• © 2007 by American Society of Clinical Oncology

Surgery After Primary Chemoradiotherapy in Squamous Cancer of the Esophagus: Is the Photon Mightier Than the Sword?

Esophageal cancer presents medical, surgical, and radiation oncologists with one of the most difficult treatment challenges. The rarity of the disease in the West and the absence of effective screening result in patients presenting with dysphagia, weight loss, and advanced stage cancers. Squamous cancer patients are usually elderly, have a history of alcohol and tobacco abuse, and have comorbid conditions that limit therapy tolerance. Esophagectomy for local disease is fraught with significant operative morbidity and operative mortality. The loss of organ function leads to protracted nutritional compromise and a slow recovery.

Primary chemoradiotherapy without surgery in the treatment of esophageal cancer emerged as a potential alternative to esophagectomy with the publication of the Radiation Therapy Oncology Group (RTOG) trial 85-01.¹ This trial, treating mainly squamous cancers, reported 2- and 5-year survival rates that were comparable with contemporary surgical series. RTOG 85-01 engendered an ongoing debate about the relative merits of esophagectomy, primary chemoradiotherapy, or multimodality therapy using preoperative chemoradiotherapy followed by surgery. The local disease recurrence rate of 45% or higher observed with primary chemoradiotherapy has led thoracic surgeons to argue that esophagectomy is required in patients undergoing chemoradiotherapy to ensure local disease control. Others argue that the addition of surgery does not improve outcome, given the overall poor prognosis, as patients with locally advanced disease ultimately develop metastatic disease in more than 50% of cases. Are the rigors of esophagectomy justified after chemoradiotherapy, if survival is defined by distant disease as well as local disease control?

The report by Bedenne and colleagues² in this issue of the Journal of Clinical Oncology represents a landmark in the management of locally advanced squamous esophageal cancer. The Fédération de Francophone de Cancérologie Digestive (FFCD) trial 9201 compares chemoradiotherapy alone with chemoradiotherapy followed by surgery in patients with esophageal cancer who responded to chemoradiotherapy. Although both squamous cell carcinoma and adenocarcinoma histologies were accrued, squamous cancer was the most common disease treated, in nearly 90% of cases. Unlike other Western countries in which adenocarcinoma of the esophagus is now more common, squamous cancer remains the predominant histology in France and accounts for 75% of all cases.

The primary end point of the trial was to demonstrate equivalence of survival in patients treated with primary chemoradiotherapy or with chemoradiotherapy followed by surgery. Secondary end points included quality of life, pattern of disease recurrence, and the need to treat subsequent dysphagia. All patients were treated with two cycles of fluorouracil and cisplatin combined with one of two schedules of radiotherapy: 30 Gy in two split courses weeks 1 and 4, or 46 Gy given continuously. Patients were assessed for response by unidimensional measurement on esophagram and by dysphagia improvement, and responding patients were then randomly assigned to surgery or to three additional cycles of chemotherapy, combined with an additional 15 Gy of radiotherapy on the split course arm and 20 Gy on the continuous arm. The investigators only randomly assigned patients responding to chemoradiotherapy to avoid continuing inactive therapy in nonresponding patients and to avoid the potential confounding of survival by crossing over of nonresponding patients to surgery. Despite the difficulty of randomly assigning patients to a surgery or no surgery treatment plan, compliance on protocol was fairly high with the majority of randomly assigned patients completing therapy on their assigned treatment arm. Compliance rates were significantly higher in the nonsurgery (97%) compared with the surgery arm (85%), as some patients assigned to surgery refused surgery or were deemed inoperable. The strengths of FFCD 9201 include adequacy of the sample size, the nearly 4-year median length of follow-up, the documentation of pattern of failure, and the long-term assessment of swallowing function.

The authors reported no difference in either median survival (17.7 v 19.3 months) or 2-year survival (34% v 40%) in patients undergoing surgery or additional chemoradiotherapy. Rates of distant metastasis were also not different in the surgery and no surgery group. However, patients undergoing surgery had a lower rate of local tumor recurrence compared with the nonsurgical group (34% v 43%) and a lesser need for palliative intervention to relieve dysphagia (24% v 46%). However, these differences in local control failed to impact survival. The vast majority of recurrences occurred either in the first year (60% of recurrences) or first 2 years (80%). Conclusions from this trial may not be applicable in the small number of patients with adenocarcinoma, although a multivariate analysis did not indicate an outcome difference as a function of histology.

Methodologic problems with the FFCD 9201 trial include the significant disparity between therapy arms. The chemoradiotherapy arm received additional cycles of both chemotherapy and radiotherapy, which could have impacted both local and distant disease recurrence. Staging did not incorporate endoscopic ultrasound, which is the most accurate current staging modality for T and N stage, but protocol arms were balanced for clinical staging. There is no report of surgical quality control, including number of lymph nodes resected. A major deficiency in the trial design was the use of a split course of radiotherapy in two thirds of all patients. Recent phase III data from France indicate that a split course of radiotherapy combined with chemotherapy results in an inferior survival outcome compared with continuous radiation therapy.³ The use of split course of radiotherapy may potentially have reduced the efficacy of chemoradiotherapy in both treatment arms. The pathologic complete response rate of 23% in the 110 patients undergoing surgery, although comparable with modern series employing preoperative chemoradiotherapy, might be expected to be higher, as only patients responding to chemoradiotherapy were randomly assigned.

Although results from this trial fail to indicate a survival benefit for surgery, the responding patients who were randomly assigned accounted for only 57% of all patients who initiated treatment. This may have led to a bias against a benefit for surgery, as some nonresponding patients are likely to gain a salvage benefit for surgery after chemoradiotherapy. Follow-up for the nonresponding patients did indicate a shorter median survival of 11 months and we await report of other outcomes in these patients.

The results reported in FFCD 9201 are consistent with results of a smaller trial recently reported by the trial of Stahl and colleagues⁴ in the Journal. One hundred seventy-two patients with squamous cancer were treated with induction chemotherapy followed by combined chemoradiotherapy, with a subsequent random assignment to surgery or additional chemoradiotherapy. All patients, not just responding patients, were randomly assigned. Consistent with the FFCD 9201 trial, although local control was enhanced with surgery (64% v 40%), there was no median or 2-year survival difference. A survival trend of 31% compared with 24% favoring surgery was not significant, although the survival curves continued to separate beyond 3 years—whether or not this difference would be significant with a larger trial is unclear. Surgery in nonresponders to chemoradiotherapy was associated with long-term survival in some patients, indicating the potential utility of surgery as a salvage procedure. Data from an earlier trial conducted by the Eastern Cooperative Oncology Group (ECOG EST-1282),⁵ also fail to support a survival benefit for surgery after patients with squamous cancer of the esophagus undergo primary chemoradiotherapy.

The results reported in the FFCD 9201 trial indicate that, with currently available systemic agents and radiation therapy to treat squamous cancer of the esophagus, adding surgery for responding patients improves local control and local palliation without an impact on survival. The morbidity and mortality of esophagectomy argue that the application of surgery in responding patients should be selective. The use of surgery as a salvage in patients not achieving a clinical complete response to chemoradiotherapy is now being evaluated by a small pilot trial conducted by the RTOG (trial 02046). However, without performing surgery, it is difficult to determine when clinical complete response to chemoradiotherapy represents a pathologic complete response. Response to chemoradiotherapy as determined by newer imaging technologies, such as positron emission tomography (PET) scan, correlates with response at surgery and is prognostic of survival.⁶ Nonetheless, PET scan does not identify pathologic complete responders. Response to induction chemotherapy, administered before the addition of radiation therapy, is also prognostic of survival.⁶ Early response to induction chemotherapy measured by PET scan may also be prognostic of survival,⁷ and may lead to a change in therapy, including earlier referral to surgery⁸ or a change in chemotherapy for nonresponders. Referral of patients destined to experience treatment failure or to have early disease recurrence for alternative or investigational therapies may spare them the rigors of ineffective surgery, chemotherapy, or radiotherapy.

For patients undergoing surgery for esophageal cancer, the debate also continues about the optimal preoperative therapy, including chemotherapy alone or combined chemotherapy and radiotherapy. It is clear from the current results for combined-modality therapy in esophageal cancer that improved systemic agents are needed, given that metastatic disease determines survival. As it is unlikely that cytotoxic chemotherapy alone will significantly improve current outcomes, the hope for the future is the development of more specific targeted therapies to eradicate or suppress the growth of micrometastatic disease. Measures to better direct therapy to patients most likely to respond are also needed, including the identification of validated, commonly occurring molecular markers of response and prognosis.