Article Text


Oesophageal II
PWE-023 Prognostic significance of circumferential resection margin status in oesophageal cancer—a systematic review and meta-analysis
  1. O Rotimi1,
  2. D Greenwood2,
  3. Y-K Tu2
  1. 1Department of Histopathology, Leeds Teaching Hospitals Trust, Leeds, UK
  2. 2Biostatistics Unit, Centre for Epidemiology & Biostatistics, University of Leeds, Leeds, UK


Introduction The status of the circumferential resection margin (CRM) in oesophageal cancer has been suggested as a prognostic factor but the reports are conflicting. Also, there are two methods of defining positive CRM—within 1 mm (Royal College of Pathologists, RCPath UK)1 and 0 mm (College of American Pathologists, CAP).2

Methods A systematic review was carried out using a pre-defined protocol and papers that met the inclusion criteria were selected. Data extracted from those with required adjusted HR for meta-analysis using STATA-11 statistical software. Assessments were made for heterogeneity, publication bias, small study effects and sensitivity analysis for influence.

Results Fourteen cohort studies3–16 were systematically reviewed but nine4 5 9 10–15 meta-analysed. Abstract PWE-023 table 1 shows the results of the pooled overall and CRM criteria sub-group estimates. There was significant heterogeneity between the studies (p value<0.001 and I2 value of 74.8%). There was evidence of publication bias and small study effects (Egger's test p value 0.029). None of the studies had undue influence.

Abstract PWE-023 Table 1

Results of fixed effects meta-analysis by CRM criteria

Conclusion This meta-analysis provides evidence that the CRM status in oesophageal carcinoma has prognostic significance. This significance is present irrespective of the criteria used for defining the margin but the estimate for the 0 mm CAP criterion is much higher than those of the within 1 mm RCPath criterion. The overall HR of 1.58 (95% CI 1.40 to 1.79) suggests patients with positive CRM have 60% more risk of death compared to patients with a negative margin. The significant heterogeneity and publication bias are limitations to the study and the former in particular requires further analysis.

Competing interests None declared.

References 1. The RCPath Dataset for the Histopathological Reporting of Oesophageal Carcinoma. 2007.

2. An Overview of the College of American Pathologists Cancer Checklists. 2009.

3. Chao Y-K, et al. Ann Surg Oncol 2011;18:529.

4. Deeter M, et al. Arch Surg 2009;144:618.

5. Dexter SP, et al. Gut 2001;48:667.

6. Khan OA, et al. BJC 2003;88:1549.

7. Sagar PM, et al. BJS 1993;80:1386.

8. Saha AK, et al. Ann Surg Oncol 2009;16:1364.

9. Scheepers JJ, et al. Dis Esophagus 2009;22:42.

10. Sujendran V, et al. BJS 2008;95:191.

11. Verhage RJJ, et al. Am J Surg Path 2011;35:919.

12. Griffiths EA, et al. EJSO 2006;32:413.

13. Pultrum BB, et al. Ann Surg Oncol 2010;17:812.

14. Mirnezami R, et al. Int J Surg 2010;8:58.

15. Thompson S, et al. Ann Surg Oncol 2008;15:3447.

16. Roh MS, et al. Korean J Pathology 2004;38:23.

Statistics from

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.