Vachon, Celine M.

Abstract Over one million biopsies are performed annually in the U.S. to evaluate suspicious breast lesions. Most biopsies lead to a diagnosis of benign breast disease (BBD) and these women have a 1.5- to fourfold increase in subsequent breast cancer (BC) risk compared with the general population. A polygenic risk score (PRS) including 313 common variants was developed to predict BC risk in the general population and has improved performance when added to existing BC prediction models. However, the value of PRS in risk prediction models among women with BBD has not been evaluated. To evaluate the role of the established PRS, BBD, and BC risk, we pooled data from 15,194 white women (6,706 BC cases and 8,488 controls) participating in five Breast Cancer Association Consortium case-control studies. BC risk associations were evaluated by self-reported BBD history and combinations of PRS expressed in tertiles using odds ratios (OR), considering the lowest category of PRS among women without BBD as the referent. Associations were also assessed in an independent case-cohort study (831 BC cases and 1,845 controls) using hazard ratios (HR). Among the case-control studies, the PRS was strongly associated with BC risk (OR=2.50 per standard deviation increase), and associations were similar in women with (OR=2.60) and without a history of BBD (OR=2.44, p-interaction=0.23). Women with BBD have a higher risk of BC than those without BBD within each level of PRS. Compared to women without BBD and in the lowest PRS category, those with BBD and PRS in the highest tertile had 5-fold increased odds of BC (OR = 5.38; 95% CI 4.73-6.13). Associations were similar in an independent case-cohort study. These findings indicate that BBD history and PRS are independent contributors to BC risk, and consideration of both simultaneously could improve breast cancer risk prediction. Future studies are needed to integrate PRS with other risk measures to refine absolute BC risk prediction among BBD patients. Associations of BC risk with combinations of PRS (tertiles) and history of BBD in five case-control studies and a case-cohort study from the Breast Cancer Association Consortium. Case-control studies Case-cohort study PRS score range History of BBD n BC cases n controls OR (95% CI)1 n BC cases n controls HR (95% CI)2 Tertile 1 (≤ -0.57) No 1009 2692 1.00 (Ref) 117 584 1.00 (Ref) Yes 508 902 1.54 (1.35-1.76) 69 163 2.05 (1.45-2.89) Tertile 2 (-0.58, -0.03) No 1469 2001 1.97 (1.78-2.17) 181 475 1.93 (1.48-2.51) Yes 760 798 2.65 (2.34-3.01) 108 170 2.99 (2.18-4.09) Tertile 3 (≥ -0.02) No 1910 1551 3.32 (3.01-3.67) 214 336 3.28 (2.52-4.26) Yes 1050 544 5.38 (4.73-6.13) 142 117 4.83 4.25-8.02) 1.Logistic regression analysis, adjusted for study site/genomic platform and age at enrollment. OR, odds ratio; CI, confidence interval.2.Cox proportional hazards regression analysis, accounting for the case-cohort study design and adjusted for age at enrollment. HR, hazard ratio; CI, confidence interval. Citation Format: Stacey J. Winham, Mark Sherman, Robert Vierkant, Bryan McCauley, Christopher Scott, Mia Gaudet, Melissa Troester, Sandhya Pruthi, Derek Radisky, Amy Degnim, Kristan Aronson, Rachel Murphy, Pascal Guenel, Therese Truong, Jenny Chang-Claude, Heiko Becher, Montserrat Garcia-Closas, Stephen Chanock, Thomas Ahearn, Xiaohong Yang, Doug Easton, Manjeet Bolla, Celine Vachon, Breast Cancer Association Consortium. Association of polygenic risk scores with breast cancer risk among women with benign breast disease. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 5241.

7577 Background: Prior research has consistently shown that patients with chronic lymphocytic leukemia (CLL) have an ~2-fold increased risk of developing melanoma compared to controls. MBL is a common condition affecting 8-10 million adults over the age of 40 years in the United States and is a precursor to CLL. Little is known about the association between MBL and melanoma. Herein, we evaluated the risk of incident melanoma in a large cohort of individuals screened for MBL. Methods: Participants from the Mayo Clinic Biobank who had no prior history of hematologic malignancy, were 40 years or older, were residents of counties surrounding Mayo Clinic, and who provided blood samples between 7/2009 and 4/2022 were included in the study. Peripheral blood mononuclear cells were screened for MBL using highly sensitive eight-color flow cytometry. Individuals with MBL were classified as CLL-like MBL (CD5+, CD20dim) or atypical MBL (CD5+, CD20+) based on immunophenotypes, and also classified as low-count MBL (LCMBL) or high-count MBL (HCMBL) based on whether the percent clonal B-cell count was < 85% or ≥85%, respectively. Diagnosis of melanoma was ascertained using ICD codes and review of the electronic health records. All individuals were followed from date of sample upon which MBL screening was performed to the earliest of incident melanoma, loss to follow-up, withdrawal from the study, or September 30, 2022. We used Cox regression to estimate hazard ratios (HR) and 95% confidence intervals (CI), adjusting for age, sex, and history of melanoma in previous 5 years. Results: A total of 7,334 participants were screened for MBL, of which 1,151 (16%) individuals had MBL. Among those with MBL, 1,107 (96.2%) had CLL-like MBL and 1098 (95.4%) had LC MBL. The median age at entry into the MBL screening cohort was 67 years (range 40-101) and 37% were male; individuals with MBL were older (median age 72 vs 65 years) and male (49% vs 35%). After a median follow-up of 2.9 years (range 0-13.2), 121 of the 7,334 (1.6%) participants developed incident melanoma after the date of MBL screening. The 10-year cumulative incidence of melanoma among individuals with MBL was 7.8% (95% CI 4.5-13.5%) compared to 3.4% (2.7-4.4%) in those who screened negative for MBL. After adjusting for age, sex, and history of prior melanoma, we found a 1.81-fold increased risk of incident melanoma (95% CI 1.19-2.76) among individuals with MBL. The increased risk of melanoma persisted when the analysis was restricted to individuals with LCMBL (HR = 1.84; 95% CI 1.21-2.82). Conclusions: In the largest MBL screening cohort to date, we found that LCMBL, a common condition, was associated with an 84% elevated risk of melanoma, similar to the ~2-fold increased risk observed in patients with CLL. These data, together with our prior findings of a heightened risk of serious infections and lymphoid malignancies, suggest that individuals with LCMBL are at increased risk for adverse health outcomes.

Abstract Pathogenic protein-truncating variants of RAD51C, which plays an integral role in promoting DNA damage repair, increase the risk of breast and ovarian cancer. A large number of RAD51C missense variants of uncertain significance (VUS) have been identified, but the effects of the majority of these variants on RAD51C function and cancer predisposition have not been established. Here, analysis of 173 missense variants by a homology-directed repair (HDR) assay in reconstituted RAD51C−/− cells identified 30 nonfunctional (deleterious) variants, including 18 in a hotspot within the ATP-binding region. The deleterious variants conferred sensitivity to cisplatin and olaparib and disrupted formation of RAD51C/XRCC3 and RAD51B/RAD51C/RAD51D/XRCC2 complexes. Computational analysis indicated the deleterious variant effects were consistent with structural effects on ATP-binding to RAD51C. A subset of the variants displayed similar effects on RAD51C activity in reconstituted human RAD51C-depleted cancer cells. Case–control association studies of deleterious variants in women with breast and ovarian cancer and noncancer controls showed associations with moderate breast cancer risk [OR, 3.92; 95% confidence interval (95% CI), 2.18–7.59] and high ovarian cancer risk (OR, 14.8; 95% CI, 7.71–30.36), similar to protein-truncating variants. This functional data supports the clinical classification of inactivating RAD51C missense variants as pathogenic or likely pathogenic, which may improve the clinical management of variant carriers. Significance: Functional analysis of the impact of a large number of missense variants on RAD51C function provides insight into RAD51C activity and information for classification of the cancer relevance of RAD51C variants.

Multiple myeloma (MM) arises following malignant proliferation of plasma cells in the bone marrow, that secrete high amounts of specific monoclonal immunoglobulins or light chains, resulting in the massive production of unfolded or misfolded proteins. Autophagy can have a dual role in tumorigenesis, by eliminating these abnormal proteins to avoid cancer development, but also ensuring MM cell survival and promoting resistance to treatments. To date no studies have determined the impact of genetic variation in autophagy-related genes on MM risk. We performed meta-analysis of germline genetic data on 234 autophagy-related genes from three independent study populations including 13,387 subjects of European ancestry (6863 MM patients and 6524 controls) and examined correlations of statistically significant single nucleotide polymorphisms (SNPs; p < 1 × 10−9) with immune responses in whole blood, peripheral blood mononuclear cells (PBMCs), and monocyte-derived macrophages (MDM) from a large population of healthy donors from the Human Functional Genomic Project (HFGP). We identified SNPs in six loci, CD46, IKBKE, PARK2, ULK4, ATG5, and CDKN2A associated with MM risk (p = 4.47 × 10−4−5.79 × 10−14). Mechanistically, we found that the ULK4rs6599175 SNP correlated with circulating concentrations of vitamin D3 (p = 4.0 × 10−4), whereas the IKBKErs17433804 SNP correlated with the number of transitional CD24+CD38+ B cells (p = 4.8 × 10−4) and circulating serum concentrations of Monocyte Chemoattractant Protein (MCP)-2 (p = 3.6 × 10−4). We also found that the CD46rs1142469 SNP correlated with numbers of CD19+ B cells, CD19+CD3− B cells, CD5+IgD− cells, IgM− cells, IgD−IgM− cells, and CD4−CD8− PBMCs (p = 4.9 × 10−4−8.6 × 10−4) and circulating concentrations of interleukin (IL)-20 (p = 0.00082). Finally, we observed that the CDKN2Ars2811710 SNP correlated with levels of CD4+EMCD45RO+CD27− cells (p = 9.3 × 10−4). These results suggest that genetic variants within these six loci influence MM risk through the modulation of specific subsets of immune cells, as well as vitamin D3−, MCP-2−, and IL20-dependent pathways.

PURPOSE Artificial intelligence (AI) algorithms improve breast cancer detection on mammography, but their contribution to long-term risk prediction for advanced and interval cancers is unknown. METHODS We identified 2,412 women with invasive breast cancer and 4,995 controls matched on age, race, and date of mammogram, from two US mammography cohorts, who had two-dimensional full-field digital mammograms performed 2-5.5 years before cancer diagnosis. We assessed Breast Imaging Reporting and Data System density, an AI malignancy score (1-10), and volumetric density measures. We used conditional logistic regression to estimate odds ratios (ORs), 95% CIs, adjusted for age and BMI, and C-statistics (AUC) to describe the association of AI score with invasive cancer and its contribution to models with breast density measures. Likelihood ratio tests (LRTs) and bootstrapping methods were used to compare model performance. RESULTS On mammograms between 2-5.5 years prior to cancer, a one unit increase in AI score was associated with 20% greater odds of invasive breast cancer (OR, 1.20; 95% CI, 1.17 to 1.22; AUC, 0.63; 95% CI, 0.62 to 0.64) and was similarly predictive of interval (OR, 1.20; 95% CI, 1.13 to 1.27; AUC, 0.63) and advanced cancers (OR, 1.23; 95% CI, 1.16 to 1.31; AUC, 0.64) and in dense (OR, 1.18; 95% CI, 1.15 to 1.22; AUC, 0.66) breasts. AI score improved prediction of all cancer types in models with density measures ( P LRT values < .001); discrimination improved for advanced cancer (ie, AUC for dense volume increased from 0.624 to 0.679, Δ AUC 0.065, P = .01) but did not reach statistical significance for interval cancer. CONCLUSION AI imaging algorithms coupled with breast density independently contribute to long-term risk prediction of invasive breast cancers, in particular, advanced cancer.

Abstract Introduction: Benign breast disease (BBD) is an important breast cancer (BC) risk factor, which may be classified as non-proliferative disease (NPD), proliferative disease without atypia (PDWA), or atypical hyperplasia (AH) for risk stratification. Data related to the frequency of specific types of BBD and their relationship to BC risk in the Hispanic American population are limited. To address this knowledge gap, we compared BBD and associated BC risk among Hispanic white (HW) and non-Hispanic white (NHW) in New Mexico (NM). Methodology: A retrospective IRB-approved study was performed of women 19 years or older residing in six counties in NM (Bernalillo, Sandoval, Sante Fe, Socorro, Torrance, Cibola/Valencia) between 1996 and mid-2007 to compare the frequency of BBD subtypes and BC risk among HW and NHW. We excluded women who had a history of BC prior to BBD diagnosis or who were diagnosed with BC within 6 months after BBD biopsy. Race and ethnicity were self-reported by women at the time of biopsy. BBD was categorized as NPD, PDWA, or AH based on medical records. Incident BC (in-situ or invasive) was ascertained via linkage to the NM Surveillance Epidemiology End Results (SEER) Registry and BC risk was assessed using standardized incidence ratios (SIRs), comparing the observed number of BC events to that expected based on the NM SEER six-county race- and ethnicity-specific incidence rates, accounting for age and calendar period. Results: Our analysis included 3,870 HW and 6,996 NHW women with BBD. The HW were younger (47.1 vs. 51 years) compared to NHW. HW women had slightly more NPD (69.4% vs. 66.6%) but less PDWA (26.2% vs. 29.4%) and similar frequency of AH (4.3% vs. 3.9%) as compared to NHW. Over a median post-BBD follow-up period of 13 years (range 6 months-17 years), 644 BCs were observed (4.81% in HW and 6.55% in NHW). The observed BC risk among women with BBD was higher than population-based expected rates (SIR 1.98, 95% CI 1.82-2.13, p&lt;0.001) and showed expected increases in risk with increasing degrees of BBD abnormality: SIR=1.90 for NPD, 2.00 for PDWA, and 3.01 for AH. Comparing BC risk by ethnic subgroups, HW women had an overall risk of BC after BBD that was statistically indistinguishable from NHW women (SIR=2.17, 95% CI 1.86-2.48 in HW women, and SIR=1.91, 95% CI 1.73-2.08 in NHW women). Within the major subgroups of the BBD findings, there were no significant differences in risk of BC after BBD for HW versus NHW women. Conclusions: In this population-based study, benign breast disease subtypes and their associated breast cancer risk were similar among the Hispanic and non-Hispanic white women. Citation Format: Kush Raj Lohani, Andrea M. Nibbe, Robert A. Vierkant, Laura M. Pacheco-Spann, Lisa R. Seymour, Celine M. Vachon, Mark E. Sherman, Amy C. Degnim, Deirdre Hill. Comparison of benign breast disease subtypes and breast cancer risk among Hispanic and non-Hispanic white women in New Mexico. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 4213.

Abstract Low count (LC) MBL is characterized by a circulating population of clonal B-cells and is a precursor state to CLL. Most individuals with LC MBL do not progress to CLL and never come to clinical attention. An active area of research is to identify factors that distinguish those individuals with MBL who progress from those who do not. Currently, 41 single nucleotide polymorphisms (SNPs) are known to be associated with CLL risk, and a portion of these SNPs have also been found to be associated with MBL risk. We hypothesize that these CLL-susceptibility SNPs associated with both MBL and CLL risk are SNPs associated with developing a B-cell clone; whereas the remaining SNPs found not to be associated with MBL risk are associated with progression of the B-cell clone (i.e., progression from MBL to CLL). To test this hypothesis, we evaluated the association of these 41 SNPs with risk of progression to CLL using a case-control study of 1,911 CLL cases and 1,579 individuals with MBL from Mayo Clinic. Individuals with MBL were identified through screening using 8-color flow cytometry assay. Individuals with LC MBL (defined by having a clonal B-cell population with the percent clonal B-cells out of total B-cell count &lt;85%) who had an immunophenotype consistent with CLL were included as controls. CLL cases were obtained through the Mayo Clinic hematology practice. DNA was genotyped from whole blood. We analyzed the 41 CLL susceptibility SNPs with risk of progression to CLL using logistic regression and estimated odds ratio (OR) and 95% confidence interval (CI), adjusting for sex and age. We evaluated the CLL polygenic risk score (CLL-PRS), comprised of a weighted average of the 41 SNPs, with risk of CLL. Out of the 41 established CLL susceptibility SNPs, 15 were associated with CLL progression from MBL to CLL (all P&lt;0.05). Nine of these 15 significant SNPs were not previously found to be associated with risk of MBL. The most significant SNPs were rs9880772 (OR=1.23, CI: 1.11-1.36, P=6.1 × 10−5), rs888096 (OR=1.25, CI: 1.12-1.38, P=3.0 × 10−5), and rs305065 (OR=1.28, CI: 1.14-1.43, P=1.5 × 10−5). One SNP (rs77551289) had a suggestive association (OR=1.21, P=0.069). The remaining 25 SNPs had no evidence of association with progression (ORs ranging between 0.9-1.1 and all P&gt;0.1). The CLL-PRS was also associated with CLL risk (continuous OR = 1.25, CI: 1.15 - 1.35, P = 6.7 × 10−8). In our study accounting for the presence of B-cell clonal population, we identified 15 out of 41 CLL-susceptibility SNPs to be significantly associated with progression from LC MBL to CLL. The remaining 25 non-significant CLL-SNPs were not associated with progression and may instead be associated with development or initiation of the B-cell clone. These findings suggest that the established CLL susceptibility SNPs may play differing role (i.e., development or progression of the clone) in the etiology of CLL. Citation Format: Raphael Mwangi, Geffen Kleinstern, Sara J. Achenbach, Dennis Robinson, Aaron D. Norman, Kari G. Rabe, Janet E. Olson, Neil E. Kay, Rosalie G. Waller, Nicholas J. Boddicker, James R. Cerhan, Esteban Braggio, Sameer A. Parikh, Curtis A. Hanson, Celine M. Vachon, Tait Shanafelt, Susan L. Slager. Genetic variants associated with progression from monoclonal B-cell lymphocytosis (MBL) to chronic lymphocytic leukemia (CLL). [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 5222.

PURPOSE Image-derived artificial intelligence–based short-term risk models for breast cancer have shown high discriminatory performance compared with traditional lifestyle/familial-based risk models. The long-term performance of image-derived risk models has not been investigated. METHODS We performed a case-cohort study of 8,604 randomly selected women within a mammography screening cohort initiated in 2010 in Sweden for women age 40-74 years. Mammograms, age, lifestyle, and familial risk factors were collected at study entry. In all, 2,028 incident breast cancers were identified through register matching in May 2022 (206 incident breast cancers were found in the subcohort). The image-based model extracted mammographic features (density, microcalcifications, masses, and left-right breast asymmetries of these features) and age from study entry mammograms. The Tyrer-Cuzick v8 risk model incorporates self-reported lifestyle and familial risk factors and mammographic density to estimate risk. Absolute risks were estimated, and age-adjusted AUC model performances (aAUCs) were compared across the 10-year period. RESULTS The aAUCs of the image-based risk model ranged from 0.74 (95% CI, 0.70 to 0.78) to 0.65 (95% CI, 0.63 to 0.66) for breast cancers developed 1-10 years after study entry; the corresponding Tyrer-Cuzick aAUCs were 0.62 (95% CI, 0.56 to 0.67) to 0.60 (95% CI, 0.58 to 0.61). For symptomatic cancers, the aAUCs for the image-based model were ≥0.75 during the first 3 years. Women with high and low mammographic density showed similar aAUCs. Throughout the 10-year follow-up, 20% of all women with breast cancers were deemed high-risk at study entry by the image-based risk model compared with 7.1% using the lifestyle familial-based model ( P < .01). CONCLUSION The image-based risk model outperformed the Tyrer-Cuzick v8 model for both short-term and long-term risk assessment and could be used to identify women who may benefit from supplemental screening and risk reduction strategies.

Abstract Mammographic breast density (MBD) has been shown to be a strong, independent risk factor for breast cancer (BC) irrespective of race/ethnicity. Given the risk association of MBD and its potential to mask tumors on a mammogram, state and federal laws have mandated that women receive information regarding their personal MBD in their mammography reports. However, concerns have been raised regarding the impact of MBD notification on patient anxiety, especially written information for women who experience health disparities such as racial/ethnic minorities, lower health literacy, limited English proficiency and lower socioeconomic status. We performed a randomized controlled clinical trial to examine the impact of three different written and interpersonal approaches to MBD notification on patient anxiety, BC worry, and self-perceived BC risk, among Latinas receiving routine mammography screening at a federally qualified medical center (FQHC). We hypothesized that interpersonal education would reduce anxiety and worry, relative to the written notifications alone. The study was performed at the Baseline Clinic of Mountain Park Health Center, a FQHC in Phoenix, AZ. Women between ages 40 and 74 years presenting for screening mammogram were eligible. After providing signed informed consent, participants were randomized equally to usual care (UC- mailed notification letter); enhanced care (notification letter and MBD educational brochure designed for this study): interpersonal care (notification letter, brochure, promotora education via telephone). A stratified block randomization procedure was used with age &gt; 50 years (yes vs no), ethnicity (Hispanic vs non-Hispanic), and language preference (Spanish vs. English) as strata. Study participants completed surveys at the time of enrollment/pre-intervention (T0), at two weeks to six months after intervention was delivered (T1), and about one year after randomization (T2). Anxiety state was measured using the state anxiety subscale of the State-Trait Anxiety Inventory scale (STAI-S) (range from 20-80). BC worry was probed using the question: “How frequently do you worry about getting breast cancer someday”. The self-perceived lifetime risk of BC was rated between 0% (no chance of BC) to 100% (definitely will get BC). The proportion of participants with an increase or persistence of higher level for each outcome between time points was compared between the three study groups. The study was approved by the Mayo Clinic Institutional Review Board. 1332 Latina women were enrolled and randomized between October 2016 and October 2019. All completed the baseline (T0) survey, 928 completed T1, 632 completed T2, and 516 completed both T1 and T2 surveys. At study entry, majority of the participants were young (64.1% between age 40-49 years), had no family history of breast cancer (81.0%), had less than high school education (68.8%), and were married or partnered (67.8%). At T0, the mean (SD) anxiety STAI score was 32.5 (11.1); 51.8% with moderate or severe anxiety (score &gt; 30). With regard to BC worry, 41.3% reported worrying “sometimes”, “often”, or “almost all the time”. Further, 25.4% reported a self-perceived lifetime risk of developing BC of &gt;10% while only 6.6% had a Gail model estimated lifetime risk score of &gt;10%. There was no significant difference in changes in anxiety, BC worry or self-perceived risk from T0 to either T1 or T2 surveys between the intervention groups. In summary, this study found high levels of baseline anxiety and BC worry (despite the majority being 40-49 years old and having no family history of BC) which persisted regardless of how notification and education were delivered. Future work is needed to improve the understanding of factors that could lower anxiety and BC worry and improve risk perception in this population. Citation Format: Karthik Ghosh, Sarah Jenkins, Jennifer Ridgeway, Jessica D. Austin, Bijan Borah, Bhavika K. Patel, Deborah Rhodes, Aaron Norman, Edna P. Ramos, Matt Jewett, Crystal Gonzalez, Valentina Hernandez, Davinder Singh, Vera Suman, Celine Vachon. Breast Density Notification: Impact on anxiety, breast cancer worry, and self-perceived risk among Latinas at a federally qualified medical center. [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 P6-05-38.

Abstract Breast parenchymal patterns on radiologic images are associated with breast cancer risk. Radiomic features have been proposed as quantitative measures of parenchymal patterns. We defined intrinsic imaging phenotypes of breast parenchymal patterns based on radiomic features extracted from full field digital mammography (FFDM) in breast screening populations and assessed whether these phenotypes are associated with breast cancer risk and masking. We selected 30,000 women with 4-view FFDM exams from Hologic machines from three institutions (Hospital of the University of Pennsylvania, Mayo Clinic, and San Francisco Mammography Registry), randomly split into a training (20,000 women) and test set (10,000 women). In total, 390 radiomic features were automatically extracted from each image using a validated software pipeline, standardized, and adjusted for site differences using ComBat. We used two methods, hierarchical clustering and Principal Components (PCs) analysis, to classify significant variation among the features in the training set and replicate among the test set. Next, we applied the replicated clusters and PCs to an independent nested case-control set [1082 invasive breast cancer (BC) cases (of which 151 were Black and 893 White women, 38 other race) matched to 2837 controls (411 Black and 2345 White women, 81 other race) on age, race, timing of images, and site]. We examined associations of the clusters and PCs with invasive breast cancer risk, as well as masking [defined as a false-negative (FN) screen (124 cases and 319 matched controls) and additionally the subset with symptomatic interval cancer (IC) within 12 months of negative screen (88 cases and 223 matched controls)] using conditional logistic regression. We evaluated their association with breast cancer alone, and with adjustment for age, body mass index (BMI) and breast density assessed by Breast Imaging Reporting and Data System (BI-RADS) using likelihood ratio tests. We estimated discrimination using area under the curve (AUC) and compared AUCs for models that included the radiomic clusters and PCs with the model that included only age, BMI and density. We also stratified analyses by race (Black/White). From hierarchical clustering, we defined six statistically significant phenotype clusters (each of at least 1000 women) in the training set which were replicated in the test set. For PC Analysis, we identified six PCs in the training set, explaining 85% of the variation in texture features and reproduced these in the test set. The six radiomic phenotype clusters (P&lt; 0.001) and six PCs (P&lt; 0.001) were both associated with invasive BC, including after adjusting for age, BMI, and density (cluster P=0.004; PCs P&lt; 0.001). Improvements in discrimination of invasive BC with inclusion of PCs or clusters were more pronounced among Black women (Table). Further, the PCs (P&lt; 0.001) and clusters (P&lt; 0.001) were significantly associated with FN overall and for symptomatic IC (PCs P&lt; 0.001; clusters P=0.001), but only PCs remained significant after adjusting for age, BMI and density (PCs P=0.004 for FN; PCs P=0.007 for symptomatic IC). Discrimination of masking also improved with inclusion of both clusters and PCs (Table). We identified reproducible radiomic phenotypes that are associated with invasive BC risk, above and beyond breast density with the strongest associations for invasive BC among Black women and symptomatic interval cancers. Citation Format: Stacey Winham, Anne Marie McCarthy, Aimilia Gastounioti, Christopher Scott, Aaron Norman, Walter C. Mankowski, Lauren Pantalone, Matthew Jensen, Eric A. Cohen, Hannah Horng, Kathleen Brandt, Emily F. Conant, Karla Kerlikowske, Despina Kontos, Celine Vachon. Radiomic phenotypes of breast texture and association with breast cancer risk and masking [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 GS4-06.