Long-term survivorship of esophageal cancer patients treated with radical intent

1Department of Radiation Oncology, British Columbia Cancer Agency, Fraser Valley Centre, Surrey; 2 Department of Population and Preventive Oncology, British Columbia Cancer Agency, Vancouver Centre, Vancouver; 3Department of Medical Oncology, British Columbia Cancer Agency, Fraser Valley Centre, Surrey, British Columbia Correspondence and reprints: Dr Alex Agranovich, Department of Radiation Oncology, British Columbia Cancer Agency, 13750 96 Avenue, Surrey, British Columbia V3V 1Z2. Telephone 604-930-4036, fax 604-930-4065, e-mail aagranov@bccancer.bc.ca Received for publication September 28, 2007. Accepted December 10, 2007 A Agranovich, CE McGahan, A Gurjal. Long-term survivorship of esophageal cancer patients treated with radical intent. Can J Gastroenterol 2008;22(4):393-398.

T he mortality of patients with esophageal cancer remains extremely high.Three-year survival of radical surgery (RS) alone varies greatly, from 6% to 26%, based on prospective randomized studies (1,2).The superiority of combined treatment regimens (CRTs), particularly those including chemotherapy, over radiation therapy (RT) alone and over RS alone is documented in only a few randomized studies (3)(4)(5)(6).Considerable controversy still exists regarding the role of neoadjuvant RT, adjuvant RT and chemotherapy, and their potential to improve the outcomes of surgical resection (2,(7)(8)(9)(10).The obectives of the present retrospective study were to evaluate the prognostic factors influencing survivorship and compare the longterm survivorship of esophageal cancer patients, treated either with RT alone or with CRTs, and to identify the optimal treatment regimen.

PATIENTS AND METHODS
The studied cohort consisted of 138 consecutive esophageal cancer patients with no overt clinical or radiological evidence of distant metastases treated with radical intent between April 1, 1995 and December 31, 2003 in a single institution (Fraser Valley Centre in Surrey, British Columbia).
The staging procedure included obtaining patient history and complete physical examination reports, laboratory results (ie, complete blood count, liver function tests, creatinine, electrolytes, carcinoembryonic antigen test for patients with adenocarcinoma), chest x-ray and computed tomography scan of the chest and abdomen.Positron emission tomography (PET) scans were not readily available in the studied time period.The stage of esophageal cancer was determined based on clinical and radiological information, and on pathology reports for a subgroup of patients treated with RS.Although functional status has not always been documented, all of the patients in the study's cohort were ambulatory and RT was delivered almost exclusively on an outpatient basis, with the exception of short admissions for chemotherapy.
RT was given alone or in combination with chemotherapy.The total dose of external beam RT was no less than 40 Gy in 15 fractions over three weeks.High dose rate endoluminal brachytherapy as a boost was optional, given at a dose of 5 Gy to 15 Gy.
The chemotherapy regimen consisted of 5-fluorouracil (5-FU) 1000 mg/m 2 /day as a 96 h continuous infusion and cisplatin 25 mg/m 2 /day as a 30 min continuous infusion.One of two schedules was used to deliver chemotherapy (Table 1).The second cycle of chemotherapy was eliminated if there was major toxicity.
RS procedures included either total thoracoabdominal esophagectomy or transhiatal esophagectomy with stomach mobilization and pull up.The patients suitable for RS were normally reassessed by computed tomography of the chest and abdomen, and underwent RS four to six weeks after completion of RT or RT plus chemotherapy.
RT alone was given to 77 patients.The CRTs consisted of RT plus chemotherapy (n=26), RT followed by RS (n=16), RT plus chemotherapy followed by RS (n=12) and RS followed by RT (n=7).The patient and treatment characteristics are summarized in Table 2.
After completing their treatment, the patients were seen (with history and physical examination) every three to four months for the first two years and then every six months until five years.Laboratory and imaging studies were performed only if dictated by clinical presentation.ChT Chemotherapy; RS Radical surgery; RT Radiation therapy Subsequently, they were followed by annual correspondence with their general practitioners.

Statistical methods
Nonparametric estimates of the survivor functions were computed using the Kaplan-Meier method (SAS 9.1.3,SAS Institute Inc, USA) for each prognostic factor, and the logrank test was used to identify potential prognostic factors predictive of survival using a 10% significance level.Multivariate analysis was performed using the Cox regression model, including only significant prognostic factors from the univariate analysis.Scaled Schoenfeld residuals were calculated in R using the cox.zph function to check the assumption of proportional hazards.Forward selection was used to determine the final predictors of survival, using a 5% significance level.

RESULTS
The median survival time of the entire cohort (Figure 1) was 13.9 months (95% CI 11.9 to 16.6 months).With a median follow-up period of 5.7 years (range 1.1 to 10.4 years), the overall survival of the entire cohort was 32% over two years and 20% over five years.In the subgroup of patients treated with RT only, the overall survival was 18% and less than 5% over two and five years, respectively, compared with 49% and 35%, respectively in the subgroup receiving CRTs.
At the end of the follow-up period, 32 patients were alive (23.2%) and 106 were deceased (76.8%).The median follow-up period for patients who were still alive was 6.2 years (range 2.2 to 10.4 years).Among the patients who died, only eight (7.5%) of the deaths resulted from causes unrelated to esophageal cancer causes.
The overall median survival is expressed in months and the factors that could potentially have an impact on it are summarized in Table 3.
The multivariate Cox regression model for survival consisted of four potential prognostic factors -sex, disease site, T stage and treatment regimen (RT versus CRT).Using the forward selection method, the final model showed that males had a significantly better survival than females (P= 0.044, hazard ratio of 0.65 [range 0.42 to 0.99]).CRTs as opposed to RT alone, was the greatest influential factor in better survivorship (P<0.0001,hazard ratio of 0.43 [range 0.29 to 0.65]).Figures 2  P<0.05 was defined as significant.*Not estimable (NE) (ie, the upper limit of the 95% CI for the median survival was not reached); † Because the number of patients in the subgroup treated with surgery followed by radiation therapy (RT) was very small (n=7), for the purpose of the analysis, it was combined with the subgroup treated with RT followed by surgery (n=16  The number of complete pathological responses were also examined, as determined by the absence of a viable tumour in the surgical specimen.In 28 patients who underwent RS after preoperative RT or RT plus chemotherapy, the percentage of pathologically proven complete responses was very similar: 37.5% (six of 16) and 33.3% (four of 12) in preoperative RT and RT plus chemotherapy subgroups, respectively.

RT toxicity
Neither esophageal perforation nor bleeding related to RT were encountered in the early and late post-RT period.
Among 25 patients who survived three years or longer, benign stricture (without evidence of tumour recurrence), which required regular or episodic esophageal dilation, developed in four patients (16%).The nutritional status of all four patients remained good.

Chemotherapy toxicity
Data were examined for patients who underwent concurrent chemotherapy and RT (n=38).An acute toxicity was more significant, with no patients being diagnosed with chronic problems that could be attributed to chemotherapy.Patients experienced varying degrees of nausea and emesis ranging from mild to severe, resulting in dehydration, and four patients required intravenous fluids.Esophagitis was also a common side effect, but it is impossible to evaluate whether chemotherapy or RT was the predominant cause.Three patients experienced coronary artery spasm and one death, secondary to myocardial infarction, occurred.Febrile neutropenia was encountered in one patient.One patient died of a cerebrovascular accident within one week of completing treatment.

DISCUSSION
Our retrospective study confirmed that the use of CRTs is superior to RT only, for the treatment of esophageal cancer.However, the question remains unanswered as to what constitutes an optimal combined regimen.
One of the potential reasons for a great variability of reported results in esophageal cancer may lie in the limitation of clinical and radiological staging methods used in selecting patients for definitive treatment.Our study and reported randomized studies of esophageal cancer were performed at a time when PET scans were not widely available for staging esophageal cancer patients.Literature reports indicate the usefulness of PET in staging esophageal cancer by detecting the locoregional extent and distant spread of esophageal cancer (11).Currently, PET scans are routinely used for staging of esophageal cancer at the British Columbia Cancer Agency.Although it may have an impact on the process of selection of patients suitable for radical treatments and improvement of their outcomes, it may not necessarily reduce the overall mortality of esophageal cancer patients because of a potential shift in staging (ie, detecting more patients with distant metastases).Alternatively, there are reports (12,13) of PET scan and endoscopic ultrasound use for RT planning in esophageal cancer, which may lead to better results of this treatment modality.
To correlate our results with the published literature, we reviewed data currently available in the Cochrane database, as well as systematic reviews of available prospective randomized trials using RT (7)(8)(9)(10)14).There was no definitive evidence, based on meta-analysis of existing randomized trials, of the benefits of preoperative and postoperative RT.However, there was good evidence for the superiority of concurrent RT plus chemotherapy over RT alone, mainly based on Radiation Therapy Oncology Group (RTOG) 85-01 study results (3,4).There is also some evidence of an advantage of choosing preoperative RT plus chemotherapy followed by RS versus RS alone, again mainly based on a single and rather small, strongly positive study reported by Walsh et al (1) and two meta-analyses (15,16) that concluded that combined regimens with preoperative RT plus chemotherapy improves three-year survival and reduces three-year mortality compared with surgery alone.In the study conducted by Walsh et al, the survivorship in surgery alone was only 6% over three years.In comparison, the three-year survivorship after surgery alone, reported by Kelsen et al (2) (which was based on an intergroup randomized study conducted by the leading North American oncology centres), was 26% (ie, it was very similar to the RT plus chemotherapy arm of the RTOG 85-01 study).The question that remains unanswered is whether RT plus chemotherapy could achieve similar results to those of RS alone.Our study could not address this question because the The present study's treatment results were compared with those achieved in the above mentioned prospective randomized controlled studies (Table 4).
The long-term treatment results in the Fraser Valley Centre cohort is favourably compared with the best results achieved in two prospective randomized trials, which set standards for treatment of esophageal cancer in North America and Europe.Because of the small number of patients in some of the treatment subgroups, we cannot demonstrate any decisive superiority of one of the CRTs.It also may reflect the heterogeneity of esophageal cancer patients and their variable suitability for more aggressive therapeutic modalities.
An optimal chemotherapy regimen for esophageal cancer is yet to be developed.The MD Anderson Cancer Center (Houston, Texas, USA) used an induction chemotherapy, with two cycles being administered before concurrent chemotherapy and RT.Patients who received this form of treatment had a significantly better three-year survival rate (59% versus 41%) (17).
To date, there are no unequivocally convincing data regarding an optimal regimen of RT for esophageal cancer (8).Perhaps even potentially superior locoregional control, which might be achieved by certain RT regimens, is obscured by a high risk of systemic failure and, thus, does not lead to a better survival rate of esophageal cancer patients.In the studied group of patients, we frequently used a three-week course of RT in patients treated with RT alone or RT followed by RS.In approximately one-half of the patients, an endoluminal brachytherapy boost was added.The pathologically complete response rate in our subgroup treated with RT followed by RS was quite respectable (37.5%) and compared favourably with that reported by Walsh et al (25%) (1).In patients treated with a combination of RT and chemotherapy, a five-week course was used, similar to the RTOG 85-01 trial.For those patients who underwent a subsequent surgery, a complete pathological response was observed in 33.3% of the patients.It demonstrates the effectiveness of the chosen RT regimens in one-third of the patients.At the same time, the locoregional control in the remaining two-thirds of the patients remains unsatisfactory and leaves room for improvement.
There is uncertainty regarding the value of endoluminal brachytherapy in the management of esophageal cancer, specifically high dose rate brachytherapy.We found only one prospective randomized study (18) addressing this issue.It demonstrated an improvement in cause-specific survival in a small subset of esophageal cancer patients (20 of 43) whose tumours were less than 5 cm in length.However, there was no difference in overall survival for the entire cohort in that study.Obviously, it is insufficient to claim that endoluminal brachytherapy is more beneficial than an extra dose of external RT.Traditionally, endoluminal brachytherapy has been used, because it is considered to be a fast and effective method to relieve obstructive symptoms.In our analysis, we also could not demonstrate any improvement in overall survival with the addition of brachytherapy.At the same time, we found it to be totally safe and was not associated with an increased risk for acute or late toxicity as reported in the RTOG phase I/II study (19).

SUMMARY
In the large group of patients with esophageal cancer treated with radical intent, apart from the patients' sex, the most important prognostic factor that positively influenced longterm survivorship was the use of CRTs.With the use of CRTs, nearly one-third of esophageal cancer patients with locoregional disease can be cured.In the present study, we were unable to identify which particular regimen is superior.This question can be answered by conducting and analyzing the results of properly designed prospective randomized trials.

ACKNOWLEDGEMENTS:
The authors thank all Fraser Valley Centre members who are involved in the management of esophageal cancer patients.Special thanks to Ms Linda Douglas for the preparation of the manuscript and to Mr Richard Sobel for editing the commentary and facilitating approval of the study by the Research Ethics Board.

Figure 1 )
Figure 1) Overall survival of the esophageal cancer patient cohort at the Fraser Valley Centre (Surrey, British Columbia), April 1, 1995 to December 31, 2003

Figure 2 )Figure 3 )
Figure 2) Overall survival of patients at the Fraser Valley Centre (Surrey, British Columbia) between April 1, 1995 and December 31, 2003, stratified by sex Long-term survival and treatments of esophageal cancer Can J Gastroenterol Vol 22 No 4 April 2008 397

TABLE 1
Two schedules used to deliver chemotherapy (ChT) to esophageal cancer patients at the Fraser Valley Centre (Surrey, British Columbia), April 1, 1995 to December 31, 2003 Each raw number in the ChT schedule represents the number of the cycle, and the numbers in parentheses indicate that the ChT cycle is optional based on the side effects.RT Radiation therapy; X Radiation therapy is given

TABLE 2
Long-term survival and treatments of esophageal cancer Can J Gastroenterol Vol 22 No 4 April 2008 395

TABLE 3
Overall median survival of patients and the factors that could potentially have an impact on survival

TABLE 4
(3)parison of the present study results with previous prospective randomized controlled studies Three-year survival was determined by Walsh et al(1)and five-year survival was determined by the Radiation Therapy Oncology Group (RTOG)(3)and the present study at the Fraser Valley Centre (FVC) in Surrey, British Columbia.CRT Combined regimens; RS Radical surgery; RT Radiation therapy *