Given the improved survival and local control with the addition of chemotherapy to radiation, the question as to whether chemoradiation followed by surgery was superior to surgery alone needed further investigation. The major randomized controlled trials comparing chemoradiation plus surgery to surgery alone are summarized in Table 10.2.

In the Walsh study, 113 patients were randomized to 4000 cGy in 15 fractions with concurrent 5-FU and cisplatin followed by surgery versus surgery alone. The median and 3-year survival rates were 16 months and 32 % for the chemoradiation group and 11 months and 6 % for the surgery-only group (P = .01) [8]. However, the outcomes of this study were criticized for short follow-up time (median 10 months), unconventional radiation dose and fraction size, as well as the poor survival outcomes in the surgery-alone arm compared to other randomized trials.

The CALGB 9781 trial randomized patients to neoadjuvant chemoradiation using a more standard radiation dose and fraction size (5040 cGy in 28 fractions) with concurrent cisplatin and 5-FU followed by surgery versus surgery alone. Although the trial was closed early due to poor accrual, the median and 5-year survival rates were 4.5 years and 39 % for patients receiving chemoradiation and 1.8 years and 16 % for patients receiving surgery alone [9].

In 2012 van Hagen et al. published the “ChemoRadiotherapy for Oesophageal cancer followed by Surgery Study” (CROSS) trial. Patients with resectable (T2–T3, N0–N1, M0) esophageal or gastroesophageal junction (GEJ) tumors were randomized to neoadjuvant chemoradiation with 4140 cGy in 23 fractions with concurrent carboplatin and paclitaxel followed by surgery versus surgery alone. A pathological complete response was achieved in 47 of 161 patients (29 %) who underwent resection following chemoradiation. Median OS was 49.4 months in the trimodality arm versus 24.0 months in the surgery-alone group. At a median follow-up of 45.4 months, the 3- and 5-year overall survival rates were 58 % versus 44 % and 47 % versus 24 % for trimodality therapy versus surgery-alone arms, respectively (P = 0.003) [10].

Multiple meta-analyses have examined the survival outcomes following chemoradiation and surgery compared to surgery alone [11–15]. All but one analysis showed a significant reduction in mortality when chemoradiation was added to surgery [16]. Sjoquist et al. reviewed 12 randomized controlled trials with a total of 1854 patients comparing chemoradiation followed by surgery versus surgery alone. The hazard ratio (HR) for all-cause mortality for patients receiving neoadjuvant chemoradiation was 0.78 (95 % CI, 0.70–0.88; P < .0001), and this benefit was irrespective of histology [15]. Urschel et al. evaluated 9 randomized trials with a total of 1116 patients and concluded that 3-year survival was improved with chemoradiation (odds ratio [OR] = 0.66; 95 % CI, 0.47–0.92; P = .016). Improvements in survival were most pronounced with concurrent chemoradiation (OR = 0.45; 95 % CI, 0.26–0.79; P = .005) compared with sequential therapy (OR = 0.82; 95 % CI, 0.54–1.25; P = .36) [13].

Stahl et al. compared chemoradiation to 4000 cGy with concurrent cisplatin and etoposide followed by surgery versus definitive chemoradiation to 6000 cGy. Although the study closed early and statistical significance was not achieved, there was a trend toward improved 3-year survival in the preoperative chemoradiotherapy arm from 27.7 to 47.4 % (HR = 0.67, 95 % CI, 0.41–1.07, P = 0.07). Postoperative mortality was not significantly increased in the chemoradiotherapy group (10.2 % vs. 3.8 %; P = 0.26) [17].

These results were confirmed by a similar randomized French trial by Bedenne et al. where 444 patients with SCC underwent concurrent chemoradiation with two cycles of cisplatin and 5-fluorouracil. Patients who demonstrated a partial or complete response to treatment were then randomized to receive either surgery or additional chemoradiation. Only 259 (58 %) of the patients treated with this regimen went on to randomization. The median survival time was 17.7 months in the surgery group versus 19.3 months in the definitive chemoradiation group. However, the trial suffered from poor accrual and suboptimal design. The randomization of these patients did not take place at the time of diagnosis, rather after they were treated with chemoradiation. Therefore, those who were unresponsive to chemoradiation went on to surgery but were not followed on the trial; this may have limited the quality of the analysis as these patients would have likely benefited most from surgery as part of the treatment regimen [18].

Squamous cell carcinomas (SCC) of the cervical esophagus that extend from the hypopharynx to the sternal notch are a unique challenge due to the associated surgical morbidity. These tumors represent 5–6 % of esophageal cancers. Surgery may require removal of a portion of the pharynx and larynx and radical neck dissection causing severe functional deficits and impairment of quality of life. Therefore, cervical esophageal tumors are treated with primary chemoradiation similarly to locally advanced cancers of the hypopharynx and larynx with surgery reserved for salvage.

Multiple series have demonstrated similar survival outcomes between surgery and chemoradiation in this population [19, 20]. In a phase II trial by Bidoli et al., 101 patients were treated with cisplatin and 5-FU with concurrent radiation to 3000 cGy. Patients with potentially resectable tumors were then assessed for curative surgery; the remaining patients received two more cycles of chemotherapy and additional radiation to a total dose of 5000 cGy. Of the 40 patients who were candidates for surgery, 32 patients underwent resection with a reported surgical mortality of 22 %. Of the 61 nonsurgical patients, 37 patients (61 %) achieved complete clinical remission, and 14 patients (23 %) achieved partial remission. The median survival for the entire group was 15 months. At 10 years, freedom from disease progression was similar in the two groups (24 %), whereas the median survival (22 months vs. 12 months) and the overall survival rates (17 % vs. 9 %) were superior in nonsurgical compared with surgical patients, a factor likely explained by the high surgical mortality rate [21].

The incidence of adenocarcinoma of the gastroesophageal junction (GEJ) continues to increase in the last few decades. Cancers of the GEJ have been included in gastric and esophageal trials, and therefore the optimal treatment approach for these patients remains unclear. Three major trials that have included patients with GEJ tumors include the US Intergroup 0116 Trial, the European MAGIC Trial, and the German POET Trial. The US Intergroup 0166 Trial provides the most compelling data for the use of adjuvant chemoradiation. Following resection, patients (n = 556) were randomly assigned to either observation alone or adjuvant chemoradiotherapy. Three-year disease-free (48 % vs. 31 %) and overall (50 % vs. 41 %) survival rates were significantly better with combined modality therapy. This benefit was confirmed with longer follow-up (5-year overall survival 43 % vs. 28 %, hazard ratio [HR] for survival 1.32 (95 % CI 1.10–1.60)) [22].

Another study of patients with gastric cancer (74 %) included patients with both distal esophageal tumors (11 %) and GEJ tumors (15 %). The Medical Research Council Adjuvant Gastric Infusional Chemotherapy (MAGIC) trial randomized 503 patients to surgery-alone or surgery plus perioperative chemotherapy consisting of 3 cycles of epirubicin, cisplatin, and infusional 5-fluorouracil (ECF). The overall survival was significantly better in the chemotherapy group (hazard ratio [HR] for death 0.75, 95 % CI 0.60–0.93) as was progression-free survival (PFS, HR for progression 0.66). The 25 % reduction in the risk of death favoring chemotherapy translated into an improvement in 5-year survival from 23 to 36 % [23].

The German multicenter POET Trial was the first randomized study exclusively for patients with adenocarcinomas of the GEJ. Patients with locally advanced (EUS-staged T3–T4, N0, M0) adenocarcinoma of the lower esophagus, GEJ, or gastric cardia were randomized to neoadjuvant chemotherapy (12 weeks) followed by surgery or induction chemotherapy followed by 3 weeks of concurrent chemoradiotherapy followed by surgery. Although the study was closed prematurely due to poor accrual, there was a trend toward improved 3-year survival with chemoradiotherapy (47 % vs. 28 %; P = 0.07). The rate of complete (R0) resection was similar in both arms (72 % vs. 70 %), but the pathologic complete response rate was higher with chemoradiotherapy (16 % vs. 2 %), as was the rate of negative lymph nodes (64 % vs. 38 %). Moreover, the local failure rate in the chemotherapy-alone group was high (41 %), a finding that has been reported by others when radiation was not a component of trimodality therapy [17].

Therefore, cancers of the GE junction can be treated with either perioperative chemotherapy as outlined in the MAGIC Trial or postoperative chemoradiation as outlined in the Intergroup 0116 study. However, as demonstrated in the POET Trial, the likelihood of a complete pathologic response, survival, and freedom from local recurrence may be higher in patients who receive neoadjuvant chemoradiation over neoadjuvant chemotherapy. The CROSS trial also favored the trimodality approach with improvement in survival compared to surgery alone. Therefore, the trimodality approach using neoadjuvant chemotherapy with weekly carboplatin and paclitaxel with radiation is the favored approach. In patients whose clinical evaluation suggests a primary esophageal origin, neoadjuvant chemoradiation using weekly carboplatin/paclitaxel is a valid option (per CROSS). For patients whose tumors are located primarily in the stomach and can tolerate ECF chemotherapy, the perioperative approach (per MAGIC) may be a favored approach. If the patients are not good candidates for the aggressive chemotherapy regimen outlined by MAGIC, surgery followed by chemoradiation (per Intergroup 0116) is recommended. Finally, patients who are not surgical candidates are considered for primary chemoradiation.

More than 50 % of patients present with unresectable or metastatic disease at the time of presentation [24]. Historically, radiotherapy has played an important role in the management of unresectable or metastatic disease, for palliation of dysphagia, pain, and bleeding.

Recently, a phase III study performed by the Trans Tasman Radiation Oncology Group (TROG) 03.01 and National Cancer Institute of Canada (NCIC) CTG ES.2 trial found that chemoradiation had increased toxicity without additional symptom control or improved survival compared to radiation alone. This study enrolled 220 patients, the majority with metastatic disease, and randomized them to palliative radiation alone (3000–3500 cGy) or concurrent chemoradiation with cisplatin and 5-fluorouracil. The patients receiving radiotherapy alone showed a dysphagia response (at any point) of 67.89 % compared to chemotherapy response in 73.87 % (P = 0.343). There was increased gastrointestinal toxicity in patients receiving chemoradiation (nausea (P = 0.0019) and vomiting (P = 0.0072)). There was no significant difference in median survival (210 days for chemoradiation and 203 days for radiation alone) [25].