BioKnow: The Fifth Biggest Killer Should Not Be Found by Accident - Authored by Dejan Terziovski
Dejan Terziovski
By 2050, chronic kidney disease is projected to become the fifth leading cause of death in the world (Lancet, 2025).
You can detect it with a routine blood and urine test, years before it turns fatal. Yet, too often patients are caught late, including in the Balkans.
We are not missing the medicine. We are missing the early diagnosis. The patient drifts between a cardiologist, an endocrinologist and a nephrologist, while the kidney quietly fails in the space between them.
An earlier screening step at the GP (serum creatinine and albuminuria) would change that trajectory for a large share of them.
In Croatia’s first national study, 17% of adults had chronic kidney disease and fewer than one in ten were aware of it. Across the Balkans the diagnostic tests are reimbursed and the treatment is funded, yet we lack a national program to drive structured screening of high-risk patients. As a result, the diagnosis keeps arriving too late and specialist nephrologists have limited options.
A region that detects its fifth biggest killer by accident is not facing a science problem.
It is facing a design problem!
Follow Dejan Terziovski on LinkedIn
Image from Dejan Terziovski
Win one of 5 tickets to the Galien Forum!
October 29, 2026 — New York City 🗽🇺🇸
We’re giving away 5 passes to the Galien Forum, and if you have ever wanted to be in a room with the people leading some of the most influential companies in pharma, this is your opportunity.
Taking place on October 29, 2026, in New York City, the Galien Forum brings together global chairs, senior executives, scientists, innovators and decision-makers for meaningful conversations about the ideas shaping the future of medicine and healthcare.Attendees will have the opportunity to hear from and connect with leaders from Johnson & Johnson, Pfizer, Roche, Regeneron, Merck and other globally recognized organizations.
These are the people operating at the cutting edge of the industry developing new treatments, advancing scientific research, building breakthrough technologies and making the decisions that influence the direction of the pharmaceutical industry at large.
For more than 20 years, Prix Galien USA has recognized the innovations transforming patient care. The Galien Forum creates an opportunity to go beyond the awards and hear directly from the leaders, researchers and organizations behind that progress.
You can see the post HERE!
Follow The Galien Foundation on LinkedIn to stay updated on announcements, speakers, and innovations and upcoming event.
How is regulation lagging behind AI-assisted drug development?
AI-assisted drug development is moving from promise to practice, but regulatory and IP rules are not keeping pace. Several reviews report that AI is already embedded across discovery, clinical development, manufacturing, and regulatory submissions [1], while no fully AI-discovered drug has yet received US marketing approval [2].
Approval Pathways
The regulatory tension starts with a mismatch: AI can accelerate candidate generation and trial preparation, but agencies still evaluate the resulting medicines through traditional safety and efficacy standards [3]. FDA has received hundreds of drug submissions involving AI, and over 100 applications in 2021 alone reportedly incorporated AI elements, showing that the issue is already operational rather than hypothetical [1, 4].
Regulators are converging on risk-based oversight, with emphasis on transparency, validation, bias control, and data integrity [4, 5]. But the standards remain incomplete. One review notes that no dedicated regulatory recommendation for AI use in drug development had yet been published, while another warns that flexible case-by-case review can encourage innovation but also leave companies uncertain about what evidence regulators will expect [6].
Black-box models complicate review because explainability and reproducibility are harder to demonstrate [7].
Cross-border divergence between FDA and EMA approaches can raise cost and delay global launches [6].
Startups face higher barriers when requirements are uncertain or duplicative across jurisdictions [6].
Intellectual Property
Patent law is the second pressure point. AI systems can generate large numbers of plausible drug candidates, but novelty checks and freedom-to-operate analysis remain difficult at scale [8]. That matters commercially because patents remain central to biotech financing and competitive strategy.
US and UK courts have now reinforced that AI cannot be named as an inventor; in the US, patentability requires a human-made significant contribution towards an invention’s conception [9]. Reviews therefore recommend meticulous documentation of human input across AI-assisted workflows, from model design to candidate selection [9, 10].
Inventorship disputes create uncertainty over ownership of AI-assisted molecules [11].
Training-data copyright adds another layer because model inputs can affect downstream patent strategy [9].
Algorithm patents are harder to secure than molecule patents, which can weaken incentives for platform developers [9].
AI-assisted drug development is advancing quickly, but the bottleneck is shifting from molecule generation to governance. In pharma and biotech, the winners will likely be firms that can show not just faster science, but cleaner evidence, clearer documentation, and defensible IP.
Superluminal Medicines bags $60M during oversubscribed series B funding round
Superluminal Medicines has completed an oversubscribed $60 million Series B funding round to support its growth from a discovery platform to a clinical-stage biotechnology company. Led by BVF Partners, L.P., the syndicate included new investors Deep Track Capital and Perceptive Advisors alongside returning backers such as RA Capital Management, Insight Partners, NVIDIA, Eli Lilly and Company, and Catalio Capital Management.
The proceeds will fund the clinical advancement of Superluminal's lead asset, a selective and biased melanocortin-4 receptor (MC4R) agonist targeting rare genetic forms of obesity and hypothalamic obesity, which is slated to enter a Phase 1 clinical trial by late 2026. By integrating Agentic Cryo-EM capabilities with proprietary machine learning and co-folding models, the company's platform designs small molecules that selectively engage target GPCR signaling pathways to maximize efficacy while minimizing off-target side effects. Beyond its lead program, Superluminal will use the capital to expand its pipeline across cardiometabolic and endocrine indications while advancing existing strategic partnerships.
Recent FDA Complete Response Letters Expose the Risks of Interdependent Filings
Above are the three most recent NDA and BLA complete response letters. The concurrent Complete Response Letters issued on July 9 to Elevar Therapeutics (rivoceranib) and Jiangsu Hengrui (SHR-1210) highlight the vulnerabilities of cross-sponsor combination therapies. Given their co-development for hepatocellular carcinoma (already approved as a combination therapy in China), interdependent NDAs and BLAs dictate that a regulatory issue in one filing may stall the partnered asset.
Data taken from FDA complete response letter database.
Pelacarsen Phase 3 Miss Raises Questions About Lp(a) as a Cardiovascular Target
Novartis and Ionis Pharmaceuticals suffered a major late-stage cardiovascular setback after pelacarsen, their lipoprotein(a) (Lp(a)) targeting therapy, missed its primary endpoint during Phase 3 trials. The negative trial readout sent both stocks tumbling in after-hours trading, with Novartis dropping 3.7% and Ionis plunging 8.6%.
The pivotal HORIZON trial evaluated pelacarsen in more than 8,000 patients with established cardiovascular disease who were receiving optimal guideline-directed care. Despite successfully achieving substantial reductions in plasma levels of Lp(a), an inherited cardiovascular risk factor largely unimpacted by diet or exercise, the drug failed to demonstrate a statistically significant reduction in major adverse cardiovascular events such as heart attack, stroke, and revascularization. The surprising disconnect between biomarker reduction and clinical outcomes raises broader questions across the pharmaceutical industry regarding whether lowering Lp(a) can effectively reduce overall cardiovascular risk.
