Neoadjuvant or adjuvant therapy for resectable esophageal cancer: a systematic review and meta-analysis

Background Carcinoma of the esophagus is an aggressive malignancy with an increasing incidence. Its virulence, in terms of symptoms and mortality, justifies a continued search for optimal therapy. The large and growing number of patients affected, the high mortality rates, the worldwide geographic variation in practice, and the large body of good quality research warrants a systematic review with meta-analysis. Methods A systematic review and meta-analysis investigating the impact of neoadjuvant or adjuvant therapy on resectable thoracic esophageal cancer to inform evidence-based practice was produced. MEDLINE, CANCERLIT, Cochrane Library, EMBASE, and abstracts from the American Society of Clinical Oncology and the American Society for Therapeutic Radiology and Oncology were searched for trial reports. Included were randomized trials or meta-analyses of neoadjuvant or adjuvant treatments compared with surgery alone or other treatments in patients with resectable thoracic esophageal cancer. Outcomes of interest were survival, adverse effects, and quality of life. Either one- or three-year mortality data were pooled and reported as relative risk ratios. Results Thirty-four randomized controlled trials and six meta-analyses were obtained and grouped into 13 basic treatment approaches. Single randomized controlled trials detected no differences in mortality between treatments for the following comparisons: - Preoperative radiotherapy versus postoperative radiotherapy. - Preoperative and postoperative radiotherapy versus postoperative radiotherapy. Preoperative and postoperative radiotherapy was associated with a significantly higher mortality rate. - Postoperative chemotherapy versus postoperative radiotherapy. - Postoperative radiotherapy versus postoperative radiotherapy plus protein-bound polysaccharide versus chemoradiation versus chemoradiation plus protein-bound polysaccharide. Pooling one-year mortality detected no statistically significant differences in mortality between treatments for the following comparisons: - Preoperative radiotherapy compared with surgery alone (five randomized trials). - Postoperative radiotherapy compared with surgery alone (five randomized trials). - Preoperative chemotherapy versus surgery alone (six randomized trials). - Preoperative and postoperative chemotherapy versus surgery alone (two randomized trials). - Preoperative chemoradiation therapy versus surgery alone (six randomized trials). Single randomized controlled trials detected differences in mortality between treatments for the following comparison: - Preoperative hyperthermia and chemoradiotherapy versus preoperative chemoradiotherapy in favour of hyperthermia. Pooling three-year mortality detected no statistically significant difference in mortality between treatments for the following comparison: - Postoperative chemotherapy compared with surgery alone (two randomized trials). Pooling three-year mortality detected statistically significant differences between treatments for the following comparisons: - Preoperative chemoradiation therapy versus surgery alone (six randomized trials) in favour of preoperative chemoradiation with surgery. - Preoperative chemotherapy compared with preoperative radiotherapy (one randomized trial) in favour of preoperative radiotherapy. Conclusion For adult patients with resectable thoracic esophageal cancer for whom surgery is considered appropriate, surgery alone (i.e., without neoadjuvant or adjuvant therapy) is recommended as the standard practice.

-Postoperative radiotherapy versus postoperative radiotherapy plus protein-bound polysaccharide versus chemoradiation versus chemoradiation plus protein-bound polysaccharide.
Pooling one-year mortality detected no statistically significant differences in mortality between treatments for the following comparisons: -Preoperative radiotherapy compared with surgery alone (five randomized trials).
-Postoperative radiotherapy compared with surgery alone (five randomized trials).
-Preoperative and postoperative chemotherapy versus surgery alone (two randomized trials).
Single randomized controlled trials detected differences in mortality between treatments for the following comparison: -Preoperative hyperthermia and chemoradiotherapy versus preoperative chemoradiotherapy in favour of hyperthermia.
Pooling three-year mortality detected no statistically significant difference in mortality between treatments for the following comparison: -Postoperative chemotherapy compared with surgery alone (two randomized trials).
Pooling three-year mortality detected statistically significant differences between treatments for the following comparisons: -Preoperative chemoradiation therapy versus surgery alone (six randomized trials) in favour of preoperative chemoradiation with surgery.
-Preoperative chemotherapy compared with preoperative radiotherapy (one randomized trial) in favour of preoperative radiotherapy.

Conclusion:
For adult patients with resectable thoracic esophageal cancer for whom surgery is considered appropriate, surgery alone (i.e., without neoadjuvant or adjuvant therapy) is recommended as the standard practice.

Background
Carcinoma of the esophagus is an aggressive malignancy that continues to kill more than 90% of people with the disease within five years [1]. The incidence of adenocarcinoma of the esophagus is rising faster than any other malignancy [2]. In 2001, there were at least 1,450 deaths due to esophageal cancer in Canada and many more people suffered because of the disease [3]. Its virulence, in terms of symptoms and mortality, justifies a continued search for optimal therapy. Surgical esophagectomy remains the preferred treatment for clinically localized thoracic esophageal carcinoma [1, [4][5][6]. Two randomized trials comparing surgery alone to radiation alone found surgery to be the better treatment for resectable cancer [5,6]. Fok et al randomly assigned 39 patients to surgery and 35 patients to 45 to 53 Gy radiation over four to five weeks [5]. The median survival time and five-year survival rate for surgery were 21.6 months and 16%, respectively, compared with 8.2 months and 7% for radiation (p < 0.05). Badwe et al compared 47 surgical patients to 52 patients undergoing 50 Gy radiation in 28 fractions plus 15 Gy boost in 8 fractions or 15 Gy brachytherapy [6]. Overall survival was better with surgery (odds ratio [OR], 2.74; 95% confidence interval [CI], 1.51 to 4.98; log-rank p = 0.002). The swallowing status was better in the surgery arm at six months after treatment (p = 0.03). Survival data from these two trials were pooled. The pooled results favoured surgery alone. There was no statistical heterogeneity (X 2 = 0.02, p = 0.9) and a 52% relative increase in the risk of death at three years with radiotherapy compared with surgery alone (relative risk ratio [RR], 1.52; 95% CI, 1.23 to 1.86; p = 0.0007).
The failure of surgery alone is attributed to the systemic nature of the disease at the time of presentation [7,8]. Early and effective systemic chemotherapy and local radiotherapy, directed at micro-metastases and added to surgical resection, could lead to increased survival. Many clinical trials have evaluated the role of adjuvant therapy, both preoperatively and postoperatively, with conflicting results. Patients with cervical esophageal cancer are generally treated with chemoradiation, either preoperatively or postoperatively, in an attempt to avoid a laryngoesophagectomy and preserve the larynx. Although the majority of studies have been performed in squamous cell carcinomas, adenocarcinomas were included in some studies, but a distinction between the two histological subtypes was not made in this guideline report because previous studies have not consistently found that they respond differently to chemotherapy or radiation [9][10][11][12][13][14][15][16][17].
The large and growing number of patients affected, the high mortality rates, the geographic variation in practice, and the large body of good quality research evidence warrants a systematic review with meta-analysis.

Methods
This systematic review was developed by the Practice Guidelines Initiative (PGI) of Cancer Care Ontario's Program in Evidence-based Care (PEBC). Evidence was selected and reviewed by two members of the PGI's Gastrointestinal Cancer Disease Site Group (DSG) and two methodologists. Members of the Gastrointestinal Cancer DSG disclosed potential conflict of interest information.
This systematic review is a convenient and up-to-date source of the best available evidence on neoadjuvant or adjuvant therapy for resectable esophageal cancer. The body of evidence in this systematic review is primarily comprised of mature randomized controlled trial data; it forms the basis of a clinical practice guideline developed by the Gastrointestinal Cancer DSG published elsewhere (18). This systematic review and companion practice guideline are intended to promote evidence-based practice in Ontario, Canada. The PGI is editorially independent of Cancer Care Ontario and the Ontario Ministry of Health and Long-Term Care.  40,2003) databases were searched with no language restrictions. "Esophageal neoplasms" (Medical subject heading (MeSH)) was combined with "chemotherapy, adjuvant" (MeSH), "radiotherapy, adjuvant" (MeSH), "immunotherapy, adjuvant" (MeSH), and each of the following phrases used as text words: "preoperative", "neoadjuvant", "chemotherapy", "radiotherapy", "radiation therapy", "irradiation", "immunotherapy", "chemoradiotherapy", "chemoradiation", and "hyperthermia". These terms were then combined with the search terms for the following study designs or publication types: practice guidelines, meta-analyses, and rand-omized controlled trials. Additionally, the conference proceedings of the 1997 to 2003 annual meetings of the American Society of Clinical Oncology (ASCO) and the 1999 to 2002 annual meetings of the American Society for Therapeutic Radiology and Oncology (ASTRO) were searched for reports of new or ongoing trials. Relevant articles and abstracts were reviewed, and the reference lists from these sources were searched for additional trials. This formal search was supplemented with published abstracts from thoracic surgery and oncology conferences, conversations with colleagues and experts in the field, and a review of textbooks related to esophageal oncology.

Study selection criteria
Articles were included in this systematic review if they were fully published reports, abstracts, or meta-analyses of randomized controlled trials (RCT) of neoadjuvant or adjuvant treatments compared with surgery alone or surgery plus another treatment in patients with resectable and operable thoracic esophageal cancer. Data on survival had to be reported. Other outcomes of interest were adverse effects and quality of life. Reports of carcinomas located in the cervical esophagus were excluded.

Synthesizing the evidence
Because diverse treatment strategies were evaluated, the eligible studies were grouped into 13 basic treatment approaches (Table 1), and each group was examined separately. Pooling was conducted using one-year mortality data for all meta-analyses except for the comparison of post-operative chemotherapy versus surgery alone, which was pooled at three years. Any time point selected for meta-analyses must be clinically credible and relevant but not so far along the survival curve that wide confidence intervals result from fewer patients contributing to the estimate. Since time points prior to the median will generally ensure that there are sufficient data to be credible, median survival times, weighted by the size of the treatment arms, were calculated to determine the time point for each meta-analysis as recommended in the literature [19]. Studies that did not provide values for survival at the time of pooling were not included in each meta-analysis, although they were included in calculating the weighted median survival time, if values for median survival were provided. . Pooled results were expressed as mortality RR with 95% CI using the random effects model. An RR less than 1.0 favours neoadjuvant or adjuvant treatment, and an RR greater than 1.0 favours surgery alone. All analyses were made based on the intent-to-treat principle, except where only evaluable patient data were available.

Potential sources of heterogeneity and sensitivity analysis
Heterogeneity of study results was assessed using a visual plot of the outcomes and by calculating the X 2 (Chisquare) statistic using a planned cut-off for significance of p < 0.05. Potential sources of heterogeneity were postulated a priori and included study quality assessed with the Jadad scale [20] (>2 versus ≤2), full article publication versus abstract publication, squamous cell versus adenocarcinoma, type of chemotherapy (cisplatin-containing versus others), type of surgery (transthoracic versus transhiatal), and radiotherapy dose (BED>48 versus BED<48). To facilitate comparison across trials, radiotherapy dose was converted to biological equivalent dose (same as biological effective dose) using the equation BED = nd (1+d/α/β), where n = number of fractions, d = dose per fraction, and it is assumed that α/β = 10 for tumour effect. Due to limitations inherent with this model, no allowances can be made for any time gaps in split-course treatments. These factors were used to explore any significant heterogeneity of results across the trials. The robustness of our conclusions was examined through subsequent sensitivity analyses using these factors. The sensitivity analysis results are not detailed, as they would not change the conclusions.

Preoperative radiotherapy and surgery versus surgery alone
Six randomized trials of preoperative radiotherapy and surgery versus surgery alone are presented in Table 2 [5,[21][22][23][24][25]. The radiotherapy regimens varied, using low to moderate doses ranging from 20 Gy in 10 fractions to 53 Gy in 20 fractions. Treatment was delivered between one to four weeks prior to surgery. Quality-of-life assessments were not conducted in any of the six trials. The Gastrointestinal Cancer DSG pooled the five trials that reported one-year mortality data [5,21,22,24,25] (Figure 1). No statistically significant difference in the risk of mortality with preoperative radiotherapy at one year compared with surgery alone was detected (RR, 1.01; 95% CI, 0.88 to 1.16; p = 0.87). No statistical heterogeneity was detected (X 2 = 3.61, p = 0.46).
A published meta-analysis [26] using updated individual patient data on 1147 patients from five trials [21][22][23][24][25] detected a hazard ratio for death of 0.89 (95% CI, 0.78 to 1.01; p = 0.062) for preoperative radiotherapy compared with surgery alone. This meta-analysis included addi- tional patients from the study by Wang et al. [23] with no description of why these patients were excluded from the published report of the trial (a total of 418 patients from this study were included in the meta-analysis versus 206 included in the trial report). The trial by Fok et al. [5] was not included in the published meta-analysis.

Postoperative radiotherapy and surgery versus surgery alone
Five randomized trials of surgery and postoperative radiotherapy compared with surgery alone are presented in Table 3[5, [27][28][29]47]. Although all studies specified the absence of distant metastases as an inclusion criterion, Zieren et al. [29] and Teniere et al. [28] included patients with celiac node involvement (M1 disease). Fok et al. [27] [28,29,47] demonstrated no significant difference in survival while another [27] found significantly shorter survival with postoperative radiotherapy and surgery compared with surgery alone.
The Gastrointestinal Cancer DSG pooled the five trials that reported one-year mortality data [5,[27][28][29],47] ( Figure   2). No significant difference in the risk of mortality with postoperative radiotherapy and surgery at one year compared with surgery alone was detected (RR, 1.23; 95% CI, 0.95 to 1.59; p = 0.11). No significant statistical heterogeneity was detected (X 2 = 7.53, p = 0.11). The rate of local recurrence with radiotherapy was lower in three of the trials [27,28,47], but two trials [27,28] noted this benefit was achieved at the expense of increased morbidity.

Preoperative radiotherapy versus postoperative radiotherapy
One randomized trial evaluated preoperative radiotherapy versus postoperative radiotherapy with curative esophagectomy as part of a four-arm study [5]. Patients in this trial, performed between 1968 and 1981, received from 24 to 53 Gy preoperatively (n = 40) or 45 to 53 Gy postoperatively (n = 42). The median survival was 11 months for both groups. No difference in the survival rate was detected, but there was increased morbidity with preoperative radiotherapy. Quality of life was not assessed in this trial.
Meta-analysis examining preoperative radiotherapy and surgery compared to surgery alone: mortality at one year

Preoperative radiotherapy and postoperative radiotherapy versus postoperative radiotherapy alone
Iizuka et al. [30] reported a randomized trial of preoperative and postoperative radiotherapy versus postoperative radiotherapy alone in 364 Japanese patients. In an analysis of 207 eligible patients (157 patients were excluded because of the extent of disease or operative complications), preoperative and postoperative radiotherapy was associated with a significantly higher mortality rate compared with postoperative radiotherapy alone (median survival was 394 days versus 648 days; p = 0.0069). The major postoperative complications were pneumonia (13.5% versus 9.7%) and leakage (11.5% versus 9.7%).

Preoperative chemotherapy and surgery versus surgery alone
Seven randomized trials of preoperative chemotherapy and surgery versus surgery alone are presented in Table  4 [24,[32][33][34][35]37,48]. Of these seven RCTs, six were available as fully published reports, and one was available as an abstract only [35]. Quality of life was not assessed in any of the trials. Additionally, three meta-analyses were obtained [38,49,50].

Preoperative and postoperative chemotherapy and surgery versus surgery alone
Two randomized trials of preoperative and postoperative chemotherapy and surgery versus surgery alone [31,36] ( Table 5) were examined. Neither Roth et al. [31] (using a now out-dated combination of cisplatin, vindesine, and bleomycin) nor the largest North American trial as reported by Kelsen et al. [36] (using cisplatin and 5-FU) detected a statistically significant difference in overall survival. The Gastrointestinal Cancer DSG pooled these two trials ( Figure 4). No significant difference in the risk of mortality with preoperative and postoperative chemotherapy and surgery compared with surgery alone was detected (RR, 0.99; 95% CI, 0.81 to 1.21; p = 0.93). No statistical heterogeneity was detected (X 2 = 0.65, p = 0.42).

Postoperative chemotherapy and surgery versus surgery alone
Three randomized trials of postoperative chemotherapy and surgery compared with surgery alone are presented in Table 6[39 -41]. All three trials used cisplatin-based regimens. Pouliquen et al. [39] found no improvement in the survival rate with postoperative chemotherapy. The patients were stratified into two groups: complete resections with or without nodal involvement and, palliative resections for positive margins or metastatic disease. Only the completely resected group was included in our analysis. Ando et al. [40] resected early (T1b) carcinomas and did not find any improvement in survival. In another study, reported in abstract form, Ando et al. found that the duration of improved dysphagia was similar for both groups.
The Gastrointestinal Cancer DSG pooled the three-year mortality data for two trials [39,40] ( Figure 5). The trial by Ando et al. [41] could not be included in the pooled analysis because the abstract did not report three-year survival. No significant difference in the risk of mortality at three years for postoperative chemotherapy compared with surgery alone was detected (RR, 0.94; 95% CI, 0.74 to 1.18; p = 0.59). There was no significant statistical heterogeneity (X 2 = 0.08, p = 0.77).

Preoperative chemotherapy and radiotherapy and surgery versus surgery alone
Eight randomized trials of combined modality neoadjuvant chemotherapy and radiotherapy are presented in reported a significant overall increase in three-year survival with combined preoperative chemoradiation but was closed prematurely following an interim analysis. This study was criticized for the lack of preoperative staging using CT  The Gastrointestinal Cancer DSG pooled the one-year mortality data for the six trials with data available at one year [24,[42][43][44][45][46] (Figure 6). No significant difference in the risk of mortality at one year for preoperative chemoradiation and surgery compared to surgery alone was detected (RR, 0.89; 95% CI, 0.76 to 1.03; p = 0.12). No significant statistical heterogeneity was detected (X 2 = 1.67, p = 0.89).
The first meta-analysis, an abstract report by Fiorica et al.
[54], pooled six RCTs comparing preoperative chemoradi- Meta-analysis examining postoperative chemotherapy and surgery compared to surgery alone: mortality at three years

Postoperative chemotherapy and radiotherapy versus surgery alone
No randomized trials have evaluated postoperative chemotherapy combined with radiation versus surgery alone.

Postoperative chemotherapy versus postoperative radiotherapy
One randomized trial evaluated postoperative chemotherapy versus postoperative radiotherapy following curative esophagectomy [56]. Patients in this Japanese trial received cisplatin and vindesine (n = 126) or radiotherapy at a dose of 50 Gy (n = 127). The median survival was 38 months for both groups. No difference in survival was detected (52% for chemotherapy versus 51% for radiotherapy at three years; log-rank p = 0.806). There were significantly more cases of decreased white blood cell counts (12 versus 3 for grade 3-4; p = 0.026), elevated blood urea nitrogen (26 versus 11 for grade 1-2; p = 0.018) and elevated creatinine concentrations (27 versus 9 for grade 1-3; p = 0.006) among patients randomized to chemotherapy compared with radiotherapy. Quality of life was not assessed in this trial.

Preoperative chemotherapy versus preoperative radiotherapy
Two randomized trials evaluating preoperative chemotherapy compared with preoperative radiotherapy were reviewed [24,57]. Kelsen et al. [57] randomly assigned 96 patients to preoperative radiotherapy or chemotherapy.
Meta-analysis examining preoperative chemoradiation and surgery compared to surgery alone: mortality at one year Figure 6 Meta-analysis examining preoperative chemoradiation and surgery compared to surgery alone: mortality at one year. Overall risk ratio = 0.89 (95% CI, 0.76 to 1.03; p = 0.12) Postoperative crossover therapy (i.e., postoperative radiotherapy for those who received preoperative chemotherapy and vice versa) was given to patients who were found to have unresectable or locally advanced disease. Only 11 of 48 chemotherapy patients and 9 of 48 radiotherapy patients did not receive additional postoperative treatment. Overall median survival was similar in both groups (10.4 months for chemotherapy versus 12.4 months for radiotherapy; p = 0.61), but the crossover design precluded proper analysis. In the fourarm trial by Nygaard et al. [24], preoperative chemotherapy was compared with preoperative radiotherapy, and the results demonstrated a significant difference in survival favouring preoperative radiotherapy (21% versus 3% at three years; p = 0.01). However, when compared to surgery alone, there was no benefit to either preoperative radiation or chemotherapy. Neither trial report included data on quality of life.

Preoperative chemoradiation versus preoperative radiotherapy
One randomized trial evaluated the role of preoperative bleomycin in addition to radiotherapy [58]. Seventy patients received preoperative chemoradiation with bleomycin and 63 patients received preoperative radiotherapy alone. The results demonstrated no significant difference in survival between the two groups (median survival was 25 weeks versus 26 weeks; survival rate was 25% versus 19% at two years; p = 0.56). There was also no benefit for bleomycin in the palliation of dysphagia. Quality of life was not assessed in this trial.

Postoperative immunotherapy in combination with radiotherapy or chemoradiation
One Japanese trial evaluated protein-bound polysaccharide (PSK) as an adjunct to postoperative radiotherapy or chemoradiation in resected esophageal cancer [59]. This trial involved 174 patients who were randomly assigned to four treatment groups. The three-year survival rates for radiotherapy, radiotherapy + PSK, chemoradiotherapy, and chemoradiotherapy + PSK were 43.3%, 45.5%, 33.5%, and 44.3%, respectively. There was no significant difference in survival when radiotherapy and radiotherapy + PSK were compared, or when chemoradiotherapy and chemoradiotherapy + PSK were compared (log-rank p = 0.19 for chemoradiotherapy versus chemoradiotherapy + PSK). Some patients randomized to PSK experienced adverse effects, including mild nausea, erythema, liver dysfunction and leukopenia, but there were no reports of toxicity that were definitely attributed to PSK. There was no assessment of quality of life.

Preoperative hyperthermia in combination with chemoradiation
One Japanese randomized trial, reported by Kitamura et al.
[60], evaluated preoperative hyperthermia and chemoradiotherapy and surgery (n = 32) versus preoperative chemoradiotherapy and surgery (n = 34). Median survival was 36 months and 20 months, respectively. The results showed a significant improvement in the survival rate (50.4% versus 24.2% at three years; p-value not reported) and local tumour control with hyperthermia compared with control. It was reported that both adjuvant treatments were well tolerated and resulted in no postoperative complications. Quality of life was not assessed.

Discussion
Most trials excluded patients with cancers located in the cervical esophagus, and therefore the interpretation of this review is limited to tumours in the more distal two thirds.
The options for neoadjuvant or adjuvant therapy for resectable thoracic esophageal cancer are many. On reviewing the results of randomized trials and meta-analyses, the Gastrointestinal Cancer DSG did not recommend neoadjuvant or adjuvant therapy based on the following: Preoperative radiotherapy does not improve survival compared with surgery alone. Postoperative radiotherapy may, in fact, be harmful [27].
Preoperative chemoradiation does not appear to improve survival compared to surgery alone. Although the pooled analysis shows that all six studies are in the direction favouring preoperative chemoradiation at one year, the pooled estimate did not achieve statistical significance. Preoperative cisplatin-based chemotherapy does not appear to improve survival. Four of the seven trials [24,[32][33][34] detected a significant survival advantage favouring preoperative cisplatin-based chemotherapy. Kok et al. [35] reported a survival advantage for chemotherapy but only reported median survival results in abstract form. The two largest trials produced conflicting results [36,37]. Kelsen et al. [36] detected no survival advantage, while the MRC OE02 trial [37] detected a significant survival advantage for preoperative chemotherapy at two years. Although all chemotherapy protocols were cisplatinbased, the varying dosages, the number of cycles completed, and the other agents used contributed to clinical heterogeneity.
The available evidence from three randomized trials does not support the use of postoperative chemotherapy over surgery alone [39][40][41]. Examination of the results of randomized trials, including the pooled analyses, fails to support the use of preoperative or postoperative adjuvant treatment of any type at this time for patients with resectable carcinoma of the thoracic esophagus. Overall, the evidence does not support the use of neoadjuvant or adjuvant therapy for patients with resectable cancer of the lower two-thirds of the esophagus. Surgical resection alone should remain the standard for localized thoracic esophageal cancer. Patient staging information can be found in Appendix 1 (Additional file 1).
Future trials should continue to assess multi-modality treatments for this patient population. For clinicians seeking guidance on treatments for patients with non-resectable esophageal cancer, the role of radiotherapy alone and chemoradiation alone without surgery is addressed in another clinical practice guideline by the Gastrointestinal Cancer DSG, Combined modality radiotherapy and chemotherapy in the non-surgical management of localized carcinoma of the esophagus [61].