Ever wonder how you teach an LLM to think like an oncologist? I wrote about the exciting work we're doing at @flatironhealth to extract clinical data from the EHR, thoroughly evaluate model performance, and advance research to benefit patients with cancer. Check it out!
In a new article Aaron Cohen deep dives into ways #AI is streamlining the clinical data journey how Flatiron is leveraging LLMs to make sense of the complex clinical information buried in medical charts.
Read the full piece now: spkl.io/6017ft8UU
Excited for our LLM work to come out in @AI_PrecisionOnc ! We used high quality labeled data to fine-tune LLMs and improve our ability to extract PD-L1 testing details across 10 different cancers! Reach out or click the link below to learn more about @flatironhealth work in #AI
New @flatironhealth research successfully utilized LLMs to extract PD-L1 details across 10 cancer types ⬇️
The methods described will enable researchers at Flatiron and beyond with the information needed to leverage LLMs for #RWD curation.
Learn more: spkl.io/6016fmSq6
After learning about @DavidFajgenbaum's unique journey with Castleman disease, Flatiron engineers used ML to better leverage Flatiron data, unlocking insights for a rare patient population.
Read our #LearnTeachGrow article for a behind the scenes look:
linkedin.com/pulse/using-mac…
I loved having the opportunity to learn more about our impressive grantees and the impactful research they are pursuing to help improve equity in cancer care.
Excited to share our paper @NatureComms on evaluating #AI when missing ground truth labels! nature.com/articles/s41467-0… Understanding how AI works in the wild is hard; we often lack ground truth labels. We introduce SUDO, a framework for evaluating AI predictions when lacking them.
This allows us to better understand the accuracy of these pseudolabels and the SUDO framework has a number of applications: we can better understand and quantify the reliability of an AI prediction in the real world and ensure model fairness in the absence of ground-truth labels.
As AI is increasingly used in healthcare, the hope is that this results in more trustworthy and ethical deployment of AI systems. @DaniKiyasseh@flatironhealth
I’m incredibly excited about @flatironhealth's continued partnership with the @AmericanCancer. This year, I had the opportunity to sit down with the three grantees and have in-depth discussions about the potential impact of their research, and how RWD helped enable it.
NEW TODAY @flatironhealth & @AmericanCancer are proud to announce the 4th annual Real-World Data Impact Awards.
Read or watch to see the grantees & learn the impact of this year's clinically-relevant research on health inequities among patients w/ cancer. page.flatiron.com/2023-real-…
I encourage all to tune into the roundtable discussion linked above, as it shows how the combination of talented investigators and RWD can accelerate cancer research and improve treatment and outcomes for patients with cancer. @ChrisManzMD
So appreciative of @MoffittNews
for having me at their annual Business of Biotech conference to speak about the opportunities and challenges of #artificialintelligence in cancer care and research. And thanks to my fellow panelists for a great discussion. @flatironhealth
During the first breakout session of #MoffittBoB, our esteemed panelists are unveiling new perspectives on AI in care optimization with a focus on tackling risks and challenges for responsible adoption.
So great to be at the @ASH_hematology annual meeting where we’re holding our annual @CureCastleman workshop and our team is giving 6 presentations (below). cdcn.org/physicians-research…
Let us know if you’re in #sandiego and want to meet up!
Identification of Autoantibodies Associated with Connective Tissue Disorders in Idiopathic Multicentric Castleman Disease
ash.confex.com/ash/2023/webp…
Clinical Characteristics, Treatment Trends, and Outcomes of Patients with HHV-8-Negative/Idiopathic Multicentric Castleman Disease Treated with Siltuximab in a Machine Learning-Selected Real-World Cohort ash.confex.com/ash/2023/webp…
Clinical Characteristics and Treatment Outcomes of Idiopathic Multicentric Castleman Disease: Data from a Pediatric Cohort
ash.confex.com/ash/2023/webp…
Exploratory Analyses of Lymph Node Tissue, Serum, and Peripheral Blood Mononuclear Cells Do Not Reveal Pathogen Signatures in Idiopathic Multicentric Castleman Disease
ash.confex.com/ash/2023/webp…
A Translational Approach to Identifying and Targeting TNF Signaling in Idiopathic Multicentric Castleman Disease
ash.confex.com/ash/2023/webp…
The Role of CABIN1 in the Pathogenesis of Idiopathic Multicentric Castleman Disease
ash.confex.com/ash/2023/webp…#castlemanwarrior#castleman
Excited to present at #ASH23: Clinical Characteristics, Treatment Trends, and Outcomes of Patients with HHV-8-Negative/Idiopathic Multicentric Castleman Disease Treated with Siltuximab in a #machinelearning -Selected Real-World Cohort.
@flatironhealth@DavidFajgenbaum
Join me tomorrow (2:45-4:15pm) in the Marriott Marquis San Diego Marina, Marriott Grand Ballroom 2-4, for the oral presentation, where I'll be discussing how patients with a rare lymphoproliferative disorder are cared for in real world practice and how they do on Siltuximab.
Our presence at #ASCO23 showcases how #RWD allows us to enhance representativeness of trials, generate insights into diverse patient populations, & unlock opportunities to improve #healthequity.
Learn more about our team’s presence at this year’s @ASCO ➡️events.flatiron.com/asco-202…
Now out in Applied Sciences, our team @flatironhealth built a high performing NLP algorithm to improve completeness of ECOG performance status across 21 different cancer types. Proud of the team, the work, and the impact it has. mdpi.com/2076-3417/13/10/620…
#ML can extract clinically relevant information from EHRs at scale, but evaluating its quality has focused on single variables.
This @flatironhealth study aims to evaluating ML's usefulness for research & RWE generation at scale: flatiron.com/resources/repli…@Cancers_MDPI