Long-term Survival Following Induction Chemoradiotherapy and Esophagectomy for Esophageal Carcinoma 

Abstract & Commentary 

Synopsis: Forty-four esophageal cancer patients participated in a phase II trial of induction with 5-fluorouracil, cisplatin, interferon alpha2b, and external beam radiotherapy prior to esophagectomy. Seventeen patients were alive at a median follow-up time of 50 months and 15 of these had no evidence of disease. Among the 14 patients with > 3 years survival, 12 had no evidence of disease. Although clinical parameters measured were not associated with long-term survival, in this data analysis, a lack of a p53 mutation revealed a trend to longer survival. In summary, further support is provided that multimodal neoadjuvant therapy may affect a cure for esophageal cancer. Moreover, it is concluded that recurrence is unlikely for those patients undergoing this neoadjuvant regimen who survive for 3 years or longer.

Source:Lew JI, et al.Arch Surg. 2001;136:737-742.

Five years have passed since the publication of the first larger phase III trial examining neoadjuvant chemoradiotherapy for esophageal carcinoma.1 Despite this promising earlier study, with 3-year survival rates increased from 6% to 32% in the treatment arm, subsequent phase III trials from other institutions did not achieve the same significantly positive results.2,3 However, given the grim prognosis and increasing incidence of esophageal tumors,4 investigators continue to pursue refinement of neoadjuvant treatment regimens as the best hope for cure.

In the current study, Lew and colleagues built upon previously published work5,6 to report "long-term survivor" (> 3 years) data and to identify clinical parameters associated with long-term survival, among a cohort with stage I-IV resectable adeno- or squamous cell carcinoma of the esophagus. Forty-four patients without evidence of metastatic disease and with American Heart Association performance status of 0-1, Eastern Cooperative Oncology Group performance status of 0-2, and normal hepatic, renal, and hematologic serology values were initially enrolled. Patients with documented tracheobronchial tree or mediastinal involvement were excluded. Pretreatment evaluation included an extensive search for metastases, including pre- and posttreatment CT staging, later reviewed by single radiologist who was blinded to posttreatment pathologic stage. In order to reduce inpatient length of stay, the induction regimen changed from 28 days of 5-fluorouracil (5FU), cisplatin, and interferon alpha2b with concurrent 40 Gy external beam radiation (n = 16) to 21 days of a similar regimen with 45 Gy of radiation (n = 28).

Thirty-six patients underwent curative resection (33 transhiatal esophagectomy; 3 with additional thoracotomy) among the 41 patients who completed the neoadjuvant course. Pathologic review of resected specimens revealed 10 "complete responders" (24%), that is, patients without evidence of residual disease after therapy. Twenty-three more patients (56%), termed "partial responders" in a definition much more rigorous than that used traditionally, had only microscopic disease in resected or nodal tissue. Remaining patients had gross residual tumor and were characterized as having "no response."

At a median follow-up time of 50 months, 17 patients were alive and 15 were without evidence of disease. Fourteen patients were "long-term survivors" and 12 of these were disease-free (median follow-up = 54 months, median survival not yet reached). Seven long-term survivors had a complete pathologic response; 6 of these remained disease-free. The other 7 long-term survivors had a partial response; again, 6 remained disease-free. Lew et al maintain that complete vs. partial pathologic response was not correlated with long-term survival.

In order to examine the value of clinical variables in predicting long-term survivorship, age, sex, 28-day vs. 21-day treatment regimen, clinical tumor stage, clinical nodal stage, and p53 mutation status were examined. Ten patients among the 14 long-term survivors had no p53 mutations; 1 was mutated. The remaining 3 may be inferred over-expressors based on previously published data.5 Lew et al found no statistically significant association between the variables listed and survival. They did note that lack of p53 mutation was associated with a trend to longer survival, but they neither supplied a supporting P value nor reported p53 results in a table that demonstrated the other potential predictor variables.

In summary, 33 of 41 patients (80%) with stage I-IV esophageal carcinoma who completed this neoadjuvant regimen had a major pathologic response. Seventeen of 44 patients enrolled (39%) were alive at a median follow-up of 50 months and 14 of 44 (32%) survived > 3 years, supporting Lew et al’s contention that multimodal therapy can cure esophageal carcinoma. Lew et al concluded that none of the clinical parameters were predictive of long-term survival, although lack of p53 mutation was suggestive. They further concluded that recurrence is unlikely for patients undergoing this regimen who survive > 3 years, noting that only 2 of 14 long-term survivors (14%) had recurred by a median follow-up time of > 54 months.

Comment by Arden Morris, MD

Esophageal carcinoma continues to be an aggressive malignancy with 5-year survival rates of 5-15%, providing impetus to resolve ongoing controversies about therapy modalities and to better predict effective treatment. As in the current phase II study, most investigators pool the histologic subtypes, adeno- and squamous-cell carcinoma, both because the mainstay of treatment continues to be anatomically based surgical resection and because adequately powered studies have been elusive. However, evidence suggests that histology may predict differing responses to chemotherapy.7 Furthermore, while the incidence of squamous cell carcinoma is in decline, the incidence of adenocarcinoma has been increasing steadily over the past 25 years.4 Among larger published phase III trials comparing neoadjuvant chemoradiotherapy to surgery alone, the first, including adenocarcinoma patients only, demonstrated a significant survival benefit with neoadjuvant therapy at 3 years.1 A second trial, including only squamous cell carcinoma patients, demonstrated no overall survival benefit but increased disease-free survival.3 Interestingly, this study used only cisplatin and no 5FU in the chemotherapy regimen, and used only 18.5 Gy of radiation. Historically, phase II studies have shown inconsistent or no benefit when less than 40 Gy radiation was used.8,9 An additional phase III trial of neoadjuvant therapy for esophageal squamous cell carcinoma showed no benefit after using 20 Gy radiotherapy.10 A more recent phase III trial included both histologic types, used 5FU, cisplatin, and vinblastine with 45Gy concurrent radiotherapy.2 Survival at 3 years was 16% in the surgery-only arm and 30% in the neoadjuvant arm, but at a median follow-up of 8.2 years, no significant survival difference could be detected. Clearly, reliable determination of patient and tumor characteristics indicating the most effective treatment regimens will require further stratification based on tumor histology.

Lew et al describe the encouraging discovery of no residual tumor in 24% of resected specimens after neoadjuvant chemoradiotherapy. This finding of complete response, consistent with other phase II and phase III trials, has led some investigators to question the use of performing a potentially highly morbid esophagectomy.11,12 Lew et al argue that available imaging modalities cannot separate complete responders from those with only microscopic disease (56% in this study), and therefore locoregional control continues to mandate resection. Indeed, clinical staging has repeatedly been < 30% accurate in studies which included histologic comparison.1,8,9 Lew et al could have bolstered their argument by reporting results from their single radiologist who retrospectively provided CT staging while blinded to pathologic results.

Although presence of a complete response in the resected specimen does not guarantee a cure, previous studies using more than 40 Gy radiotherapy have indicated a significant association with or at least a trend toward longer survival.1,13 The current study reports a lack of association between complete pathologic response and long-term survival, despite 7 of 10 complete responders (70%) going on to > 3 year survival while only 7 of 23 partial responders (30%) enjoyed > 3 year survival. A p-value is not provided by Lew et al but, by Fischer’s exact test, P = 0.057. Regardless of complete response status, patients will continue to require systemic treatment, but if complete response enhances prognosis, it should certainly continue to be a goal.

Those who argue for trials including nonresection treatment arms will surely agree that more accurate predictors of complete response must be identified. In previously published work from this cohort, p53 mutation correlated positively with residual disease, disease-free interval, and survival.5 The current study examined lack of p53 mutation and long-term survival, reporting no significant association but a trend toward improved survival. Because data provided are incomplete, no P value can be determined by the reader. Another esophageal carcinoma study of biological markers predicting response to neoadjuvant therapy found that p53 was nonprognostic.14 However, Lew et al determined that c-erb B-2 showed a significant favorable association and may prove a fruitful area for future research.

In conclusion, Lew et al have provided further interim data about their cohort of esophageal carcinoma patients subjected to neoadjuvant chemoradiotherapy, especially focusing on patients surviving > 3 years. They add to the literature supporting better outcomes for patients undergoing neoadjuvant therapy but, although data were gathered prospectively, fail to provide a solid denominator of all patients presenting with disease that might allow conclusions to be drawn more accurately. No correlation between the clinicopathologic parameters measured (age, sex, 28-day vs 21-day treatment regimen, clinical tumor stage, clinical nodal stage, and p53 mutation status) and long-term survival were identified, reassuringly in the case of the change in treatment regimen. Ultimately, Lew et al presented excellent clinical results for their 32% long-term survivors (median survival > 54 months) and advocated effectively for ongoing resection, but were unable to identify clinical correlates of long-term survival.


1. Walsh T, et al. N Engl J Med. 1996;335:462-467.

2. Urba SG, et al. J Clin Oncol. 2001;19:305-313.

3. Bosset JF, et al. N Engl J Med. 1997;337:161-167.

4. Blot WJ, et al. JAMA. 1991;265:1287-1289.

5. Ribiero U, et al. Cancer. 1998;83:7-18.

6. Posner MC, et al. Cancer J Sci Am. 1998;4:237-246.

7. Wolfe WG, et al. J Thorac Cardiovasc Surg. 1993;105: 749-755.

8. Terz JJ, et al. Surgery. 1993;114:71-75.

9. Hoff SJ, et al. Ann Thorac Surg. 1993;56:282-287.

10. LePrise EL, et al. Cancer. 1994;73:1779-1784.

11. al-Sarraf M, et al. J Clin Oncol. 1997;15:277-284.

12. Coia LR, et al. J Clin Oncol. 2000;18:455-462.

13. Forastiere AA, et al. J Clin Oncol. 1993;11:1118-1123.

14. Duyhaylongsod FG, et al. Ann Surg. 1995;221(6): 677-683.