A new scoring system for the chances of identifying a BRCA1/2 mutation outperforms existing models including BRCAPRO

Abstract

Purpose: To develop a simple scoring system for the likelihood of identifying a BRCA1 or BRCA2 mutation. Methods: DNA samples from affected subjects from 422 non-Jewish families with a history of breast and/ or ovarian cancer were screened for BRCA1 mutations and a subset of 318 was screened for BRCA2 by whole gene screening techniques. Using a combination of results from screening and the family history of mutation negative and positive kindreds, a simple scoring system (Manchester scoring system) was devised to predict pathogenic mutations and particularly to discriminate at the 10% likelihood level. A second separate dataset of 192 samples was subsequently used to test the model's predictive value. This was further validated on a third set of 258 samples and compared against existing models. Results: The scoring system includes a cut-off at 10 points for each gene. This equates to .10% probability of a pathogenic mutation in BRCA1 and BRCA2 individually. The Manchester scoring system had the best trade-off between sensitivity and specificity at 10% prediction for the presence of mutations as shown by its highest C-statistic and was far superior to BRCAPRO. Conclusion:The scoring system is useful in identifying mutations particularly in BRCA2. The algorithm may need modifying to include pathological data when calculating whether to screen for BRCA1 mutations. It is considerably less time-consuming for clinicians than using computer models and if implemented routinely in clinical practice will aid in selecting families most suitable for DNA sampling for diagnostic testing.T he prevalence of BRCA1 and BRCA2 mutations in families with breast/ovarian cancer depends on the type of cancer found in the families, the number of cases in a family, the age at onset, and the ethnic background. The Breast Cancer Linkage Consortium (BCLC) has carried out analysis in high risk families to ascertain the proportion of familial breast cancers attributable to BRCA1 or BRCA2, or neither.1 In 237 families with at least four cases of breast cancer diagnosed at ,60 years of age (or one or more cases of male breast cancer (MBC) or ovarian cancer contributing to four or more cancers in total) 52% were linked to BRCA1 and 32% to BRCA2, with 16% linked to neither. Of the breast-ovarian families, 81% were linked to BRCA1 and 14% were linked to BRCA2. However, 76% of families with MBC but no ovarian cancer were linked to BRCA2. Of families with four or five cases of female breast cancer only, 67% were not linked to BRCA1 or BRCA2, suggesting the presence of other breast cancer predisposing genes. From this information, it is therefore substantially less likely that families with only female breast cancer cases will harbour BRCA1/2 mutations.

A clinical and genetic analysis of multiple primary cancer referrals to genetics services

“…The MTS was designed to be analogous to the Manchester score for prioritising genetic testing in familial breast/ovarian cancer kindreds. 16 This system assigns a score for each cancer of a particular type that occurs in a single lineage, with higher scores for tumours more characteristic of pathogenic BRCA1/BRCA2 variants (Supplementary Table 1). The sum of scores is incorporated into clinical guidelines, with Z15 providing a threshold for testing in breast/ovarian cancer cases.…”

Family History of Pancreatic Cancer in a High‐Risk Cancer Clinic: Implications for Risk Assessment

“…Secondly, a BRCAPRO estimate performed on case and control probands showed that cases in fact had a slightly lower risk estimate than controls, and therefore should have had no greater likelihood of being offered testing. Although at least two current decision tools factor a history of PC into account in BRCA1/2 risk modeling, neither was readily available or used by our group at the time of clinical decision-making in this study (Couch et al 1997;Evans et al 2004). Nonetheless, it must be emphasized that the decisions surrounding referral and testing were not included in this analysis, and that families were not necessarily offered testing based solely on an elevated BRCAPRO estimation.…”

“…For known carriers of BRCA1/2 mutations, the relative risk of PC is elevated [BRCA1: RR 2.26 (1.26-4.06); BRCA2: RR 3.51 (1.87-6.58)] though absolute risks are small (anonymous, 1999;Thompson et al 2002). Several tools designed to quantify an individual's risk of carrying a BRCA1/2 mutation incorporate a family history of PC, though BRCAPRO, one of the best-known and most widely used, does not (Berry et al 2002;Couch et al 1997;Evans et al 2004). …”

Optimal selection for BRCA1 and BRCA2 mutation testing using a combination of ‘easy to apply’ probability models

“…Additionally, an explanation may be found in the fact that mutations in either of the two genes are related to different life time risks for nonbreast cancers such as ovarian cancer and prostate cancer (Antoniou et al, 2003;Van Asperen et al, 2005). Evans et al (2004) developed the so-called Evans2 model for the selection of BRCA2 mutation-positive families. In our study, this model was analysed for different combinations and cutoff levels.…”

“…To obtain optimal ascertainment, many risk assessment models and prior probability models have been developed and evaluated (de la Hoya et al, 2003;Domchek et al, 2003). Four such models, the Claus, Gilpin, Frank and Evans model (Claus et al, 1998;Gilpin et al, 2000;Frank et al, 2002;Evans et al, 2004) are empirically derived scoring systems, easy to apply in daily practice with the use of a pencil and a paper and easy to understand for both counsellor and patient.With the Claus tables the probability of developing breast cancer can be determined, but not the likelihood of detecting a BRCA mutation (as in prior probability models). These tables are based on series of unselected women with breast cancer.…”