Abstract Materials and Methods: A total of 32,024 women were offered the CRA survey at the time of their screening mammograms between January 2017 and June 2018. 8,438 women who declined the CRA and 7 symptomatic women were excluded. Of the remaining 23,579 women, 16,548 had one screening, and 7,031 had two screenings in this time period, leading to 30,610 total screening mammograms in women who filled out the CRA. Lifetime risk of breast cancer was assessed using the BRCAPRO, and Tyrer-Cuzick v7 (TC7) models, while the risk of BRCA mutation was estimated using the BRCAMUTATION model. Patients were identified as HR if either their BRCAPRO or TC7 lifetime risk was greater than 20%, or if their BRCAMUTATION risk was higher than 5%. Otherwise, patients were identified as LR. Pathologies were broken down into four groups: “Invasive Cancer”, “In Situ Cancer” (excluding lobular carcinoma in situ without pleomorphic features), “Benign Excision” (which often are surgically excised, including atypia and complex sclerosing lesion), and “Benign”. The Call Back Rate (percent of BIRADS 0 per screening mammogram, CBR), Cancer Detection Rate (number of “Invasive” plus “In Situ” cancers per 1,000 screening mammograms, CDR), and percent (as a fraction of screening mammograms) of each of the four pathology groups were calculated for HR and LR patients. Results: 13.6% (4,150) of screens were HR, and 86.4% (26,460) were LR. CBR was 11.1% (460) and 8.1% (2,133) for HR and LR, respectively (p<0.0001). On subsequent diagnostic imaging, 2.4% (101) and 1.7% (442) received a BIRADS 4 or 5 (HR & LR respectively, p<0.05). Of these, only 4 HR and 18 LR did not undergo biopsy, in part due to comorbidities. The CDR was 7.7 and 4.6 (HR & LR respectively, p<0.05). The number of “Invasive Cancer”, “In Situ Cancer”, “Benign Excision”, and “Benign” pathologies were 26 (0.63%), 6 (0.14%), 9 (0.22%), and 56 (1.35%) for HR and 91 (0.34%), 32 (0.12%), 70 (0.26%) and 231 (0.87%) for LR with p-values of p<0.05, p=0.73, p=0.63, and p<0.05, respectively. Conclusion: The cancer detection rate was greater in high risk vs. low risk patients (7.7 vs. 4.6, HR vs. LR, p<0.05). High risk patients had a higher percent of “Invasive Cancer” compared to low risk patients (0.63 vs. 0.34, HR vs. LR, p<0.05). These results indicate that this self-administered breast cancer risk stratification is clinically relevant.
Read full abstract