Matching Platform Research Articles

Research Buddy—From a Framework for Overcoming Language Barriers to the Development of a Qualitative Citizen Science Platform

Abstract Students benefit from international study programmes in fields such as HCI. However, user-centred research in a foreign country has some challenges, especially language hurdles when going into the field and conducting ethnographic and participatory studies. A particular obstacle is added when targeting groups that do not speak English, such as in Germany, where English was not taught in schools until the 1970s. The task and challenge of the present project, ‘Research Buddy’, is to bring together young international students and older people as co-researchers in Germany. We describe the participatory development of the concept of a matching platform and its design. In focus groups with older people, university faculty and students, requirements for a collaborative research infrastructure were jointly identified and prototypically implemented. The empirical results focus on the possibilities the platform offers for qualitative citizen science in the field of age tech.

Effectiveness of an HIV peer support precision matching platform: a randomized controlled trial protocol

BackgroundPeer support is a promising and significant way for people living with HIV (PLWH) to improve their health outcomes. Matching PLWH with HIV peer volunteers can effectively stimulate and mobilize the function of peer support in health management. We develop the online platform Aspark, which features a precision matching function for PLWH and volunteers. This protocol provides a comprehensive overview of the randomized controlled trial design based on Aspark.MethodA parallel randomized controlled trial using block randomization will be conducted at the Beijing YouAn Home of Loving Care. The intervention group (n = 70) will have access to all 12 features of Aspark, including precise volunteer matching, health management tools, drug interaction queries, and medication reminders. The control group (n = 70) will have access to two features of informational support on Aspark. The outcomes of both groups will be evaluated at baseline and at 3, 6, 9, and 12 months after platform registration. The primary outcome is perceived social support. The secondary outcomes include mental symptoms, quality of life, self-efficacy, medication adherence, and health behaviors. Linear mixed models or generalized linear mixed models will be applied to test the intervention effects.DiscussionThe Aspark trial is a randomized controlled trial designed to examine the influence of a mobile health (mHealth) application that integrates peer support and precision matching to strengthen social support for PLWH in China. This trial has the potential to significantly enhance social support for PLWH while also optimizing the delivery of peer support services and fully mobilizing the role of peer support in the overall health management of PLWH.Trial registrationClinical registration number: ChiCTR2300077454.

Optimal analysis for bandit learning in matching markets with serial dictatorship

The problem of two-sided matching markets is well-studied in computer science and economics, owing to its diverse applications across numerous domains. Since market participants are usually uncertain about their preferences in various online matching platforms, an emerging line of research is dedicated to the online setting where one-side participants (players) learn their unknown preferences through multiple rounds of interactions with the other side (arms). Sankararaman et al. [23] provide an Ω(Nlog⁡(T)Δ2+Klog⁡(T)Δ) regret lower bound for this problem under serial dictatorship assumption, where N is the number of players, K(≥N) is the number of arms, Δ is the minimum reward gap across players and arms, and T is the time horizon. Serial dictatorship assumes arms have the same preferences, which is common in reality when one side participants have a unified evaluation standard. Recently, the work of Kong and Li [10] proposes the ET-GS algorithm and achieves an O(Klog⁡(T)Δ2) regret upper bound, which is the best upper bound attained so far. Nonetheless, a gap between the lower and upper bounds, ranging from N to K, persists. It remains unclear whether the lower bound or the upper bound needs to be improved. In this paper, we propose a multi-level successive selection algorithm that obtains an O(Nlog⁡(T)Δ2+Klog⁡(T)Δ) regret bound when the market satisfies serial dictatorship. To the best of our knowledge, we are the first to propose an algorithm that matches the lower bound in the problem of matching markets with bandits.

Impact of Exigent Research Network’s partnership with the Tempus AI TIME Program on the screening and matching of patients for clinical trials.

e13599 Background: Exigent Research is comprised of independent community oncology practices with a central patient data repository to support clinical research. The Tempus AI TIME program provides clinical trials solutions including a diverse trial portfolio, TApp an AI powered patient matching platform, patient tracking LINK portal, and a rapid trial activation process. This partnership facilitated Tempus AI to continuously search the data repository and enabled Exigent to increase trial enrollment. Methods: TApp is a clinical trial matching platform that utilizes patient data, trial eligibility criteria, and NLP models to execute AI powered matching of subjects to trials. TApp searches the entire patient population for TIME trial matches. New patient clinical information or changes to trial eligibility criteria triggers a new TApp search and matches are reviewed by Tempus AI nurses. Trials were opened in an accelerated “just in time” manner if a subject was ready immediately or prospectively prior to a first patient being identified. Data collected included trial activations, number of patients, available studies, patient and trial updates, TApp searches, nursing effort, matches, and consents. Interventional Phase I to III trials and observational studies were included. Results: From 2022 - 2023, there were 16 Exigent sites in the TIME network. The patient population screened was 244,986 matched to 189 clinical trials. There were 3,179,079 changes to individual patient records and 2,152 trial updates. This triggered 216,712,093 unique TApp searches and 33,083 matches reviewed by nurses. In total, 1,312 patients matched to observational studies and 6,040 patients for interventional trials. Cumulatively, nurses spent 6,617 hours on patient screening resulting in a $330,830 cost. There were 71 trial activations resulting in 312 patient consents (169 observational, 143 interventional). The average duration for study activation was 12.3 days for a just in time activation and 33.8 days for a prospective activation. Conclusions: Exigent Research leveraged a central data repository with the Tempus AI TIME program to continuously screen a large patient population for 185+ trials. During a two-year period, 215M+ unique TApp searches occurred which resulted in over 310 patient consents with rapid activation times. Future clinical trial strategies should partner AI screening technologies with robust trial infrastructure solutions to maximize enrollment success. [Table: see text]

Bridging gaps with navigator support and its influence on breast cancer clinical trial matching and accessibility.

e13509 Background: Breast cancer patients face challenges navigating clinical trials. Online trial matching platforms like BreastCancerTrials.org (BCT) and MetastaticTrialSearch (MTS) are important resources to assist patients in identifying suitable clinical trials. However, data collected from users shows a concerning underrepresentation of ethnic minorities using this technology. Acknowledging the need to address this inequality through personalized support, we have incorporated a Clinical Trial Navigator. The navigator has engaged in partnerships with diverse community organizations and is committed to offering continuous support to patients throughout their experience. Methods: We analyzed data from our online trial matching sites spanning the period 2009–2023 (BCT) and data from our new platform MTS covering the years 2018-2023. Users visiting these sites provided details about their breast cancer type and medical history to identify appropriate clinical trials. Demographic information including race, gender, and education level was collected. The utilization and effectiveness of these features in facilitating patient access to relevant clinical trials were evaluated. Results: Analysis of the data reveals information about the characteristics of patients actively using the BreastCancerTrials.org matching service. Although BCT and MTS strives to reach a diverse patient population, 87% of BCT users identified as white women, with African American users making up only 5%. Education level of all users was high with 73% stating a college degree or higher. On the MTS site, 80% of users identified as white, while 7% identified as African American.This emphasized the need for specific interventions designed to tackle the underrepresentation of ethnic minorities and underserved populations when utilizing BCT and MTS. Conclusions: For more than 15 years, BCT has offered a patient-centric trial matching service. Within the last six years, with the enhancements to the website, BCT has served over 281,197 users. To improve minority enrollment in trials, we have developed a comprehensive clinical trial breast cancer navigation program. We plan to use this program to expand our outreach to underserved populations. Navigators can enhance patient engagement, help patients make informed decisions, and ultimately improve access to participation in breast cancer research.

Volunteer and NGO Matching Platform

According to a poll conducted in 2021, during the epidemic, the number of donors in India increased by 43%. During the pandemic, Indian non-governmental organizations (NGOs) volunteered their services in every manner conceivable, and they still do. However, over the years, the proportion of volunteers has remained consistent. This demonstrates a disconnect between the general population and nonprofits as well as volunteerism. Establishing a platform that connects these volunteers with other charitable organizations would be crucial. The "Volunteer Matching Platform" project may be able to bridge this gap. This web-based platform is intended to facilitate volunteer connections between non-governmental organizations and local groups. NGOs will use this platform to promote their events to the nation's citizens. On the other hand, volunteers will be able to browse through the available opportunities, filtering based on their interests, location, and availability.

Prediction algorithms in matching platforms

We follow the future trajectory of more targeted wage formation in labor matching platforms, such as freelancing, crowd-sourcing, home-delivery, and ride-hailing, where local job search is coordinated by improving prediction algorithms. A labor matching platform is modelled as a directed search and matching market. We observe that targeted wage setting promotes efficient matching and longer employment spells. However, because a higher employment rate accentuates any disparities between available workers and vacancies, the effects of targeted wage setting on firm competition depend on prevailing market tightness. The impact of targeted wage formation on workers is positive when the vacancy-to-worker ratio is intermediate but turns negative at both extremes. Our results suggest that targeted wage setting may benefit occasional workers while potentially posing drawbacks for full-time platform workers.

Signaling Effects Under Dynamic Capacity in Online Matching Platforms: Evidence from Online Health Consultation Communities

Online Health Consultation Communities (OHCCs) have emerged as vital platforms connecting patients with physicians for online consultations. However, finding the right match between patients and physicians can be tricky due to physicians’ changing capacity to consult on the OHCC. Our study delves into how signals provided by OHCCs on physicians’ professional status and behaviors can help make these matches successful, especially when considering their capacity fluctuations. We differentiate between two types of signals pertaining to physicians–owned, pertaining to their professional status, and earned, pertaining to their OHCC activity and patient reviews of the physician. Employing a Hidden Markov Model to analyze data from a large OHCC on physicians’ voluntary online consultations with patients, we find the role of the signals in efficient matching to be contingent on the capacity state of the physician. Physicians’ professional status is particularly important when they have less time available, and showing active participation in the community can make the status even more impactful. Conversely, when physicians have more availability, patient feedback becomes crucial, even diminish the importance of professional status. These insights suggest that OHCCs should tailor how information on physicians’ professional status and patient feedback are presented depending on physicians’ availability as this can help patients to make better choices. By being active in the OHCC and earning favorable patient feedback, physicians with more availability can improve their attractiveness to patients, even offsetting concerns that can stem from the lack of seniority of the physician. The findings underscore the need for OHCCs to develop signaling and matching mechanisms that consider the capacity of physicians, thereby fostering efficient and satisfactory patient-physician consultations.

Mr. Right or Mr. Best: The Role of Information Under Preference Mismatch in Online Dating

The rise of two-sided matching platforms such as Uber, Airbnb, Upwork, and Tinder has changed the way we commute, travel, work, and even date. The success of these platforms depends on the role of information: What information and how much information should be provided? In this study, we focus on a defining characteristic of two-sided matching markets—that is, a match depends on the possibly different preferences of the two sides—and argue that the optimal amount of information released depends on the extent to which the preferences of the two sides are mismatched. Specifically, in an empirical context of online dating, we find that when there exists preference mismatch between the two sides, having less match-relevant information about the other side leads to a better matching outcome. Our study provides insights into how the amount of information available to each side affects matching outcomes on two-sided platforms and offers guidance on information design strategies. Additionally, our findings are not confined to dating websites and can be extended to other matching platforms, such as Airbnb and Upwork, where misaligned preferences can exist between the two sides.

Digital Labour Platforms as Shock Absorbers: Evidence from the COVID-19 Pandemic in Mozambique

Abstract Digital labour platforms have grown five-fold over the past decade, enabling significant expansion of gig work worldwide. We interrogate the critique that these platforms tend to amplify aggregate shocks for registered workers. Based on the universe of records from a matching platform for manual freelancers in Mozambique, we analyse how task supply and demand altered with the onset of COVID-19. Treating the pandemic as a structural break, which extends to an event study analysis, we find it was associated with a net increase in tasks demanded per worker, but no clear change in supply growth (new registrations). While these general trends are evident across multiple market segments, there is also evidence of relatively higher demand for female as compared to male workers during the pandemic. This suggests digital labour markets can help some workers adjust to economic shocks in low-income contexts.

Compatibility and Information Asymmetry in Online Matching Platforms

Firms seeking business partners and individuals seeking life partners face several challenges in addition to finding available candidates. One of these challenges is uncertainty about a candidate’s compatibility with the match-seeker on various subjective criteria. Another challenge is authentication of objective quality features, particularly in online settings where credentials can be easily misrepresented. Online matching platforms can address these challenges through counseling services that help assess mutual compatibility of match-seekers and authentication services to validate match-seekers’ credentials. Using a stylized model, we show that the platform’s decisions to offer counseling and authentication services are influenced by match-seeker behavior in response to factors such as the platform’s counseling capability, the composition of the matching market, and the interactions between the two services. This leads to optimal strategies for the platform that are not obvious and can even be counterintuitive. For instance, offering better counseling does not necessarily result in higher revenues for the platform. And it may be optimal for the platform to reduce its access fee when it decides to offer an authentication service. We also show that when the platform’s counseling capability is modest, an increase in this capability increases the value of authentication (i.e., counseling complements authentication); however, when the platform’s counseling capability is high, counseling and authentication become substitutes, because narrowing the search to only high-value candidates based on authentication can exclude some compatible candidates. Our analysis and results provide valuable insights for platform operators on the relative value of counseling and authentication, and for researchers studying matching platforms for vertically and horizontally differentiated markets. This paper was accepted by Hemant Bhargava, information systems. Supplemental Material: The online appendix and data files are available at https://doi.org/10.1287/mnsc.2021.03807 .

Migratory destinations of endangered humpback whales, Megaptera novaeangliae (Cetartiodactyla: Balaenopteridae), from El Salvador

Introduction: The study of many aspects of cetacean ecology is made possible by identifying individuals through space and time. Humpback whales (Megaptera novaeangliae) can be easily identified by photographing their ventral tail flukes’ unique shape and pigmentation patterns. The small and endangered distinct population segment (DPS) of Central America humpback whales visit El Salvador seasonally each winter; however, dedicated research has been extremely limited there. Before 2018, only 11 individual whales had been photo-identified, and the migratory destinations of Salvadoran humpback whales were unknown. In recent years, photo-identification efforts have increased, and today there are 92 individually identified humpback whales from El Salvador.
 Objective: To identify the main high-latitude feeding areas of Salvadoran humpback whales.
 Methods: Using the online matching platform Happywhale, Salvadoran whales were matched via automated image recognition to a global humpback whale fluke photo-identification catalog of 66 043 individuals.
 Results: In total, 80 (87.0 %) of the whales photographed in El Salvador were matched to individuals seen in North Pacific feeding areas. Sighting histories of Salvadoran whales resighted in feeding areas ranged from two to 29 years (average = 12.1, SD = 5.8). While we note that survey effort was likely very different between regions, the main feeding area of Salvadoran humpback whales on Happywhale was Central California (n = 70, 76.1 %). Of these whales, 21 (22.8 %) had also been sighted in Southern California, while just three (3.3 %) individual whales were registered only in Southern California. Additionally, two whales (2.2 %) were sighted in Southern British Columbia, Canada, and one whale was matched to a humpback whale from Southeast Alaska. This whale (of unknown sex) has a sighting history of 27 years but no prior documentation in a breeding area and is the first published sighting of a Southeast Alaskan humpback whale in the breeding area of the endangered Central America DPS.
 Conclusions: Our study shows that while Salvadoran humpback whales were matched to various feeding areas in the Eastern North Pacific, their primary migratory destinations are in Southern and Central California.

The determinants of online matching platforms for freight services

With the development of information technology and the success of E-commerce, online freight exchange platforms have expanded to trucking freight services by matching truckers with shippers directly. Despite the endeavors of online freight platforms to emulate the success of other platform services, they face difficulties in facilitating successful matches. By analyzing the transactional data of an existing freight service platform, we seek to identify the salient factors that improve platform matching rates using a linear regression with splines for freight rate, capturing non-linearities, and a RUM discrete outcome model for matching. Both models are estimated jointly (simultaneously) using full information maximum likelihood (FIML). The empirical findings indicate that successful matching is more likely if the online freight platforms allow making substantial price adjustments once a job has been posted, as well as allowing shippers to have a longer lead time for loading their shipments, particularly in cases where finding a suitable trucker is challenging. Furthermore, allowing customers to revise shipment information provides greater flexibility in accommodating customer needs and ultimately enhances the likelihood of successful matching. Our analysis provides actionable and pragmatic suggestions to improve matching rates for online platforms of freight services.

POMF: A Privacy-preserved On-chain Matching Framework

With the growth of the Internet, an increasing number of online matching platforms have emerged. However, current online matching platforms need to address three open challenges to improve user experience: (1) transparency and inherent issues with centralized platforms; (2) privacy concerns of users’ sensitive information; and (3) high service fees for matching services. In this paper, we propose a Privacy-preserved On-chain Matching Framework (POMF) that is suitable for a broad range of matching scenarios to overcome the challenges above simultaneously. In order to address the first challenge, POMF is constructed via the decentralized blockchain, based on which smart contracts are utilized to execute matching services automatically and publicly. To address the second challenge, we combine Order-Preserved Encryption (OPE) with Software Guard Extensions (SGX) to design a privacy-preserved on-chain matching mechanism that can enable information matching among different users in the encrypted domain, while revealing only order information and maintaining confidentiality. To address the third challenge, we optimize the OPE algorithm to reduce its computing and communication overhead, which further reduces the matching service fee. Besides, we ensure the fairness of POMF by designing a reasonable price mechanism that prices the contact information based on the matching degree between the involved matched users. Finally, we prove the availability of POMF via an comprehensive analysis and demonstrate the practicality of POMF by implementing a prototype and deploying it on a Fabric test network with extensive simulations.

Tailored recommendations on a matching platform

Tailored recommendations on a matching platform

Sharing economy in the food sector: A systematic literature review and future research agenda

In recent years, the sharing economy (SE) has proliferated in various domains, causing a shift from traditional consumption patterns. However, within the hospitality sector, studies have predominantly addressed the accommodation domain while neglecting the ‘food’ sector. This low focus may have been compounded by the lack of a standard definition of SE. Therefore, using a five-faceted standard definition, this study extends the SE lens to the food sector in the hospitality domain. A literature review was conducted to identify relevant studies on food-based SE services aligned with the five defining characteristics of SE: technology-based matching platform, access-oriented, crowdsourced supply, enhanced customer role, and economically substantive. Ninety-six articles were selected and synthesized to identify three food-sector SE services: on-demand food delivery (ODFD), meal-sharing, and food-sharing, along with their associated stakeholders (customers, service providers, complementors, and platform providers). Extant literature on each food-sector SE service was synthesized to obtain their research profile, and extract the antecedent & consequence variables, which led to research gaps. This study makes a novel contribution to the SE domain by consolidating the literature on one of its relatively unaddressed sectors, and setting a future research agenda.

Investigation of FRMPD4 variants associated with X-linked epilepsy

BackgroundThe etiology of unexplained epilepsy in most patients remains unclear. Variants of FRMPD4 are suggested to be associated with neurodevelopmental disorders. Therefore, we screened for disease-causing FRMPD4 variants in patients with epilepsy. MethodsTrios-based whole-exome sequencing was conducted on a cohort of 85 patients with unexplained epilepsy, their parents, and extended family members. Additional cases with FRMPD4 variants were identified from the China Epilepsy Gene Matching Platform V.1.0. The frequency of variants was analyzed, and their subregional effects were predicted using in silico tools. The genotype–phenotype correlation of the newly defined causative genes and protein stability were analyzed using I-Mutant V.3.0 and Grantham scores. ResultsTwo novel missense variants of FRMPD4 were identified in two families. Using the gene matching platform, we identified three additional novel missense variants. These variants presented at low or no allele frequencies in the gnomAD database. All the variants were located outside the three FRMPD4 main domains (WW, PDZ, and FERM). In silico analyses revealed that the variants were damaging and were predicted to be the least stable. All patients eventually became seizure-free. Eight of the 21 patients with FRMPD4 variants had epilepsy, of which five (63%) had missense variants located outside the domains, two had deletions involving exon 2, and one had a frameshift variant located outside the domains. Patients with epilepsy caused by missense variants were often free of intellectual disabilities (4/5), whereas patients with epilepsy caused by truncated variants had intellectual disabilities and structural brain abnormalities (3/3). ConclusionsThe FRMPD4 gene is potentially associated with epilepsy. The genotype–phenotype correlation of FRMPD4 variants indicated that differences in variant types and locations of FRMPD4 may explain their phenotypic variation.

The Tempus TIME trial portal and clinical trial enrollment.

1583 Background: The TIME Trial Program is a research network that screens > 1 million patients daily via EMR integrations and NGS testing through an automated trial matching platform and human review. There are enormous complexities following patients longitudinally as their clinical status changes while accurately matching them with hundreds of studies. The TIME Trials Portal, which stores both clinical trial documents and patient status updates, allows for real-time communication between the TIME team and Clinical Sites. Methods: After both an automated and nursing review, eligible trial patients are uploaded into the TIME Trials Portal. This database tracks the patient’s name, date of birth, biomarkers, clinical status, trial matches, free text comments, and upcoming physician, radiology, and laboratory appointments. To protect Personal Health Information, each Clinical Site can only view their own patients. Patient status is classified as Needs Review, Monitoring, Enrolled, Inactive. Each update is time-stamped and tracked. Enrolled patients have the trial name, consent date, and other pertinent information documented. Results: The data from 2020 to 2022 includes 71 TIME Sites. Over this period, 6,816 patients were added to the TIME Trials Portal for an average of 219 per month. Of these, 1,891 patients were no longer eligible due to worsening performance status, trial closure, death, and other causes. The TIME team provided 8,643 patient updates compared to 7,442 Clinical Site updates. Clinical Sites spent 1,231.7 hours using the TIME Therapies Portal resulting in 623 total consents. Data from 2022 alone, demonstrated that 30 (42.3%) Clinical Sites with EMR integrations were responsible for 4,491 (76.3% of the total) patient updates translating into 288 (82.1%) of all consents. Also in 2022, of the 27 (38%) Clinical Sites that updated their patient’s status > 50 times consented 291 (82.9% of the total) patients. Conclusions: By 2022, the TIME Trials Portal was averaging approximately 400 new patients monthly resulting in a 13.1X increase in patients added from the program’s inception. Clinical Sites updated the patient status over 7,400 times spending over 1,200 hours using the TIME Trials Portal. The highest-performing Clinical Sites, which enrolled approximately 80% of the total patients, included those with EMR integrations and/or > 50 patient updates yearly. The TIME Trials Portal enabled patient tracking and communication to facilitate clinical trial enrollment.[Table: see text]

Learning Equilibria in Matching Markets with Bandit Feedback

Large-scale, two-sided matching platforms must find market outcomes that align with user preferences while simultaneously learning these preferences from data. Classical notions of stability (Gale and Shapley, 1962; Shapley and Shubik, 1971) are, unfortunately, of limited value in the learning setting, given that preferences are inherently uncertain and destabilizing while they are being learned. To bridge this gap, we develop a framework and algorithms for learning stable market outcomes under uncertainty. Our primary setting is matching with transferable utilities, where the platform both matches agents and sets monetary transfers between them. We design an incentive-aware learning objective that captures the distance of a market outcome from equilibrium. Using this objective, we analyze the complexity of learning as a function of preference structure, casting learning as a stochastic multi-armed bandit problem. Algorithmically, we show that “optimism in the face of uncertainty,” the principle underlying many bandit algorithms, applies to a primal-dual formulation of matching with transfers and leads to near-optimal regret bounds. Our work takes a first step toward elucidating when and how stable matchings arise in large, data-driven marketplaces.

Abstract 1067: MatchMiner: An open-source AI precision medicine trial matching platform

Abstract As the number of precision medicine (PM) trials and patient genomic data has grown, it has become challenging for clinicians and trial staff to identify PM trial options for patients. Several trial matching software platforms have been developed to match genomic data from patients with PM trials, but these existing platforms are proprietary and are not easily accessible for adoption by institutions. At Dana-Farber Cancer Institute (DFCI), we have addressed this challenge by developing our own open-source institutional trial matching software, MatchMiner. MatchMiner algorithmically matches patient genomic and clinical data with PM trial eligibility data. Trial eligibility data is manually curated into a human-readable markup language, called clinical trial markup language (CTML), for matching with patient genomic data. MatchMiner has 2 main modes of clinical use: (1) patient-centric, where clinicians search for trial matches for individual patients and (2) trial-centric, where trial staff identify patients that match their trial’s genomic eligibility. We recently described MatchMiner’s usage at DFCI and since our report, we have added 90 additional trial consents facilitated by MatchMiner (>250 trial consents, called MatchMiner consents [MMC]). Here, we describe new characteristics of our MMC including which user mode (patient-centric or trial-centric) was used to match the consent, genomic alterations and cancer types that matched to eligibility criteria, and whether the patient went onto trial. MMCs were mostly identified by patient-centric mode (70%), genomic alterations and cancer types among MMC were diverse (n=55 genes and n=20 cancer types), and 87% of MMC went on trial. Among MMCs, the most common altered genes leading to trial eligibility were ERBB2 and KRAS in breast cancer and lung cancer, which is consistent with the number of therapies targeting ERBB2 and KRAS. MMCs also included patients with rare cancer types, like extraskeletal myxoid chondrosarcoma, as well as rare genomic alterations, such as NTRK fusions. Thus, MatchMiner has been successful at facilitating PM trial matching for a broad range of genomic alterations and cancer types at DFCI. MatchMiner matches patients to trials as soon as their genomic report is available, however, many patients are not yet ready to enroll onto a trial because their cancer is responding to the standard of care or they are in a remission period. To address this problem, we are evaluating the use of artificial intelligence (AI) to identify patients that may be ready for a new treatment option. After trial matches have been generated by MatchMiner, radiology scan text from patients’ tumor scans is run through a natural language processing (NLP) model to identify patients who are more likely to be ready to enroll onto a trial. By using NLP to filter trial matches, we hope to improve MatchMiner’s efficiency of finding trial matches and provide more timely trial options for patients. Citation Format: Harry Klein, Tali Mazor, Matthew Galvin, Jason Hansel, Emily Mallaber, Pavel Trukhanov, Joyce Yu, James Lindsay, Kenneth Kehl, Michael Hassett, Ethan Cerami. MatchMiner: An open-source AI precision medicine trial matching platform [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 1067.