Aims: To assess the validity of grading in the Edinburgh Breast Conservation Series; a consecutive cohort of 1812 early breast cancer patients treated by breast conservation and radiotherapy between 1981 and 1998 in a single specialist centre with > or =9 years' follow-up and full staging data.
Methods and results: A single pathologist (J.St.J.T) graded 1650 cases using the Elston and Ellis method (EE) with particular reference to the component data: acinar differentiation, nuclear pleomorphism and mitotic counts. The original method was then compared with binary scoring of the same components and the relationship to prognosis reassessed. EE grades and individual grade components were prognostic (P < 0.0001) with 10-year cause-specific survival of 95.6%, 86.4% and 74.7% for EE grades 1, 2 and 3, respectively. A binary scoring of grade components produces four groups, splitting EE grade 2 tumours into two groups with different outcomes--10-year survival rates for the four revised grades were 96.0%, 89.0%, 79.7% and 75.4%, respectively.
Conclusions: Existing grading methodology is fully applicable in the narrower context of a conservation series but can be simplified. Subdivision of EE grade 2 into a true intermediate prognosis group and a second group with a worse prognosis also adds benefit.