Improving the accuracy of BRCA1/2 mutation prediction: validation of the novel country-customized IC software

Abstract

Inherited mutations of the BRCA1/2 genes confer a significantly increased risk for breast and/or ovarian cancer development. Several models were elaborated to help genetic counsellors in selecting individuals with high probability of being mutation carriers. The IC software, a country-customized version of the Brcapro model, was recently shown to be particularly accurate in the prediction of carrier probability status in the Italian population. Here, we used our independent series of 70 breast/ovarian cancer families to analyze the performances of the IC software and compare it to widely used models, such as Brcapro and the Myriad mutation prevalence tables. Analysis of the areas under the receiver operator characteristics (ROC) curves indicated that overall the models performed well. However, the IC software and Myriad tables were more efficient in predicting mutated cases, showing a higher sensitivity (94 and 88%, respectively) and negative predictive value (NPV, 94 and 92%, respectively) compared to Brcapro (sensitivity 71 and NPV 83%). IC software also appeared particularly accurate in the identification of families belonging the low mutation risk group (o10%). Finally, most Brcapro failures occurred in the hereditary breast cancer (HBC) family subset, and in 75% of the cases, the IC software corrected them. Our data suggest that the country-customized implementation operated on the Brcapro software generated a more accurate tool for the prediction of BRCA1/2 gene mutation. Whether the IC or other countrycustomized models might improve BRCA1/2 mutation prediction also in non-Italian families needs to be further explored.

Validity of Models for Predicting BRCA1 and BRCA2 Mutations

“…The high-risk samples reflect genetic counseling clinics, and the population-based samples reflect the broader spectrum of patients seen in general oncology practice. The c-statistics reported for the high-risk population are similar to those of previous studies performed in similar settings (53)(54)(55)(56)(57). The c-statistic of models is generally greater in the population-based samples than in the high-risk samples (P = 0.036) (Table 2), which suggests that the models can be applied to more broadly representative settings than high-risk clinics.…”

Prevalence of BRCA1 and BRCA2 genomic rearrangements in a cohort of consecutive Italian breast and/or ovarian cancer families

“…This resulted in the identification of two BRCA1 deletions. The one involving exon 23 and 24 has been recently found in Italian [21,30] and Spanish [22] families, but it had not been characterized at the molecular level. In this paper we described its molecular characterization and provided the breakpoint coordinates suggesting the involvement of an erroneous homologous recombination event between two perfectly matching Alu repeats.…”

Novel BRCA1 and BRCA2 germline mutations and assessment of mutation spectrum and prevalence in Italian breast and/or ovarian cancer families

“…Nevertheless we got a low mutation rate (18.3%) compared to the other subsets. In addition we noted that most of the incorrect evaluations made by the prediction models specifically occurred in this family subset [22]. The lowest mutation rate occurred in HBC families selected according to our first criteria (at least three female breast cancer cases diagnosed at any age, Table 1) compared to other HBC subgroups.…”