In the news

Three papers in the November issue of the Journal of Clinical Oncology deal with the vexed question of whether there are differences in tumour characteristics and survival between breast cancer patients who carry germline mutations in the BRCA1 or BRCA2 genes, and those who do not (see Breast cancer summary for background information). Verhoog et al analysed the overall and disease-free survival rates of 28 BRCA2-carrier breast cancer patients from 14 consecutively ascertained families, and compared them with 112 patients, matched for age and date of diagnosis, selected from a cancer registry [Verhoog, L.C. et al. (1999) J Clin Oncol 17, 3396-3402 (Abstract)]. They found no significant difference. However, BRCA2 carriers were significantly more likely to suffer contralateral breast within five years of initial diagnosis. Phillips et al review the accumulating evidence that bears on the question of cancer prognosis and survival in mutation carriers [Phillips, K-A. et al (1999) J Clin Oncol 17, 3653-3663 (Abstract)]. They conclude that the available data are inadequate to answer the question and discuss the problems associated with finding answers. For example, studies need to use incidence rather than prevalence cases to avoid excluding those with a worse prognosis. Problems also arise if there is incomplete follow-up of a significant percentage of patients, if sample sizes are too small, or if different ascertainment methods are used for the patient and control populations (for example, a family cancer clinic for patients and a population cancer registry for controls). In an editorial commenting on both papers, Haber highlights some of the clinical questions that cannot be answered until further information is available [Haber, D. (1999) J Clin Oncol 17, 3367-3370]. For example, is prophylactic drug treatment effective in BRCA1/2 mutation carriers? When is prophylactic mastectomy indicated? What is the optimal frequency of mammographic surveillance? Should cancers in mutation carriers be managed differently from sporadic cases?

Comment: Large, well-designed population-based studies, perhaps involving multinational collaboration, are the only way to resolve these issues. The discussion of the methodological problems by Phillips et al is particularly useful, as similar questions about prognosis, treatment and prophylaxis may well arise in future if genetically distinct subsets of other common diseases are discovered.