text  France; 1997–2011  Chem: 5-fluorouracil + cisplatin; Radi: a total dose of 45 Gy  257  122 (13/109)135 (17/118)  SCC/ADC  I–III  <7 weeks≥7 weeks  Wang et al.25  Full-text  China; 2008–2011  Chem: NA; Radi: an average total dose of 45.8 Gy  665  90 (5/85)385 (18/367)141 (3/138)49 (3/46)  SCC  I–III  <30 days30–59 days60–89 days≥90 days  ADC, adenocarcinoma; Chem, chemotherapy; F, female; M, male; NA, not available; SCC, squamous cell carcinoma; Radi, radiotherapy. View Large Primary outcomes of five-year OS Six studies involving 1206 patients (621 in the prolonged interval, 585 in the short interval) contributed the outcome of five-year OS, and the pooled estimate of OR showed that the prolonged interval between nCRT and esophagectomy was not associated with five-year OS (OR 0.87, 95% CI 0.66 to 1.14, P = 0.30; Fig. 2), with low heterogeneity (P for heterogeneity = 0.78, I2 = 0%). The results remained consistent in sensitivity analysis by defining the cutoff as a different interval, as shown in Table 2. Fig. 2 View largeDownload slide Forest plot for 5-year overall survival. Fig. 2 View largeDownload slide Forest plot for 5-year overall survival. Table 2 Sensitivity analysis for primary outcomes according to different factors Sensitivity analysis  No. of study  No. of patients  I2  PH  OR (95%CI)  P value    5-year OS  All studies  6  1206  0%  0.78  0.87 (0.66–1.14)  0.30  Cutoff of 7–8 weeks  5  1077  0%  0.97  0.81 (0.61–1.08)  0.16  Cutoff of 8 weeks–60 days  4  861  0%  0.52  0.85 (0.62–1.17)  0.31  Cutoff of 30 days  1  129  –  –  10.7 (0.62–184.37)  0.10    Incidence of anastomotic complication  All studies  6  1274  11%  0.35  1.71 (1.15–2.54)  0.008  Anastomotic leak  5  1084  29%  0.23  1.73 (1.02– 2.94)  0.04  Cutoff of 7–8 weeks  4  989  0%  0.71  1.37 (0.89–2.09)  0.15  Cutoff of 30 days  2  285  0%  0.81  3.32 (1.59 to 6.96)  0.001  Sensitivity analysis  No. of study  No. of patients  I2  PH  OR (95%CI)  P value    5-year OS  All studies  6  1206  0%  0.78  0.87 (0.66–1.14)  0.30  Cutoff of 7–8 weeks  5  1077  0%  0.97  0.81 (0.61–1.08)  0.16  Cutoff of 8 weeks–60 days  4  861  0%  0.52  0.85 (0.62–1.17)  0.31  Cutoff of 30 days  1  129  –  –  10.7 (0.62–184.37)  0.10    Incidence of anastomotic complication  All studies  6  1274  11%  0.35  1.71 (1.15–2.54)  0.008  Anastomotic leak  5  1084  29%  0.23  1.73 (1.02– 2.94)  0.04  Cutoff of 7–8 weeks  4  989  0%  0.71  1.37 (0.89–2.09)  0.15  Cutoff of 30 days  2  285  0%  0.81  3.32 (1.59 to 6.96)  0.001  PH, P for heterogeneity; OR, odds ratio; OS, overall survival. View Large Secondary outcomes of the incidence of anastomotic complication Six studies involving 1274 patients reported the outcome of anastomotic complications. The combined estimate showed that the prolonged interval was significantly associated with increased incidence of anastomotic complications (OR 1.71, 95% CI 1.15 to 2.54, P = 0.008; Fig. 3), with low heterogeneity (P for heterogeneity = 0.35, I2 = 11%). Sensitivity analysis found that a significantly increased risk of the anastomotic leak was associated with the prolonged interval (OR 1.73, 95% CI 1.02 to 2.94, P = 0.004). The results of the sensitivity analysis are shown in Table 2. Fig. 3 View largeDownload slide Forest plot for anastomotic complication. Fig. 3 View largeDownload slide Forest plot for anastomotic complication. Secondary outcome of perioperative mortality In-hospital mortality, 30-day mortality, and postoperative mortality were treated as perioperative mortality. The pooled estimate of six studies with 1783 patients suggested that prolonged interval was not correlated with perioperative mortality (OR 1.20, 95% CI 0.79 to 1.83, P = 0.40; Fig. 4), with low heterogeneity (P for heterogeneity = 0.88, I2 = 0%). Fig. 4 View largeDownload slide Forest plot for perioperative mortality. Fig. 4 View largeDownload slide Forest plot for perioperative mortality. Secondary outcomes Prolonged interval was associated with decreased R0 resection rate (OR 0.60, 95% CI 0.41–0.88, P = 0.009; Appendix 4 (Fig. S1)), and increased positive CRM rate (OR 2.20, 95% CI 1.44–3.36, P < 0.001; Appendix 5 (Fig. S2)). However, no difference was observed in pCR rate (OR 1.02, 95% CI 0.78 to 1.33, P = 0.89; Appendix 6 (Fig. S3)). DISCUSSION Current meta-analysis suggests that the prolonged interval between nCRT and esophagectomy has no impact on five-year OS and perioperative mortality. Moreover, it has been associated with increased risk of anastomotic complications. The results are consistent in sensitivity analysis. Additionally, prolonged interval fails to increase the PCR rate, and, even worse it was correlated with decreased R0 resection rate and increased positive CRM rate. Four eligible studies with unavailable data are described in Table 3. Three of them found that the prolonged interval could increase the probability of pCR,18,27,30 and one found that it was associated with better relief in dysphagia and weight gain without an increase in postoperative complications.30 However, no effects on OS and DFS were found,18,30 and even worse it could increase the risk of postoperative complications18,22 and anastomotic leaks.22 In other studies, prolonged interval was associated with an increased trend of mortality. 22,27 The results of these studies are similar, and further consolidate our findings. They also show that the inconsistency in the increase of pCR was due to different statistical power caused by various sample sizes and other baseline characteristics. Table 3 Baseline characteristics and results of the four studies without available data Study ID  Article type  Patient  Grouping  Results  Gupta et al.30  Meeting article  117; mean interval 44.36 days  Group 1: <30 days; Group 2: 30–60 days; Group 3: 60–90 days  No effect on OS (P = 0.6) and DFS (P = 0.2); Better relief in dysphagia, better weight gain, and higher pCR rate without increase of postoperative complication  Muller et al.27  Meeting article  106; median interval 40 days  Group 1: ≤40 days; Group 2: >40 days  Trend toward a higher mortality risk (P = 0.064); Trend towards a higher pCR rate (P = 0.097).  Shapiro et al.18  Full text  325; median interval 48 days  Group 1: ≤45 days; Group 2: >45 days  Increased probability of pCR (OR 1.35 per additional week, P = 0.0004); Increased risk of postoperative complications (OR 1.20, per additional week, P < 0.001); No effect on DFS (HR 1.00, per additional week, P = 0.976) and OS (HR per additional week 1.06 P = 0.139)  Teman et al.22  Meeting article  457; Mean interval 49.99 days  Continuous/quintiles  Continuous: no impact on any of the outcome measures. Quintiles: increased postoperative pulmonary complications (P = 0.05), anastomotic leaks (P = 0.02) and a trend toward mortality (P = 0.09)  Study ID  Article type  Patient  Grouping  Results  Gupta et al.30  Meeting article  117; mean interval 44.36 days  Group 1: <30 days; Group 2: 30–60 days; Group 3: 60–90 days  No effect on OS (P = 0.6) and DFS (P = 0.2); Better relief in dysphagia, better weight gain, and higher pCR rate without increase of postoperative complication  Muller et al.27  Meeting article  106; median interval 40 days  Group 1: ≤40 days; Group 2: >40 days  Trend toward a higher mortality risk (P = 0.064); Trend towards a higher pCR rate (P = 0.097).  Shapiro et al.18  Full text  325; median interval 48 days  Group 1: ≤45 days; Group 2: >45 days  Increased probability of pCR (OR 1.35 per additional week, P = 0.0004); Increased risk of postoperative complications (OR 1.20, per additional week, P < 0.001); No effect on DFS (HR 1.00, per additional week, P = 0.976) and OS (HR per additional week 1.06 P = 0.139)  Teman et al.22  Meeting article  457; Mean interval 49.99 days  Continuous/quintiles  Continuous: no impact on any of the outcome measures. Quintiles: increased postoperative pulmonary complications (P = 0.05), anastomotic leaks (P = 0.02) and a trend toward mortality (P = 0.09)  DFS, disease-free survival; HR, hazard ratio; OR, odd ratios; OS, overall survival; pCR, pathologic complete response. View Large Table 4 summarizes four large-sample, database-based studies14–16,20 evaluating the effect of prolonged intervals between nCRT and esophagectomy on clinical outcomes. These studies were all performed by analyzing the data from the National Cancer Data Base, and sample sizes ranged from 4564 to 7900. In these four studies, the benefit of higher pCR14–16 and higher odds of tumor downstaging16 in the prolonged interval group were found, but one study also suggested that it had no effect on long-term OS, 30-day mortality and 30-day readmission.16 Moreover, prolonged interval was associated with a significantly increased risk of mortality and worse OS in three studies.14,15,20 Although, the studies reported a significant detrimental effect of a prolonged interval on mortality and OS with an advantage of large sample sizes (more than 4200 vs. 2383), important outcomes of anastomotic complications, R0 resection rate, and positive CRM rate were not assessed in any of them. Moreover, potential selection biases could not be accounted for because of the inherent nature of database-based study: their results could only indicate the associations and trends rather than practice-altering conclusions.20 Additionally, two of them14,16 were conference abstracts without detailed information. Combining the results of these studies with our own suggests that a prolonged interval between nCRT and esophagectomy might be associated with increased risk of mortality and worse OS, and the null association might be attributed to chance error induced by the limited sample size. Table 4 Baseline characteristics and results of the four database researches Study ID  Article type/Data source  Patient  Grouping  Results  Franko et al.20  Full-text; 2003–2011; NCDB  4284; median interval 7.1 weeks  <5 weeks; ≥5 and <7 weeks; ≥7 and <9 weeks; ≥9 weeks  Higher odds of 30-day mortality (aOR, 2.10; 95%CI, 1.10–4.22; P = 0.025); Higher odds of 90-day mortality (aOR, 1.91; 95%CI, 1.29–2.84; P < 0.001); Increased mortality risk (aHR, 1.19; 95%CI, 1.03–1.38; P = 0.017); Shorter OS in adjusted analyses  Hanna et al.14  Abstract; NR; NCDB  7900; NR  short vs. long  Higher rate of pCR; Worse OS  Lee et al.15  Full-text; 2003–2012; NCDB  5393; median interval 50 day  ≤40 days; 41–50 days; 51–63 days; ≥64 days  Increased pCR rate (12.3% for ≤40 days, 15.7% for 41–50 days, 16.8% for 51–63 days, and 18.3% for ≥64 days, P < 0.001) Worse OS in ≥64 days (HR, 1.16 95%CI: 1.01–1.33; P = 0.03)  Probst et al.16  Abstract; 2006–2013; NCDB  4564; median interval 51 day  <4 weeks; 4–6 weeks; 7–10 weeks; ≥11 weeks  Higher odds of tumor downstaging; Higher odds of pCR; No effect on long-term OS, 30-days mortality and 30-days readmission  Study ID  Article type/Data source  Patient  Grouping  Results  Franko et al.20  Full-text; 2003–2011; NCDB  4284; median interval 7.1 weeks  <5 weeks; ≥5 and <7 weeks; ≥7 and <9 weeks; ≥9 weeks  Higher odds of 30-day mortality (aOR, 2.10; 95%CI, 1.10–4.22; P = 0.025); Higher odds of 90-day mortality (aOR, 1.91; 95%CI, 1.29–2.84; P < 0.001); Increased mortality risk (aHR, 1.19; 95%CI, 1.03–1.38; P = 0.017); Shorter OS in adjusted analyses  Hanna et al.14  Abstract; NR; NCDB  7900; NR  short vs. long  Higher rate of pCR; Worse OS  Lee et al.15  Full-text; 2003–2012; NCDB  5393; median interval 50 day  ≤40 days; 41–50 days; 51–63 days; ≥64 days  Increased pCR rate (12.3% for ≤40 days, 15.7% for 41–50 days, 16.8% for 51–63 days, and 18.3% for ≥64 days, P < 0.001) Worse OS in ≥64 days (HR, 1.16 95%CI: 1.01–1.33; P = 0.03)  Probst et al.16  Abstract; 2006–2013; NCDB  4564; median interval 51 day  <4 weeks; 4–6 weeks; 7–10 weeks; ≥11 weeks  Higher odds of tumor downstaging; Higher odds of pCR; No effect on long-term OS, 30-days mortality and 30-days readmission  aHR, adjusted hazard ratio; aOR, adjusted odd ratios; OS, overall survival; pCR, pathologic complete response. View Large A previous meta-analysis found that the prolonged interval could decrease the R0 resection rate, but failed to impact the five-year OS, pCR rates, postoperative mortality, and anastomotic leak.34 In general, our study agreed with and further extended the partial results with six additional cohorts and 1,367 cases involved. Nevertheless, the risk of anastomotic complications, including anastomotic leak was significantly higher in the prolonged group, the inconsistency of anastomotic leak risk was mainly because of only three studies and 799 patients included in the previous meta-analysis. Additionally, a decreased R0 resection rate and concertedly increased positive CRM rate were found in our study. NCRT-induced tumor necrosis and regression may be strengthened by increasing the interval, and thus improved resectability and maximal effect of nCRT could be achieved. However, a long time delay may increase the risk of tumor repopulation, recurrence, and lose the benefit of tumor staging.21 Additionally, waiting longer could theoretically increase radiation-induced fibrosis,35 and subsequently make dissection more difficult. Difficult dissection in esophagectomy might be associated with an increased positive CRM rate, decreased the R0 resection rate, and the anastomosis complications. These detrimental effects of a longer interval were all observed in our meta-analysis. In rectal cancer, evidence12,36,37 has demonstrated that a longer interval could result in higher pCR rate, decreased postoperative morbidity and mortality, improved tumor down-staging, improved disease-free survival at a critical point, and decreased recurrence. Consistently, database research14–16 in our research, enrolling more than 4000 cases, supported a longer interval in increasing pCR. However, this point was not verified by the estimate of meta-analysis. We ascribed this controversy to the relatively small total sample size and various cutoff points for intervals in the included studies. In our meta-analysis, the cutoff points for intervals included 30 days, 50 days, 60 days, seven weeks, and eight weeks. Chiu et al.21 suggested that an eight-week interval was enough to get a maximal radiotherapy response and recovery from nCRT, since there was no increased rate of pCR, frequent closer CRM and non-R0 resection, increased amount of residual cancer, and decreased five-year OS in patients with complete pathological remission observed while the interval was more than eight weeks. Unlike rectal cancer, although an increased pCR rate was found in several reports; a longer interval not only failed to generate a favorable clinical outcome in patients with esophageal cancer,16,17 but also it could also lead to increased mortality and worsened OS.20 This suggests that delayed surgery might not be beneficial for other disease sites, and this organ-dependent response was also supported by another study, which concluded that lung surgery should be performed within six weeks after NCRT and that the longer interval resulted in worse clinical outcomes among patients with IIIA N2 non-small cell lung cancer.38 Several other reasons should be considered for potential failure of adopting longer interval for esophageal cancer. First, delaying surgery after nCRT is inevitable for various reasons, including reversible factors, such as surgery capacity and patient indecision, and medical objectives reasons (e.g., complications of nCRT and patient's condition).9,20,21 These medical factors could worsen clinical outcomes themselves and may mask the benefit of a longer interval. Second, significantly more advanced stage of esophageal cancers were noted in the delayed surgery group,11,25 and advanced stage cancer itself was associated with poor outcomes, which could confound the effect of the longer interval. Third, evidence indicated that esophageal squamous-cell carcinoma (SCC) was more chemoradiosensitive than adenocarcinoma (ADC).4 However, ADC and SCC were not analyzed separately. Owing to various confounding factors about the issue, a prospective randomized study may be necessary. Some limitations should be addressed. First, current meta-analysis was based on retrospective cohorts, and potential bias was always involved in the observation study. However, no substantial heterogeneity existed and our results remain similar to the large-sample database researches and conference abstracts included in the systematic review. Second, the cut-off point for the interval was diverse, which could confound the pooled estimate. Sensitivity analysis according to various intervals was performed, but the results of sensitivity analysis should be treated with caution because of decreasing the sample size. While in our study, all the results of sensitivity analysis, except the cut-offs of 7–8 weeks for anastomotic complications, remained still with the pooled estimate. Third, although the effect of nRCT was correlated to the type of esophageal cancer, both adenocarcinoma and squamous-cell cancers were enrolled in several included studies, and subgroup analysis could not be performed because of the sparse number of studies included. Considering the above limitations, the results should be interpreted with caution. In summary, current systematic review and meta-analysis suggested a prolonged interval between esophagectomy, and nCRT failed to result in better outcomes. Even worse, it could increase the risk of anastomotic complications and positive CRM rate and decrease the R0 resection rate. However, this conclusion should be interpreted with caution due to the limitations of retrospective observational study and substantial clinical heterogeneity. Accordingly, prospective randomized studies are urgently needed before a definitive conclusion can be reached. Notes Specific author contributions: Hongtao Tie, Feng He, and Jianfei Shen contributed equally to this work. Financial support: None. Conflicts of interest: The authors declare that they have no conflict of interest. Ethical approval: This article does not contain any studies with human participants performed by any of the authors. All analyses are based on previous published studies, thus ethical approval and patient consent are waived. SUPPLEMENTARY DATA Supplementary data are available at DOTESO online. Appendix 1 Search strategy Appendix 2 Table S1 Outcome data of the ten cohorts included in meta-analysis Appendix 3 Table S2 Quality assessment of the ten cohort included in meta-analysis Appendix 4 Fig S1 Forest plot for R0 resection rate Appendix 5 Fig S2 Forest plot for positive CRM rate Appendix 6 Fig S3 Forest plot for pathologic complete response References 1 Siegel R L, Miller K D, Jemal A. Cancer statistics, 2016. CA Cancer J Clin 2016 ; 66: 7– 30. CrossRef Search ADS   2 Bosset J F, Gignoux M, Triboulet J P et al.   Chemoradiotherapy followed by surgery compared with surgery alone in squamous-cell cancer of the esophagus. N Engl J Med  1997; 337: 161– 7. Google Scholar CrossRef Search ADS PubMed  3 Tepper J, Krasna M J, Niedzwiecki D et al.   Phase III trial of trimodality therapy with cisplatin, fluorouracil, radiotherapy, and surgery compared with surgery alone for esophageal cancer: CALGB 9781. J Clin Oncol  2008; 26: 1086– 92. Google Scholar CrossRef Search ADS PubMed  4 van Hagen P, Hulshof M C, van Lanschot J J et al.   Preoperative chemoradiotherapy for esophageal or junctional cancer. N Engl J Med  2012; 366: 2074– 84. Google Scholar CrossRef Search ADS PubMed  5 Gebski V, Burmeister B, Smithers B M et al.   Survival benefits from neoadjuvant chemoradiotherapy or chemotherapy in oesophageal carcinoma: a meta-analysis. Lancet Oncol  2007; 8: 226– 34. Google Scholar CrossRef Search ADS PubMed  6 Sjoquist K M, Burmeister B H, Smithers B M et al.   Survival after neoadjuvant chemotherapy or chemoradiotherapy for resectable oesophageal carcinoma: an updated meta-analysis. Lancet Oncol  2011; 12: 681– 92. Google Scholar CrossRef Search ADS PubMed  7 National Comprehensive Cancer Network. 2015 NCCN guidelines: Esophageal and esophagogastric junction cancers. Available at: https://www.nccn.org/professionals/physician_gls/f_guidelines.asp. 8 Little A G, Lerut A E, Harpole D H et al.   The Society of Thoracic Surgeons practice guidelines on the role of multimodality treatment for cancer of the esophagus and gastroesophageal junction. Ann Thorac Surg.  2014; 98: 1880– 5. Google Scholar CrossRef Search ADS PubMed  9 Kathiravetpillai N, Koeter M, van der Sangen M J et al.   Delaying surgery after neoadjuvant chemoradiotherapy does not significantly influence postoperative morbidity or oncological outcome in patients with oesophageal adenocarcinoma. Eur J Surg Oncol.  2016; 42: 1183– 90. Google Scholar CrossRef Search ADS PubMed  10 Burmeister B H, Smithers B M, Gebski V et al.   Surgery alone versus chemoradiotherapy followed by surgery for resectable cancer of the oesophagus: a randomised controlled phase III trial. Lancet Oncol  2005; 6: 659– 68. Google Scholar CrossRef Search ADS PubMed  11 Tessier W, Gronnier C, Messager M et al.   Does timing of surgical procedure after neoadjuvant chemoradiation affect outcomes in esophageal cancer? Ann Thorac Surg  2014; 97: 1181– 9. Google Scholar CrossRef Search ADS PubMed  12 Tulchinsky H, Shmueli E, Figer A et al.   An interval >7 weeks between neoadjuvant therapy and surgery improves pathologic complete response and disease-free survival in patients with locally advanced rectal cancer. Ann Surg Oncol  2008; 15: 2661– 7. Google Scholar CrossRef Search ADS PubMed  13 Kalady M F, de Campos-Lobato L F, Stocchi L et al.   Predictive factors of pathologic complete response after neoadjuvant chemoradiation for rectal cancer. Ann Surg  2009; 250: 582– 9. Google Scholar PubMed  14 Hanna A, Chuong M D, Søren B M et al.   An analysis of treatment timing parameters for patients receiving neoadjuvant chemoradiation followed by surgery for esophageal adenocarcinoma: a national cancer data base study. J Am Coll Surg  2015; 221: e126. Google Scholar CrossRef Search ADS   15 Lee A, Wong A T, Schwartz D et al.   Is there a benefit to prolonging the interval between neoadjuvant chemoradiation and esophagectomy in esophageal cancer? Ann Thorac Surg  2016; 102: 433– 8. Google Scholar CrossRef Search ADS PubMed  16 Probst C P, Aquina C T, Hensley B J et al.   Is more patience required between time from neoadjuvant therapy to esophagectomy? J Am Coll Surg  2015; 221: S150. Google Scholar CrossRef Search ADS   17 Shaikh T, Ruth K, Scott W J et al.   Increased time from neoadjuvant chemoradiation to surgery is associated with higher pathologic complete response rates in esophageal cancer. Ann Thorac Surg  2015; 99: 270– 6. Google Scholar CrossRef Search ADS PubMed  18 Shapiro J, van Hagen P, Lingsma H F et al.   Prolonged time to surgery after neoadjuvant chemoradiotherapy increases histopathological response without affecting survival in patients with esophageal or junctional cancer. Ann Surg  2014; 260: 807– 13. Google Scholar CrossRef Search ADS PubMed  19 Ruol A, Rizzetto C, Castoro C et al.   Interval between neoadjuvant chemoradiotherapy and surgery for squamous cell carcinoma of the thoracic esophagus: does delayed surgery have an impact on outcome? Ann Surg  2010; 252: 788– 95. Google Scholar CrossRef Search ADS PubMed  20 Franko J, Voynov G, Goldman C D. Esophagectomy timing after neoadjuvant therapy for distal esophageal adenocarcinoma. Ann Thorac Surg  2016; 101: 1123– 30. Google Scholar CrossRef Search ADS PubMed  21 Chiu C H, Chao Y K, Chang H K et al.   Interval between neoadjuvant chemoradiotherapy and surgery for esophageal squamous cell carcinoma: does delayed surgery impact outcome? Ann Surg Oncol  2013; 20: 4245– 51. Google Scholar CrossRef Search ADS PubMed  22 Teman N R, Silski L, Zhao L et al.   Delaying surgery for esophageal cancer increases postoperative complications. J Am Coll Surg  2013; 217: S35– 6. Google Scholar CrossRef Search ADS   23 Roh S, Iannettoni M, Keech J et al.   Timing of esophagectomy after neoadjuvant chemoradiation therapy affects clinically significant anastomotic leak rates. Ann Surg Oncol  2016; 23: S150. 24 Parekh K, F.T. V N T, Lynch W, Lannettoni M, Timing of esophagectomy after completion of neoadjuvant therapy affects anastomotic leak rates. Interactive Cardiovasc Thorac Surg  2008; S185. 25 Wang B Y, Chen H S, Hsu P K et al.   Clinical impact of the interval between chemoradiotherapy and esophagectomy in esophageal squamous cell carcinoma patients. Ann Thorac Surg  2015; 99: 947– 55. Google Scholar CrossRef Search ADS PubMed  26 Singla S, Kukar M, Alnaji R M et al.   Complete pathologic response is independent of the timing of esophagectomy and is predictive of improved survival following neoadjuvant chemoradiation for esophageal cancer. Ann Surg Oncol  2015; 22: S14. 27 Muller A K, Lenschow C, Palmes D et al.   Timing of esophagectomy in multimodal therapy of esophageal cancer: impact of time interval between neoadjuvant therapy and surgery on outcome and response. Chirurg  2015; 86: 874– 80. Google Scholar CrossRef Search ADS PubMed  28 Kim J Y, Correa A M, Vaporciyan A A et al.   Does the timing of esophagectomy after chemoradiation affect outcome? Ann Thorac Surg  2012; 93: 207– 13. Google Scholar CrossRef Search ADS PubMed  29 Howard T, Hsu H, Goldman C D et al.   Timing of esophagectomy after neoadjuvant treatment in squamous cell carcinoma. J Am Coll Surg  2015; 221: S17. Google Scholar CrossRef Search ADS   30 Gupta R, Shenvi S, Babu Y et al.   Impact of delayed surgery on outcome after neoadjuvant chemoradiotherapy in patients with carcinoma esophagus. J Gastroenterol Hepatol  2013; 28: 527. Google Scholar CrossRef Search ADS   31 Stroup D F, Berlin J A, Morton S C et al.   Meta-analysis of observational studies in epidemiology: a proposal for reporting. Meta-analysis of observational studies in epidemiology (MOOSE) group. JAMA  2000; 283: 2008– 12. Google Scholar CrossRef Search ADS PubMed  32 Wells G, Shea B, O’connell D et al.   The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses. in 3rd Symposium on Systematic Reviews: Beyond the Basics . 2000. 33 Higgins J P, Thompson S G, Deeks J J et al.   Measuring inconsistency in meta-analyses. BMJ  2003; 327: 557– 60. Google Scholar CrossRef Search ADS PubMed  34 Lin G, Han S Y, Xu Y P et al.   Increasing the interval between neoadjuvant chemoradiotherapy and surgery in esophageal cancer: a meta-analysis of published studies. Dis Esophagus  2015; 96: 75– 7. 35 Delanian S,, Lefaix J L. Current management for late normal tissue injury: radiation-induced fibrosis and necrosis. Semin Radiat Oncol  2007; 17: 99– 107. Google Scholar CrossRef Search ADS PubMed  36 Foster J D, Jones E L, Falk S et al.   Timing of surgery after long-course neoadjuvant chemoradiotherapy for rectal cancer: a systematic review of the literature. Dis Colon Rectum  2013; 56: 921– 30. Google Scholar CrossRef Search ADS PubMed  37 Kerr S F, Norton S, Glynne-Jones R. Delaying surgery after neoadjuvant chemoradiotherapy for rectal cancer may reduce postoperative morbidity without compromising prognosis. Br J Surg  2008; 95: 1534– 40. Google Scholar CrossRef Search ADS PubMed  38 Gao S J, Corso C D, Wang E H et al.   Timing of surgery after neoadjuvant chemoradiation in locally advanced non-small cell lung cancer. J Thorac Oncol.  2017; 12: 314– 32. Google Scholar CrossRef Search ADS PubMed  © The Authors 2017. Published by Oxford University Press on behalf of International Society for Diseases of the Esophagus. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com

Journal

Diseases of the EsophagusOxford University Press

Published: Jan 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 12 million articles from more than
10,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Unlimited reading

Read as many articles as you need. Full articles with original layout, charts and figures. Read online, from anywhere.

Stay up to date

Keep up with your field with Personalized Recommendations and Follow Journals to get automatic updates.

Organize your research

It’s easy to organize your research with our built-in tools.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve Freelancer

DeepDyve Pro

Price
FREE
$49/month

$360/year
Save searches from
Google Scholar,
PubMed
Create lists to
organize your research
Export lists, citations
Read DeepDyve articles
Abstract access only
Unlimited access to over
18 million full-text articles
Print
20 pages/month
PDF Discount
20% off