Advertise on podcast: Beyond the Clinic: Innovation in Healthcare

The first drug being called "AI-designed" is due an FDA decision in early 2027, and the company that originally designed it says that's not the term it would use. In this solo episode, I go stage by stage through drug development, from finding a target to selling the drug, and give AI a verdict at each one: 1) Where AI is useful. Regulatory writing is the clearest win. Novo Nordisk says a clinical study report draft dropped from 10+ weeks to ~10 minutes, and Pfizer drafted statistical analysis plans in minutes instead of days. Narrow tools are working in trials too, from patient screening to the FDA's first qualified AI drug development tool. 2) Where it's still hype. AI-discovered molecules pass Phase 1 at 80 to 90%, but only ~40% pass Phase 2, the same as the industry norm. Virtual cell models still struggle to beat simple baselines, and vendor claims like "2 years faster to market" are self-reported. How to test any AI claim. 3) Three questions: is it about speed or success, whose data is it built on, and who checks the output? Featuring insights from past guests including Sam Windsor (Ignota Labs), Jenny Yang (Outpost Bio) and Hugh O'Brien (Lir Therapeutics). Hear the full conversations with Sam, Jenny and Hugh in EP27, EP31 and EP37. Follow me on LinkedIn for more episodes with the founders, investors and operators building the next generation of life sciences companies: linkedin.com/oscar--vallejo

On 26th August the FDA approved daraxonrasib (Rasonque) in metastatic pancreatic cancer, roughly doubling median survival to 13.2 months against 6.7 on chemotherapy. It's about as clean as a trial result gets, and this solo episode uses it to show you how to read any readout: 1) The 7 questions. Who was in the trial and who the label will cover, what they said they'd measure, what they compared against, how big the effect is, how confident you should be, whether a patient would notice, and who stopped taking the drug. Asked in that order, they work in any disease. 2) RASolute 302 in full. The hazard ratio of 0.40 and what it does and doesn't mean, 53.3% of patients alive at 1 year against 18.7%, a median follow-up of 8.5 months that limits what the survival curve can tell us yet, and a 1.2% discontinuation rate against 11.2% on chemotherapy. 3) Two that didn't hold up. Amgen's sotorasib in lung cancer, where the FDA said no 14 months after a positive Phase 3, and Biogen's aducanumab in Alzheimer's, where a 22% slowing of decline was 0.39 points on an 18 point scale. Every sentence both companies published was accurate. Follow me on LinkedIn for more episodes with the founders, investors and operators building the next generation of life sciences companies: linkedin.com/oscar--vallejo

Gene therapy's delivery vectors still fail on 3 fronts at once: dose, tissue targeting and immune response. Hugh O'Brien, CTO and Co-Founder of Lir Therapeutics, which announced a capsid engineering partnership with VectorBuilder in August, covers what it takes to fix all three together: 1) Open source and a venture-backed drug company can coexist. Building Lir's AI from first principles would take world-class engineers and enough capital to make it unviable as a startup, so open work from academia and industry is what makes it feasible. Hugh's position is that a secret architecture is a weak moat. If something better gets published, Lir moves to it and retrains on its own data. 2) Where AI helps, and where the big models don't. Asking where all the AI-designed drugs are is the wrong test, since co-folding models were in use inside pharma before AlphaFold. For virus engineering and T cell recognition, the data a generalist model would need doesn't exist, so Hugh backs specialist models over general ones. 3) Immunogenicity is an engineering problem. Antibodies bind many sites on a capsid whatever region you edited, and a degraded capsid still gives a T cell an epitope to find. Lir engineers the whole capsid surface rather than one region, and is building an immune map to do it. More on Lir Therapeutics at lirtx.com. Follow me on LinkedIn for more episodes with the founders, investors and operators building the next generation of life sciences companies: linkedin.com/oscar--vallejo

~90% of women tear during childbirth, and a proven fix has sat in the clinical literature for a decade without anyone commercialising it. Eviatar Natan, Co-Founder and CEO of PeriPear, is bringing that fix to market, and covers how he got there: 1) Building in a market investors write off. He was told it was too risky, that it was yucky, that it was too niche. 100M women tear during birth every year, it is not niche 2) Choosing where to validate, not just where to sell. 5 NHS trusts wanted the product, then it disappeared into finance and legal for 12 months. He ran usability testing in Israel instead for under £1,000, sequencing Israel, Singapore and the US to match each stage of the product 3) Choosing investors as carefully as they choose you. He opens every conversation with the least attractive thing about the company, turned down an aggressive angel while nobody was taking a salary, and books intro calls with funds a full round before he needs them so he's rarely pitching anyone cold More on PeriPear at ⁠peri-pear.com⁠. Follow me on LinkedIn for more episodes with the founders, investors and operators building the next generation of life sciences companies: ⁠linkedin.com/in/oscar--vallejo/⁠

For years the assumption has been that if the science is strong enough, everything else will follow. In reality it rarely works that way. Thousands of papers are published every day, and good science gets lost in the noise unless someone advocates for it. No matter how novel the biology, if investors don't get it, if partners can't see the value, and if patients don't understand why it matters, progress stalls. I sat down with Ben McLeod, Founder of Convey Bio, who helps biotech companies articulate what they're building and why it deserves attention. We get into why great science is rarely enough on its own, and where branding and communication change outcomes in an industry built on data and IP. Ben explains why you have to be the advocate for your own science, how to build real trust by being useful long before you ever pitch, and why a personal brand reaches 10 to 20 times more people than a company page. We cover the most common mistakes founders make when they try to explain their work, how to separate an executive's personal brand from the company without touching the IP, and a simple framework for getting started: speak to your audience's pain, pick a cadence you can keep, and pre-empt the objections before you're in the room. A practical conversation about a part of biotech that's often overlooked, and increasingly hard to ignore. Follow Ben on LinkedIn: https://www.linkedin.com/in/benjamin-mcleod/ Follow Oscar on LinkedIn: https://www.linkedin.com/in/oscar--vallejo/

Most founders only get 1 shot in front of an early-stage VC. What decides whether it counts? At pre-seed in health and bio, there's almost no traction and often limited data, so the pitch is almost entirely a test of the founder. The single most common failure is founders who are more obsessed with their solution than with the problem they're solving. Dr. Luka Nićin is a Health & Bio VC Investor at Pace Ventures, an early-stage (pre-seed and seed) health and bio VC fund. He has reviewed ~1,000 startups and invested in only a handful. In this episode we cover: * What Luka evaluates within the first 20 minutes of a pitch * Why 90% of the pre-seed decision comes down to the founding team * How he tests whether a founder is solving a real commercial problem or just advancing interesting biology * What makes an AI platform pitch actually investable * The mistakes health and biotech founders consistently make in front of investors * Where the asymmetric opportunities in health and bio are over the next 3 to 5 years If you are raising, or planning to raise, in health and bio, this is the episode to hear before your next investor meeting. Follow Luka on LinkedIn: ⁠linkedin.com/in/lukanicin/⁠ Follow me for weekly LinkedIn content: linkedin.com/in/oscar--vallejo/

Most biotechs never get the deal their science deserves. Not because the science isn't good enough, because the story isn't clear enough. In a field like liver disease, where 1 in 6 patients responds to approved therapies and the models the industry has relied on for decades break when you leave the mouse, that cost compounds fast. Longer fundraising timelines, weaker partnerships, worse valuations. The science loses before it gets a chance to prove itself. Jack Castle is the Chief Business Officer at Ochre Bio, a company running discovery and validation entirely in human biology, from primary cells and bariatric tissue through to keeping whole donor organs alive for 5 to 10 days to test therapies in a fully human physiological context. In this episode, we cover: - Why framing yourself as a therapeutic company rather than a platform company changes everything with investors - What most biotechs get wrong when translating science into a commercial story - How to use partnerships to validate your platform and bring in non-dilutive capital - Why knowing your red lines before entering a pharma negotiation is the most underused piece of advice in biotech - How biology-first companies are producing targets nobody else can see Jack Castle on LinkedIn: ⁠linkedin.com/in/jackcastle/⁠ Ochre Bio: ⁠ochre-bio.com⁠ For more Beyond the Clinic content: ⁠linkedin.com/in/oscar--vallejo/⁠

Hormone testing is one of the most overlooked bottlenecks in modern healthcare. For IVF patients, it means taking time off work, every day, for weeks, just to get same-day results that clinics can barely keep up with. Mingoo Kim, Founder and CEO of Kompass Diagnostics, is fixing that. His company has cut a 2-4 hour hormone test down to 10 minutes, changing what's clinically possible inside a single appointment. In this episode, we cover: x Why this problem has been accepted as normal for decades despite the technology to fix it existing x The impossible position clinics are caught in between in-house labs and outsourced testing x Why navigating regulatory, reimbursement, and distribution complexity is the real moat, not the science x What founder-problem fit actually looks like when someone chooses a hard, unglamorous problem deliberately x Advice for founders and investors entering the diagnostics space Mingoo studied at Stanford before choosing to tackle one of healthcare's most underappreciated infrastructure problems. This conversation gets into why that choice matters, and what it takes to actually see it through. Mingoo Kim on LinkedIn: https://www.linkedin.com/in/mingookim/ Kompass Diagnostics: https://kompassdx.com/ For more Beyond the Clinic content: https://www.linkedin.com/in/oscar--vallejo/

Dr. Jenny Yang, CEO and Co-Founder of Outpost Bio, shares how AI is unlocking the microbiome’s potential to transform drug development. From overcoming the biggest barriers in the field to translating insights into real-world impact, Jenny explains the science, strategy, and vision behind Outpost Bio, plus her hot take on where the industry is heading. Check out Outpost Bio: ⁠outpost.bio⁠ Follow Jenny on LinkedIn: https://www.linkedin.com/in/jennyyang259/ Follow me on LinkedIn for more Beyond the Clinic content: https://www.linkedin.com/in/oscar--vallejo/

In this episode, David Ogman shares the deeply personal journey that reshaped his life; navigating his son’s diagnosis with an ultra-rare genetic disease (TECPR2) and confronting the realities families face when science has few answers. We explore how that experience led David to co-found the Jordan Avi Ogman Foundation, his efforts to drive rare disease research forward, and what his family’s story reveals about the gaps, and opportunities, in how we approach rare diseases today. To learn more or support the work of the Jordan Avi Ogman Foundation, visit: SavingJordan.org

Imagine a world where people with diabetes are having their doors taken off their hinges because they slipped into a coma after missing a dose of insulin, or constantly having to do finger-prick glucose tests at traffic lights mid-commute to the office. With today's glucose monitoring technology it's almost hard to believe, but this is the world that Melissa Holloway experienced as a person with diabetes throughout her time in university and her early career. Now, she's committed to helping people like her to better understand their diabetes and how to get the most out of their continuous glucose monitors (CGMs). In this episode, I sit down with Melissa to learn about her diabetes journey, how that led to founding SmartStart Health, how to avoid £200+ K mistakes when building a tech startup... and much more! Follow Melissa on LinkedIn: https://www.linkedin.com/in/mpfholloway/ Check out SmartStart Health: https://smartstart.health/ Visit Beyond the Clinic on LinkedIn for more content: https://www.linkedin.com/company/beyondtheclinicpodcast/?viewAsMember=true Oscar's LinkedIn: https://www.linkedin.com/in/oscar--vallejo/

In this episode, we sit down with Felix Lorenzen to talk about his latest venture, Autonome—a platform designed to put your health on autopilot. Autonome takes a proactive approach, managing your health behind the scenes so you can stay ahead, not just react. The future of healthcare is clearly preventive, personalised, and always in your pocket - and Felix is building exactly that. Last year, Felix made the move from Europe to San Francisco, drawn by the culture and momentum he needed to bring Autonome to life. Since then, he’s launched the product in beta and is working closely with early users to refine the experience ahead of its full release.

We’re joined by Sam Windsor, CEO and Co-Founder of Ignota Labs - a company tackling one of the biggest hidden inefficiencies in drug development: the lack of insight into why drugs fail. Ignota Labs rescues promising but failing drugs, bringing new life to abandoned projects and new hope to patients. The company works with R&D teams to revisit failed assets using machine learning and biological context to suggest new hypotheses for why those drugs failed and how they could be rescued or repositioned, reviving the assets and most importantly getting them into the hands of patients. Earlier this year, Ignota Labs raised a $6.9 million seed round, and in this episode Sam shares what they’re building, how they collaborate with partners, and why now is the time for a company like this to exist. Follow Sam on LinkedIn to keep up to date with Ignota Labs’ journey: ⁠www.linkedin.com/in/sam-windsor-002993b2/⁠ For extra content, follow us on LinkedIn: Beyond the Clinic: ⁠www.linkedin.com/company/beyondtheclinicpodcast/⁠ Oscar: ⁠www.linkedin.com/in/oscar--vallejo/⁠ Liz: ⁠www.linkedin.com/in/elizabethannedoyle/⁠

70% of chronic conditions are preventable through lifestyle. Want to prevent these? Well, you need data. Wearable data alone won’t tell you the full picture. It’s only when you incorporate this with clinical results from labs or scans that your data paints a real picture. We sit down with Boris Lipchin, Head of Product at Junction Health who are on a mission to do just that. Activate our otherwise siloed health data and enable an industry to do the same. Listen how Junction aims to power the future of personalised care after recently raising $18m Series A round. Check out Boris on LinkedIn: https://www.linkedin.com/in/blipchin/

David is developing a treatment for CMT1A, a rare disease affecting 1.6 million people worldwide, including his 4-year-old son, Ari. In this episode, we dive deep into Ari's condition, David's motivation for starting Shark Tooth Biotech and his plan for finding a cure for CMT1A. It's not one to miss... Follow David on LinkedIn: https://www.linkedin.com/in/appledavid/