Neoadjuvant Radiation Therapy

Trials that have evaluated the use of preoperative radiation have largely reported no benefit.

Kelsen and colleagues performed a randomized trial comparing preoperative radiation to preoperative chemotherapy in 96 patients with esophageal cancer. Although there was no increase in operative morbidity or mortality for patients treated with preoperative therapy compared with historical controls treated with surgery alone, there was also no additional treatment benefit. Another randomized trial involving 176 patients also failed to identify a benefit for preoperative radiation.

A prospective, multicenter Scandinavian trial reported by Nygaard and colleagues randomized 186 patients who had esophageal SCC to four treatment groups: surgery alone, preoperative chemotherapy (cisplatin/bleomycin) and surgery, preoperative radiation and surgery, or preoperative chemotherapy and radiation, followed by surgery. The 3-year OS was significantly higher in the pooled groups receiving radiation compared with the nonradiation groups. The results indicated an intermediate-term survival benefit for preoperative radiation but found that chemotherapy did not influence survival.

A subsequent meta-analysis, however, was unable to establish a significant benefit for preoperative radiation. With a median follow-up of 9 years, an analysis of more than 1100 patients from five randomized trials suggested a survival benefit of 3% at 2 years and 4% at 5 years that was not statistically significant ( P = .062).

Adjuvant Therapy

Combined-modality therapy in esophageal carcinoma long has focused on preoperative strategies. The role of adjuvant therapy has not been studied extensively, and the data that are available suggest equivocal results.

Postoperative chemotherapy without preoperative therapy was studied in two Japanese randomized trials, where patients who had SCC histology were randomized to receive two cycles of chemotherapy with cisplatin/vindesine or cisplatin/5-FU respectively. Although the trial with cisplatin/vindesine did not show any survival benefit, an unplanned subset analysis of the trial with cisplatin/5-FU revealed a survival benefit for patients who had lymph node involvement (5-year DFS 52% versus 38%).

The possible benefit for postoperative therapy suggested by the previously mentioned trials led to a subsequent Japanese trial that randomized 330 patients who had SCC histology to surgery and either two cycles of pre- or postoperative cisplatin/5-FU. Data recently presented in abstract form revealed that preoperative chemotherapy was associated with a significant improvement in OS compared with postoperative chemotherapy (hazard ratio [HR] 0.64, 95% CI, 0.45 to 0.91), further questioning the role of adjuvant chemotherapy for SCC. A significant number of patients on this trial, however, never received postoperative chemotherapy, making the results difficult to interpret. Another unexpected finding is that an unplanned subset analysis suggested that the benefit for preoperative therapy over postoperative therapy was seen only in patients without lymph node involvement, in contrast to the previously mentioned study where a benefit for adjuvant chemotherapy over observation was noted in patients who had lymph node involvement.

The overall lack of benefit for adjuvant chemotherapy suggested by the Japanese trials is consistent with the results of a randomized French trial, which also found no survival benefit for 6 to 8 months of adjuvant chemotherapy with cisplatin/5-FU. In fact, there were significantly more complications in the chemotherapy group.

In contrast, a pilot Eastern Cooperative Oncology Group (ECOG) trial recently evaluated four cycles of postoperative paclitaxel/cisplatin in patients who had node-positive esophageal or GE junction adenocarcinoma. Two-year OS was 60%, which is statistically superior compared with the historical control (38%, derived from Intergroup 113 trial).

Trials involving adjuvant radiotherapy generally have reported negative results. A French study randomized 221 patients to surgery alone versus surgery followed by radiation and found no survival benefit from radiation.

Another randomized study of 130 patients from Hong Kong actually demonstrated increased mortality with postoperative radiation (8.7 versus 15.2 months, in favor of the no adjuvant therapy group), with the difference attributed to radiation-related deaths and early metastatic disease.

Finally, a large prospective Chinese study also failed to detect an OS benefit among 495 patients randomized to adjuvant radiation or no further therapy. A subgroup analysis of node-positive patients, however, did show a 5-year OS benefit favoring the radiation group (35.1% versus 13.1%).

Although trials of adjuvant radiotherapy alone have not suggested significant benefit, there may be benefit from adjuvant concurrent chemoradiotherapy, as suggested the results of the Intergroup trial 116 in gastric adenocarcinoma. This trial revealed a significant improvement in OS and DFS for the delivery of postoperative therapy with 5-FU/leucovorin and radiation compared with surgery alone. As a relatively modest 20% of the patients treated had proximal gastric cancers (with involvement of the GE junction) and primary GE junction cancers, these data may justify the use of postoperative therapy in such patients who have not received preoperative therapy. It should be noted that the results of this trial have been questioned because of the relatively inadequate surgical resections that were performed; 54% of patients had a D0 resection, which is less than a complete dissection of the involved lymph nodes. It has been argued that radiation in this setting compensated for inadequate surgery and that its benefits may not be seen if a more complete or extensive D1 or D2 surgical resection is undertaken.

Combined Neoadjuvant Chemoradiotherapy

Although recent pre- and peri-operative chemotherapy trials have indicated a survival benefit, the low rate of pCR and the inconsistent impact on rates of operability have led researchers to investigate neoadjuvant chemoradiotherapy.

Chemoradiotherapy typically involves regimens of cisplatin or mitomycin and continuous infusion 5-FU, with radiotherapy dosages from 30 to 40 Gy and up to 60 Gy in more recent trials. Such therapy results in pCR rates of 20% to 40%, with long-term survival of no more than 25% to 35%. Superior survival is achieved consistently, though in patients achieving a pCR to chemoradiotherapy (up to 50% to 60% at 5 years).

These results are at the expense of significant toxicities, primarily hematologic and gastrointestinal (GI), which have been greatest in trials employing a higher dose of or twice-daily radiation or in which radiotherapy overlapped all cycles of preoperative chemotherapy. The GI toxicity associated with cisplatin/5-FU and radiation includes nausea, mucositis, and esophagitis, leading some investigators to mandate placement of enteral feeding tubes before treatment initiation.

The seminal phase 3 United States Radiation Therapy Oncology Group (RTOG) trial 85-01 demonstrated the superiority of chemoradiotherapy over radiation alone. This nonoperative study compared standard-fractionation radiation (64 Gy) with radiation (50 Gy) plus concurrent cisplatin/5-FU. The trial was stopped when data from 121 patients showed an improved median OS in favor of chemoradiotherapy (12.5 months versus 8.9 months). Two-year survival was also improved in the chemoradiotherapy group (38% versus 10%), as was 5-year survival (21% versus 0%). Although most patents treated on this trial had SCC, long-term survival also was seen in the small number of adenocarcinoma patients on the trial, with 13% of patients alive at 5 years.

In addition to a survival benefit, disease recurrence was reduced significantly by the addition of chemotherapy to radiation. At 1 year, recurrent disease was observed in 62% of the group that received radiation versus 44% in the chemoradiotherapy arm. Distant recurrence rates were 38% and 22%, respectively. Based on this study, chemoradiotherapy was established as the standard of care in the nonsurgical management of locally advanced esophageal SCC.

Building on these results, more intensive treatment strategies have been investigated. In the nonoperative RTOG 90-12 chemoradiotherapy study, induction chemotherapy with cisplatin/5-FU followed by chemoradiotherapy with the same regimen did not appear to afford any additional benefit. The RTOG 94-05 study compared a total radiation dose of 64.8 Gy versus 50.4 Gy during concurrent cisplatin/5-FU and also failed to demonstrate superior results with the more intense regimen. This study confirmed 50.4 Gy as the standard radiation dose when given in combined therapy with cisplatin/5-FU. Finally, the phase 1/2 RTOG 92-07 trial, which attempted to boost radiation with brachytherapy following external beam radiation, revealed significant toxicity, including a 12% incidence of treatment-related fistulas.

Phase 3 Trials of Chemoradiotherapy

Five contemporary randomized trials have compared preoperative chemoradiotherapy followed by surgery versus surgery alone. The results are summarized in Table 2 .

Urba and colleagues from the University of Michigan randomized 100 patients to preoperative cisplatin/5-FU/vinblastine and radiation or to surgery alone. Despite a statistically significant decrease in the rate of local recurrence favoring preoperative therapy (19% versus 42%), 3-year OS trended toward improvement but was not statistically significant (30% versus 16%, P = .15). Rates of curative resection were equivalent in both groups (90%). Most patients treated on this trial had adenocarcinoma.

Walsh and associates from Ireland randomized 113 patients who had esophageal adenocarcinoma to preoperative cisplatin/5-FU/radiation or surgery alone. Rates of negative margin resection were not reported, although it was noted that the preoperative therapy group had a significantly lower incidence of positive lymph nodes or metastatic disease at surgery (42% versus 82%). A significant improvement in 3-year OS was noted (32% versus 6%). Interpretation of this study is confounded by the very poor survival of the surgical control arm—6% at 3 years—which is inconsistent with the approximate 20% 5-year survival rates reported for modern surgical series. Other shortcomings of this trial include inadequate pretherapy staging that could have led to an imbalance in prognostic factors between both groups, the variable surgical procedures used, premature termination based on an unplanned interim analysis, and the relatively short follow-up period for surviving patients (18 months).

Bosset and colleagues, on behalf of the European Organization for Research and Treatment of Cancer (EORTC), randomized 282 patients with esophageal SCC to preoperative cisplatin and concurrent split-dose radiation or surgery. Compared with the surgery-only group, the chemoradiotherapy group had a significantly higher rate of curative resection (81% versus 69%), and an improvement in disease-free survival and a decreased risk of local recurrence. OS (the primary trial endpoint), however, was not significantly different. It might be that the significantly higher postoperative mortality in the chemoradiotherapy arm (12% versus 4%) outweighed any potential survival benefit for the chemoradiotherapy group.

In an Australian trial, Burmeister and colleagues randomized 256 patients to one cycle of preoperative cisplatin/5-FU and radiation or to surgery alone. Although the trial failed to show a survival advantage for patients who received chemoradiotherapy, they did have a significantly higher curative resection rate compared with the surgery-only patients (80% versus 59%). In this study, the administration of a single chemotherapy cycle may represent suboptimal delivery of chemotherapy, as suggested by the unexpectedly low pCR rate in the adenocarcinoma patients of only 9%.

Finally, results of the Cancer and Leukemia Group B (CALGB) trial 9781 recently were published. This trial randomized patients to two cycles of preoperative cisplatin/5-FU and radiation or to surgery alone. Fifty-six patients were randomized before the trial was closed for poor accrual. Nevertheless, patients assigned to chemoradiotherapy had substantially improved median survival (4.5 versus 1.8 years) and 5-year OS (39% versus 16%) compared with patients undergoing surgery alone.

Overall, these randomized trials are associated with methodological concerns (including the lack of rigorous pretherapy staging with endoscopic ultrasound or laparoscopy), are significantly smaller than randomized preoperative chemotherapy trials, and produce conflicting results. They do suggest improved curative resection rates and decreased local recurrence, however. A survival advantage for preoperative chemoradiotherapy over surgery alone is not demonstrated clearly, although several studies suggest such a trend.

These observations are supported further by a recent meta-analysis, in which 10 randomized trials of preoperative chemoradiotherapy versus surgery alone and eight trials of preoperative chemotherapy versus surgery alone were analyzed. Preoperative chemoradiotherapy was associated with a hazard ratio of all-cause mortality of 0.81 versus surgery alone (95% CI, 0.70 to 0.93, P = .002), which translated to a 13% absolute difference in mortality at 2 years. This benefit was irrespective of histology. Preoperative chemotherapy was associated with a hazard ratio of 0.90 (95% CI, 0.81 to 1.00, P = .05) compared with surgery alone, which related to a 2-year absolute survival benefit of 7%. There did not appear to be any benefit of preoperative chemotherapy for patients who had squamous histology (hazard ratio 0.88; 95% CI, 0.75 to 1.03, P = .12), although there was a benefit for patients who had adenocarcinoma histology (hazard ratio 0.78; 95% CI, 0.64 to 0.95, P = .014).

The possible superiority of preoperative chemoradiotherapy over preoperative chemotherapy also has been suggested by a randomized study recently presented in abstract form. In this study by Stahl and colleagues for the German Esophageal Cancer Study Group, patients were randomized to preoperative chemotherapy with cisplatin/5-FU/leucovorin followed by surgery versus cisplatin/5-FU/leucovorin followed by chemoradiotherapy with cisplatin/etoposide and then surgery. One hundred-twenty eligible patients were randomized before the trial was closed because of poor accrual. The results revealed a trend toward improved local progression-free survival (PFS; 77% versus 59%), median OS (32.8 versus 21.1 months), and 3-year survival (43% versus 27%) for the chemoradiotherapy over chemotherapy group, but these results were not statistically significant ( P = .14). Both the pCR rate (16% versus 2%) and node-negative status (64% versus 37%) were significantly higher in the chemoradiotherapy group. Strengths of this trial include the careful pretherapy staging (which included endoscopic ultrasound and laparoscopy), the enrollment only of high-risk patients with at least T3 or node-positive tumors and the careful balancing of pretherapy stage between the two treatment arms.

Definitive Chemoradiotherapy Versus Chemoradiotherapy Followed by Surgery

Two recent randomized trials have compared definitive chemoradiotherapy versus chemoradiotherapy followed by surgery. The first study was performed by the German Esophageal Cancer Study Group, which assigned 172 patients who had SCC to preoperative therapy (three cycles of cisplatin/5-FU/leucovorin/etoposide, then cisplatin/etoposide and concurrent radiation to 40 Gy) followed by surgery or to the preoperative therapy alone with a higher radiation dose (to at least 65 Gy) in lieu of surgery. Although local PFS was improved with the addition of surgery (HR for chemoradiotherapy-only group versus surgery group 2.1, 95% CI, 1.3 to 3.5, P = .003), there was only a nonsignificant trend toward improvement in 3-year OS (31.3% versus 24.4%). Treatment-related mortality was also significantly higher in the surgery group compared with the chemoradiotherapy-only group (12.8% versus 3.5%). Ten-year survival data for this trial recently was presented in abstract form, reaffirming the absence of a significant difference between both groups.

The second study is the French FFCD 9102 trial, where 444 eligible patients who had mostly SCC histology underwent initial chemoradiotherapy with cisplatin/5-FU. Those who responded to initial therapy then were randomized either to undergo surgery or to receive an additional three cycles of cisplatin/5-FU with radiation, as the authors felt that it would be inappropriate to continue chemoradiotherapy in patients not responding to therapy. Of the 444 patients, 259 were randomized. The 2-year survival rate was not significantly different between both groups (34% in surgery group versus 40% in chemoradiotherapy-only group, P = .44). Locoregional recurrence, however, was higher in the chemoradiotherapy-only group (43% versus 34%), and there was also a higher incidence of stent placement in this group (32% versus 5%). Three-month mortality was significantly higher in the surgery group (9.3% versus 0.8%). Based on these data, the authors concluded that patients who have tumors, especially of SCC histology, that respond to initial chemoradiotherapy did not derive any survival benefit from subsequent surgery. Patients who underwent surgery did have improved local control of their disease, albeit at the cost of increased treatment-related mortality.

An interesting question that arises from this study is whether patients who do not respond to initial therapy benefit from subsequent surgery. In a recent abstract, the authors discussed the outcome of the 192 of the 451 registered patients from the previous study who were not randomized to further protocol therapy after initial chemotherapy primarily because of a lack of response but also because of medical contraindication or patient refusal. Of these nonrandomized patients, 112 subsequently underwent surgery, with 80 undergoing R0 resections. The median OS for the patients who underwent surgery was significantly superior to the median OS of those who did not (17.3 versus 6.1 months) and was comparable to the median OS of the patients who were randomized. Although there are clear limitations and potential strong confounders to such an analysis, these data may suggest that salvage esophagectomy can be beneficial for a subset of patients who do not respond to initial therapy.

As a related issue, definitive chemoradiotherapy alone versus surgery alone also recently was compared in a Scandinavian phase 3 trial of 91 patients with adenocarcinoma and SCC who were randomized to receive either cisplatin/5-FU and radiation alone or surgery. At a median follow-up of 51.8 months, there was no survival difference between both groups. Although this study may be underpowered to detect small survival differences, the data collectively support definitive chemoradiotherapy as an acceptable approach for patients who have contraindications to surgery.