Translational Science Techniques

I am tremendously excited about the real-world impact of our latest publication on #AI #Biomarkers in Nature Medicine: https://lnkd.in/dv-7aS7Y Even in the US barely half of #lungcancer patients are tested for #EGFR mutations, for which targeted therapies readily exist. We have worked for many, many years now to try to overcome this gap with AI for H&E slides to offer patients a fast and cost-effective solution to get the right treatment. The point of this work is not only that we actually built it, but that Gabriele Campanella and Chad Vanderbilt organized a consortium and created the infrastructure for the first real-world, real-time deployment of a fine-tuned pathology foundation model for lung cancer biomarker detection. 𝙋𝙧𝙤𝙨𝙥𝙚𝙘𝙩𝙞𝙫𝙚𝙡𝙮!   𝐌𝐞𝐞𝐭 𝐄𝐀𝐆𝐋𝐄 (EGFR AI Genomic Lung Evaluation): ✅ 𝟎.𝟖𝟗 𝐀𝐔𝐂 in a 𝐩𝐫𝐨𝐬𝐩𝐞𝐜𝐭𝐢𝐯𝐞 silent trial with clinical-grade performance. 🌍 Generalizes 𝐚𝐜𝐫𝐨𝐬𝐬 𝐡𝐨𝐬𝐩𝐢𝐭𝐚𝐥𝐬 𝐚𝐧𝐝 𝐜𝐨𝐧𝐭𝐢𝐧𝐞𝐧𝐭𝐬 with robustness and reproducibility. 🔬 Validated on 𝐢𝐧𝐭𝐞𝐫𝐧𝐚𝐭𝐢𝐨𝐧𝐚𝐥 𝐜𝐨𝐡𝐨𝐫𝐭𝐬, 𝐦𝐮𝐥𝐭𝐢𝐩𝐥𝐞 𝐢𝐧𝐬𝐭𝐢𝐭𝐮𝐭𝐢𝐨𝐧𝐬, 𝐚𝐧𝐝 𝐬𝐜𝐚𝐧𝐧𝐞𝐫𝐬. 🧪 𝟒𝟑% 𝐫𝐞𝐝𝐮𝐜𝐭𝐢𝐨𝐧 𝐢𝐧 𝐫𝐚𝐩𝐢𝐝 𝐦𝐨𝐥𝐞𝐜𝐮𝐥𝐚𝐫 𝐭𝐞𝐬𝐭𝐬, preserving biopsy tissue for full genomic profiling. ⚡ 𝐃𝐞𝐥𝐢𝐯𝐞𝐫𝐬 𝐫𝐞𝐬𝐮𝐥𝐭𝐬 𝐢𝐧 𝐮𝐧𝐝𝐞𝐫 𝟏 𝐡𝐨𝐮𝐫, compared to 2–3 weeks for NGS. 🚀 A foundational step toward regulatory approval and 𝐀𝐈-𝐢𝐧𝐭𝐞𝐠𝐫𝐚𝐭𝐞𝐝 𝐜𝐥𝐢𝐧𝐢𝐜𝐚𝐥 𝐰𝐨𝐫𝐤𝐟𝐥𝐨𝐰𝐬.   We have worked on Computational Biomarkers in Pathology continuously for over a decade starting with AI for predicting SPOP in prostate cancer from H&E in 2015, but seeing everything come to fruition at such a scale in 2025 is very humbling. AI, when done right, can give real, tangible help to cancer patients. 𝑰𝒕 𝒊𝒔 𝒐𝒖𝒓 𝒓𝒆𝒔𝒑𝒐𝒏𝒔𝒊𝒃𝒊𝒍𝒊𝒕𝒚 𝒕𝒐 𝒎𝒂𝒌𝒆 𝒊𝒕 𝒂 𝒓𝒆𝒂𝒍𝒊𝒕𝒚! I am deeply grateful to everyone on this most amazing team: Gabriele Campanella, Neeraj Kumar, Ph.D., Swaraj Nanda, Siddharth Singi, Eugene Fluder, Ricky Kwan, Silke Mühlstedt, Nicole  Pfarr, Peter Schüffler, Ida Häggström, Noora Neittaanmäki, Levent Akyürek, Alina Basnet, Tamara Jamaspishvili, Michel Nasr, Matthew Croken, Fred Hirsch, Arielle Elkrief, Helena Yu, Orly Ardon, Greg Goldgof, Meera Hameed, Jane Houldsworth, Maria E. Arcila, Chad Vanderbilt #AI #ComputationalPathology #Biomarkers #AIinHealthcare #DigitalPathology #PrecisionMedicine #LungCancer #EGFR #NatureMedicine #FoundationModels #EAGLEModel #EAGLE #Oncology

Hot off our recent transformer paper, we're excited to share another AI model for precision medicine! Biological data collected from patients has exploded in recent years, presenting a challenge: how do we decipher that data to understand which patients will benefit most from specific therapies?  We in the Applied Data Science team at AstraZeneca are thrilled to share our paper in Cancer Cell called "AI-Driven Predictive Biomarker Discovery with Contrastive Learning to Improve Clinical Trial Outcomes." Here, we introduce the *Predictive Biomarker Modeling Framework (PBMF)*, a neural network-powered contrastive learning process that: 🔍 Explores vast multimodal datasets to uncover predictive biomarkers in an automated, systematic, and unbiased manner  🧠 Distinguishes predictive biomarkers (which indicate a likely benefit from a specific therapy) from prognostic biomarkers (which indicate general disease outlook)  💡 Distills its outputs into an interpretable decision tree, showing what drives treatment response In our studies, the PBMF:  📊 Surpassed existing methods in finding predictive biomarkers for immunotherapy success across various cancers in clinical trial and real-world data  📈 Discovered a predictive biomarker in an early-stage trial that boosted efficacy by 15% when retrospectively applied to the corresponding phase 3 clinical trial  📈 Discovered predictive biomarkers in single-arm early phase trial data with synthetic control arms, retrospectively improving the efficacy of the corresponding phase 3 trials by at least 10% We believe the PBMF has the potential to improve the way we design clinical trials and match patients to the right therapies. It can integrate with other models like our Clinical Transformer, creating exciting possibilities to someday discover biomarkers of adverse events, dosing strategies, and even to back-translate new drug targets. Read the full paper here: https://lnkd.in/eveAnVRY   Thanks to all the co-authors: Gustavo Arango, Damian Bikiel, Gerald Sun, Elly Kipkogei, Kaitlin Smith, Sebastian Carrasco Pro, Elizabeth Choe #PrecisionMedicine #ClinicalTrials #AIinHealthcare #Biomarkers #Immunotherapy

Ladies and gentlemen - Our Atlas of Human Inflammation today out in Nature Medicine 💫 Welcome on stage: The Cell as a Living Biomarker 🩸 We Built an #Inflammation Atlas of Circulating Immune Cells. By profiling more than 6.5 million peripheral blood mononuclear cells from 1,047 patients across 19 inflammatory and immune-mediated diseases using single-cell RNA sequencing, we generated a unified, high-resolution reference of immune cell states in human disease. Despite decades of research, our understanding of Inflammation has largely remained fragmented, studied one disease at a time and often confined to specific tissues. We set out to take a different approach: to ask whether inflammation leaves a shared, measurable fingerprint at the systems level of circulating immune cells and whether this signal could be leveraged for #Diagnostics and ultimately #PrecisionMedicine. Blood, unlike tissue biopsies, is accessible, repeatable, and scalable, making it an ideal substrate for #liquidbiopsy approaches. The key question was whether it also contains sufficient biological signal to capture disease-driving inflammatory mechanisms❓ Using interpretable #machinelearning approaches, specifically gradient boosted decision trees (GBDT) combined with SHAP explanations, we asked whether single cells, and ultimately patients, could be classified based solely on their blood immune transcriptomes. !!! Disease-associated gene expression signatures enabled accurate classification within the study cohort, reinforcing the concept that immune cells in circulation function as “Living Biomarkers”🩸 We are now working toward a diagnostic prototype that brings these concepts closer to clinical reality. Together with hospital and industry partners, we aim to develop Foundation Models of immune cell plasticity that can operate across diseases and settings, enabling patient stratification, therapy monitoring, and outcome prediction from blood alone. Our long-term vision is a future in which a simple blood draw provides a systems-level readout of immune health, guiding diagnosis and treatment across inflammatory diseases. Immune cells are already doing the sensing. The challenge and opportunity is to learn how to listen. Please find the Nature Medicine Paper here: https://lnkd.in/eYV7pwrN

AI Repurposing of Drugs Unlocks Life-Saving Treatment for Rare Disease >> 🦓An AI tool analyzed 4,000 existing drugs and identified Adalimumab as a potential life-saving treatment for a patient with idiopathic multicentric Castleman’s disease (iMCD) 🦓 iMCD is a rare and deadly cytokine storm disorder (excessive immune response that causes inflammation and organ damage) with limited treatment options, often leading to multi-organ failure 🦓 Researchers at the University of Pennsylvania used machine learning to pinpoint TNF inhibition as a key target, aligning with lab findings that showed elevated TNF signaling in severe iMCD cases 🦓 The patient, once in hospice care, has been in remission for nearly two years after receiving Adalimumab. Researchers estimate hundreds with iMCD worldwide could benefit annually, with potential applications to other rare diseases 🦓 This breakthrough demonstrates the potential power of AI in drug repurposing, a strategy that identifies existing medications with untapped potential for treating different diseases 🦓 The study was led by David Fajgenbaum, MD, an iMCD patient who discovered his own life-saving treatment and co-founded Every Cure to use AI in repurposing drugs for rare diseases 🦓 A new clinical trial is set to begin this year, testing another repurposed drug, a JAK1/2 inhibitor for iMCD, potentially further expanding treatment options 👇Link to article in comments #digitalhealth #rare #ai

The Michael J. Fox Foundation helped validate a biomarker for Parkinson's. Alpha-synuclein seeding amplification assay. Detects abnormal protein in spinal fluid. Before clinical symptoms appear. This is the Parkinson's biomarker we've needed. What it does: Finds misfolded alpha-synuclein - the "Parkinson's protein." In people diagnosed with Parkinson's: 88% positive. In people at high risk without symptoms: detects pathology years early. Like finding cancer Stage 1 instead of Stage 4. Why this matters: Parkinson's damage starts 10-20 years before tremor. By diagnosis, 60-80% of dopamine neurons already dead. Can't bring those back. But if we catch it earlier? When only 20% are damaged? Could slow or stop progression with the right drugs. The technology: Takes tiny amount of abnormal alpha-synuclein from spinal fluid. Amplifies it until detectable. Like PCR for COVID. Same principle. Binary result: pathology present or not. Who this helps now: Clinical trials can recruit earlier patients. Test disease-modifying drugs at stages where they might actually work. Current trials test drugs on people with 80% neuron loss. Of course they fail. Too late. With biomarker: test drugs on people with 20% loss. Actually have neurons left to save. What's coming: Optimizing the test to measure amount of pathology. Not just yes/no. How much. Track if treatments are working. See if protein levels decrease with therapy. Why now: New disease-modifying drugs in trials. Prasinezumab (anti-alpha-synuclein antibody) entering Phase 3. Ambroxol (boosts protein clearance) starting Phase 3 UK trial. Exenatide (diabetes drug) showing promise. All need early diagnosis to work. The diagnostic revolution: This is biology's century for Parkinson's. Move from clinical diagnosis to biological diagnosis. From "you have symptoms" to "you have pathology." Allows treatment before disability. The challenge: Spinal tap required. Not blood test yet. People hesitate. It's invasive. But researchers working on blood-based version. Alpha-synuclein appears in blood at lower levels. Harder to detect. But possible. Timeline: Biomarker validated now. Widespread clinical use: 3-5 years. Blood test version: 5-10 years. 💬 Would you want to know if you had Parkinson's pathology before symptoms? ♻️ Repost if early detection enables early treatment 👉 Follow me (Reza Hosseini Ghomi, MD, MSE) for biomarker breakthroughs that change medicine Citation: Siderowf A et al. α-synuclein seed amplification assay results from the Parkinson’s Progression Markers Initiative (PPMI) study. Lancet Neurol. 2023. Orrú CD, et al. Diagnostic and prognostic value of α-synuclein seed amplification assay kinetic measures in Parkinson’s disease including PPMI data. Lancet Neurol. 2025.

Many approaches aim to identify biomarkers, but this one stands out. A recent article in Nature Aging shows that disease classification can be achieved not just by measuring protein abundance or detecting specific epitopes, but by analyzing protein structural states in plasma. Instead of asking how much of a protein is present, the study asks whether the surface accessibility of its amino acid residues has changed. These conformational signatures were sufficient to accurately distinguish stages of Alzheimer’s disease. This work reinforces an important idea: ⬛ Protein confirmations and modifications reflect important information not captured in protein concentrations. As proteomics technologies mature, looking beyond abundance toward higher-order structure could significantly expand what we can detect — especially in complex neurodegenerative diseases. It’s an exciting direction for biomarker science.

What makes a biomarker…a good biomarker? Each year I teach a course on biomarkers for clinicians. Here are my main takeaways: 🎯 A biomarker is, by FDA and EMA definition, a well-defined, measurable indicator of some biological ground truth. ➡️ the biomarker is only a proxy. This means uncertainty and variation are built into every measurement. ⚖️ Variation comes in three flavours: -      pre-analytical (sample handling) -      analytical (method accuracy) -      biological (within or between individuals). 🫸 🫷 If the normal reference range is wide (large between-subject variation), but the within variation is low, then personal reference ranges are better suited for monitoring changes than population reference ranges. 🟡 An example: A patient may easily increase significantly in their creatinine before exceeding the upper limit of normal. Not noticing an individual’s habitual average but exclusively focusing on whether results are inside or outside the reference range will lead you to miss important deviances until it may be too late. 📏 Accuracy tells us how well a test separates sick from healthy. Precision tells us how closely a biomarker’s risk estimates hits to the mark. A good biomarker needs both. ⚗️ A test that shines in a specialist clinic can flop in primary care. Always validate in the population it is meant for if you want to avoid spectrum bias. 🧮 Remember Thomas Bayes: if disease prevalence is 3%, the negative predictive value of a coin toss is 97% by default. It looks impressive, but is really not. In contrast, positive predictive values struggle in low-prevalence cohorts: Even a test with 95% sensitivity & specificity would only reach a positive predictive value of 37% when pre-test probability is 3%. 💡 Concluding remark: The value of a biomarker lies not in its p-value or AUC, but in whether it helps a clinician make a better decision for a real person. As decisions are dichotomous in nature, cut-offs remain among the top important features of any test.

🚀 New preprint alert! One major reason for why many promising cancer drugs fail in the clinic is on-target, off-tumor toxicity, when a drug attacks healthy tissue alongside the tumor. To solve this, we need human model systems that are more "patient-true" than conventional in vitro models. In our new preprint from the Institute of Human Biology (IHB) at Roche, we establish a predictive platform using human lung explants. By using fresh tissue resections from patients, we can now see how drugs behave in a complex, native environment before they reach clinical trials. Our key findings: * Quantifiable killing: We successfully measured T-cell mediated killing of target cells within the complex architecture of human tumor tissue upon addition of T cell bispecifics. * Predicting safety: The model accurately recapitulated the toxicities seen in past clinical trials for drugs like Solitomab. * Real-world insights: We found that while T-cells activate in both tumor and healthy tissue, actual tumor killing is often hampered by the local environment, a critical distinction for drug efficacy. Using explants to bridge the gap between simplified in vitro models and the human body, we are building a faster, safer path for the next generation of cancer therapies. Huge thanks to the brilliant Elisa D'Arcangelo for her leadership on this project. Grateful to all contributors Manuel Tschan, Tania Jetzer, Melanie Obenloch (pRED), Annika Blank (pRED), Carmen Yong (pRED), Nitya Nair (pRED), and Lauriane Cabon for their creativity and teamwork. Read the full study here: https://lnkd.in/e-_5zZUd #HumanBiology #TranslationalResearch #CancerResearch #PharmaRD #Immunotherapy #LungCancer #DrugSafety #RocheIHB

New research in JACC: Basic to Translational Science highlights why women are twice as likely to develop microvascular dysfunction. Chronic estrogen exposure can disrupt the delicate balance of ceramide and sphingosine-1 phosphate in the microvasculature, shifting protective nitric oxide signaling to damaging hydrogen peroxide. This helps explain why long-term hormone therapy may increase cardiovascular risk in some populations. The study also points to solutions: targeting ceramide pathways and reducing oxidative stress could protect blood vessel function. Practically, lifestyle choices like managing blood pressure, maintaining a healthy weight, staying active, and avoiding smoking remain critical for heart health. AI can accelerate this research by analyzing complex signaling pathways, identifying at risk patients, and predicting who might benefit most from therapies that prevent microvascular dysfunction. Combining human insight with AI driven modeling could personalize cardiovascular care like never before. Read the full study here: https://lnkd.in/gDRFcB3m Follow Zain Khalpey, MD, PhD, FACS for more on Ai & Healthcare. #HeartHealth #Cardiology #WomenInMedicine #MicrovascularHealth #Estrogen #HormoneTherapy #CardiovascularDisease #NO #H2O2 #Ceramide #S1P #EndothelialFunction #PrecisionMedicine #AIinHealthcare #TranslationalResearch #PreventiveHealth #VascularHealth #HealthTech #MedicalResearch #CardiovascularRisk