Neryssa

You will always be my baby. I miss you, my Max.🤍🐶

GU Research Digest for the week of May 16 - 22: a selection of publications about GU cancers in high-impact journals. Search the PMIDs on pubmed.ncbi.nlm.nih.gov to read!

Excited to be presenting at #AUA26 and catching up with colleagues and friends! @michaelabern @DukeCancer @Hoomandjaladat @siadaneshmand @AmerUrological

Wait, wait, wait...You're telling me there's another study showing prophylactic abx don't help with infections? And might increase them? This is my shocked face...😐 Antibiotic prophylaxis and risk of orbital cellulitis following orbital floor fracture tandfonline.com/doi/10.1080/01…

If intracellular NAD+ can increase ~53% without any rise in circulating NAD+ then plasma NAD+ may not reflect cellular NAD+ biology Yet most human studies still anchor on circulating NAD+ We may be measuring the wrong compartment #NAD #Metabolism

If intracellular NAD+ availability—and metabolic flux—are the key variables… then delivery becomes a pharmacologic question, not just a dosing one. The goal may not be raising NAD+ everywhere— but enabling its utilization where it matters most. This is the framework behind our work on LNAD+. #NAD #CellBiology

A shift in how we think about NAD+ biology: • From exposure → utilization • From plasma → intracellular pools • From levels → metabolic flux Much of the field still anchors on circulating NAD These distinctions may matter more than total NAD+ levels alone. #NAD #Metabolism

@ReznikMD @ReznikMD This 2017 piece on patient empathy still resonates strongly, Dr. Reznik. Empathy drives better adherence and outcomes, it's a real edge in value-based care. What’s one practical change that’s given you the best ROI on patient trust?

Putting Yourself in Your Patient's Place drreznik.com/download_file/…

Access to obesity care should not be a matter of luck. Tracy Zvenyach, Vice President of Advocacy and Research at the OAC, highlights how policy gaps are driving inconsistent coverage and limiting access to care. Obesity is a chronic disease that deserves comprehensive, evidence-based treatment. Learn more now: bit.ly/4tVbIxl

Our residents hard at work in a recent wet lab—sharpening hand‑eye coordination for advanced procedures including cataract surgery, corneal transplants, and retinal surgery. @NUFeinbergMed

Don’t walk, run to see @APMiracle’s production of Dear Evan Hansen. Simply amazing! And you’ll get to see my fabulous nephew, Paul Tuaty.

Dr. Alcina Lidder moderating the inaugural Resident Research Forum at the American Glaucoma Society meeting is exactly the kind of initiative that drives innovation forward in ophthalmology. Lowering barriers for residents and med students to present and discuss meaningful glaucoma research = stronger pipeline of ideas and future breakthroughs in one of the most important therapeutic areas in eye care. When the field invests in its young researchers, patients (and long-term returns) win. Kudos to Northwestern Medicine and the AGS for making this happen. #Ophthalmology #Glaucoma #MedTech #HealthcareInnovation #LifeSciences

Claude Mythos Preview scores 94.5% on GPQA Diamond. That benchmark tests PhD-level questions in biology, chemistry, and physics — questions where domain experts outside their subfield score below 35%. This is graduate-level scientific reasoning across disciplines, exceeding most specialists operating outside their narrow lane. What the benchmark card means for clinical medicine. 97.6% on USAMO — competition mathematics requiring multi-step logical chains where a single error cascades into failure. Differential diagnosis works the same way. The reasoning chain that rules out Wilson's disease before anchoring on hepatitis requires structured, sequential logic across multiple variables. Claude Opus 4.6 scored 42.3%. Mythos scores 97.6%. That is not incremental. That is the difference between a model that occasionally reasons well and one that almost never drops a step. 80% on GraphWalks BFS 256K-1M — long-context reasoning over complex interconnected data up to one million tokens. GPT-5.4 scored 21.4%. This maps to longitudinal patient care. A complex patient with 15 years of records across four systems, twelve specialists, and 40 medications generates exactly this kind of dense interconnected data. The model that traverses a million-token graph without losing the thread can hold an entire patient history in context and find the interaction no single clinician could. 93.2% on CharXiv Reasoning — interpreting scientific charts and data visualizations from research papers. A model at 93.2% reads every figure in every paper and catches the methodological signal in the visualization that contradicts the abstract. 56.8% on Humanity's Last Exam without tools — designed to contain the hardest questions across all human knowledge. No model was expected to break 10% at launch. Mythos solves more than half. And the number that should reframe every health system's AI timeline: Mythos accelerated internal AI research by up to 400x on tasks equivalent to 40 hours of expert work. Not 4x. 400x. Every benchmark maps to a clinical capability. Differential diagnosis. Longitudinal record synthesis. Drug interaction analysis. Literature interpretation. Multi-step reasoning under uncertainty. The model scoring 97.6% on complex reasoning and 80% on million-token graph traversal is not a future capability. It exists now. The 17-year adoption lag assumes capability advances generationally. This model went from 42.3% to 97.6% on complex reasoning in one generation. The lag isn't outdated. It's dangerous.

"From Knowledge to Power" — Read and discuss on ManyStories via @manystories_app #ManyStories manystories.com/story/from-kno…

The 2025 @ClevelandClinic 17th Annual International Symposium on #SBRT & #SRS will be held February 21-23, bringing together world leaders in brain, spine and body-targeted stereotactic radiation modalities and techniques. Register here: ccfcme.org/SBRT

Significant progress for patients with sarcoma at #ASCO25 with positive results from 3 phase III trials. Pimicotinib for TGCT, anlotinib for STS and catequentinib for LMS. All are TKIs. @CleClinicMD #CleClinicCancer

Classic for the ages

@jimsabogal This 442% surge in vishing attacks is alarming. Patient data has become a prime target. We’re now pushing portfolio companies toward phishing-resistant MFA and behavioral monitoring. What defenses are proving most effective right now, especially for smaller providers?

DYK voice phishing attacks rose 442% between the first and second half of 2024? Discover more important findings for #Healthcare organizations in the new @CrowdStrike Global Threat Report. dy.si/f2ZP6

@RoupenMD @RoupenMD Excellent template, Roupen! Smart, proactive AI policies like this are exactly what healthcare innovation needs to train the next generation of physicians responsibly, especially in oncology. Well done.

AI in medical education is here! I wrote an article on creating AI policy for medical education programs. It's important to guide learners in the right direction for using AI rather than reacting to it late! meducationai.com/blog/creating-…

@Jmortonmd @Jmortonmd Spot on, Dr. Morton. This is high-ROI medicine. Nutrient-dense meals = faster recovery & lower costs. Food as medicine should be the standard.

For Healthy Patients, we need Healthy Hospitals including food for patients and providers!

CCNE amplified tumors need specific agents to address the underlying biology. CDK2 selective blockade has not been achieved by any investigational drugs IMO. We designed an absolutely selective molecular glue degrader to achieve this target product profile. I’m thankful to @GileadSciences for believing in this program early on and now for taking KT-200, @KymeraTx discovered DC, into development!