A comparison of the three referral guidelines has revealed differences among the criteria with each of the guidelines. The criteria an individual at any age with a known pathogenic/ likely pathogenic variant in a cancer susceptibility gene within the family and male breast cancer in family identify the same number women from our study group as high risk for breast and ovarian cancer with all three analyzed guidelines. Using other criteria, different guidelines identify a varied number of women as high risk for HBOC.
NCCN guidelines are the most inclusive amongst the guidelines, identifying more than double the number of women compared to the other two guidelines. A group of 95 women have been identified as high-risk only by NCCN, with 67 of those women identified due to having a FDR/SDR with pancreatic or metastatic prostate cancer in the family. NCCN guidelines included this criterion in a 2019 update (12) based on the research that 2–5% of unselected adenocarcinoma patients harbor BRCA1/2 pathogenic variant (28), making BRCA1/2 pathogenic variants the most common genetic cause of pancreatic cancer (29) and so, an attractive candidate gene for genetic testing. Moreover, pancreatic cancer has a high mortality rate, and the possibility to test the affected relative is, therefore, time limited (30). Prostate cancer is the second most common cancer in men (31) and has a high rate of heritability as well (32). A study has shown a 1.2% prevalence rate of pathogenic BRCA1/2 variants in unselected prostate cancer patients (33). Patients with metastatic prostate cancer are known to harbor pathogenic variants in a cancer predisposition genes, including BRCA1 and BRCA2, far more frequently than patients with localized prostate cancer (34), therefore genetic testing of all patients with metastatic prostate cancer may be beneficial, especially since screening and treatment options are available (29). NCCN guideline is also the only guideline that identifies women with a SDR with any of the following as high-risk: ovarian cancer, breast cancer before age of 50 years, breast cancer primaries or two SDRs with breast cancer (1 before 50 years). Those women with SDRs account for the remaining 39 women exclusively identified by NCCN. A previously published study revealed that including SDRs in cancer risk assessment is frequently beneficial, therefore, this criterion has important implications (35).
ACMG/NSGC and SGO guidelines identify 7.1% and 7.0% of women as high-risk, respectively, with some differences amongst the guidelines. The criterion that uniquely identifies women as high-risk with ACMG/NSGC guidelines is having a FDR with breast cancer at ages between 45 and 50 years present in the family. This criterion results in the identification of six additional women. Testing first-degree relatives of women with the diagnosis of breast cancer between ages 45 and 50 years might be reasonable, since a recent study showed that the peak incidence of breast cancer in BRCA1 pathogenic variant carriers occurred between the ages 41 and 50 (36). SGO guidelines identify one woman that is not identified by the other two guidelines, since SGO are the only guidelines that consider two cases of breast cancer (one in a patient younger than 50 years old) in a family sufficient for the identification of high-risk women if one of the cases is third-degree relative (TDR). In our cohort, they additionally identify a woman that has a mother (FDR) and a maternal cousin (TDR) with breast cancer, where the cousin was younger than 50 years at diagnosis. Identification of only one additional woman in our group is supported by a previous study that has shown that information on the breast cancer history of cousins or other TDRs rarely improves the accuracy of risk assessment in a family, therefore, the additional effort of incorporating data on the cancer history of TDRs is seldom beneficial and the inclusion of this criterion in the identification process may have a low yield (37).
Application of various combinations of criteria to our study group reveals that 6.2% of women from our cohort were concordantly identified as high-risk by all guidelines. The criteria that selected these women as high-risk group were a BRCA1/2 pathogenic variant present in the family, patient or a FDR with two or more primary breast cancers, ovarian cancer, breast cancer before age 45 or male breast cancer, or had a combination of three or more specific cancers present in the family. These shared criteria may represent the core criteria of the guidelines, identifying the highest risk women, since it has already been suggested that agreement of multiple guidelines might be considered for the selection of women with the highest risk for the pathogenic variant, viewing each guideline as an ‘expert’ and all guidelines as an ‘expert panel’ (22, 23).
It has recently been observed that the prevalence of BRCA pathogenic variant carriers is higher than previously estimated (38) and that current clinical guidelines fail to identify an important proportion of patients carrying the pathogenic variants (18, 19, 21, 39). Consequently, recent updates of the referral guidelines recommend referral of women with less remarkable family history of cancer (8). Studies that examined older versions of NCCN guidelines for HBOC referral of women from the general population identified more women with each update (17, 22, 23). Our study reveals that the NCCN 2019 guidelines identify 16.7% of women from the general Slovenian population as high risk for HBOC. This is an increase from the referral rates in the aforementioned studies (17, 22, 23) and might account for the expansion of the referral criteria in the recent NCCN guidelines.
The aim of the identification of high-risk women in the population is to reduce morbidity and mortality in those women by referring them to appropriate screening and prevention (40). Our study identified 17.4% of women from the general population older than 18 years as being at an increased risk for HBOC based on their personal and family history of cancer. This subset of women represents a high-risk population with possible implications for cancer prevention, as recommended by the guidelines (40). In the Slovenian population of 2 million, that means more than 150.000 women may need more frequent breast and ovarian cancer screening than women from the general population, which would present an important public health burden. This finding may support further considerations about healthcare implications in Slovenia.
The limitation of our study is that we cannot disclose the specificity and sensitivity of each of the guidelines for the identification of pathogenic variant carriers in high-penetrance cancer genes due to several reasons. Most importantly, the 1000 women from our study population have not been molecularly tested and their BRCA status has not been determined. Because of that, we cannot assess the sensitivity of the guidelines, since we are unable to determine whether some of our average risk population might be asymptotic carriers of a pathogenic variant in a cancer predisposition gene without a family history of cancer and, therefore, part of the high-risk population. This can be partly explained by the fact that pathogenic variant carriers sometimes lack indicative family of cancer because of different reasons: imperfect reporting of cancer disease in the family, presence of adoption or risk-reducing surgeries in a family, families with few female relatives or a small family size (12). Secondly, it is important to note that only a subset (20–30%) of familial breast cancer is explained by a pathogenic variant in BRCA1/2 or another highly penetrant cancer gene (41, 42). Put differently, even when performed, a molecular analysis does not account for breast and ovarian cancer susceptibility in high-risk women with familial breast cancer without an identifiable pathogenic variant, making the specificity of the guidelines for the identification of high-risk population difficult to assess. Further research is needed to determine the rate of cancer predisposition in this high-risk population. Another limitation of our study is self-reporting of personal and family history of cancer diagnosis by the participants in the study without the ability to confirm the diagnoses by consulting the patient records, however, patient-reported data of family history of breast and ovarian cancer has been found reliable in a previous study (43).