Onkologie - spezialisierte Tumortherapie

Medical Research Council Oesophageal Cancer Working Party*

---

*Members and collaborators listed at end of report

---

Correspondence to:Dr D J Girling, MRC Clinical Trials Unit, Cancer Division, 222 Euston Road, London NW1 2DA, UK (e-mail:OE02@ctu.mrc.ac.uk)

Summary
Introduction
Patients and methods
Results
Discussion
References

Background
The outlook for patients with oesophageal cancer undergoing surgical resection with curative intent is poor. We aimed to assess the effects of preoperative chemotherapy on survival, dysphagia, and performance status in this group of patients.

Methods
802 previously untreated patients with resectable oesophageal cancer of any cell type were randomly allocated either two 4-day cycles, 3 weeks apart, of cisplatin 80 mg/m²by infusion over 4 h plus fluorouracil 1000 mg/m²daily by continuous infusion for 4 days followed by surgical resection (CS group, n=400), or resection alone (S group, 402). Clinicians could choose to give preoperative radiotherapy to all their patients irrespective of randomisation. Primary outcome measure was survival time. Analysis was by intention to treat.

Findings
No patients dropped out of the study. Resection was microscopically complete in 233 (60%) of 390 assessable CS patients and 215 (54%) of 397 S patients (p<0·0001). Postoperative complications were reported in 146 (41%) CS and 161 (42%) S patients. Overall survival was better in the CS group (hazard ratio 0·79; 95% CI 0·67-0·93; p=0·004). Median survival was 512 days (16·8 months) in the CS group compared with 405 days (13·3 months) in the S group (difference 107 days; 95% CI 30-196), and 2-year survival rates were 43% and 34% (difference 9%; 3-14).

Interpretation
Two cycles of preoperative cisplatin and fluorouracil improve survival without additional serious adverse events in the treatment of patients with resectable oesophageal cancer.

Lancet2002;359:1727-33

The outlook for patients with oesophageal cancer who undergo surgical resection with curative intent is poor, with only about 20-30% survival at 2 years. Factors that contribute to this dismal outlook include presence of locally advanced disease and undetected metastatic cancer at diagnosis. Because of the high rates of locoregional and distant failure, there is much interest in the combination of systemic chemotherapy and local surgical treatment.

There is increasing evidence that oesophageal cancer responds to combination chemotherapy regimens based on cisplatin. Response rates before surgery of 40-60% for squamous carcinoma and 30-40% for adenocarcinoma, and 2-year survival rates of 30-40% with preoperative chemotherapy plus surgery, have been reported.^1-6However, what is not clear from the results of these studies is whether these survival rates are attributable to selection of patients or whether they are an effect of chemotherapy.

We therefore did a pragmatic randomised trial in patients thought to have resectable cancer of the oesophagus, comparing surgical resection as locally practised with or without preoperative chemotherapy, to investigate whether chemotherapy first lengthens survival and second affects dysphagia and performance status. We chose the combination of cisplatin and fluorouracil since they were two of the most active single agents in both squamous carcinoma and adenocarcinoma,⁷and because of reported efficacy in combination.^1,3,7-9We decided to give only two cycles at modest doses in view of the reported rapidity with which responses to chemotherapy arose, to keep delay of definitive surgical treatment to a minimum, to avoid excessive toxicity and morbidity before major surgery in patients who are often old and undernourished, and to encourage acceptability of the trial design by patients.

Patients were eligible for the study if they had previously untreated cancer of the oesophagus that was judged resectable and had been microscopically confirmed as squamous carcinoma, adenocarcinoma, or undifferentiated carcinoma. Patients were invited to take part in the study by the clinician responsible for their care. We included tumours of the upper, middle, or lower third of the oesophagus and of the cardia, but not post-cricoid tumours. Patients were not eligible if they had other previous or concomitant malignant disease, apart from basal-cell carcinoma, or if there was cervical lymph-node involvement or evidence of metastases. Patients had to have no contraindication to surgery or chemotherapy, normal renal function, white-cell count >3·5x10⁹/L, and platelet count >100x10⁹/L. Local ethics committees approved the protocol, and patients' informed consent was obtained.

We randomly allocated eligible patients to either immediate preoperative chemotherapy followed by surgical resection (CS group) or immediate surgical resection alone (S group). Randomisation was done by telephone call to the Cancer Division of the Medical Research Council Clinical Trials Unit. We randomly assigned patients by minimisation, with criteria of surgeon, site of tumour (upper, middle, lower, cardia), WHO performance status,¹⁰and histology (squamous, adenocarcinoma, undifferentiated). Because of the nature of the treatment, randomisation was not masked.

Chemotherapy comprised two 4-day cycles of cisplatin 80 mg/m²by intravenous infusion over 4 h on day 1 and fluorouracil 1000 mg/m²daily as a continuous infusion over 96 h, with an interval of 3 weeks between the first day of each cycle. We gave CS patients antiemetics and hydration intravenously before, during, and after cisplatin infusion. For patients in this group, surgical resection was done 3-5 weeks after the start of the second cycle of chemotherapy, and for patients in the S group, surgery was done as soon as possible after randomisation. The local surgeon decided the surgical procedure for patients in both treatment groups, in accordance with site of the tumour and local practice.

Clinicians chose whether to give preoperative external-beam radiotherapy, but if they did they had to use it for all their trial patients irrespective of randomisation, and the dosage and fractionation policy had to be the same in both treatment groups. If radiotherapy was given, we recommended 25 Gy in five fractions over 1 week, 32·5 Gy in ten fractions over 2 weeks, or a biologically equivalent dose.

Primary outcome measure was survival time. Secondary outcome measures were dysphagia and performance status. Patients were assessed before start of treatment, on completion of therapy, at 3, 6, 9, and 12 months from date of randomisation, and then 6-monthly until death. Assessments before treatment included chest radiography, bronchoscopy for upper-third and middle-third tumours, and liver scan by ultrasonography or computed tomography. At all assessments, the clinician recorded the patient's degree of dysphagia and WHO performance status, with the definitions in table 1. At every follow-up assessment, details of any local recurrence, metastases, and anticancer treatment were given.

---

Sample-size calculations based on the log-rank test showed that we needed 800 patients to detect an increase in 2-year survival from 20% to 30%, with 90% power and two-sided of 0·05. We did analyses by intention to treat with SAS (version 6.12). Survival was calculated from date of randomisation to date of death; surviving patients were censored at the date they were last known to be alive. We calculated disease-free survival from a landmark time of 6 months after date of randomisation to allow for the difference in timing of surgery between the two treatment groups. In this analysis, events--including incomplete resection, local and distant recurrence, and death--arising within the first 6 months were regarded as events at this landmark time. We produced survival curves by the Kaplan-Meier method and made treatment comparisons with the log-rank test. To assess proportions between treatments we used the ²test, and continuous data were compared with the large-scale normal approximation of the Mann-WhitneyUtest. ²tests for interaction, trend, or both were applied to assess whether treatment effect varied according to prognostic factors.

An independent data monitoring and ethics committee monitored data and interim analyses yearly. Conventional survival analyses were supplemented by a Bayesian analysis,¹¹which showed the probable effect our results would have on groups of clinicians with various levels of enthusiasm for chemotherapy. This analysis protected against unduly early termination of the trial, emphasising the need to continue until we obtained a convincing result for a broad range of clinicians.

The sponsor appointed the writing committee, who interpreted data, wrote the report, and submitted it for publication.

Between March, 1992, and June, 1998, 802 patients from 42 European centres were randomly assigned either chemotherapy followed by surgery (CS group, n=400) or surgery alone (S group, 402; figure 1). 69 patients from China were also randomly allocated (34 to CS, 35 to S) but, because of changes in funding of their health-care system, recruitment ceased, and we could not obtain follow-up data. These patients were therefore excluded from the analysis.

We have shown that two cycles of preoperative cisplatin and fluorouracil improve overall and disease-free survival compared with surgical resection alone in patients with resectable oesophageal cancer. Adherence to chemotherapy was good, and survival benefit seemed to be consistent for squamous cancers and adenocarcinomas. In patients who received chemotherapy, surgical resection was more often complete, and resection specimens showed less extension into surrounding tissue and less lymph node involvement, than in patients who had surgery alone, although the proportion of patients with involved resection margins and sites of primary failure were closely similar in the two treatment groups. Resection was more often complete after chemotherapy even though surgery was undertaken later in this group of patients, clearly showing an effect of chemotherapy on the tumour.

The chemotherapy regimen used in our trial proved to be highly acceptable to patients. It required few changes for toxic effects and substantially improved dysphagia before surgery. Carey and colleagues¹also reported that patients were fitter for resection after two cycles of cisplatin 100 mg/m²on day 4 and fluorouracil 1000 mg/m²by continuous infusion for 4 days than before chemotherapy in a non-randomised study of 24 patients with squamous cancer.

Our results differ from those reported for the North American Intergroup by Kelsen and colleagues.¹²These investigators did a randomised trial comparing surgical resection with or without preoperative and postoperative chemotherapy in 467 patients. They reported no significant difference between the two treatment groups in overall or disease-free survival. Comparison of this trial with ours, and identification of differences between them that might help to explain the discrepant results, is important.

Although comparisons across trials must be made with caution, results of surgery alone were closely similar between our trial (in which surgeons decided the surgical procedure according to local practice) and the Intergroup trial (in which surgical procedures were prescribed in the protocol). Survival differences between the two trials were seen in the CS groups: median survival was 2 months longer and 2-year survival 8% better in our trial than in the Intergroup trial.

The populations recruited to the two trials were closely similar. In the Intergroup trial compared with our trial, 84% and 75% of patients were male; 54% and 66% had adenocarcinoma; median age was 62 and 63 years; and the two treatment groups were well balanced with respect to pretreatment characteristics of patients in both trials. The small differences in patients' characteristics are unlikely to explain the different outcomes.

Although the same anticancer drugs were used in both trials, in the Intergroup trial, patients were offered three cycles of preoperative cisplatin 100 mg/m²on day 1 and fluorouracil 1000 mg/m²daily as a continuous infusion from day 1 for 5 days every 4 weeks. Patients with stable or responding disease and resection margins free of tumour were offered a further two cycles of chemotherapy postoperatively, but with the cisplatin dose reduced to 75 mg/m². In our trial, patients were offered only two cycles of preoperative cisplatin 80 mg/m²on day 1 and fluorouracil 1000 mg/m²daily from day 1 for 4 days at an interval of 3 weeks and no postoperative chemotherapy. Thus, total prescribed preoperative doses were cisplatin 300 mg/m²and fluorouracil 15 000 mg/m²over 8 weeks in the Intergroup trial compared with cisplatin 160 mg/m²and fluorouracil 8000 mg/m²over 3 weeks in our trial. Total postoperative doses were cisplatin 150 mg/m²and fluorouracil 10 000 mg/m²in the Intergroup trial compared with no dose in our trial.

In the Intergroup trial, 71% of patients who were assigned chemotherapy received all three preoperative cycles, 13% received two cycles, 16% one cycle, and 1% did not receive any chemotherapy. Although 100% of the planned doses were given for the first cycle, this proportion dropped to 84% for the second, and 70% for the third cycle. Of patients eligible for postoperative chemotherapy, 48% received none, 38% received both cycles, and the remaining 14% only one cycle. In our trial, 90% of patients received both preoperative cycles, 6% one cycle, and 4% no chemotherapy. For these differences to account for the different outcomes, if only in part, one would have to argue that any benefit from the large individual and total doses of chemotherapy given in the Intergroup trial was nullified by toxic effects, and that the benefit seen in our trial was related to choice of a less-demanding regimen in patients who are frequently undernourished.

Other differences between the trials were seen in timing and extent of surgical resection, and could help to explain the different outcomes. In the CS group, there was, as expected, a longer delay before surgery in the Intergroup trial (median 93 days) than in our trial (63 days). In the Intergroup trial, fewer patients underwent surgery: 80% of eligible (not defined) CS patients and 92% of S patients compared with 92% and 97% in our trial, respectively. Resection was deemed microscopically complete, however, in very similar proportions in the two trials: in 62% of CS and 59% of S patients in the Intergroup trial, and in 60% and 54% in our trial, respectively (60% and 53% of patients who did not receive preoperative radiotherapy).

In the Intergroup trial, no preoperative radiotherapy was given. In our trial, centres could choose to give it, but only 9% of patients in each group had radiotherapy, and the regimens used were very similar in the two treatment groups. Radiotherapy could not account for the different trial outcomes. Indeed, survival benefit from chemotherapy in our trial was not altered by removal from the analysis of the 9% of patients who received preoperative radiotherapy.

We have done a systematic search for other randomised trials, published and unpublished, comparing resection with or without preoperative chemotherapy; most of these studies were too small to detect a survival difference that could have been clinically worthwhile (table 6).¹³With the methods of Parmar and colleagues,²⁰we estimated hazard ratios for nine of the ten completed trials. Four had hazard ratios less than 1, implying a survival benefit with chemotherapy, and five had hazard ratios of 1 or greater. With the exception of our trial and the Intergroup trial, in which hazard ratios are stated, we should emphasise that these are approximate estimates. Hazard ratios range from 0·69 to 1·34, and there is substantial heterogeneity (p=0·02).

---

Writing committee--J Bancewicz, P I Clark, D B Smith, R J Donnelly, P M Fayers, S Weeden, D J Girling, T Hutchinson, A Harvey, J Lyddiard. From MRC Clinical Trials Unit, Cancer Division, London (SW, PMF, DJG, TH, AH, JL); Hope Hospital, Salford (JB); Cardiothoracic Centre, Liverpool (RJD); and Clatterbridge Centre for Oncology, Liverpool (PIC, DS).

Members of the working party at the time the trial was planned-- A N Al-Jilaihawi, J Bancewicz, S G Bown, P I Clark, B Cottier, M E Cowen, M H Cullen, D Cunningham, R J Donnelly, P M Fayers, D J Girling, A Harvey, K Jeyasingham, R E Lea, H R Matthews, K Moghissi, G N Morritt, M W Myskow, K M Pagliero, C R Rowland, D B Smith, H M A Yosef.

Aberdeen Royal Infirmary, Aberdeen, UK (A W Hutcheon, M C Nicolson); Royal Victoria and City Hospitals, Belfast (V A Anikin, K McManus, J A McGuigan); Heartlands and Queen Elizabeth Hospitals, Birmingham (A G Casson, D R Ferry, Y F Khalil-Marzouk, J G Temple, D C T Watson); Bradford Royal Infirmary and Airedale General Hospital, Bradford (W A F McAdam, A J Mearns); Frenchay Hospital, Oncology Centre, and Royal Infirmary, Bristol (D Alderson, S J Falk, K Jeyasingham); Essex County Hospital, Colchester (D Menzies, P A Murray, B Sizer); Ninewells Hospital, Dundee (A Cuschieri, J A Dewar, M H Lyall, S Shimi); Hairmyres Hospital, Royal Infirmary, Southern General Hospital, and Western Infirmary, Glasgow (A N Al-Jilaihawi, J R Anderson, D Dodds, A Faichney, E J Junor, A J B Kirk, J McEwan, D Prakash, M A Shahin, H M A Yosef); Castle Hill Hospital, Hull (M E Cowen); Cookridge, Killingbeck, and St James' University Hospitals, Leeds (A M Crellin, J N Primrose, N Saunders, D Sebag-Montefiore, M T Seymour); Aintree Hospital, Cardiothoracic Centre, Clatterbridge Centre for Oncology, and Fazakerley Hospital, Liverpool (P I Clark, R J Donnelly, D D Kerrigan, P G McCulloch, R D Page, D B Smith); Bolton General, Christie, and Hope Hospitals, Oldham Royal Infirmary, and Wythenshawe Hospital, Manchester (H Anderson, J Bancewicz, A H K Deiraniya, D J Flook, M T Jones, N Thatcher, M J S Wilkinson); Middlesbrough General Hospital, Middlesbrough (W M Cooke, P R C Dunlop, A J Rathmell); Norfolk and Norwich Hospital, Norwich (W M C Martin, R Vaughan, D C T Watson); Derriford Hospital, Plymouth (F N F Daniel, Y Rahamin) Royal Berkshire Hospital, Reading (T C B Dehn, J Q Gildersleve); Northern General and Royal Hallamshire Hospitals, Sheffield, and Doncaster Royal Infirmary, Doncaster (G Jacob, C J Stoddard, J A C Thorpe); Leiden Academisch Ziekenhuis, Leiden, Netherlands (H J Keizer, M Nooy, K W E Welvaart).

Furness General Hospital, Barrow-in-Furness, UK (C S Ball); Royal Sussex County Hospital, Brighton (A W Clark, P C Hale); Burnley General Hospital, Burnley (E Gross); Addenbrooke's and Papworth Hospitals, Cambridge (D Gilligan, A Price, F C Wells); Cheltenham General Hospital, Cheltenham (S A G Elyan); Derbyshire Royal Infirmary, Derby (S Y Iftikhar); Gloucestershire Royal Hospital, Gloucester (H Barr, M Vipond); Royal Surrey County Hospital, Guildford (R E Sayer); Raigmore Hospital, Inverness (D Whillis); Royal Lancaster Infirmary, Lancaster (M B McIllmurray); Leicester Royal Infirmary, Leicester (F J F Madden); St George's Hospital, London (M J Knight); North Middlesex Hospital, London (S J Karp); King Edward VII Hospital, Midhurst (R E Sayer, S J Whitaker); Northampton General Hospital, Northampton (C H Macmillan); Poole General Hospital, Poole (V M Laurence); Royal South Hants Hospital, Southampton (C A Baughan, D Weeden, R E Lea); Southend General Hospital, Southend (A Lamont, A C R Robinson); St Alban's City Hospital, St Alban's (G R Sagor); North Staffordshire Royal Infirmary, Stoke-on-Trent (M Deakin, J B Elder, C J Smallpiece); Royal Marsden Hospital, Sutton (D Cunningham, S R Ebbs); Morriston Hospital, Swansea (T H Brown); Treliske Hospital, Truro (S S Hasan); Worthing Hospital, Worthing (G H Dickson).

None declared.

The trial was funded by the British Medical Research Council.

1 Carey RW, Hilgenberg AD, Wilkins EW, Choi NC, Mathisen DJ, Grillo H. Preoperative chemotherapy followed by surgery with possible postoperative radiotherapy in squamous carcinoma of the esophagus: evaluation of the chemotherapy component.  J Clin Oncol 1986; 4: 697-701. [PubMed]

2 Forastiere AA, Gennis M, Orringer MB, Agha FP. Cisplatin, vinblastine and mitoguazone chemotherapy for epidermoid and adenocarcinoma of the esophagus.  J Clin Oncol 1987; 5: 1143-49. [PubMed]

3 Hilgenberg AD, Carey RW, Wilkins EW Jr, Choi NC, Mathisen DJ, Grillo HC. Preoperative chemotherapy, surgical resection, and selective post-operative therapy for squamous cell carcinoma of the esophagus.  Ann Thorac Surg 1988; 45: 357-63. [PubMed]

4 Roth JA, Pass HI, Flanagan MM, Graeber GM, Rosenberg JC, Steinberg S. Randomized clinical trial of preoperative and postoperative adjuvant chemotherapy with cisplatin, vindesine and bleomycin for carcinoma of the esophagus.  J Thorac Cardiovasc Surg 1988; 96: 242-48. [PubMed]

5 Ajani JA, Roth JA, Ryan B, et al. Evaluation of pre- and post-operative chemotherapy for resectable adenocarcinoma of the esophagus or gastroesophageal junction.  J Clin Oncol 1990; 8: 1231-38. [PubMed]

6 Kelsen DP, Minsky B, Smith M, et al. Preoperative therapy for esophageal cancer: a randomized comparison of chemotherapy versus radiation therapy.  J Clin Oncol 1990; 8: 1352-61. [PubMed]

7 Carey RW, Hilgenberg AD, Grillo HC, et al. Esophageal cancer: long term follow-up of patients treated by neoadjuvant chemotherapy, surgery and possible post-operative radiation and/or chemotherapy.Proc Am Soc Clin Oncol1990;9:105.

8 Shields TW, Rosen ST, Hellerstein SM, Tsang T, Ujiki GT, Kies MS. Multimodality approach to treatment of carcinoma of the esophagus.  Arch Surg 1984; 119: 558-62. [PubMed]

9 Kelsen DP, Hilaris B, Martini N. Neoadjuvant chemotherapy and surgery of cancer of the esophagus.  Semin Surg Oncol 1986; 2: 170-76. [PubMed]

10 World Health Organization. WHO handbook for reporting results of cancer treatment: WHO offset publication No 48. Geneva: WHO, 1979.

11 Fayers PM, Ashby D, Parmar MKB. Bayesian data monitoring in clinical trials.  Stat Med 1997; 16: 1413-30. [PubMed]

12 Kelsen DP, Ginsberg R, Pajak TF, et al. Chemotherapy followed by surgery compared with surgery alone for localized esophageal cancer.  N Engl J Med 1998; 339: 1979-84. [PubMed]

13 Sutton P, Clark P. Neo-adjuvant treatment for oesophageal cancer.  GI Cancer 2000; 3: 231-38. [PubMed]

14 Nygaard K, Hagen S, Hansen HS, et al. Pre-operative radiotherapy prolongs survival in operable esophageal carcinoma: a randomized, multicenter study of pre-operative radiotherapy and chemotherapy--the second Scandinavian trial in esophageal cancer.  World J Surg 1992; 16: 1104-10. [PubMed]

15 Schlag PM, for the Chirurgische Arbeitsgemeinschaft fuer Onkologie der Deutschen Gesellschaft fuer Chirurgie Study Group. Randomized trial of preoperative chemotherapy for squamous cell cancer of the esophagus.  Arch Surg 1992; 127: 1446-50. [PubMed]

16 Maipang T, Vasinanukorn P, Petpichetchian C, et al. Induction chemotherapy in the treatment of patients with carcinoma of the esophagus.  J Surg Oncol 1994; 56: 191-97. [PubMed]

17 Kok TC, van Lanschot J, Siersema PD, van Overhagen H, Tilanus HW, for the Rotterdam Esophageal Tumor Study Group. Neoadjuvant chemotherapy in operable esophageal squamous cell cancer: final report of a phase III multicenter randomized controlled trial.Proc Am Soc Clin Oncol1997;16:277a, abstr 984.

18 Law S, Fok M, Chow S, Chu K-M, Wong J. Preoperative chemotherapy versus surgical therapy alone for squamous cell carcinoma of the esophagus: a prospective randomized trial.  J Thorac Cardiovasc Surg 1997; 114: 210-17. [PubMed]

19 Ancona E, Ruol A, Santi S, et al. Prognostic role of downstaging after neoadjuvant chemotherapy in potentially resectable (T2-3 Nx M0) esophageal squamous cell carcinoma.Can J Gastroenterol1998;12(suppl B):82B, abstr 243.

20 Parmar MKB, Torri V, Stewart L. Extracting summary statistics to perform meta-analyses of the published literature for survival endpoints.  Stat Med 1998; 17: 2815-34. [PubMed]