Abstract P3-02-02: Real-World Time-to-Treatment Discontinuation in Hormone-Receptor-Positive Metastatic Breast Cancer Patients following CDK4/6 Inhibitor Treatment, Based on Observational Data Collected Through Patient-Partnered Research

Background: Cancer treatment decisions are often made without specific and representative information that can inform personalized treatment. The aim of this study was to determine if we can predict, based on clinical features, which treatment regimen may maximize real-world time-to-treatment discontinuation (rwTTD) after a hormone-receptor-positive (HR+) metastatic breast cancer (MBC) patient stops responding to a first CDK4/6 inhibitor in any line.

Methods: We used patient reported data (PRD) about diagnosis and treatment and medical records from 1,777 patients across the U.S. and Canada from Count Me In’s Metastatic Breast Cancer Project (MBCproject). We interviewed 17 people, academic and community based medical oncologists and MBC patients, to inform the analysis plan. Patient eligibility criteria were prior HR+ MBC diagnosis, received exactly one prior CDK4/6 inhibitor (CDK4/6) containing regimen, start date of any subsequent regimen within four months of the end date of the CDK4/6-containing regimen, and completion of MBCproject’s follow-up questionnaire at least one month after the start date of the subsequent regimen. We processed RWD from the follow-up questionnaire, performed chart review in ambiguous cases of patient eligibility, performed conformance, completeness, and plausibility verification checks to determine the dataset’s fit-for-use, and described treatment variation seen in real-world settings. We designed a new user, active-comparator cohort study with rwTTD as the continuous outcome measure, used known and hypothesized confounders to control for treatment-by-indication bias, assessed covariate balance across cohorts, and conducted Cox proportional hazards (PH) outcome regressions to identify clinically relevant associations and estimate treatment effects across regimens. The analysis plan was publicly registered with the Center for Open Science prior to performing the analysis.

Results: 261 eligible HR+ MBC patients were identified, with 110 unique pairs of CDK4/6-containing and subsequent regimens. The most common CDK4/6-containing regimen was Letrozole and Palbociclib (n=98) and subsequent regimen was Capecitabine (n=63). Three mutually exclusive and clinically relevant groupings of subsequent regimens chosen for analysis were chemotherapy only (n=99), fulvestrant-containing (n=53), and everolimus-containing (n=42). Among patients in these three groups, 93.9%+ are white race, 95%+ are non-hispanic, 2.7-9.4% live in a medically underserved area, 7.1-13.1% have HR+/HER2+ MBC, mean age at subsequent treatment was 52.6-53.8 years, 17-36% had bone-only metastasis and 14.3-25.3% had liver metastasis at MBC diagnosis, median number of past treatment regimens was one, and median time on CDK4/6-containing regimen was 9-14 months. The median rwTTD was 9, 15, and 5 months in the three groups, respectively. Out of 11 covariates, nine covariates failed to reject the null hypothesis that the distribution of values are the same across the three cohorts (p>0.05). Outcome regression Cox PH revealed rwTTD hazard ratio (HR) of 2.52 [1.53-4.15; 95% confidence interval (CI)] for presence of liver metastasis, HR of 1.09 [0.63-1.89; 95% CI] for presence of bone-only metastasis, HR of 2.00 [1.20-3.33; 95% CI] for everolimus-containing regimen vs. chemotherapy only, HR of 0.85 [0.50-1.46; 95% CI] for fulvestrant-containing regimen vs. chemotherapy only, and HR of 0.82 [0.65-1.00; 95% CI] for every six months rwTTD on previous CDK4/6-containing regimen.

Conclusion: In this cohort, chemotherapy was the most common treatment regimen following CDK4/6 even in second and third line settings and in patients with bone-only metastasis, which is a deviation from guideline-based treatment for many HR+ MBC patients. PRD helps develop hypotheses about patient response to treatment following CDK4/6 that can be further evaluated in larger, more diverse observational studies and clinical trials.

Table 1. Characteristics of eligible patients who received chemotherapy only, everolimus containing, or fulvestrant containing regimens.

Citation Format: Ariel B. Carmeli, Seth A. Wander, Mary McGillicuddy, Caroline Block, Nikhil Wagle. Real-World Time-to-Treatment Discontinuation in Hormone-Receptor-Positive Metastatic Breast Cancer Patients following CDK4/6 Inhibitor Treatment, Based on Observational Data Collected Through Patient-Partnered Research [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 P3-02-02.

Abstract OT1-19-01: The metastatic breast cancer project: Generating the clinical and genomic landscape of metastatic breast cancer through patient-partnered research

The Metastatic Breast Cancer Project (MBCproject) is an ongoing research study that directly engages patients (pts) through social media and advocacy groups, and empowers them to share their samples, clinical information, and experiences. The goal is to create a publicly available dataset of linked genomic, clinical, and pt-reported data to enable research. In collaboration with pts, advocates, and advocacy groups, a website (MBCproject.org) was developed that allows pts with metastatic breast cancer (MBC) anywhere in the US or Canada to register. From 10/20/15-6/1/21, 6100 patients with MBC registered for the MBCproject. Registered pts are sent an online consent form that asks for permission to obtain and analyze their medical records and samples. Consented pts are sent a saliva and/or blood kit and asked to mail back a saliva sample, which is used to extract germline DNA, and/or a blood sample, which is used to extract germline DNA and cell free DNA (cfDNA). We contact participants’ medical providers to obtain medical records and a portion of their stored tumor biopsies. 3456 pts receiving care at over 1700 different institutions have consented to share medical records and tumor/saliva/blood samples and to have genomic analysis performed. Whole exome sequencing (WES) is performed on tumor DNA, germline DNA, and cfDNA; transcriptome sequencing (RNA-seq) is performed on tumor RNA. Medical records and pt-reported data are abstracted to create a detailed clinical record for each pt. Table 1 highlights clinical data collection, biospecimen acquisition, and genomic data generation to date. Examples of clinicogenomic analyses are shown in Table 2. De-identified linked genomic, clinical, and pt-reported data is shared regularly via public databases (mbcproject.org, cBioPortal, dbGaP, NCI Genomic Data Commons). To date, this data has been cited in over 40 publications. Study updates are shared with participants regularly. The MBCproject continues to enroll new patients, generate additional data, and perform integrated clinical and genomic analyses with the goal of building a dataset that is representative of patients with MBC. We have partnered with over 30 non-profit breast cancer advocacy groups. We also have several community engagement efforts underway to more directly reach patients in underrepresented communities, including partnerships with faith-based organizations and colleges/universities, as well as targeted engagement with the African American community. In addition, in partnership with Latinx patients, advocates, and researchers, a Spanish-language version of the MBCproject was launched in June 2021. Partnering directly with pts rapidly enables thousands of pts to remotely share tumors, blood, saliva, and medical records to accelerate research. The resulting publicly shared clinically annotated dataset is a resource that allows researchers to identify patients with specific phenotypes, who have often been challenging to identify with traditional approaches.

Clinical data collection, biospecimen acquisition, and genomic data generation:NumberConsent signed (US & CA)3456 ptsPatient-reported data collected (demographics, diagnosis details, receptor status, clinical experiences, pathology details, sites of metastasis, treatments with start and stop dates3456 ptsMedical record received from clinical institution1365 ptsSaliva sample received from pt2124 ptsBlood sample received from pt1114 ptsTumor samples received from clinical institution631 tumor samples from 398 ptsWES from germline complete505 germline samplesWES from tumor (primary and metastatic) samples complete429 tumor samplesRNA-seq from tumor (primary and metastatic) samples complete351 tumor samplesULP-WGS from cfDNA (taken in metastatic setting) complete953 blood samplesWES from circulating tumor DNA (taken in metastatic setting) complete144 blood samples

CohortConsented (US & CA)Tumor WES completeTumor RNA-seq completePts diagnosed < 40 yrs of age114615292De novo MBC1207158109Late recurrence (>5 years after dx)9099141Long term survivors (MBC > 10yrs)163138Resistance to CDK4/6 inhibitors70914839NED at time of f/u survey4305445Triple Negative Breast Cancer3304632Patients with 2 or more tumor biopsies/cfDNA samples collected by the MBCproject29810882

Citation Format: Nikhil Wagle, Corrie Painter, Elana Anastasio, Mary McGillicuddy, Esha Jain, Tania G. Hernandez, Brett N. Tomson, Beena Thomas, Daniel Abravanel, Dewey Kim, Sara Balch, Alyssa L. Damon, Shahrayz Shah, Rafael Ramos, Delia Sosa, Ilan Small, Colleen Nguyen, Sarah Winnicki, Taylor Cusher, Parker Chastain, Michael Dunphy, Jorge Gomez Tejeda Zanudo, Netsanet Tsegai, Lauren Sterlin, Ulcha F. Ulysse, Imani Boykin, Oyin Alao, Todd R. Golub. The metastatic breast cancer project: Generating the clinical and genomic landscape of metastatic breast cancer through patient-partnered research [abstract]. In: Proceedings of the 2021 San Antonio Breast Cancer Symposium; 2021 Dec 7-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2022;82(4 Suppl):Abstract nr OT1-19-01.

Abstract PD6-02: The metastatic breast cancer project - Expanding the clinical, genomic, and transcriptomic landscape of metastatic breast cancer through patient-partnered research

Background: The Metastatic Breast Cancer Project (MBCproject) is an ongoing research study that directly engages patients (pts) through social media and advocacy groups, and empowers them to share their biospecimens, clinical history, and experiences. The goal is to create a publicly available dataset of linked clinicogenomic and pt-reported data to enable research. From 2015-10-20 to 2020-3-31, 3,245 MBC pts who received treatment at >1,700 institutions, consented to share medical records, pt-reported data, and tumor/saliva/blood samples and to have genomic analysis performed. Here, we describe an analysis of the clinical and genomic features in this initial MBCproject cohort. Methods: We performed whole exome sequencing (WES) on 379 tumors (with matched germline) from 301 pts, 377 germline samples and RNA sequencing (RNA-seq) from 200 tumors from 141 pts. In 14 pts, we characterized 2 or more serial tumor biopsies. WES data was analyzed for mutations and copy number variants. RNA-seq data was used to call fusions, research-grade PAM50, and gene set enrichment scores. Medical records and pt-reported data were abstracted to create detailed clinical record for each pt. Results: WES of 249 metastatic tumors identified 34 cancer genes (e.g., TP53, PIK3CA, CDH1, PTEN, AKT1, NF1, ESR1) that were significantly recurrently altered. Potential clinically actionable alterations were identified in 39% of metastatic tumors. 45 tumors (22.5%) had fusions with known functional effects and 24 (12%) had in-frame fusions in key cancer genes like FANCD2 (3), FGFR3 (2), ESR1 (1), BRAF (1), and NCOR1 (1). PAM50 classification suggested depletion of Luminal A subtype in this cohort compared to TCGA breast cancer cohort (p-value <0.05). Of 29 pts with paired RNA-seq biopsies, 10 pts showed a PAM50 subtype switch [LumB to Basal = 1, Her2 to Basal =1, Basal to LumA = 1, Her2 to LumA = 2, LumA/B to Her2 = 3, LumA to LumB = 2]. Germline analysis showed that 30.2% (114/377) of pts had at least one pathogenic variant in a cancer predisposition gene, including BRCA1/BRCA2 (5%), NF1 (5.8%), ATM (1.06%) and PALB2 (1.59%). In 10.9% pts (33/301), pathogenic variants in a cancer predisposition gene was accompanied by an apparent somatic loss of function event in the same gene. There was ~ 90% concordance between abstracted features (e.g., receptor status, histology, diagnosis dates, metastatic sites) and pt-reported data, enabling the use of both data types for integrated analyses. The utility of integrated analyses is illustrated by a case study of a pt who had an initial diagnosis of ductal carcinoma in situ and developed metastatic disease 5.5 yrs later. Analysis of 3 metastatic tumors (WES and RNA-seq) and a circulating tumor DNA sample (WES only) from this pt collected over the course of therapy revealed evidence of tumor evolution and multiple mechanisms of resistance to endocrine therapies and CDK4/6 inhibitors, including 2 distinct acquired ESR1 mutations, an activating BRAF fusion, and estrogen receptor (ER) loss by immunohistochemistry with concomitant development of ESR1-PLEKHG1 in-frame fusion. RNA-seq data also showed a PAM50 luminal phenotype and a persistent ER. signature throughout despite the apparent ER loss, suggesting compensation by the ESR1 fusion. Analyses of 13 additional pts with serial biopsies will be presented. Integrated analyses of several additional cohorts of interest, such as de novo MBC (91 pts), resistance to CDK4/6 inhibitors (39 pts), and pts diagnosed with breast cancer <40 yrs, will also be presented. Conclusion: This clinicogenomic dataset generated by partnering directly with pts was uniquely built with pt-reported data, medical record data, and multi-omic characterization, and serves as a powerful tool for researchers to harness to accelerate discoveries in MBC.

Citation Format: Esha Jain, Dewey Kim, Jorge Gomez Tejeda Zanudo, Sara Balch, Mary McGillicuddy, Brett N. Tomson, Tania G. Hernandez, Beena S. Thomas, Daniel L. Abravanel, Shahrayz Shah, Rafael Ramos, Delia Sosa, Ilan Small, Lauren Sterlin, Sarah Winnicki, Colleen Nguyen, Micheal Dunphy, Elana Anastasio, Todd R. Golub, Corrie A. Painter, Nikhil Wagle. The metastatic breast cancer project - Expanding the clinical, genomic, and transcriptomic landscape of metastatic breast cancer through patient-partnered research [abstract]. In: Proceedings of the 2021 San Antonio Breast Cancer Symposium; 2021 Dec 7-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2022;82(4 Suppl):Abstract nr PD6-02.

Abstract OT-18-01: The metastatic breast cancer project: Generating the clinical and genomic landscape of metastatic breast cancer through patient-partnered research

The Metastatic Breast Cancer Project (MBCproject) is an ongoing research study that directly engages patients (pts) through social media and advocacy groups, and empowers them to share their samples, clinical information, and experiences. The goal is to create a publicly available dataset of linked genomic, clinical, and pt-reported data to enable research. In collaboration with pts, advocates, and advocacy groups, a website (MBCproject.org) was developed that allows pts with metastatic breast cancer (MBC) anywhere in the US or Canada to register. From 10/20/15-3/31/20, 5708 women and men with MBC registered for the MBCproject. Registered pts are sent an online consent form that asks for permission to obtain and analyze their medical records and samples. Consented pts are sent a saliva and/or blood kit and asked to mail back a saliva sample, which is used to extract germline DNA, and/or a blood sample, which is used to extract germline DNA and cell free DNA (cfDNA). We contact participants’ medical providers to obtain medical records and a portion of their stored tumor biopsies. 3245 pts receiving care at over 1700 different institutions have consented to share medical records and tumor/saliva/blood samples and to have genomic analysis performed. Whole exome sequencing (WES) is performed on tumor DNA, germline DNA, and cfDNA; transcriptome sequencing (RNA-seq) is performed on tumor RNA. Medical records and pt-reported data are abstracted to create a detailed clinical record for each pt. Table 1 highlights clinical data collection, biospecimen acquisition, and genomic data generation to date. Examples of clinicogenomic analyses are shown in Table 2. De-identified linked genomic, clinical, and pt-reported data is shared regularly via public and semi-public databases (mbcproject.org, cBioPortal, dbGaP, NCI Genomic Data Commons). To date, this data has been cited in over 20 published journal articles. Study updates are shared with participants regularly. The MBCproject continues to enroll new patients, generate additional data, and perform integrated clinical and genomic analyses with the goal of building a dataset that is representative of patients with MBC. We have partnered with over 30 non-profit breast cancer advocacy groups. We also have several community engagement efforts underway to more directly reach patients in underrepresented communities, including partnerships with faith-based organizations and colleges/universities, as well as targeted engagement with the African American community. In addition, in partnership with Latinx patients, advocates, and researchers, the project has been translated into Spanish and is expected to launch in late 2020. Partnering directly with pts rapidly enables thousands of pts to remotely share tumors, blood, saliva, and medical records to accelerate research. The resulting publicly shared clinically annotated database is a resource that allows researchers to identify patients with specific phenotypes, who have often been challenging to identify with traditional approaches.

Clinical data collection, biospecimen acquisition, and genomic data generation:NumberConsent signed (US & CA)3245 ptsSurvey #1 submitted(demographics, diagnosis details, receptor status, clinical experiences)3245 ptsSurvey #2 submitted(pathology details, sites of metastasis, treatments with start and stop dates)1638 ptsMedical record received1352 ptsSaliva sample received2004 ptsBlood sample received1121 ptsTumor samples received585 tumor samples from 424 ptsDigital image of tumor slide H&E generated585 tumor samplesWES from germline complete458 germline samplesWES from tumor (primary and metastatic) samples complete343 tumor samplesRNA-seq from tumor (primary and metastatic) samples complete228 tumor samplesULP-WGS from cfDNA (taken in metastatic setting) complete993 blood samplesWES from circulating tumor DNA (taken in metastatic setting) complete143 blood samples

CohortConsented (US & CA)Tumor WES completeTumor RNA-seq completePts diagnosed < 40 yrs of age107312071De novo MBC112712183Late recurrence (>5 years after dx)8307752Long term survivors (MBC > 10yrs)158115Resistance to CDK4/6 inhibitors70914839NED at time of f/u survey4238939Triple Negative Breast Cancer3107531Patients with 2 or more tumor biopsies / cfDNA samples collected by the MBCproject2876138

Citation Format: Nikhil Wagle, Corrie Painter, Elana Anastasio, Michael Dunphy, Mary McGillicuddy, Esha Jain, Brett Tomson, Tania G. Hernandez, Beena Thomas, Dewey Kim, Alyssa L. Damon, Shahrayz Shah, Rafael Ramos, Colleen Nguyen, Lee O'Neil, Sarah Winnicki, Sara Balch, Rachel Stoddard, Taylor Cusher, Parker Chastain, Jorge Gomez Tejeda Zanudo, Jorge Buendia-Buendia, Ofir Cohen, Netsanet Tsegai, Lauren Sterlin, Ulcha F. Ulysse, Imani Boykin, Kate Sine, Oyin Alao, Jacqueline Lucia, Eric S. Lander, Todd R. Golub. The metastatic breast cancer project: Generating the clinical and genomic landscape of metastatic breast cancer through patient-partnered research [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 OT-18-01.

Abstract PO-015: The Metastatic Breast Cancer Project in Spanish: Developing a Spanish language patient-partnered research project with the Latinx community through culturally appropriate materials

Breast cancer is the leading cause of cancer-related mortality for Latinx women in the US. Compounding this challenge is low representation of Latinx patients (pts) in cancer research studies. Neglecting Latinx participation limits scientific understanding and the development of effective therapies for all pts. To address this disparity and increase research accessibility, we collaborated with Latinx community (comm) members and translated all aspects of an existing patient-partnered research project, The Metastatic Breast Cancer Project (MBCproject), into Spanish.

The MBCproject (MBCproject.org) is a research study that directly engages pts through social media and advocacy groups, empowering pts to share their samples, clinical information, and experiences. Despite having 3,300 consented participants, only 3.7% (123) identify as Hispanic, Latinx, or Spanish. With the Latinx population increasing in the US and cancer’s impact on the comm, it is imperative that Latinx individuals participate in cancer research. Native Spanish speaking members on the MBCproject team leveraged the existing MBCproject model of direct collaboration with pt communities to translate pt-facing materials. Pts and pt advocates provided feedback on site content, project name, surveys, and biospecimen kit instructions through video calls with team staff. Over 17 months, 6 focus groups were held with 9 members of the comm, assessing 6 different material items. Iterations of the materials discussed were shared via email to complete the process. Proposed Spanish language dialect, pronoun formality, and transparency in project content were also reviewed to ensure messaging resonated with the broader Latinx comm.

Following comm review, modifications were integrated into materials. A Mexican- Spanish dialect and formal pronouns were chosen to be used across all materials. The project name remained to maintain continuity across languages. Distinct feedback included simplified terminology and materials that could be shared with their health care provider. Clarity in tissue acquisition, privacy of data collected, and participation requirements was also requested. Attendance of study staff at Latinx cancer disparities research conferences revealed interest amongst investigators for more comprehensive demographic data collection of Latinx pts in studies, leading to the development of more detailed race and ethnicity questions in our intake survey. Finalized project materials have been submitted for IRB approval and expected to be launched at mbcprojectenespanol.org in Fall 2020. This partnered approach has led to an inclusive set of materials embraced by the Latinx pts and pt advocates who have helped us develop the MBCproject in Spanish. This is a foundational pilot that will allow Spanish speaking participants to engage in cancer research without a language barrier and contribute to the acceleration of scientific discoveries that impacts their community.

Citation Format: Tania G. Hernandez, Sara Balch, Jorge Buendia- Buendia, Jorge Gomez Tejeda Zanudo, Mary McGillicuddy, Colleen M. Nguyen, Foxy Robinson, Corrie Painter, Elana Anastasio, Michael Dunphy, Nikhil Wagle. The Metastatic Breast Cancer Project in Spanish: Developing a Spanish language patient-partnered research project with the Latinx community through culturally appropriate materials [abstract]. In: Proceedings of the AACR Virtual Conference: Thirteenth AACR Conference on the Science of Cancer Health Disparities in Racial/Ethnic Minorities and the Medically Underserved; 2020 Oct 2-4. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2020;29(12 Suppl):Abstract nr PO-015.

Abstract PD8-01: The metastatic breast cancer project: Generating the clinical and genomic landscape of metastatic breast cancer through patient-partnered research

Background: The Metastatic Breast Cancer Project (MBCproject) is a research study that directly engages patients (pts) through social media and advocacy groups, and empowers them to share their samples, clinical information, and experiences. The goal is to create a publicly available dataset of linked genomic, clinical, and pt-reported data to enable research.

Methods: In collaboration with pts, advocates, and advocacy groups, a website (MBCproject.org) was developed that allows pts with metastatic breast cancer (MBC) anywhere in the U.S. or Canada to register. Registered pts are sent an online consent form that asks for permission to obtain and analyze their medical records and samples. Once enrolled, pts are sent a saliva kit and a blood kit and asked to mail back a saliva sample, which is used to extract germline DNA, and/or a blood sample, which is used to extract germline DNA and cell free DNA (cfDNA). We contact participants’ medical providers and obtain medical records and a portion of their stored tumor biopsies. Whole exome sequencing (WES) is performed on tumor DNA, germline DNA, and cfDNA; transcriptome sequencing (RNA-seq) is performed on tumor RNA. Medical records and pt-reported data are abstracted to create a detailed clinical record for each pt. All de-identified data are shared regularly via public databases (cbioportal.org, mbcproject.org, dbGaP, NCI Genomic Data Commons) without restrictions. Study updates are shared with participants regularly.

Results: From 10/20/15-7/8/19, 5357 women and men with MBC registered. 3290 pts receiving care at over 1700 different institutions consented to share medical records and tumor/saliva/blood samples, and to have genomic analysis performed. Details of clinical data collection, biospecimen acquisition, and genomic data generation to date are outlined in the Table. WES from 463 tumors obtained from 326 pts have been generated (with matched germline WES), including 61 pts with 2 timepoints, 19 pts w 3 timepoints, and 11 pts w 4+ timepoints. 278 tumor exomes were from the breast/regional lymph nodes, 63 from distant metastatic sites and 122 from cfDNA. 110 tumor exomes were from samples obtained before the diagnosis of MBC, 258 from after the diagnosis of MBC, and 95 to be determined (TBD). 161 tumor exomes were obtained prior to any therapy, 204 following some therapy, and 98 TBD. Clinically annotated genomic data are used to study specific pt cohorts (including rare subsets and outliers) and to identify mechanisms of response and resistance to therapies. Examples of the clinical and genomic analyses that will be presented include: - Pts diagnosed <40 yrs of age (1108 pts enrolled; 120 with tumor WES) - de novo MBC (1122 pts enrolled; 121 with tumor WES) - Late recurrence, >5 yrs after diagnosis (830 pts enrolled; 77 with tumor WES) - Long-term survivors, >10 yrs with MBC (159 pts enrolled; 11 with tumor WES) - Resistance to CDK4/6 inhibitors (709 pts enrolled; 148 with tumor WES)

Conclusions: Partnering directly with pts enables rapid identification of thousands of pts willing to share tumors, blood, saliva, and medical records to accelerate research. This approach allows for identification of patients with specific phenotypes, who have been challenging to identify with traditional approaches. Remote acquisition of medical records and saliva/blood/tumor samples is feasible. This clinically annotated dataset is a shared resource for the research community.

Table 1Clinical data collection, biospecimen acquisition, and genomic data generation:NumberConsent signed3290 ptsSurvey #1 submitted3290 pts(demographics, diagnosis details, receptor status, clinical experiences)Survey #2 submitted1435 pts(pathology details, sites of metastasis, treatments with start and stop dates)Medical record received1307 ptsSaliva sample received1976 ptsBlood sample received1121 ptsTumor samples received482 tumor samples from 346 ptsDigital image of tumor slide H&E generated482 tumor samplesWES from germline complete310 germline samplesWES from tumor sample complete341 tumor samplesRNA-seq from tumor sample complete229 tumor samplesULP-WGS from cfDNA complete947 blood samplesWES from circulating tumor DNA complete122 blood samples

Citation Format: Nikhil Wagle, Corrie Painter, Elana Anastasio, Michael Dunphy, Mary McGillicuddy, Esha Jain, Tania G Hernandez, Sara Balch, Beena Thomas, Dewey Kim, Alyssa L. Damon, Shahrayz Shah, Brett N. Tomson, Rachel Stoddard, Colleen Nguyen, Jorge Buendia-Buendia, Ofir Cohen, Jorge Gomez Tejeda Zanudo, Netsanet Tsegai, Lauren Sterlin, Ulcha Fergie Ulysse, Kathryn Sine, Oyin Alao, Jacqueline Lucia, Eric S. Lander, Todd R. Golub. The metastatic breast cancer project: Generating the clinical and genomic landscape of metastatic breast cancer through patient-partnered research [abstract]. In: Proceedings of the 2019 San Antonio Breast Cancer Symposium; 2019 Dec 10-14; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2020;80(4 Suppl):Abstract nr PD8-01.

Abstract PD9-03: The genomic landscape of de novo metastatic breast cancer (MBC)

Background: Approximately 5-10% of newly diagnosed breast cancers (BC) are de novo MBC, which means that metastatic disease was identified at the time of initial diagnosis. Patients with de novo MBC are underrepresented in currently available genomic studies. In The Cancer Genome Atlas (TCGA) dataset, only 15 out of ˜980 BC patients can be classified as having de novo MBC. The objective of this study is to analyze the genomic landscape of de novo MBC and to study the genomic differences of this cohort with early stage BC. To enhance our ability to study de novo MBC, we utilized data from the Metastatic Breast Cancer Project (MBCproject), a patient-partnered research project that aims to generate a large public database of clinical, genomic, and patient reported data (PRD) from patients with MBC.

Methods: We defined de novo MBC as patients diagnosed with metastatic disease less than 4 months after their initial diagnosis with BC.Out of 127 patients in the MBCproject with publicly released whole exome sequencing (WES) data, we identified 33 patients with de novo MBC. We combined this data with 15 de novo MBC patients in TCGA. For patients with de novo MBC with multiple tumor biopsies available, we used WES from breast biopsies to enable appropriate comparison to the early stage biopsies. Somatic mutations were evaluated and significantly recurring genes were identified using MutSig2CV. We compared the mutations found in the de novo cohort with early stage tumors. 10 patients in the de novo MBC cohort had paired simultaneous primary and metastatic biopsies; genomic alterations in these samples were compared. Finally, we used RNA sequencing data to compare cytolytic signatures among the de novo and early stage biopsies.

Results: Among the 48 patients analyzed the receptor subtype distribution was: HR+/HER2-(23), HR+/HER2+(13), HR-/HER2+(2), HR-/HER2-(3), HR+/HER2 unknown(5), and HR-/HER2 unknown(2). Histology subtype distribution was as follows: IDC(39), MDLC(6), ILC(2) and Other (1). Significantly recurrent genes in the de novo MBC cohort (q<0.1) included TP53(27%), PIK3CA(30%), CDH1(8%) and MAP3K1(11%). Mutations in PTEN, EGFR, and MDM4 were significantly enriched (p <0.05) in the de novo cohort when compared to early stage BC Evolutionary analysis of paired primary and metastatic biopsies for de novo MBC patients demonstrated the presence of shared clonal mutations, indicating that these were highly evolutionarily related. RNA-seq immune cytolytic signature was downregulated in de novo MBC as compared to early stage BC (p <0.2).

Gene% Mutation Rate in De Novo MBC (N=48)% Mutation rate in Early Stage BC (N= 997)p-valuePTEN10.403.510.0324EGFR6.250.500.00435MDM44.170.300.0189

Conclusions: Our initial results highlight genomic differences between de novo MBC and early stage BC, including increased frequency of PTEN, EGFR, and MDM4 mutations. Enrichment of PTEN mutations (implicated in tumor immune surveillance), and downregulation of cytolytic signature potentially suggests that de novo MBC may have immunosuppressive tumor microenvironment. To date, ˜1200 patients with self-reported de novo MBC have registered for the MBCproject. We anticipate that additional study of genomic and clinical data from these patients will greatly improve our understanding of de novo MBC.

Citation Format: Jain E, Kim D, Buendia JB, Cohen O, Sousa RB, Anastasio E, Dunphy M, McGillicuddy M, Stoddard R, Balch S, Thomas B, Di Lascio S, Tomson BN, Nguyen C, Painter C, Wagle N. The genomic landscape of de novo metastatic breast cancer (MBC) [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr PD9-03.

Abstract 5384: The Angiosarcoma Project: Generating the genomic landscape of an exceedingly rare cancer through a nationwide patient-driven initiative

Angiosarcoma (AS) is an exceedingly rare soft tissue sarcoma, with an incidence of 300 cases/yr and a 5-year disease-specific survival of 30%. The low incidence has impeded large-scale research efforts that may lead to improved clinical outcomes. To address this, we launched a nationwide clinical-genomics study in order to empower patients to accelerate research by sharing their normal and tumor samples and clinical information remotely. Patients can access the study through an online portal (ASCproject.org). Enrolled patients are mailed saliva and blood draw kits. The study team obtains medical records and stored FFPE tumor samples. All received FFPE samples are examined by an expert pathologist to confirm a diagnosis of angiosarcoma. In order to validate that our processes would enable the generation of a robust dataset from tissues acquired from multiple institutions, we sought to characterize previously described genes known to be altered in angiosarcoma (e.g., TP53, NF1, KDR, BRCA2, MET, ARID1A, POT1, BRCA1, ASXL1, KDM6A, BRAF, SETD2, PTPRB, NRAS). A total of 251 patients have enrolled since the project launched in March of 2017. Primary locations of AS are primary breast 59 (25%), breast with prior radiation 45 (19%), head/face/neck/scalp 52 (22%), bone/limb 26 (11%), abdomen 5 (2%), heart 5 (2%), lung 2 (1%), liver 1 (1%), lymph 1 (0.4%), multiple locations 25 (11%), and other locations 12 (5%); 107 (52%) reported being disease free at the time of enrollment. To date, we have received 129 saliva kits, 106 medical records, 19 blood samples, and 36 tissue samples. Whole-exome sequencing (WES) was performed on 21 FFPE/saliva matched pairs with a goal mean target coverage of 150x for tumors. Ultra-low pass whole-genome sequencing (0.1x) was performed on cell free DNA (cfDNA) from plasma in order to determine tumor fraction. Of 10 cfDNA samples sequenced, 4 samples met criteria to perform WES. Additionally, transcriptome sequencing was performed on 9 FFPE samples. Sequence data processing and analysis has been completed on the first 10 samples and is in progress for the subsequent samples. Alterations were detected in genes previously described to be affected in angiosarcoma. Recurrent mutations in TP53 were detected in 50% (5/10) of analyzed samples, comprising 3 missense mutations, 1 frameshift deletion, and 1 frameshift insertion. Alterations were seen in at least one sample in all other genes selected for this initial analysis. This initiative demonstrates the feasibility of studying tissues from geographically dispersed patients and serves as proof of concept that patient-driven genomics efforts can democratize research for exceedingly rare cancers. Enrollment is still in progress, and additional samples will be sequenced and analyzed at scale. The data generated from these studies will be deposited into the public domain in six-month intervals.

Citation Format: Michael Dunphy, Esha Jain, Elana Anastasio, Mary McGillicuddy, Rachel Stoddard, Beena Thomas, Sara Balch, Kristin Anderka, Katie Larkin, Niall Lennon, Yen-Lin Chen, Andrew Zimmer, Esme O. Baker, Simone Maiwald, Jen Hendrey Lapan, Jason Hornick, Chandrajit Raut, George Demetri, Eric Lander, Todd Golub, Nikhil Wagle, Corrie Painter. The Angiosarcoma Project: Generating the genomic landscape of an exceedingly rare cancer through a nationwide patient-driven initiative [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 5384.

Abstract 5371: The Metastatic Breast Cancer Project: Partnering with patients to accelerate progress in cancer research

The Metastatic Breast Cancer Project (MBCproject) is a research study that directly engages patients (pts) through social media and advocacy groups, and empowers them to share samples, clinical data, and experiences. The goal is to create a publicly available database of genomic, molecular, clinical, and patient-reported data to enable research. Working with pts and advocates, a website (MBCproject.org) was developed that allows pts with metastatic breast cancer (MBC) to register. Registered pts are sent an online consent form that asks for permission to obtain and analyze their medical records and samples. Once enrolled, pts are sent a saliva kit and asked to mail back a saliva sample, which is used to extract germline DNA. We contact participants' medical providers and obtain medical records and a portion of their stored tumor biopsies. Pts may be asked to mail in a blood sample, which is used to extract cell free DNA (cfDNA). Whole-exome sequencing (WES) is performed on tumor DNA, germline DNA, and cfNDA; transcriptome sequencing is performed on tumor RNA. Clinically annotated genomic data are used to study specific pt cohorts (including outliers) and to identify mechanisms of response and resistance to therapies. All de-identified data are shared via public databases. Study updates are shared with participants regularly. From 10/2015-11/2017, 4237 MBC pts registered, representing over 1,000 institutions. 95% answered the 16-question survey about their cancer, treatments, and demographic information. 2471 (58%) completed the consent form. 2,136 saliva kits were mailed to pts and 1,523 saliva samples were sent in (71%). 408 blood kits were mailed to pts and 175 blood samples have been received for cfDNA analysis. To date, we have obtained medical records from 311 pts and 190 tumors from 127 pts. In 10/2017, all data generated so far were publicly released on cbioportal.org, including WES for 103 tumors from 78 pts linked to clinical data including pathology (22 elements), medical record abstraction including all treatments and timelines/durations (67 elements), and patient-reported data (11 elements). 81% of biopsies included in this release were from the breast and 19% from metastatic sites. 75% were obtained prior to any therapy, 24% following therapy. New data will be released 4/2018 and every six months thereafter, as they are generated. Additional patient-reported data, including treatments, side effects, quality of life, family history, pregnancies, and sites of metastasis, will also be collected and shared. In summary, a patient-driven approach enabled rapid identification of thousands of MBC pts willing to share samples and clinical data. Remote acquisition of medical records, saliva, blood, and tumor tissue for pts across the U.S. is feasible. This shared clinico-genomic database should enable research in MBC and may serve as a model for patient-driven research in other cancers.

Citation Format: Nikhil Wagle, Corrie Painter, Elana Anastasio, Michael Dunphy, Mary McGillicuddy, Rachel Stoddard, Esha Jain, Dewey Kim, Simona Di Lascio, Brett N. Tompson, Sara Balch, Beena Thomas, Priti Kumari, Shawn Johnson, Jamie Holloway, Ofir Cohen, Erik H. Knelson, Katie Larkin, Sam Pollock, Alicia Wong, Samira Bahl, Simone Maiwald, Andrew Zimmer, Esme O. Baker, Jen Hendry Lapan, Scott Sutherland, Scott Sassone, Viktor Adalsteinsson, Eric S. Lander, Todd R. Golub. The Metastatic Breast Cancer Project: Partnering with patients to accelerate progress in cancer research [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 5371.

The Angiosarcoma Project: Generating the genomic landscape of a rare cancer through a direct-to-patient initiative

1519

Background: Angiosarcoma (AS) is a rare soft tissue sarcoma, with an incidence of 300/yr and a 5-year DSS of 30%. The low incidence has impeded large-scale research efforts that may lead to improved clinical outcomes. To address this, we launched a nationwide study, which seeks to empower patients (pts) to accelerate research by sharing their samples and clinical information remotely. Methods: With pts and advocacy groups we developed a website to allow AS pts to participate across the US. Pts are mailed a saliva and blood draw kit for germline and cell free (cf) DNA analysis. We then obtain medical records and stored tumor samples. Whole exome sequencing will be performed on tumor, cfDNA and saliva samples. Transcriptome analysis will be performed on tumor samples. A clinically annotated genomic database will be generated and shared widely to identify genomic drivers and mechanisms of response and resistance to therapies. Study updates will be shared with pts regularly. Results: We conducted a 3-week pilot study to test the feasibility of enrolling geographically dispersed AS pts through a direct-to-patient (DTP) approach. Through social media, we identified 100+ pts willing to participate, 90 within the first day of outreach. We enrolled 15 pts from 10 states to test our ability to remotely obtain pt reported data, online consent, and samples. The average age of pts is 48, ranging 23-71 yrs. Primary locations of AS are breast 6 pts (40%), cardiac 4 pts (27%), scalp 2 pts (13%), liver 1 pt (6%), bladder 1 pt (6%), forehead 1 pt (6%). 9 pts (60%) reported being disease free, 4 pts (27%) reported having AS spread to lung, lymph, bone, and hip. Requests for medical records and tissue samples are underway, and initial saliva samples have been received. We are now opening this study to all AS pts in the USA. Conclusions: A DTP approach enabled rapid identification of an initial cohort of AS pts willing to share tumors, saliva, blood and medical records. We were able to obtain detailed clinical experiences and samples to perform genomic analysis. This study serves as proof of principle that DTP genomics efforts can democratize cancer research for exceedingly rare cancers, which to date have been disproportionately understudied.

The Metastatic Breast Cancer (MBC) project: Accelerating translational research through direct patient engagement

1076

Background: The Metastatic Breast Cancer Project is a nationwide research study, launched in Oct 2015 in collaboration with patients (pts) and advocacy groups, that directly engages pts through social media and seeks to empower them to share their experiences, clinical information, and samples to accelerate research. Methods: MBC pts enroll by providing their information at mbcproject.org. Pts are sent a saliva kit and asked to mail back a sample which is used to extract germline DNA. We contact pts medical providers and obtain medical records (MRs) and stored tumor samples. Pts may also submit a blood sample, used to extract cell free DNA (cfDNA). Whole exome sequencing (WES) is performed on tumor, germline, and cfDNA; transcriptome sequencing is performed on tumor. Clinical and genomic data are used to generate genomic landscapes in pt subgroups and to identify mechanisms of response and resistance to therapies. Data are shared widely through public databases. Pts receive regular study updates. Results: In 12 months, 2908 MBC pts from 50 states enrolled. 95% completed the 16-question survey about their cancer, treatments, and demographics. 1730 (60%) completed the online consent form. 100-200 pts continue to enroll monthly. To date, 1539 saliva kits were mailed and 1120 samples were received (73%). 992 unique treating institutions were reported by pts, including 733 institutions reported by only 1 pt each and 5 institutions reported by more than 40 pts each. We have obtained MRs from 253 patients (67% yield) and tumor samples from 85 pts (67% yield). WES was successfully completed for 79 tumors of 88 attempted (90%). WES has been performed on initial cfDNA samples. Conclusions: A direct-to-patient approach enabled rapid identification of thousands of MBC pts willing to share MRs, saliva, and tumor samples, including many with rare phenotypes. Remote acquisition of MRs, saliva, tumor, and blood for pts located throughout the US is feasible. We estimate that for ~33% of consenting patients, we can obtain medical records, saliva, and tumor tissue. Genomic analysis of tumor and cfDNA from subgroups including young pts, pts with extraordinary responses, and pts with de novo MBC will be presented.