Overall survival curve.
Cumulative survival in patients with and without residual disease.
Tamim WZ, Davidson RS, Quinlan RM, O'Shea MA, Orr RK, Swanson RS. Neoadjuvant Chemoradiotherapy for Esophageal CancerIs It Worthwhile?. Arch Surg. 1998;133(7):722-726. doi:10.1001/archsurg.133.7.722
Copyright 1998 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.1998
With promising results from several institutions, many centers began treating patients with esophageal cancer with neoadjuvant chemoradiotherapy (NC) followed by esophagectomy. This approach is demanding for the patient and has not been proved to be better than esophagectomy alone.
To assess survival time and measures of quality of life associated with NC.
A retrospective review during 1990 to 1996.
The 3 tertiary academic hospitals affiliated with the University of Massachusetts Medical School, Worcester.
All patients (N=51) with cancer of the middle or lower esophagus who were treated with NC followed by esophagectomy during this period.
Main Outcome Measures
Median and 1-, 2-, and 3-year survival times; median preoperative treatment time (first office visit for surgical consultation before beginning NC to the date of surgery), median hospital stay, and postoperative swallowing function.
The median survival time of all patients was 16.3 months; 1-, 2-, and 3-year overall survival rates were 67%, 46%, and 39%, respectively. The median hospital stay was 12 days. The median postoperative treatment time was 3.3 months, which was 20% of the median survival time. Of the 51 patients, 19 were alive with a median follow-up time of 2.5 years. Twenty-nine percent of the patients had a complete pathological response with median and 1-, 2-, and 3-year survival rates of 17.5 months, 73%, 57%, and 57%, respectively. Palliation of dysphagia was excellent, with 44 (93%) of 47 operative survivors taking either a soft diet (18 ) or a regular(26 ) diet by the first postoperative visit.
Median survival time with NC followed by esophagectomy for resectable cancer of the esophagus does not appear to be significantly better than that reported for esophagectomy alone. Further, treatment time with NC consumed 20% of survival time. Examining only these outcome variables suggests that NC is not worthwhile. However, examining a longer-term outcome survival variable, such as 3-year survival time, suggests that NC followed by esophagectomy may result in greater long-term survival than that reported for esophagectomy alone. We conclude that further randomized, controlled studies are necessary before NC followed by esophagectomy is considered superior to esophagectomy alone for the treatment of resectable esophageal cancer.
ESOPHAGEAL carcinoma is an aggressive disease with a poor prognosis. Its incidence in the United States has been increasing; it now causes about 9800 deaths per year.1 Many institutions have been using various combinations of multimodality treatment to improve the chance of survival over that achieved with the standard of care for this disease—surgical resection alone. In the late 1980s, neoadjuvant chemoradiotherapy (NC) followed by esophagectomy became a popular approach after several favorable pilot studies were reported.
The University of Massachusetts Medical Center, Worcester, began using this approach in the late 1980s. The other 2 tertiary care affiliates of the University of Massachusetts Medical School—St Vincent's Hospital and Memorial Hospital—began using this approach in 1990. We reviewed the experience of the 3 tertiary care affiliates of the University of Massachusetts Medical School with preoperative chemoradiotherapy followed by esophageal resection for the treatment of esophageal carcinoma to assess the practicality of its application, specific survival end points, and the relief of dysphagia.
Patients with carcinoma of the middle and lower thirds of the esophagus treated between July 1, 1990, and June 30, 1996, at the 3 tertiary care affiliates of the University of Massachusetts Medical School (University of Massachusetts Medical Center, St Vincent's Hospital, and Memorial Hospital) were identified retrospectively by reviewing tumor registry data and office files at the 3 institutions. The 51 patients with localized and resectable (stages I, II, and III) esophageal adenocarcinoma and squamous cell carcinoma who were treated with preoperative cisplatin, fluorouracil, and radiotherapy followed by surgery are the subjects of this study. During this period, some patients were treated with alternative combinations of surgery, chemotherapy, and radiation; they will not be included. From 1995 to the present, patients at University of Massachusetts Medical Center have been treated according to a different multimodality protocol, and they will not be included in this study.
Although this was not a multi-institutional protocol, chemotherapy was similar at all 3 institutions. All patients had 2 courses of preoperative chemotherapy with external beam radiotherapy between courses. Each course of chemotherapy consisted of a 4-hour infusion of cisplatin at 80 to 100 mg/m2 and then a 96-hour infusion of fluorouracil at 900 mg/m2. Patients typically were hospitalized for 4 to 5 days for each course.
External beam radiotherapy was administered between courses of chemotherapy for 3 to 5 weeks at total doses of 3000 to 5400 cGy. The total dose varied according to the preference of the treating hospital: 3000 cGy at St Vincent's Hospital, 4500 cGy at Memorial Hospital, and 5400 cGy at University of Massachusetts Medical Center.
Four to 6 weeks after the end of neoadjuvant therapy, 1 of several methods of surgical resection was performed according to the surgeon's preference. The majority of patients (33 ) had an Ivor Lewis thoracoabdominal esophagogastrectomy; others had either a left thoracoabdominal esophagogastrectomy or a transhiatal esophagectomy.
The main outcome measures included the median survival time, for which survival was defined as the interval from the first surgical consultation before neoadjuvant therapy until the date of the last follow-up; the median preoperative treatment time, which was the interval between the first surgical consultation before neoadjuvant therapy and the date of admission for esophageal resection; the median hospital stay, which was the time spent in the hospital for surgery; and the median follow-up time for the survivors, which was determined by using the interval between the first surgical consultation before neoadjuvant therapy and the date of last follow-up for the survivors. These outcome measures were determined for the group as a whole and for those who had a pathological complete response as well as those who had residual disease in the specimen.
Dysphagia relief was assessed by noting the diet at the first postoperative office visit.
Kaplan-Meier analysis was used to calculate survival times. The log-rank test was used to compare survival rates for those with and without pathological complete response.
Of the 51 patients, 48 (94%) were male. The average age was 61 years (range, 34-81 years). Forty-three patients (84%) presented with dysphagia; 32 (63%) presented with weight loss. Forty-five patients (88%) smoked tobacco and 34 (67%) drank alcohol, with 6 (12%) of the group admitting to alcohol abuse. Forty-three tumors (84%) were adenocarcinomas; 8 (16%) were squamous cell carcinomas. Barrett esophagus was noted in 11 (22%) of the surgical specimens. In 43 patients (84%), the tumor was located more than 30 cm from the incisors by endoscopy; in 8 (16%) the tumors were between 25 and 30 cm from the incisors (Table 1).
Neoadjuvant chemoradiotherapy was tolerated well without major interruption except in 2 patients who had myelosuppression and 1 patient who had severe dysphagia. The median preoperative treatment time was 3.3 months.
The surgical mortality rate (deaths in the hospital after surgery or within 30 days of surgery) was 7.8% (4 deaths). There were no surgery-related deaths for the final 26 patients. The median hospital stay after surgery was 12 days. Dysphagia relief was excellent. Of the survivors of surgery, 44 (93%) were taking a soft (18 ) or a regular (26 ) diet at the first postoperative visit. Median follow-up time for the survivors was 2.5 years.
Of the 51 patients, 15 (29%) had a pathological complete response and 36 (71%) had residual cancer in the operative specimen. The median and 3-year survival rates were 16.3 months and 38% overall, 17.5 months and 53% for those with a pathological complete response, and 15.1 months and 34% for those with residual disease in the operative specimen. The log-rank test did not demonstrate significance (P=.5) for the difference in survival rates between those with a pathological complete response and those with residual disease in the operative specimen (Figure 1 and Figure 2).
Although median follow-up time was relatively short, of 36 patients with residual disease, 6 were alive more than 3 years from diagnosis, and 3 of these patients were alive more than 4.5 years after diagnosis. The longest-surviving patient of the entire group had residual disease at esophagectomy and was disease free 7.6 years after diagnosis. Interestingly, he developed a large squamous cell carcinoma of the proximal deep right thigh, which was his only site of metastasis (or an unusual second primary); it eventually required a modified hemipelvectomy 2.8 years after the initial diagnosis.
When mean survival times were compared for patients with adenocarcinoma vs squamous cell carcinoma, there was not a significant difference (Table 2). Similarly, when mean survival times were compared for those who had an Ivor Lewis esophagogastrectomy vs transhiatal esophagogastrectomy, there was not a significant difference (Table 2).
Esophageal carcinoma is one of the most aggressive visceral cancers, with a poor prognosis and overall estimated 5-year survival rate in the range of 10%.2,3 The standard treatment for localized disease is surgical resection, but survival rates associated with surgery alone are poor. Swisher et al4 reported a median survival time of 13 months and a 3-year survival rate of 25% for patients treated with surgery alone. Ellis5 reported an actuarial 5-year survival rate adjusted for operative mortality and noncardiac deaths of 23.3% for patients with esophageal cancer treated with resection alone. Similar survival data for esophagectomy alone are presented in Table 3.5- 8
Faced with low survival rates with esophagectomy alone as the treatment for esophageal cancer, several institutions have been examining methods to improve survival by using combinations of chemotherapy and radiotherapy. Some have used chemotherapy and radiation without surgery to treat this disease. At Wayne State University in Detroit, Mich, this approach yielded a median survival time of 22 months and a prohibitive toxic reaction from the chemotherapy.9 Al-Sarraf et al10 reported a median survival time of 17.2 months with chemotherapy and radiation without surgery. At Fox Chase Cancer Center, Philadelphia, Pa, 57 patients received chemoradiotherapy without surgery, and the median survival time was 18 months while the 5-year survival rate was 18%.11 Herskovic et al12 randomly assigned patients with resectable esophageal cancer to radiation vs chemoradiation (neither group had surgery). The median survival time was 12.5 months, and the 2-year survival rate was 38% in the chemoradiation group. Interestingly, 10 of 61 patients died before completing the 100 days of chemoradiotherapy, suggesting a 16.4% treatment-associated mortality rate. John et al13 reviewed the survival rates for 30 patients with esophageal cancer treated with chemoradiotherapy. The median survival time was 15 months and the 2-year survival rate was 29%.
Other investigators have combined all 3 modalities, using chemoradiation as a neoadjuvant therapy before surgical resection. Bates et al14 reported a 25.8-month median survival time and a 46% 3-year survival rate with this approach. Forastiere et al15 reported a median survival time of 29 months and a 3-year survival rate of 46%. Others did not report such favorable survival results with this method (Table 4).4,14- 19
Our series from University of Massachusetts Medical School included 51 patients with localized and resectable esophageal cancer who were treated with NC followed by esophagectomy. Our median survival time of 16.3 months is similar to that reported by other institutions using this approach, and it is also not distinctly different from that reported for surgery alone (Table 3 and Table 4). Using median survival time as the survival end point to determine the worth of this treatment suggests that NC is not obviously beneficial. The "costs" of NC are also important. Although the morbidity rate is small, NC consumed 20% of survival time. It is difficult to propose a treatment (NC) that does not seem to improve survival while the time to complete its treatment consumes one fifth of survival time.
At the same time, the value of NC may not be in improving median survival time; it may be in prolonging long-term survival. Urba et al19 recently reported preliminary results of a randomized, controlled study at the University of Michigan, Ann Arbor, comparing surgical resection with NC followed by surgical resection. Although median survival times were similar (16.9 and 17.5 months), 3-year survival rates were different: 15% for surgery alone and 32% for NC followed by surgery. Our 3-year survival rate of 38% compares favorably with the University of Michigan data for NC followed by surgery. Certainly, our data do not prove that long-term survival is improved with NC. They suggest that maturation of the University of Michigan data and further randomized, controlled trials are necessary before we can consider NC followed by esophagectomy to be superior to surgery alone.
One published randomized, controlled prospective trial did report a significant improvement in both the median survival time and the 3-year survival rate associated with NC followed by surgery as compared with surgery alone.16 In this study by Walsh et al16 from Dublin, Ireland, the median survival time was 16 months for the multimodality group as compared with 11 months for the surgical alone group; 3-year survival rates were 32% and 6%, respectively. This is an important study, but there are 2 important considerations. First, only patients with adenocarcinoma of the esophagus were eligible; patients with squamous cell carcinoma were excluded. Second, this was a small study with less than 60 patients in each arm. Clearly, the results of this study need to be reproduced before NC can be accepted as the standard of care.
In our study, 29% of patients had a pathological complete response. This value compares favorably with that reported by others.4,15,16,18 Our data suggest that patients who had a pathological complete response tended to live longer than those who did not. This finding was seen by other authors.4,15,18 It is not clear whether improved survival with a complete response to NC is because the chemoradiotherapy was effective or whether response to chemoradiotherapy is simply a marker for favorable disease (and the subset who responded to NC would have done as well with surgery alone).
Some argue that resection is not necessary after chemoradiotherapy because those who have a complete response do not need it and those who have residual disease do so poorly that resection does not help. Our data cannot definitively address this issue, but it is interesting that our longest current survivors are those who had residual disease removed at surgery. Walsh et al16 reported that the omission of resection would have left 75% of the patients in their multimodality group with residual disease, which in 19 cases (33%) appeared to be confined to the esophagus.
Quality-of-life issues are difficult to assess when complicated treatment regimens are used to manage difficult diseases. Certainly, although most patients can complete combined-modality treatment, it is demanding for the individual patient. In our series, patients spent almost as much time in the hospital for NC (10 days for inpatient chemotherapy) as they did for the surgery (median hospital stay of 12 days). The decision to pursue NC demanded an additional 3.3 months of treatment time before surgical resection. If NC is not worthwhile—or only marginally beneficial—it would be difficult to recommend NC on the basis of the length of treatment. A short treatment time is appealing when survival time is relatively short. As most patients presented with dysphagia, an important quality-of-life issue was relief of dysphagia. In our series, most patients had relief of dysphagia after resection. Our results with relief of dysphagia compare favorably with those reported for surgery alone8: NC did not appear to worsen (or improve) these results.
From our data and those cited above, it is clear that NC followed by esophagectomy may improve survival over esophagectomy alone for patients with adenocarcinoma or squamous cell carcinoma of the esophagus, but the results of further randomized, controlled, prospective trials will be necessary to confirm this impression. Until we have results of these trials, NC followed by esophagectomy should be considered investigational and not the standard of care.
Presented at the New England Surgical Society Annual Meeting, Bolton Landing, NY, September 19, 1997.
Corresponding author: Richard S. Swanson, MD, Department of Surgery, University of Massachusetts Medical Center, 55 Lake Ave N, Worcester, MA 01655.