10604 Background: Although germline and somatic genetic testing is increasingly recommended for many cancers including gynecologic cancers, uptake remains low. Barriers include consenting processes that are time consuming for patients and providers. We sought to test uptake and acceptability of a digital, self-directed genetic testing consent process. Methods: We developed a digital, self-guided consenting process for MSK-IMPACT tumor-normal targeted sequencing called “read and sign” (R&S) that included an educational video and information about genetic testing along with a direct digital consent. We performed a pilot study in gynecologic surgery (GYNS) and medical oncology (GMO) clinics to evaluate the uptake of genetic testing and other metrics including time to consent, clinic time saved, and patient satisfaction. Briefly, adults with gynecologic cancers eligible for MSK-IMPACT were sent R&S consent via electronic health record. Research staff followed up with patients at 2, 7, and 14 days. If patients did not consent after 14 days, the clinical team was notified to follow-up. This study was approved under IRBs 12-245 and X25-012. Results: Between 10/6/2025 and 12/31/2025, we identified 173 patients (122 GYNS and 51 GMO) eligible for MSK-IMPACT and offered them R&S consent. This represented 8% of total MSK-IMPACT consents across the institution during this time. Median age was 61 years. 106 (61%) patients identified as White, 27 (16%) as Asian, 17 (10%) as Black, and 24 (14%) as Hispanic. Most common cancer types were endometrial (n=75, 43%) and ovarian (n=39, 23%). In GYNS, 102/122 (84%) patients consented with R&S, and 59 (58%) consented on the same day the R&S consent was sent. On average, 8 minutes of in-person clinic time per patient was saved using R&S, but 35 (29%) patients had 0 time saved, and 7 (7%) consents were later completed in-person after initial incomplete R&S attempt. One patient who completed the R&S consent later withdrew consent. In GMO, 50/51 (98%) patients consented via R&S, with 38 (76%) consented on the same day. On average, 14 minutes of clinic time per patient was saved, and 4 (8%) patients ultimately required in-person consent. Average time to consent was 3.3 days in GYNS and 1.7 days in GMO. Among 11 patients who completed surveys, 10 (91%) were very/somewhat satisfied with R&S. More in-depth surveys and semi-structured interviews with a subset of patients and staff are ongoing, and data will be forthcoming. Conclusions: A digital, self-guided consent for MSK-IMPACT genetic testing was feasible and acceptable with some differences in workflow between surgical and medical oncology clinics. The R&S consent process empowers patient decision-making and should be scaled to other cancer types.

e21001 Background: Clinical trial timelines are significantly influenced by the efficiency of data transfer from sites’ Electronic Health Records (EHRs) to sponsors’ Electronic Database Capture (EDC) systems. However, optimizing workflows and tools to assist Clinical Research Coordinators (CRCs) in this process remains underprioritized. Previous studies show that data abstraction roles turnover at rates up to 40%, further exacerbating data transfer delays and escalating costs. Methods: The objective of the study was to investigate shared CRC challenges for the data abstraction process to establish a common understanding of CRC challenges. Forty remote, 60-minute, semi-structured interviews were conducted between 01/22 to 11/24 across six academic cancer centers in the US (4) and the UK (2). Participants included were responsible for transferring data from EHRs to EDCs for clinical trials. To analyze the qualitative data gathered, ground theory-based thematic coding was employed using Dovetail software. Results: Key pain points for CRCs identified across institutions include: Labor-intensive data abstraction: Time-consuming processes to extract data from unstructured EHR notes (e.g., medical history, QoL surveys) and manually enter dozens of data points (e.g., labs, concomitant medications) per cycle into EDCs. A CRC typically manages 5–8 protocols simultaneously. Disjointed system workflow: Navigating multiple EHR tabs and systems, particularly when abstracting PI-verified data like Adverse Events, disrupts workflow efficiency. Five of six sites used the same EHR vendor. Task and query management challenges: Tracking and prioritization of tasks related to data entry, administrative, and queries increase burden. Data gaps from external institutions: Missing or incomplete external records often require follow-up with other institutions. Upstream data reconciliation: Incomplete data capture frequently necessitates additional input or clarification from PIs. Limitations in CRF design: Sponsor-designed case report forms (CRFs) in EDCs often lack CRC input, leading to inefficiencies and excessive auto-queries. Variations in workflows were largely driven by differences in system architectures, EHR configurations, and whether templated data capture processes were in place. Conclusions: Current systems are not optimized for CRC workflows for efficient data abstraction. Addressing these pain points through collaboration between sites and sponsors could improve operational efficiency, shorten study timelines, and reduce costs through fewer data inconsistencies, decreased query resolution volume, and mitigated turnover-related expenses. Establishing integrated systems and workflows will drive systemic improvements in clinical trial operations, advance the pace of oncology research and ultimately improve the quality of patient care.

Introduction:Clinical trial data is still predominantly manually entered by site staff into Electronic Data Capture (EDC) systems. This process of abstracting and manually transcribing patient data is time-consuming, inefficient and error prone. Use of Electronic Health Record to Electronic Data Capture (EHR-To-EDC) technologies that digitize this process would improve these inefficiencies. Objectives:This study measured the impact of EHR-To-EDC technology on the data entry workflow of clinical trial data managers. The primary objective was to compare the speed and accuracy of the EHR-To-EDC enabled data entry method to the traditional, manual method. The secondary objective was to measure end user satisfaction. Materials and Methods:Five data managers ranging in experience from 9 months to over 2 years, were assigned an investigator-initiated, Memorial Sloan Kettering-sponsored oncology study within their disease area of expertise. Each data manager performed one-hour of manual data entry, and a week later, one-hour of data entry using IgniteData's EHR-To-EDC solution, Archer, on a predetermined set of patients, timepoints and data domains (labs, vitals). The data entered into the EDC were compared side-by-side and used to evaluate the speed and accuracy of the EHR-To-EDC enabled method versus traditional, manual data entry. A user satisfaction survey using a 5-point Likert scale was used to collect feedback regarding the selected platform's learnability, ease of use, perceived time savings, perceived efficiency, and preference over the manual method. Results:The EHR-To-EDC method resulted in 58% more data entered versus the manual method (difference, 1745 data points; manual, 3023 data points; EHR-To-EDC, 4768 data points). The number of data entry errors was reduced by 99% (manual, 100 data points; EHR-To-EDC, 1 data point). Regarding user satisfaction, data managers either agreed or strongly agreed that the EHR-To-EDC workflow was easy to learn (5/5), easy to use (4.6/5), saved time (5/5), was more efficient (4.8/5), and preferred it over the manual entry workflow (4/5). Conclusion:EHR-To-EDC enabled data entry increases data manager productivity, reduces errors and is preferred by data managers over manual data entry.

BACKGROUND:Mirvetuximab soravtansine-gynx (MIRV) is a FOLR1-binding antibody-drug conjugate (ADC) with a microtubule inhibitor payload. We investigated MIRV's efficacy, toxicity profile, and determinants of resistance in a cohort of patients with recurrent/persistent high FOLR1-expressing high-grade serous ovarian cancer (HGSOC). METHODS:This retrospective study included 170 patients with recurrent/persistent FOLR1-high (≥75 % of tumor cells with ≥2+ membranous staining intensity) HGSOC who received standard-of-care MIRV monotherapy. We evaluated progression-free survival (PFS) and overall survival (OS) using the Kaplan-Meier method and multivariable Cox proportional hazards models. We classified adverse events using CTCAE v5.0. RESULTS:Overall, median PFS was 3.5 months (95 % CI, 3.0-4.1). However, 22.4 % had PFS ≥6 months and were less likely to have progressed on or within one month of prior taxane-based therapy (P = 0.008). Patients with previous progression on a taxane had worse PFS (HR, 1.69; 95 % CI, 1.19-2.40; adjusted P = 0.003) and OS (HR, 2.34; 95 % CI, 1.45-3.77; adjusted P = 0.0005). FOLR1 expression was lower in post-MIRV samples (n = 12; P = 0.005). New or worsening neuropathy was observed in 37.6 % of patients. Among the 34.1 % who experienced ocular toxicity, median onset was 42.5 days. Treatment was discontinued in 5.3 % of patients due to toxicity. DISCUSSION:MIRV confers meaningful PFS benefit for a subset of individuals with HGSOC. Resistance may be associated with decreased FOLR1 target expression or payload resistance. FOLR1-targeted ADCs with a different payload should be evaluated for patients who progress on MIRV but retain high tumor FOLR1 expression.

Highest OncoKB level of evidence per (a) tumor type and per (b) tumor type and gene for a subset of solid tumor samples from the MSK-IMPACT cohort with n≥100.

5565 Background: BRCA1/2-associated ovarian cancer (OC) exhibits homologous recombination (HR)-deficiency (HRD) and a better prognosis than BRCA1/2-wildtype (WT) OC. However, it is unclear if patients with germline pathogenic variants (gPV) in other HR-related genes have a similar tumor phenotype. We sought to define OC molecular features from patients with gPV in other HR genes and analyze survival compared to patients with BRCA1/2 and WT OC. Methods: We identified patients with OC treated at our institution who underwent tumor-normal sequencing using MSK-IMPACT targeting 341-505 cancer-related genes with germline analysis of ≥76 genes from 7/2015-12/2020. Biallelic inactivation was inferred via assessments of loss of heterozygosity (LOH) at gPV in non- BRCA1/2 HR-related genes ( ATM, BARD1, BRIP1, FANCA, FANCC, NBN, PALB2, RAD50, RAD51B, RAD51C, and RAD51D). Clinical characteristics were compared to patients with BRCA1/2-associated and WT OC using non-parametric tests. Progression-free (PFS) and overall-survival (OS) were calculated from date of pathologic diagnosis using the Kaplan-Meier method. Left truncation at date of MSK-IMPACT consent was applied. Whole-exome sequencing (WES) was performed in a subset of OCs. Results: Of 882 patients with OC, 56 (6.3%) had gPV in non- BRCA1/2 HR-related genes compared to 95 (10.8%) patients with BRCA1-associated OC (58 germline, 37 somatic) and 59 (6.7%) patients with BRCA2-associated OC (40 germline, 19 somatic). Patients with a deleterious genetic alteration were diagnosed with OC at younger age and more likely to have high-grade serous histology compared to the WT group (p < 0.01). High rates of biallelic alterations were observed amongst gPV in BRIP1 (11/13), PALB2 (3/4), RAD51B (3/4), RAD51C (3/4), and RAD51D (8/10), and WES was performed in a subset (n = 27) of tumors from these patients with adequate tumor purity (>30%). We observed a higher tumor mutational burden (TMB), median 2.5 (1.1-6.0) vs. 1.2 (0.6-2.6) mut/Mb, and enrichment of HRD mutational signatures in tumors associated with PALB2 and RAD51B/C/D compared with BRIP1 (p < 0.01), although markers of telomeric-allelic imbalance (TAI), large-scale state transitions (LST) and fraction of genome altered (FGA) were similar. PFS and OS varied by gene group with best survival in BRCA1/2-associated OC, even after adjustment for clinical covariates in multivariable models (p < 0.01). Although we observed heterogeneity in PFS and OS for those with gPV in other HR-related genes by biallelic status and HRD phenotype, none had significantly better survival than those with WT OC. Conclusions: OCs associated with gPV in non- BRCA1/2 HR-related genes represent a heterogenous group. OCs in those with gPV in PALB2 and RAD51B/C/D preferentially harbored biallelic alterations and displayed an HRD phenotype.

Among genes found to have limited to no clinical actionability (OncoKB Levels 3B or 4 as the highest level of clinical actionability, or no alteration with an OncoKB level) and altered in >1% of samples in the MSK-IMPACT subset of the AACR Project GENIE dataset, left: the breakdown of genes into different gene-function categories and right: the breakdown of genes as oncogenes, tumor suppressors, both or neither per www.oncokb.org.

e13651 Background: Manual abstraction of data from a site’s clinical systems to a biopharmaceutical firm’s electronic data capture (EDC) system is inefficient and error prone. In partnership with data managers (DMs), we used a human-centered design thinking methodology to create a web application, CTDataHub, to reduce the effort associated with this process. CTDataHub extracts and consolidates adverse events (AE) and concomitant medications (ConMed) data from clinical systems and displays it in a user-friendly view for easy entry into EDCs. Methods: CTDataHub was launched to DMs in July 2023 at a large, high-volume academic cancer center. To evaluate the App’s impact on DMs, we assessed: 1) App usage, 2) data entry efficiency, 3) trust in data (data is correct to enter in sponsor EDC) assessed via a 5-point Likert attitudinal trust scale, and 4) satisfaction using Net Promoter Score (NPS) by comparing two surveys. Survey 1 (S1) was sent to 313 DMs prior to receiving access to CTDataHub to establish a baseline. Survey 2 (S2) was sent to meaningful users of CTDataHub; those who used the App and viewed data pages at least 10 times between 09/01/23-10/31/23. Behavioral usage on CTDataHub was tracked using Heap Analytics. Results: App Use: S1 had a 72% response rate (n=225), and S2 had a 51% response rate (n=30); with all respondents of S2 completing S1. Meaningful use DMs sent S2 (59) used CTDataHub to view data 3,094 times, with 76% copying data values directly from the App (1,033 times) for abstraction. Data Entry Efficiency: On average, S1 respondents spend 10.2 hours/week abstracting AEs and ConMeds. In S2, the App increased DM efficiency by reducing average data abstraction time 18% (2h). Most respondents (70%, n=21) reported CTDataHub increased efficiency, 20% (n=6) noted no impact, and 10% (n=3) reported decreased efficiency by 18%. Notably, time savings were greater for DMs who used the App who were new to their role (<1 year) versus seasoned DMs (>1 year); 24% vs. 13%. Trust in Data:DMs responded that CTDataHub was more likely to correctly retrieve ConMed and AE data when compared to their manual methods (4.07 mean (n=30) vs. 3.4 mean (n=225) on a 5-point Likert trust scale. App User Satisfaction:NPS for CTDataHub versus DM standard workflows using various clinical systems (on average 2.5 systems) used for data entry was 23 compared to 4, likely due to its ease-of-use. Conclusions: CTDataHub decreased data abstraction times for AEs and ConMeds for 70% of meaningful users by an average of 2h/week relative to other clinical systems. In addition to DM satisfaction, the App increased trust in data being abstracted. The App will be enhanced with additional data sets (site local labs, vitals, tumor response) and automated EHR2EDC capabilities in 2024.

Table S6: Gene-alteration-tumor types associated with an OncoKB Therapeutic Level of Evidence (Levels 1-4) in March 2017, and changes made to the dataset to ensure 2017 OncoKB clinically actionable variants are consistent with the OncoKB Standard Operating Procedure v2.2.

A list of precision oncology therapies (per the definition provided in the manuscript) approved by the US FDA between January 1998 and November 2022, as well as the year each drug was first FDA approved, the FDA-recognized biomarker(s), the method of biomarker detection, and classification of each drug as first-in-class, mechanistically-distinct, follow-on, or resistance (per definitions provided in the manuscript).

Supplementary Note 1: The Methods on FDA-drug curation taken directly from Olivier T, Haslam A, Prasad V. Anticancer drugs approved by the US food and drug administration from 2009 to 2020 according to their mechanism of action. JAMA Netw Open. 2021;4:e2138793. Supplementary Note 2: The Methods and References on FDA-drug curation taken directly from Sun J, Wei Q, Zhou Y, Wang J, Liu Q, Xu H. A systematic analysis of FDA-approved anticancer drugs. BMC Syst Biol. 2017;11:87.

A breakdown of the number and percentage of tumor type-specific samples included in the MSK-IMPACT subset of the AACR Project GENIE dataset.

For solid tumor samples from the MSK-IMPACT subset of the AACR Project GENIE dataset with n≥100, (a) the median number of oncogenic and actionable (gray), oncogenic and non-actionable (light blue), or total (orange) mutations per sample per cancer type, and (b) the percentage of samples that carry zero, one, two, or three or more actionable mutations (as defined by OncoKB version October 2022) per sample.

Table S5: Genes associated with an OncoKB Therapeutic Level of Evidence (1, 2 or 3A) in March of 2017

Table S4: Genes found to have limited to no clinical actionability (OncoKB Levels 3B or 4 as the highest level of clinical actionability, or no alteration with an OncoKB level) were categorized by 1. the gene/protein function and 2. assignment as an oncogene, tumor suppressor, both or neither per www.oncokb.org.

A list of oncology drugs first approved by the US FDA between January 1998 and November 2022, as well as the year each drug was first FDA approved, the class of agent and mechanism of action, and categorization of each drug as a targeted therapy (Y/N) and precision oncology therapy (Y/N) per definitions provided in the manuscript.

Abstract There is a continuing debate about the proportion of cancer patients that benefit from precision oncology, attributable in part to conflicting views as to which molecular alterations are clinically actionable. To quantify the expansion of clinical actionability since 2017, we annotated 47,271 solid tumors sequenced with the MSK-IMPACT clinical assay using two temporally distinct versions of the OncoKB knowledge base deployed 5 years apart. Between 2017 and 2022, we observed an increase from 8.9% to 31.6% in the fraction of tumors harboring a standard care (level 1 or 2) predictive biomarker of therapy response and an almost halving of tumors carrying nonactionable drivers (44.2% to 22.8%). In tumors with limited or no clinical actionability, TP53 (43.2%), KRAS (19.2%), and CDKN2A (12.2%) were the most frequently altered genes. Significance: Although clear progress has been made in expanding the availability of precision oncology-based treatment paradigms, our results suggest a continued unmet need for innovative therapeutic strategies, particularly for cancers with currently undruggable oncogenic drivers. See related commentary by Horak and Fröhling, p. 18. This article is featured in Selected Articles from This Issue, p. 5

PURPOSE Although germline genetic testing (GT) is recommended for all patients with ovarian cancer (OC) and some patients with endometrial cancer (EC), uptake remains low with multiple barriers. Our center performs GT in parallel with somatic testing via a targeted sequencing assay (MSK-IMPACT) and initiates testing in oncology clinics (mainstreaming). We sought to optimize our GT processes for OC/EC. METHODS We performed a quality improvement study to evaluate our GT processes within gynecologic surgery/medical oncology clinics. All eligible patients with newly diagnosed OC/EC were identified for GT and tracked in a REDCap database. Clinical data and GT rates were collected by the study team, who reviewed data for qualitative themes. RESULTS From February 2023 to April 2023, we identified 116 patients with newly diagnosed OC (n = 57) and EC (n = 59). Patients were mostly White (62%); English was the preferred language for 90%. GT was performed in 52 (91%) patients with OC (seven external, 45 MSK-IMPACT) and in 44 (75%) patients with EC (three external, 41 MSK-IMPACT). GT results were available within 3 months for 100% and 95% of patients with OC and EC, respectively. Reasons for not undergoing GT included being missed by the clinical team where there was no record that GT was recommended, feeling overwhelmed, financial and privacy concerns, and language barriers. In qualitative review, we found that resources were concentrated in the initial visit with little follow-up to encourage GT at subsequent points of care. CONCLUSION A mainstreaming approach that couples somatic and germline GT resulted in high testing rates in OC/EC; however, barriers were identified. Processes that encourage GT at multiple care points and allow self-directed, multilingual digital consenting should be piloted.

Recurrent cases in MSKCC vs. TCGA cohort