Mediation analysis seeks to determine whether an independent variable affects a response directly or whether it does so indirectly, by way of a mediator or mediators. Scenarios that assume a single mediation are often overly simplistic, and analyses that include multiple mediators are becoming more common, particularly with the incorporation of high-dimensional data. Surprisingly, however, little attention has been given to multiple mediator and interaction effects. In this article, we propose new methods for testing the null hypothesis of no indirect effect with multiple mediators and interaction effects. We allow the estimators of the path effects to be possibly correlated; we also consider the practice of using confidence intervals to determine whether a mediation effect is zero. We compare the performance of our proposed method with existing methods through extensive simulation studies. Finally, we provide an application to data from the Coronary Artery Risk Development in Young Adults (CARDIA) study.

ImportanceEnhanced breast cancer screening with magnetic resonance imaging (MRI) is recommended to women with elevated risk of breast cancer, yet uptake of screening remains unclear after genetic testing.ObjectiveTo evaluate uptake of MRI after genetic results disclosure and counseling.Design, Setting, and ParticipantsThis multicenter cohort study was conducted at the University of Southern California Norris Cancer Hospital, the Los Angeles General Medical Center, and the Stanford University Cancer Institute. Patients were recruited from July 1, 2014, through November 30, 2016. Following multiplex gene panel testing and genetic counseling, patients responded to surveys about breast MRI screening at 3, 6, 12, and 24 months and to a final survey between 3 and 4 years after counseling. Participants met standard clinical criteria for genetic testing or had a 2.5% or greater probability of inherited cancer susceptibility. Patients were categorized based on breast cancer risk from genetic testing results and Tyrer-Cuzick model-calculated risk as having (1) a BRCA or other high-risk pathogenic variant (PV), (2) a moderate-risk PV, (3) a higher lifetime breast cancer risk (≥20%), or (4) a lower lifetime breast cancer risk (<20%). Analysis was conducted from September 28 to November 9, 2023.InterventionsGenetic testing with a 25- or 28-gene panel, and pretest and posttest genetic counseling by a genetic counselor or an advanced practice genetics nurse practitioner, which included cancer-specific screening recommendations.Main Outcomes and MeasuresMRI screening adherence over time across risk groups was estimated using Cox proportional hazards regression modeling. Likelihood of screening adherence (odds ratios [ORs] with 95% CIs), controlling for potential confounders, was estimated using logistic regression.ResultsThis study included 638 patients, with a mean (SD) age of 50.7 (13.3) years at testing. There were 43 patients (6.7%) with a BRCA or other high-risk PV, 16 (2.5%) with a moderate-risk PV, 146 (22.9%) with higher lifetime breast cancer risk, and 433 (67.9%) with lower lifetime breast cancer risk. A total of 52 patients (8.2%) identified as Asian, 21 (3.3%) as Black, 271 (42.5%) as Hispanic, and 255 (40.0) as White. Compared with patients with lower lifetime breast cancer risk, patients with a BRCA or other high-risk PV and those with a moderate-risk PV were approximately 10 times (OR, 9.81 [95% CI, 4.05-23.86]; P < .001) and 4 times (OR, 4.12 [95% CI, 1.10-14.35]; P = .03) as likely to undergo MRI, respectively. Patients with a BRCA or other high-risk PV were nearly 16 times (OR, 15.81 [95% CI, 5.17-48.31]) as likely to report consistent yearly MRI screening compared with patients with lower lifetime risk.Conclusions and RelevanceIn this study, women with inherited PVs conferring increased breast cancer risk had higher and more consistent MRI uptake than women with lower estimated risk. These findings emphasize the importance of genetic cancer risk assessment for effective enhanced breast cancer screening.

Causal mediation analysis seeks to determine whether an independent variable affects a response variable directly or whether it does so indirectly, by way of a mediator. The existing statistical tests to determine the existence of an indirect effect are overly conservative or have inflated type I error. In this article, we propose two methods based on the principle of intersection-union tests that offer improvements in power while controlling the type I error. We demonstrate the advantages of the proposed methods through extensive simulation. Finally, we provide an application to a large proteomic study.

Purpose: Multiplex gene panel testing (MGPT) is used to identify individuals with an inherited susceptibility to cancer. However, little is known about the uptake of screening and surveillance among patients after MGPT and genetic counseling. The purpose of this study was to measure the uptake of guideline-concordant breast cancer screening after genetic testing and counseling. Patients and Methods: 2,000 patients who met NCCN testing guidelines or had ≥2.5% probability of a pathogenic/likely pathogenic variant (PV) were recruited at three cancer genetics clinics (University of Southern California (USC) Norris Comprehensive Cancer Center, Los Angeles County + USC Medical Center, Stanford Cancer Institute) from July 2014 through November 2016. All patients had 25- or 28-gene MGPT and results were disclosed by a genetic counselor, who provided screening recommendations to patients based on their risk. Post-test surveys were administered at three months, six months, one year, two years, and three years. Results: 1,614/2,000 (80.7%) patients were female and 1,147/1,614 (71.7%) completed at least one survey regarding MRI screening for breast cancer over the three years of longitudinal follow-up. Of these, 94/1,147 (8.2%) patients tested positive for at least one PV in a breast cancer risk gene; 58/94 (61.7%) tested positive for PVs in a high-risk breast cancer gene (BRCA1/2 (n=53), CDH1, PALB2, TP53 (n=5)), and 34/94 (36.2%) of patients tested positive for a PV in a gene characterized as moderate-risk at the time of disclosure (CHEK2, ATM, NBN). MRIs were recommended to 43/58 (74.1%) patients with a high-risk breast cancer gene PV, 20/34 (58.8%) patients with a moderate-risk gene PV, and 171/1,053 (16.2%) patients without a breast cancer risk gene PV. Multivariate logistic regression models revealed that patients with a high-risk gene PV were more likely to undergo MRI screening within 3 months of receiving genetic test results (OR=6.54 95% CI [3.09 - 14.43], p< 0.001), within one year (OR=1.34 95% CI [1.18 - 1.52], p< 0.001), two years (OR=1.43 95% CI [1.24 – 1.65], p< 0.001), and three years (OR=1.44 95% CI [1.25 – 1.66], p< 0.001) when compared to patients without a PV. Patients with a moderate-risk PV were also more likely to have undergone MRI within 3 months of receiving genetic test results (OR=2.89 95% CI [1.05 - 7.81], p=0.036), within one year (OR=1.33 95% CI [1.10 - 1.62], p=0.004), two years (OR=1.31 95% CI [1.09 - 1.59], p=0.004), and three years (OR=1.44 95% CI [1.18 - 1.76], p< 0.001), compared to those without a PV (Table 1). Conclusions: After three years of longitudinal follow up of 2000 patients in this multicenter prospective cohort study, patients with a PV in a breast cancer susceptibility gene were more likely to undergo guideline concordant breast MRI compared to patients without a PV. Carriers of high-risk breast cancer gene PVs were over six times as likely to have undergone MRI compared to patients without PVs within the first three months after genetic results disclosure and counseling. These results demonstrate the effectiveness of MGPT and genetic counseling in guiding patients with PVs in breast cancer susceptibility genes to the appropriate adoption of guideline-concordant screening. Odds ratios of MRI screening in patients carrying PV in breast cancer risk genes. Odds in relation to patients who do not carry a PV High risk gene PV: BRCA1/2, CDH1, PALB2, TP53; Moderate Risk PV: CHEK2, ATM, NBN. Percent of patients having undergone an MRI at the specified time points High risk gene PV: BRCA1/2, CDH1, PALB2, TP53; Moderate Risk PV: CHEK2, ATM, NBN. Citation Format: Leah A. Naghi, Charite N. Ricker, Duveen Sturgeon, Julie Culver, Kerry Kingham, Rachel Hodan, Nicolette M. Chun, John Kidd, Joseph Bonner, Christine Hong, Meredith Mills, Sidney S. Lindsey, Kevin McDonnell, Uri Ladabaum, James M. Ford, Stephen Grube, Allison W. Kurian, Gregory E. Idos. Uptake of Breast Cancer MRI Screening in Patients After Multiplex Gene Panel Testing [abstract]. In: Proceedings of the 2022 San Antonio Breast Cancer Symposium; 2022 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2023;83(5 Suppl):Abstract nr P6-02-07.

Purpose Young age at breast cancer (BC) diagnosis and family history of BC are strongly associated with high prevalence of pathogenic variants (PVs) in BRCA1 and BRCA2 genes. There is limited evidence for such associations with moderate/high penetrance BC-risk genes such as ATM , CHEK2 , and PALB2 . Methods We analyzed multi-gene panel testing results (09/2013–12/2019) for women unaffected by any cancer (N = 371,594) and those affected with BC (N = 130,151) ascertained for suspicion of hereditary breast and/or ovarian cancer. Multivariable logistic regression was used to test association between PV status and age at BC diagnosis (≤ 45 vs. > 45 years) or family history of BC after controlling for personal/family non-BC histories and self-reported ancestry. Results An association between young age (≤ 45 years) at diagnosis and presence of PVs was strong for BRCA1 (OR 3.95, 95% CI 3.64–4.29) and moderate for BRCA2 (OR 1.98, 95% CI 1.84–2.14). Modest associations were observed between PVs and young age at diagnosis for ATM (OR 1.22, 95% CI 1.08–1.37) and CHEK2 (OR 1.34, 95% CI 1.21–1.47) genes, but not for PALB2 (OR 1.12, 95% CI 0.98–1.27). For women with BC, earliest age of familial BC diagnosis followed a similar pattern. For unaffected women, earliest age of family cancer diagnosis was significantly associated with PV status only for BRCA1 (OR 2.34, 95% CI 2.13–2.56) and BRCA2 (OR 1.25, 95% CI 1.16–1.35). Conclusions Young age at BC diagnosis is not a strong risk factor for carrying PVs in BC-associated genes ATM, CHEK2, or PALB2 .

10517 Background: TNBC represents ̃15% of invasive BC. Ancestry-specific variabilities in TNBC risk are well-described, with African American (AA) women experiencing higher incidence and mortality from TNBC than women of other races/ethnicities. Increased risk of TNBC has been associated with both rarer (e.g. RAD51C/D, BARD1) and more commonly detected (e.g. BRCA1/2, PALB2) germline PV in hereditary CA predisposition genes, but less is known about ancestry-specific TNBC risks for PV carriers. Methods: We examined clinical and genetic records from women referred for multigene CA panel testing (9/2013-5/2020). Multivariable logistic regression was used to test associations of PV in 13 genes with risk of TNBC after accounting for age, ancestry, and personal/family CA history. We analyzed each gene in the full cohort, and in subcohorts defined by self-reported ancestry. Effect sizes are expressed as odds ratios with 95% confidence intervals. Seven genes are not reported in ancestry-stratified analyses due to small numbers of PV carriers with TNBC. Results: From 627,219 individuals referred for multigene panel testing, 115,337 (18.4%) women with personal history of BC were identified, of whom 17,951 (15.6%) reported TNBC. Personal history of TNBC was reported more frequently in women of African ancestry (26.9%) than in women of European (13.9%) or Asian (11.7%) or Latinx ancestry (14.9%). Ancestry-stratified risks of TNBC associated with germline PVs are seen in the Table. Conclusions: While small samples sizes limit some gene-specific analyses, comparable ancestry-specific risks of TNBC were seen across the racial/ethnic groups examined here. [Table: see text]

Background Although several hereditary cancer predisposition genes have been implicated in pancreatic ductal adenocarcinoma (PDAC) susceptibility, gene-specific risks are not well defined and are potentially biased because of the design of previous studies. More precise and unbiased risk estimates can result in screening and prevention better tailored to genetic findings. Methods This is a retrospective analysis of 676 667 individuals, 2445 of whom had a personal diagnosis of PDAC, who received multigene panel testing between 2013 and 2020 from a single laboratory. Clinical data were obtained from test requisition forms. Multivariable logistic regression models determined the increased risk of PDAC because of pathogenic variants (PVs) in various genes as adjusted odds ratios (ORs) with 95% confidence intervals (CIs). Multivariable odds ratios were adjusted for age, personal and/or family cancer history, and ancestry. Results Overall, 11.1% of patients with PDAC had a PV. Statistically significantly elevated PDAC risk (2-sided P < .05) was observed for CDK2NA (p16INK4a) (OR = 8.69, 95% CI = 4.69 to 16.12), ATM (OR = 3.44, 95% CI = 2.58 to 4.60), MSH2 (OR = 3.17, 95% CI = 1.70 to 5.91), PALB2 (OR = 3.09, 95% CI = 2.02 to 4.74), BRCA2 (OR = 2.55, 95% CI = 1.99 to 3.27), and BRCA1 (OR = 1.62, 95% CI = 1.07 to 2.43). Conclusions This study provides PDAC risk estimates for 6 genes commonly included in multigene panel testing for hereditary cancer risk. These estimates are lower than those from previous studies, possibly because of adjustment for family history, and support current recommendations for germline testing in all PDAC patients, regardless of a personal or family history of cancer.

There is an increasing interest in using multiple types of omics features (e.g., DNA sequences, RNA expressions, methylation, protein expressions, and metabolic profiles) to study how the relationships between phenotypes and genotypes may be mediated by other omics markers. Genotypes and phenotypes are typically available for all subjects in genetic studies, but typically, some omics data will be missing for some subjects, due to limitations such as cost and sample quality. In this article, we propose a powerful approach for mediation analysis that accommodates missing data among multiple mediators and allows for various interaction effects. We formulate the relationships among genetic variants, other omics measurements, and phenotypes through linear regression models. We derive the joint likelihood for models with two mediators, accounting for arbitrary patterns of missing values. Utilizing computationally efficient and stable algorithms, we conduct maximum likelihood estimation. Our methods produce unbiased and statistically efficient estimators. We demonstrate the usefulness of our methods through simulation studies and an application to the Metabolic Syndrome in Men study.

The prevalence, penetrance, and spectrum of pathogenic variants that predispose women to two or more breast cancers is largely unknown. We queried clinical and genetic data from women with one or more breast cancer diagnosis who received multigene panel testing between 2013 and 2018. Clinical data were obtained from provider-completed test request forms. For each gene on the panel, a multivariable logistic regression model was constructed to test for association with risk of multiple breast cancer diagnoses. Models accounted for age of diagnosis, personal and family cancer history, and ancestry. Results are reported as odds ratios (ORs) with 95% confidence intervals (CIs). This study included 98,979 patients: 88,759 (89.7%) with a single breast cancer and 10,220 (10.3%) with ≥ 2 breast cancers. Of women with two or more breast cancers, 13.2% had a pathogenic variant in a cancer predisposition gene compared to 9.4% with a single breast cancer. BRCA1, BRCA2, CDH1, CHEK2, MSH6, PALB2, PTEN, and TP53 were significantly associated with two or more breast cancers, with ORs ranging from 1.35 for CHEK2 to 3.80 for PTEN. Overall, pathogenic variants in all breast cancer risk genes combined were associated with both metachronous (OR 1.65, 95% CI 1.53–1.79, p = 7.2 × 10–33) and synchronous (OR 1.33, 95% CI 1.19–1.50, p = 2.4 × 10–6) breast cancers. This study demonstrated that several high and moderate penetrance breast cancer susceptibility genes are associated with ≥ 2 breast cancers, affirming the association of two or more breast cancers with diverse genetic etiologies.

Abstract BACKGROUND: TNBC is among the most aggressive subtypes of invasive breast cancer (BC), and accounts for approximately 10-15% of incidental BC diagnoses. TNBC is associated with early age of onset (median age of diagnosis <50) and disproportionately affects African American women. Breast MRI is currently recommended to screen for BC in women with at least a moderate-to-high lifetime risk of BC (a 2-fold or higher increased risk), and may also be superior to mammogram to screen for TNBC. TNBC has been most closely associated with germline PVs in BRCA1. However, recent studies have suggested that PVs in other genes previously associated with invasive BC may specifically confer high risks of the TNBC subtype. METHODS: Results were analyzed from 627,219 women undergoing clinical multi-gene panel testing at a single US-based commercial laboratory between 5/2013 and 2/2020, including genes associated with hereditary BC and other cancers. Demographic and personal/family history data were collected on a test requisition form. Individuals who had single- or founder-site testing, or prior BRCA1 or BRCA2 testing, were excluded. Multivariable regression analysis was used to examine the association between PVs/suspected PVs and personal history (PHx) of TNBC. Models were adjusted for age, personal/family cancer history, and ancestry. Odds ratios (OR) with 95% confidence intervals (CI) excluding 1.0 were considered significant. RESULTS: In total, 22.4% (140,467/627,219) of women tested reported PHx of BC, of whom 12.8% (17,951/140,467) reported PHx of TNBC. Elevated risks of TNBC were identified in carriers of PVs in 10 genes (see Table). While the highest TNBC risk was associated with PVs in BRCA1 (OR 21.24, 95% CI 19.71-22.88), high risks were also seen for BARD1 (OR 7.05, 95% CI 5.71-8.71), TP53 (OR 5.64, 95% CI 3.08-10.33), PTEN (OR 5.52, 95% CI 2.35-13.00) and PALB2 (OR 5.27, 95% CI 4.55-6.10). Moderate-to-high risks (2-5-fold increased risk) of TNBC were also seen for carriers of PVs in RAD51C, RAD51D, BRCA2, and CDKN2A/P16. By contrast, PVs in NBN, ATM, and CHEK2 were all associated with an apparent decreased risk of TNBC. CONCLUSIONS: PVs in several hereditary cancer genes routinely tested on multi-gene panel tests are associated with high risks (OR>5.0) and moderate-to-high risks (OR 2.0-5.0) of TNBC. These findings can inform practice guidelines about which genes to test when evaluating breast cancer risk and which PV carriers may benefit from intensive breast screening with magnetic resonance imaging (MRI). Odds ratios for TNBC in germline carriers of PV in hereditary cancer risk genesRisk GenePV Positive with TNBCOR95% CIp-valueHigh RiskBRCA1119321.2419.71-22.88<0.001BARD11257.055.71-8.71<0.001TP53125.643.08-10.33<0.001PTEN65.522.35-13.00<0.001PALB22315.274.55-6.10<0.001Moderate-to-High RiskRAD51C864.923.86-6.26<0.001RAD51D454.643.34-6.45<0.001BRCA24884.434.02-4.89<0.001CDKN2A (p16)172.521.52-4.18<0.001Moderate-to-Low RiskBRIP1771.921.51-2.44<0.001Protective EffectNBN140.550.32-0.940.030ATM400.510.37-0.70<0.001CHEK2490.440.33-0.58<0.001HOXB1350.330.14-0.810.015 Citation Format: Michael J Hall, Eric Rosenthal, Susana San Roman, Ryan Bernhisel, John Kidd, Elisha Hughes, Thomas Slavin, Allison Kurian. Triple-negative breast cancer (TNBC) risk with pathogenic variants (PV) in hereditary cancer predisposition genes [abstract]. In: Proceedings of the 2020 San Antonio Breast Cancer Virtual Symposium; 2020 Dec 8-11; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2021;81(4 Suppl):Abstract nr PD10-03.

e13653 Background: The HOXB13 c.251G > A (p.G84E) variant is associated with increased prostate cancer risk, possibly at younger ages. Studies on this variant have focused primarily on men, although genetic testing for hereditary cancer risk is most often performed in women. There remains a need to assess whether male and female carriers of this variant have increased risk for other cancers. Methods: We identified male and female carriers of the HOXB13 c.251G > A (p.G84E) variant among individuals referred for hereditary cancer panel genetic testing from October 2018 – December 2019. Non-carriers had no pathogenic variants in any gene or variants of uncertain significance in HOXB13. Personal and family (1 st - and 2 nd -degree relative) cancer histories were obtained from provider-completed test request forms. Multivariable logistic regression (MLR) models were conducted separately for males and females to estimate cancer risks for the variant as odds ratios (ORs), and 95% Wald confidence intervals (CIs) adjusted for age, ancestry and personal/family cancer history. Results: The analysis included 197,978 patients: 4.5% (8,998/197,978) male and 95.5% (188,980/197,978) female. The HOXB13 variant was present in 0.44% (40/8,998) of tested males, 45% (18/40) of whom had a diagnosis of prostate cancer. The variant was present in 0.32% (621/188,980) of tested females. Male carriers with prostate cancer were diagnosed at younger ages (median, 56; Interquartile ratio [IQR], 52, 62) than non-carriers (median, 63; IQR, 57,70), but this difference was not statistically significant. The MLR model calculated a 3.30 OR for prostate cancer in male carriers of c.251C > A ( P = 0.01; 95% CI 1.30-8.39). In an analysis combining males and females, carriers were significantly more likely to report prostate cancer in a family member than non-carriers ( P = 3.5 x 10 −7 ; OR 1.61, 95% CI 1.34-1.93). There was no apparent association with increased risk for other cancers among carriers versus non-carriers, or among relatives of carriers compared with relatives of non-carriers. Conclusions: The HOXB13 c.251G > A (p.G84E) variant was associated with significantly increased risk of prostate cancer, confirming previously published studies. We found no evidence of association with other cancers. For unaffected male carriers, who may frequently be identified through testing of a female relative, identification of this HOXB13 variant provides an opportunity for more precise prostate cancer risk stratification and screening.

1538 Background: BRIP1/ FANCJ participates in DNA replication and repair via interactions with BRCA1 and possibly MLH1. Previous studies have reported that pathogenic variants (PV) in BRIP1 are associated with an ~2-fold increase in risk for ovarian cancer (OC) and triple-negative breast cancer (TNBC). Although multigene panel testing for hereditary cancer (CA) has identified BRIP1 PV and uncertain variants (VUS) in patients with diverse CAs including breast (BC), colorectal (CRC) and melanoma (Mel), association with these CA types has not been established. Methods: We examined BRIP1 risks in two independent populations: Fox Chase Cancer Center (FCCC) and Myriad Genetics (MGL). At FCCC, pedigrees of BRIP1 PV ( N= 10) and VUS families ( N= 47) were reviewed. The MGL population included patients referred for testing by multigene panel (9/2013-12/2019) ( N= 586,740). Multivariable logistic regression analysis estimated BRIP1 CA risks as odds ratios (ORs) and 95% confidence intervals (CIs). Models were adjusted for age, sex, ancestry, personal CA history (PHX), and family CA history. Results: In the FCCC cohort, BRIP1 PV carriers ( N= 12) reported PHX of early-onset ( < 50) BC, CRC, and bladder CA. BRIP1 VUS were also identified among several patients with striking PHX and negative panel testing: BC < 40 ( N= 3), bilateral BC ( N= 4), TNBC ( N= 2), CRC < 40 ( N= 3), and a patient with 3 CAs < 40 (CRC, BC, and Mel). All FCCC families with a BRIP1 PV and select VUS families ( N= 6) are seen in the Table. In the MGL population, 0.3% (1,678/586,740) carried a BRIP1 PV. Logistic regression analyses found that female BRIP1 PV carriers have significantly increased risk for OC (OR 2.40, 95% CI 1.93-2.98) and TNBC (OR 1.93, 95% CI 1.52-2.46). Data were insufficient for testing risk of bladder or prostate CA. Findings did not support associations of BRIP1 with CRC, melanoma, endometrial, pancreatic or gastric CA. Conclusions: BRIP1 PV and VUS may be identified in patients with diverse CA histories. These results confirm studies showing that BRIP1 PV are associated with an ~2-fold increased risk of OC and TNBC, but do not support increased risks of CRC, melanoma or endometrial CA in BRIP1 PV carriers. [Table: see text]

The role of genetic predisposition in male breast cancer (MBC) patients who test negative for a BRCA mutation is unclear. The aim of this study is to define the association between MBC and family history of breast cancer in patients without mutations in BRCA1 or BRCA2. We conducted an unmatched case–control study with men who received commercial testing for germline mutations in cancer susceptibility genes, including 3,647 MBC cases who tested negative for deleterious mutations in BRCA1/BRCA2, and 4,269 men with a personal history of colorectal cancer who tested negative for mutations in DNA mismatch repair genes to serve as controls. Associations between family history of breast cancer and MBC were estimated using unconditional multivariable logistic regression with adjustment for age, race/ethnicity and year of testing. Breast cancer in a first- or second-degree relative was associated with a four-fold increased odds of MBC (OR 4.7; 95% CI 4.1, 5.3). Associations with MBC were strongest for family history of breast cancer in 2 or more first-degree relatives (FDR) (OR 7.8; 95% CI 5.2, 11.6), for probands and FDR diagnosed at age < 45 years (OR 6.9; 95% CI 3.9, 12.4), and for family history of MBC (OR 17.9; 95% CI 7.6, 42.1). Findings were confirmed in a sensitivity analysis of MBC cases who tested negative on a 25-gene pan-cancer panel. MBC patients without mutations in BRCA1/2 have significantly higher odds of a family history of breast cancer, suggesting the existence of unidentified MBC susceptibility alleles.

Background Little is known about the psychological outcomes of germline multigene panel testing, particularly among diverse patients and those with moderate‐risk pathogenic variants (PVs). Methods Study participants (N = 1264) were counseled and tested with a 25‐ or 28‐gene panel and completed a 3‐month postresult survey including the Multidimensional Impact of Cancer Risk Assessment (MICRA). Results The mean age was 52 years, 80% were female, and 70% had cancer; 45% were non‐Hispanic White, 37% were Hispanic, 10% were Asian, 3% were Black, and 5% had another race/ethnicity. Approximately 28% had a high school education or less, and 23% were non–English‐speaking. The genetic test results were as follows: 7% had a high‐risk PV, 6% had a moderate‐risk PV, 35% had a variant of uncertain significance (VUS), and 52% were negative. Most participants (92%) had a total MICRA score ≤ 38, which corresponded to a mean response of “never,” “rarely,” or only “sometimes” reacting negatively to results. A multivariate analysis found that mean total MICRA scores were significantly higher (more uncertainty/distress) among high‐ and moderate‐risk PV carriers (29.7 and 24.8, respectively) than those with a VUS or negative results (17.4 and 16.1, respectively). Having cancer or less education was associated with a significantly higher total MICRA score; race/ethnicity was not associated with the total MICRA score. High‐ and moderate‐risk PV carriers did not differ significantly from one another in the total MICRA score, uncertainty, distress, or positive experiences. Conclusions In a diverse population undergoing genetic counseling and multigene panel testing for hereditary cancer risk, the psychological response corresponded to test results and showed low distress and uncertainty. Further studies are needed to assess patient understanding and subsequent cancer screening among patients from diverse backgrounds. Lay Summary Multigene panel tests for hereditary cancer have become widespread despite concerns about adverse psychological reactions among carriers of moderate‐risk pathogenic variants (mutations) and among carriers of variants of uncertain significance. This large study of an ethnically and economically diverse cohort of patients undergoing panel testing found that 92% “never,” “rarely,” or only “sometimes” reacted negatively to results. Somewhat higher uncertainty and distress were identified among carriers of high‐ and moderate‐risk pathogenic variants, and lower levels were identified among those with a variant of uncertain significance or a negative result. Although the psychological response corresponded to risk, reactions to testing were favorable, regardless of results.

Currently, CDKN2A and CDK4 have primarily been implicated in familial melanoma. However, pathogenic variants (PVs) in these genes only represent a small percentage of familial melanoma cases, indicating that other genes may be involved. By looking at the patient family history data of over 500,000 individuals tested with a pan-cancer panel because they met National Comprehensive Cancer Network (NCCN) hereditary breast and ovarian cancer and/or Lynch syndrome criteria, we identified 27,946 patients with a personal history of melanoma or first-degree relative with melanoma. Of these patients, 1932 (6.9%) patients carried a PV in a cancer predisposition gene, which was significantly higher than in individuals with no personal history of any cancer and no first-degree relative with melanoma (5.1%, P < .0001). These PVs were most commonly identified in CHEK2 (n = 356), BRCA2 (n = 313), ATM (n = 268), BRCA1 (n = 242), PALB2 (n = 137), and CDKN2A (n = 121). In this cohort, only one patient had a PV in CDK4. Of the 1520 patients first diagnosed with melanoma who went on to develop an additional primary cancer, 148 (9.7%) had a PV. These individuals went on to develop breast, colorectal, ovarian, or another cancer. Early identification of PV-carrying patients with melanoma would present the provider and patient with the opportunity to prevent the occurrence of a second primary cancer through increased surveillance or risk-reducing surgeries. Identification of patients with concerning family cancer histories eligible for hereditary cancer testing according to NCCN guidelines during dermatology appointments may lead to positive health outcomes by preventing future cancers.

Introduction:Common low-penetrance genetic variants (SNPs) has been consistently associated with colorectal cancer (CRC) risk, but the risk conferred by each of these variants is usually modest.The relevance of these SNPs in the development of colorectal adenomas has been scarcely evaluated.Aims: We aimed to evaluate the potential of a genetic risk score (GRS) to identify individuals with an increased risk of colorectal adenomas.METHODS.Patients were selected from the Spanish colorectal cancer screening registries in Aragon and The Canary Islands.Patients with hereditary non-polyposis CRC, familial adenomatous polyposis, history of CRC or inflammatory bowel disease were excluded.We included 858 patients with no precancerous colorectal lesions and 642 with adenomas.Genomic DNA from participants was genotyped by the MassArray™(Sequenom) platform for a panel of 99 SNPs previously associated with CRC risk.We derived a GRS based on the 11 SNPs significantly associated with adenoma development in our study.The number of risk alleles was coded as 0, 1 or 2 for each SNP assuming a log-additive genetic effect.To test the relevance of our score, we estimated the impact of having each additional risk allele by computing the odds ratio (OR) and the associated 95% confidence interval (95%CI).In addition, the predictive value for colorectal adenoma risk was calculated as the area under the ROC curve (AUC).PredictABEL R package was used to compute the genetic risk score.RESULTS.Five SNPs were identified using a multivariate logistic model that included the 11 SNPs that were associated with colorectal adenomas in our study: rs10505477, rs11255841, rs13181, rs4779584 and rs8180040.The distribution of risk alleles for cases and controls are showed in Figure 1.We observed that the risk of developing adenomas increased with the number of risk alleles (per-allele OR=1.183, 95%CI 1.101 to 1.270 p<0.001).Five risk alleles were considered as reference since it was the median number in patients without adenomas.See Table 1.Individuals with ≥8 risk alleles had nearly a 2.4-fold increase in adenoma risk compared with those with 5 risk alleles.The risk score showed a discriminating power of 57% (AUC = 0.571, 95%CI 0.541 to 0.601).CONCLUSIONS.An increased GRS based on CRC-associated SNPs was associated with increased prevalence of colorectal adenomas.These findings may help to identify better a subgroup of patients with increased risk of adenomas who would benefit from stricter cancer screening programs.

Purpose Multiplex gene panel testing (MGPT) allows for the simultaneous analysis of germline cancer susceptibility genes. This study describes the diagnostic yield and patient experiences of MGPT in diverse populations. Patients and Methods This multicenter, prospective cohort study enrolled participants from three cancer genetics clinics—University of Southern California Norris Comprehensive Cancer Center, Los Angeles County and University of Southern California Medical Center, and Stanford Cancer Institute—who met testing guidelines or had a 2.5% or greater probability of a pathogenic variant (N = 2,000). All patients underwent 25- or 28-gene MGPT and results were compared with differential genetic diagnoses generated by pretest expert clinical assessment. Post-test surveys on distress, uncertainty, and positive experiences were administered at 3 months (69% response rate) and 1 year (57% response rate). Results Of 2,000 participants, 81% were female, 41% were Hispanic, 26% were Spanish speaking only, and 30% completed high school or less education. A total of 242 participants (12%) carried one or more pathogenic variant (positive), 689 (34%) carried one or more variant of uncertain significance (VUS), and 1,069 (53%) carried no pathogenic variants or VUS (negative). More than one third of pathogenic variants (34%) were not included in the differential diagnosis. After testing, few patients (4%) had prophylactic surgery, most (92%) never regretted testing, and most (80%) wanted to know all results, even those of uncertain significance. Positive patients were twice as likely as negative/VUS patients (83% v 41%; P < .001) to encourage their relatives to be tested. Conclusion In a racially/ethnically and socioeconomically diverse cohort, MGPT increased diagnostic yield. More than one third of identified pathogenic variants were not clinically anticipated. Patient regret and prophylactic surgery use were low, and patients appropriately encouraged relatives to be tested for clinically relevant results.

Genome-wide association studies (GWASs) have identified thousands of genetic loci associated with cardiometabolic traits including type 2 diabetes (T2D), lipid levels, body fat distribution, and adiposity, although most causal genes remain unknown. We used subcutaneous adipose tissue RNA-seq data from 434 Finnish men from the METSIM study to identify 9,687 primary and 2,785 secondary cis-expression quantitative trait loci (eQTL; <1 Mb from TSS, FDR < 1%). Compared to primary eQTL signals, secondary eQTL signals were located further from transcription start sites, had smaller effect sizes, and were less enriched in adipose tissue regulatory elements compared to primary signals. Among 2,843 cardiometabolic GWAS signals, 262 colocalized by LD and conditional analysis with 318 transcripts as primary and conditionally distinct secondary cis-eQTLs, including some across ancestries. Of cardiometabolic traits examined for adipose tissue eQTL colocalizations, waist-hip ratio (WHR) and circulating lipid traits had the highest percentage of colocalized eQTLs (15% and 14%, respectively). Among alleles associated with increased cardiometabolic GWAS risk, approximately half (53%) were associated with decreased gene expression level. Mediation analyses of colocalized genes and cardiometabolic traits within the 434 individuals provided further evidence that gene expression influences variant-trait associations. These results identify hundreds of candidate genes that may act in adipose tissue to influence cardiometabolic traits.

1525 Background: Guidelines recommend consideration of prophylactic surgery for patients with a germline pathogenic variant in some cancer predisposition genes. We assessed surgery utilization in a prospective, multi-institutional cohort study of MGPT. Methods: 2000 patients had MGPT and completed questionnaires at 3, 6, and 12 months. Patients reported surgical utilization and indication (treatment or prevention). Surgery utilization was assessed according to cancer history and MGPT test results: Positive, pathogenic variant; VUS, variant of uncertain significance; Negative, benign variants. Results: Overall, 12.9% (198/1537) of patients reported surgery after MGPT (median follow-up 13 months). Only 31.3% (62/198) of patients specified that their surgery was preventive. Preventive surgery utilization was significantly higher among patients who tested positive (n=30, 14.9%) compared to those testing negative (n=20, 2.3%, p<0.001) or VUS (n=12, 2.2%, p<0.001). Preventive surgery was very low among patients testing negative or VUS who had no personal history of cancer in the relevant organ (Table). For example, mastectomy was not reported among any patients testing negative or VUS who had no personal history of breast cancer (Table). Conclusions: More than one year after MGPT, prophylactic surgery use was low among patients with VUS or negative results, especially among those with no personal history of cancer at the relevant site. Surgery utilization. [Table: see text]

Background Healthcare providers increasingly use information about pathogenic variants in cancer predisposition genes, including sequence variants and large rearrangements (LRs), in medical management decisions. While sequence variant detection is typically robust, LRs can be difficult to detect and characterize and may be underreported as a cause for hereditary cancer risk. This report describes the outcomes of hereditary cancer genetic testing using a comprehensive strategy that employs next-generation sequencing (NGS) for LR detection, coupled with LR confirmation using repeat hybrid capture NGS, microarray comparative genomic hybridization (microarray-CGH), and/or multiplex ligation-dependent probe amplification (MLPA). Methods Sequencing and LR analysis were conducted in a consecutive series of 376,159 individuals who received clinical testing with a hereditary pan-cancer gene panel from September 2013 through May 2017. NGS dosage analysis was used to evaluate potential deletions or duplications, with controls in place to exclude pseudogene reads. Samples positive for a putative LR based on NGS were confirmed using a comprehensive approach that included targeted microarray-CGH and/or MLPA analysis, with further examination as needed to ascertain the nature of the LR. Results A total of 3461 LRs were identified and classified as a deleterious mutation (DM), suspected deleterious mutation (SDM) or variant of uncertain significance. Pathogenic LRs (DM/SDM) accounted for the majority of LRs (67.7%), the largest proportion of which were deletions (86.1%), followed by duplications (11.3%), insertions (1.8%), triplications (0.5%), and inversions (0.3%). Several cases presented illustrate that the laboratory approach employed here can ensure consistent identification and accurate characterization of LRs. In the absence of this comprehensive testing strategy, 9% of LRs identified in this testing population might have been missed, potentially leading to inappropriate medical management in as many as 210 individuals referred for hereditary cancer testing. Conclusions These data show that copy number analysis using NGS coupled with confirmatory testing reliably detects and characterizes LRs. Further, LRs comprise a substantial proportion (7.2%) of pathogenic variants identified by the test. A robust and accurate LR identification strategy is an essential component of a high-quality genetic testing program, enabling clinicians to optimize patient medical management decisions.