Past Forum 2026

Senator Shelley Moore Capito joined David Brown MD for a conversation about biomedical research, healthcare access and the long-term strength of the US scientific enterprise, following an introduction by Michael Benson DPhil.

Capito described research funding as an area where bipartisan cooperation is still possible, but emphasized that institutions need more than dollars. They need predictability. Uncertainty makes it harder to plan laboratories, support postdoctoral researchers and give young scientists confidence that they can build careers in the United States.

Alzheimer’s disease made the discussion personal. With both of her parents having lived with the disease, Capito discussed efforts to expand access to blood-based biomarker testing as earlier diagnosis and treatment become possible. The conversation also touched on pediatric cancer research, rural healthcare, telehealth and the financial pressures facing academic research institutions.

Congressman Richard Neal joined Suffolk CEO John Fish for a conversation that connected healthcare policy with the economic role hospitals, universities and research institutions play in their communities.

Neal pointed to Massachusetts as an example of what can happen when government, healthcare and academic institutions work together around insurance coverage and access. He also discussed the growing cost of long-term care and the pressure it places on families, hospitals and Medicaid.

Artificial intelligence brought the conversation into a different kind of policy question. Neal was not arguing against AI or calling for innovation to slow. He focused instead on who shapes technologies with this much influence and cautioned against allowing AI to become a substitute for human judgment and critical thinking.

Mass General Brigham President and CEO Anne Klibanski MD and Cleveland Clinic CEO Tom Mihaljevic MD joined a conversation introduced by Scott Sperling that explored how major academic health systems can keep evolving under persistent financial, workforce, and research pressures.

Mihaljevic opened with what sounded like a current diagnosis of American healthcare: hospitals under pressure, research funding constrained and reimbursement declining. Then he revealed that the comments were more than two decades old. His point was not that today’s challenges are small, but that academic medicine has always operated under pressure and that nostalgia for a simpler healthcare system rarely survives close examination.

Klibanski emphasized research as central to Mass General Brigham’s role, pointing to work in areas including genome editing and AI-enabled clinical trial matching. Both leaders also returned to the question of scale. Digital tools and artificial intelligence may finally allow academic medical centers to extend expertise beyond a single geography. Still, Mihaljevic offered an important warning: technology layered onto a fragmented organization can produce a more expensive fragmented organization. The care model has to change with it.

Martin Mellish moderated a discussion with Adel AlMarshad MD, Haipeng Xiao MD, PhD and Bjorn Zoega PhD on how health systems in different parts of the world are trying to scale innovation while improving access, workforce resilience and patient outcomes.

Zoega described leadership as removing barriers for clinicians closest to the patient. He used robotic surgery to illustrate the point, including teams that pursued a fully robotic heart transplant and robotic brain tumor procedure. Xiao described a different-scale challenge in China, where AI, robotics, clinical partnerships, and workforce training are being used to extend specialist capabilities into rural communities and reduce unnecessary referrals to major centers.

AlMarshad brought the conversation back to international collaboration, arguing that physicians who train across health systems can become important bridges for transferring clinical standards, expertise and new models of care between countries. Across very different systems, the common theme was that transformation depends less on importing a single technology than on building organizations that can adopt innovation quickly and sustainably.

Robert Lustig MD moderated a conversation with Bahia Jaffer, Joan LaRovere MD and Carter Williams about a question healthcare has struggled with for decades: how to intervene before metabolic disease becomes established rather than waiting to treat its consequences.

Lustig argued that GLP-1 therapies have transformed obesity and metabolic care but do not eliminate the need to address the upstream food and environmental conditions that contribute to disease. Jaffer brought that idea into the consumer market, describing efforts to reformulate products around the Metabolic Matrix framework, including a reduced-sugar ice cream tested with patients with diabetes.

Williams focused on the economics of prevention and the importance of making the healthier option financially sustainable for companies, payers and investors. LaRovere emphasized that the solution cannot sit with healthcare alone. Pediatrics, food companies, agriculture, payers and health systems all influence the environments in which metabolic risk develops. The discussion made prevention feel less like a slogan and more like a design challenge spanning science, business, and everyday life.

Medtech is in a period of extraordinary scientific opportunity, but getting a new technology from idea to patient is becoming harder. Moderated by Brian Johnson, the conversation brought together John Bloom MD, Elazer Edelman MD, PhD, Michelle Fox RN, Paul LaViolette and Nadim Yared to look at the state of medical device innovation from the perspectives of clinicians, company builders, investors and strategic buyers.

The science and technology are moving quickly. Bloom pointed to AI as an opportunity to improve outcomes and create more value from healthcare technology. At the same time, Edelman described a field where the tools and ideas have never been stronger, but investors and customers now demand much more evidence before committing. LaViolette and Yared focused on the financing challenge, particularly the shrinking pool of capital available to early-stage medtech companies that often feed the industry’s future pipeline.

Fox brought the conversation closer to how technologies actually reach patients, highlighting reimbursement, market access, point-of-care screening and a broader shift from reactive treatment toward earlier detection. Drawing on her experience evaluating acquisitions, she also stressed that strategic buyers are looking beyond the product itself to a company’s culture, people, and ability to adapt as it grows.

One message kept resurfacing: breakthrough technology is not enough on its own. It has to solve a real clinical problem, generate convincing evidence, find a viable path to reimbursement, and fit into healthcare without creating more work for already stretched clinicians. Bloom’s experience at Podimetrics captured that last point particularly well. The company’s biggest shift came when it stopped asking physicians to do more and redesigned its model to take work off their hands.

Siemens Healthineers CEO Bernd Montag PhD joined Giles Boland MD and Milton Hsu MD for a conversation about artificial intelligence, medical imaging, and how technology changes the work clinicians actually do.

Montag recalled reading about the arrival of the pocket calculator and the fear that it would diminish professional expertise. Instead, technology helped clarify that a job’s value was never the mechanical task alone. A CFO is not a calculator, and a radiologist is not simply a pattern recognizer.

He sees the same issue playing out with AI in healthcare. Hospitals may have sophisticated tools available, but adoption stalls when organizations introduce technology without helping clinicians redefine how their work changes around it. Montag also challenged the idea that imaging should be viewed primarily as a radiology function, connecting it instead to earlier detection, treatment selection and therapy across cancer, cardiovascular disease, stroke and Alzheimer’s disease.

Scale came with a condition. Larger healthcare systems should deliver better, less expensive care. If consolidation produces only greater negotiating leverage, Montag argued, it has missed the point.

Toshio Takiguchi joined Jim Brink MD to trace Canon’s long history in healthcare and discuss how medical imaging, AI and portable technology are changing where diagnostic care can be delivered.

Takiguchi connected Canon’s current medical business to a history that began with an X-ray camera developed in 1940 and later expanded through ophthalmic imaging, digital radiography and the acquisition of Toshiba’s medical business. What interested him most, however, was not the product timeline but what happens between invention and broad use.

Canon’s approach emphasizes validating clinical outcomes before global commercialization. Takiguchi illustrated that philosophy through portable X-ray systems paired with AI for remote settings, along with a mobile lung cancer screening program in rural West Virginia.

AI also has to prove value beyond accuracy. Takiguchi pointed to staffing and operational benefits, including reductions in CT overtime, as outcomes that matter to hospitals facing workforce shortages. His larger point was that building an AI algorithm is becoming easier. Integrating it into healthcare in a way that clinicians can actually use is much harder.

Philips CEO Roy Jakobs joined Michael Gee MD, PhD and Milton Hsu MD for a conversation about connected care, clinical workflow and the role artificial intelligence can play in reducing friction across the patient journey.

Jakobs started with a simple observation: every handoff in a patient’s care creates another opportunity for something to be lost. Hospitals facing rising demand and workforce shortages are therefore looking for simpler, more integrated workflows that connect monitoring, imaging, diagnosis and intervention rather than adding another disconnected point solution.

He described the goal as people-led and AI-powered. Artificial intelligence should absorb repetitive data gathering and routine work so clinicians can spend more time on complex decisions and direct patient care. That human connection remains important even as automation expands.

Jakobs expects areas such as medical imaging to become increasingly autonomous, but with clinicians retaining responsibility for consequential decisions. The challenge is not whether AI can perform more tasks. It is making sure the technology is embedded in the workflow well enough to make care easier, not more complicated.

Astellas Chief Research and Development Officer Tadaaki Taniguchi MD, PhD joined Rishi Singh MD to discuss how drug development is changing across ophthalmology, oncology, rare disease and other areas where new therapeutic modalities are moving quickly.

Taniguchi described R&D as a sequence of connected choices: understand the biology, select the modality that best addresses it and focus on an unmet need meaningful enough to justify the work. Astellas’ expansion into ophthalmology illustrates that strategy. Its acquisition of Iveric Bio added more than a near-market product. It brought scientific, clinical and commercial capabilities that complemented the company’s existing work in areas such as cell therapy.

Artificial intelligence is also helping accelerate discovery. Taniguchi described computational tools and high-performance computing reducing timelines in parts of R&D while noting that complex modalities such as protein degraders, antibody-drug conjugates and bispecifics still require substantial human judgment.

That complexity makes partnership fundamental. Modern drug development demands too many capabilities for any single organization to own.

Gaurav Singal MD moderated a panel with Bea Bakshi MD, Alexandra Goncalves, Faisal Mahmood PhD and Nandini Nayer on an increasingly important use of healthcare AI: finding disease signals in data that already exists before patients enter a formal screening pathway.

Bakshi noted that many cancers are discovered because a patient presents with a problem, not because routine screening found the disease first. Records, symptoms and prior encounters may contain signals that are easy to miss in a short primary care visit. Goncalves described a similar lesson from an AI program designed to identify hypertrophic cardiomyopathy. Some apparent false positives turned out to have other significant cardiac conditions, showing that a model can surface clinically important risk even when it misses the intended label.

Mahmood and Nayer pushed the discussion toward what happens next. As models get better at identifying high-risk patients, health systems need the capacity to act on those findings without turning every alert into another specialist referral. The next phase of opportunistic screening will depend as much on redesigning the downstream clinical pathway as on improving the algorithm itself.

Adele Gulfo moderated a conversation with Kristen Fortney PhD, Raj Panjabi MD, Elisabeth Roider MD, PhD and David Scadden MD about one of the most hyped areas in medicine: whether the biology of aging itself can become a therapeutic target.

The panel quickly separated lifespan from healthspan. Fortney focused on the years people spend living with an accumulating burden of chronic disease, while Panjabi pointed to cancer, brain health, immune health and metabolic health as major drivers of unhealthy aging. Scadden brought the perspective of a physician-scientist working across cancer and regenerative biology, reinforcing how interconnected those systems become over time.

Roider focused on the translation problem. The science is advancing quickly, but aging itself is not treated as a disease, which complicates funding, regulation and the path from promising biology to an approved therapy.

The conversation was much more confident about extending healthy years than dramatically extending maximum human lifespan. Near-term progress is more likely to come from delaying disease and preserving function than from a single treatment that fundamentally resets human aging.

Kara Brotemarkle joined Tracy Batchelor MD for a conversation about what has to happen around an Alzheimer’s therapy for it to deliver its full value.

Brotemarkle used the large number of people living with Alzheimer’s disease who are never formally diagnosed to explain Genentech’s focus on innovation beyond the molecule. Better therapies must be paired with better diagnostic pathways and easier ways for eligible patients to enter care.

Formulation matters too. Brotemarkle pointed to multiple sclerosis, where shorter subcutaneous administration can reduce time spent in infusion centers, lessen caregiver burden and free clinical capacity. Batchelor extended the discussion to the scientific challenge of getting large molecules across the blood-brain barrier, an issue with implications well beyond Alzheimer’s disease.

The same systems question is now emerging around blood-based biomarkers. Easier testing will only help if reimbursement, interpretation and referral pathways evolve with it. The technology is one part of the solution. The care pathway has to be ready for what the technology finds

Merck CMO Eliav Barr MD joined David Ryan MD and Jason Gerberry for a conversation about antibody-drug conjugates, global clinical development and personalized cancer vaccines.

Barr used antibody-drug conjugates to make a broader point about modern oncology: treatments that appear similar on paper can behave very differently in patients. The antibody, linker, payload and mechanism of release all matter. Those differences contributed to Merck’s interest in sacituzumab tirumotecan, where early activity across tumor types and the tolerability profile suggested potential for broader use.

The discussion also explored why China has become increasingly important to global drug development. Barr pointed to clinical sites that can enroll large patient populations quickly and generate substantial experience with new therapies.

Personalized cancer vaccines offered another glimpse of where oncology may be heading. Barr focused on the Moderna and Merck program and the possibility that individualized vaccination could extend beyond tumors traditionally considered highly immunogenic. Manufacturing remains more complicated than for a conventional drug, but he expects the process to become increasingly manageable as the field matures.

Entering the US healthcare market takes more than opening an office or running a pilot. Moderated by Yonatan Keschner MD, the conversation brought together Or Benifla, Liat Primor, Davidi Vortman and Andrei Soran to explore what helps international healthcare companies move from visiting an innovation ecosystem to becoming part of it.

Trust and local relationships surfaced again and again. Vortman described the importance of UltraSight’s work with Mayo Clinic, where partnership extended beyond funding to clinical expertise, data and technology development. Benifla discussed how the ARC Landing network helped CustoMed establish credibility and connect with Boston Children’s Hospital, while the broader panel reinforced that strong clinical results alone are not enough to secure adoption.

Health systems also want to understand workflow, cost, reimbursement and operational impact. Primor’s experience building FeelBetter and Soran’s perspective from the health system and investment side added to a larger point: embedding means learning how decisions are actually made inside healthcare organizations and developing relationships beyond the clinicians who may initially champion a product. Operations, finance and other decision makers often determine whether an innovation ultimately reaches patients.

What if companies could understand a clinical problem before they started designing the solution? Mark Richardson MD, PhD moderated a conversation with Atsuhiro Nakagawa MD, Masaki Kato, Akiko Okubo and Heinz Lubenau PhD about a model at Tohoku University Hospital built around exactly that idea.

Since 2014, Tohoku has brought corporate R&D teams directly into clinical settings to observe care, work with physicians and identify unmet needs before large amounts of time and capital are committed to product development. Kato and Okubo described how that immersion changed how their companies think about development, including moving away from building prototypes internally and then asking clinicians for feedback. Lubenau discussed the value of academic partnerships for clinical development, scientific advice and faster experimentation.

Nakagawa framed the hospital itself as innovation infrastructure. The discussion quickly became relevant to Boston: could Mass General Brigham create a similar environment for industry co-creation? Doing so would require practical solutions around clinical space, protected physician time, research funding, intellectual property and incentives. Tohoku’s experience suggests the first step can remain relatively simple, with deeper research and commercialization relationships developing once the partners understand the problem they are trying to solve.

Rishi Singh MD moderated a wide-ranging ophthalmology discussion with Lance Baldo MD, Jason Comander MD, PhD, Maha Radhakrishnan MD, Richard Small and Cheng Zhang on how gene therapy, cell-based approaches, genetics and better clinical trial design are changing treatment for retinal disease.

Baldo discussed positive Phase III gene therapy results in X-linked retinitis pigmentosa and the lessons behind them. Better vector and capsid design mattered, but so did patient selection, surgical delivery and choosing endpoints that capture changes patients actually experience. Comander put that progress into perspective through patients with inherited retinal disease who are now seeing functional improvements that would once have been difficult to imagine.

Zhang described how Character Bio is combining genetic, imaging and longitudinal clinical data from thousands of patients with age-related macular degeneration to understand why disease progresses differently across patients. Small discussed the path from encapsulated cell technology to commercialization, while Radhakrishnan brought the investor perspective, emphasizing the importance of clearly defined patient populations, meaningful endpoints and differentiated benefits.

Looking ahead, the panel pointed to improved gene therapy, durable delivery, cell-based treatments and better trial design. Zhang offered another possibility: the biggest advance may come from treating retinal disease earlier, before irreversible vision loss has occurred.

JF Formela MD moderated a conversation with Jonathan Carlson MD, PhD, Viswa Colluru PhD, Rajiv Kaul and Al Sandrock MD, PhD on how artificial intelligence is changing drug discovery and how investors should think about the opportunities emerging from new chemistry and new data.

Colluru focused on the enormous amount of useful chemistry that already exists in nature but remains poorly characterized. Enveda uses AI to interpret mass spectrometry data from complex biological samples, treating chemical signals in context much like language models interpret words. That approach is intended to expose drug leads conventional screening may never see.

Carlson brought the academic discovery perspective, while Kaul and Sandrock added the views of capital markets and biotech development. The discussion moved beyond the idea that AI speeds up existing drug discovery. The larger opportunity is expanding the chemical and biological space scientists can interrogate, then determining which discoveries are strong enough to justify clinical development and investment.

One Enveda program has already progressed into the clinic after researchers moved from ethnobotanical observations to a new chemical scaffold and a previously unknown mechanism affecting inflammation. It offered a concrete example of what AI-enabled discovery can look like when computation, biology and drug development stay connected.

Maurizio Fava MD moderated a panel with Mill Brown MD, Paul English, Luis Voloch and Sabine Wilhelm PhD on how artificial intelligence could expand mental healthcare in a system that does not have enough clinicians to meet current demand.

The panel drew an important line between access and clinical care. Wilhelm distinguished general-purpose AI systems from tools designed and validated specifically for mental health. A consumer chatbot may be willing to discuss depression, anxiety or obsessive-compulsive disorder without the evidence, safety testing or privacy standards expected of a healthcare product.

Brown brought the clinical operating perspective, English the entrepreneurial one and Voloch the challenge of building and paying for new models of care. Together, they pushed the conversation beyond whether AI can talk to patients and toward whether tools can be validated, reimbursed and integrated responsibly.

Wilhelm also noted that clinicians may increasingly need to ask patients how they use AI. In some cases it can provide useful support. In others, it can reinforce harmful patterns, such as repeatedly seeking reassurance in obsessive-compulsive disorder. The opportunity is significant, but human oversight, clinical evidence and payment will determine whether AI actually improves mental healthcare.

Steve Walsh moderated a conversation with Colin McHugh, Christine Schuster RN and Allen Smith MD about how community health systems can adopt sophisticated technology without losing the accessibility and personal relationships that make local care valuable in the first place.

Smith described the competitive reality for South Shore Health, located close enough to Boston that patients can readily reach globally recognized academic hospitals. Community providers therefore have to deliver high-quality care while being clear about which services belong locally and which need referral. Schuster brought the perspective of a community hospital leader balancing innovation, workforce realities and the high-touch care patients expect.

McHugh offered an example of how operational problems often extend beyond medicine. By convening community leaders in greater Nashua, his organization surfaced guardianship delays that were keeping patients in hospital beds even after their clinical needs had been addressed.

He also closed with something decidedly low-tech. Southern New Hampshire Health asks employees to walk with people who appear lost rather than simply pointing down a hallway. The lesson was not technology versus people. Community healthcare needs both.

Dennis Selkoe MD and Rudy Tanzi PhD moderated a discussion with John Sims, Reisa Sperling MD and Donna Masterman on how Alzheimer’s disease is beginning to move from late diagnosis and disease management toward earlier detection, prevention and more effective treatment.

Sims compared current Alzheimer’s care to waiting for someone to have a heart attack before checking blood pressure or cholesterol. The biology can be developing for years before symptoms become visible, which makes the growing availability of blood-based biomarkers especially important. Sperling pointed to prevention trials aimed at treating patients before cognitive symptoms develop and discussed the possibility of identifying people at meaningful risk much earlier.

The conversation also looked beyond today’s first generation of anti-amyloid therapies. Selkoe and Tanzi framed the broader scientific challenge. At the same time, Masterman and Sperling discussed the potential of new delivery approaches, including brain-shuttle technologies designed to move drugs across the blood-brain barrier more efficiently. Future treatment will likely involve multiple targets, including amyloid, tau, inflammation, and other mechanisms.

Even with that progress, one major gap remains. Removing disease-associated proteins is not the same as restoring brain function. The field still needs therapies that protect neurons and help synapses work better.

Harold Lee moderated a conversation with Tal Liani, Jane Moran and Julia Strandberg about the other side of healthcare’s AI transformation: the security and privacy infrastructure required before increasingly connected and autonomous technologies can be trusted at scale.

Liani focused on the rise of agentic AI. Software is beginning to act on behalf of users rather than provide information, which means AI agents may need credentials, system access and permission to complete transactions. Those capabilities also create a new attack surface.

Moran brought the question inside the health system. As agents begin communicating with other agents and clinical systems, organizations will need to define what they are authorized to do, monitor their behavior and detect when something changes. Strandberg added the connected-care perspective, where security and privacy have to extend across devices and settings rather than stop at the hospital wall.

The cybersecurity architecture is evolving in response. The panel expects continued movement away from collections of isolated tools and toward broader platforms that connect identity, endpoints and data protection.

David Ting MD moderated a discussion with Adam Kaye, Sandra McAllister PhD, Diarmuid Moran and Rahul Aras PhD on how precision oncology is moving beyond the first generation of matching a genetic alteration with a targeted drug.

Aras used the Wnt beta-catenin pathway to show how deeper biological understanding is turning previously undruggable targets into therapeutic opportunities. The pathway has been understood for decades, but earlier approaches could not separate cancer-driving biology from functions normal cells need. Kaye brought the economics into the discussion, noting that increasingly narrow cancer subtypes can support highly targeted development but also raise difficult questions about pricing and payer sustainability.

Moran’s precision medicine and diagnostics perspective underscored how much more complicated patient selection becomes as oncology moves into antibody-drug conjugates, immune therapies and other modalities that may depend on multiple biomarkers. McAllister added age to the equation, describing how the tumor microenvironment and systemic biology can change across the lifespan even when two patients appear to have the same cancer.

Precision oncology is becoming less about finding one molecular answer and more about understanding the full biological context of each patient’s disease.

Daniela Lamas MD moderated a discussion with AJ Blood MD, Calum MacRae MD, PhD, Holly Maloney and Ian Tong MD about what happens as more care moves outside traditional hospitals and into homes, virtual settings and new consumer models.

Lamas framed the session as an audit of the future: which parts of the hospital are genuinely durable, and which are likely to move elsewhere. Blood described cardiometabolic care that already works differently, using AI and team-based delivery to reach patients asynchronously rather than waiting months for the next office visit.

MacRae was more cautious about how quickly existing institutions will change, arguing that many of the biggest barriers are structural rather than technical. Maloney brought the investor perspective to the same question of where value will shift as new care models emerge. At the same time, Tong added a clinical view of what more autonomous and distributed care could look like.

The common thread was that healthcare already knows how to do many things better than it consistently delivers them. The disruption may come less from a single technology than from organizations built to close that gap without depending on the traditional hospital visit.

Patients are already using artificial intelligence to understand symptoms, interpret test results and prepare for conversations with clinicians. Mark Bohen brought together Ken Callahan, Allison Chang MD, PhD, Pete Clardy MD and Eric Rosenthal MD to examine what that growing use of AI means for trust, safety and the patient experience.

Chang emphasized that strong model performance is not enough for high-stakes clinical decisions. The level of evidence should rise with the potential risk to the patient, and clinicians need to know whether a model works for the population and setting in which it is being used. Rosenthal extended that thinking to health systems, arguing that trust requires ongoing monitoring as data, treatments and clinical practice change.

Clardy described patients using AI for increasingly complex health questions, while Callahan discussed the role standards and evidence can play as adoption expands. The conversation ultimately kept the clinician-patient relationship at the center, with AI adding information and decision support rather than replacing human judgment.

Artificial intelligence is advancing globally, but bringing an AI-enabled healthcare company into the US takes more than a strong algorithm. Abi Barrow PhD brought together Frederik Bender, Declan Kelly MD, Harald Schnidar PhD and Jukka Tapaninen MSc to discuss what their European companies have learned as they expand into American health systems.

Their technologies address very different parts of care, from 3D ultrasound and clinical knowledge access to smartphone-based skin assessment and digital pathology. Across the panel, the focus stayed on practical value: improving workflow, reducing time, increasing consistency and helping clinicians make better use of the information already around them.

The US market offers significant opportunity, but the panelists also described the importance of trusted reference sites, clinical validation and proving that a technology works inside real healthcare workflows. European regulatory experience can help, but adoption still depends on building credibility with US health systems and demonstrating clear value to clinicians and organizations.

Every AI application discussed at the Forum runs on silicon, which is the connection Ahmad Bahai PhD built this panel around. Nanotechnology is not adjacent to healthcare AI; it is what the algorithms, sensors and imaging systems sit on.

Paru Deshpande PhD broke silicon’s contribution into three parts: compute, sensing, and the unglamorous infrastructure of battery life, wireless and edge computing that determines whether a sensor can live in a watch. He expects sequencing to extend from genomics into proteomics, metabolomics and microbiome, and named the consequence plainly. The gap will be between the existence of data and the existence of insight.

Paul Gerrish made the case for continuous monitoring through blood pressure. A cuff reading in a clinic is a point in time, usually unrepresentative because the patient is stressed while it is taken. Measuring across day and night produces something different in kind.

Alex Shalek PhD added the sequencing that matters. Comprehensive measurement comes first, because it establishes what is worth monitoring at all. Only then can continuous sensors be pointed at the right thing.

Immunology has become very good at suppressing inflammation. The next question is whether treatment can go further and reset the immune system deeply enough to give patients durable, treatment-free remission. Ellen Gravallese MD led that discussion with Alyssa Johnsen MD, PhD, Eric Mortensen and Scott Requadt.

Johnsen emphasized that cytokine-targeted therapies still have an important role, but that deeper remission may require thinking about what happens after immune depletion, including whether tolerance can be restored. Requadt focused on B-cell depletion and the possibility that reaching pathogenic cells in tissue, not just in the blood, could help explain why CAR T cells and T-cell engagers may produce more durable responses than chronic immunosuppression.

Mortensen, who joined after the session began, pushed the conversation toward a more tunable version of immune reset. Different autoimmune diseases may require different depths of depletion, different targets and different therapeutic platforms. Across the panel, the promise of CAR T cells, T-cell engagers and other emerging approaches was balanced by the same unresolved questions: how much immune depletion is enough, how long remission will last, which biomarkers can guide patient selection and how to achieve that depth of response without bringing oncology-level toxicity into chronic autoimmune disease.

Healthcare AI is moving into a more competitive phase, and Brenton Fargnoli MD asked Ada Glover, Will Gordon MD, Terrence Chen PhD and Eric Podradchik what will separate durable solutions from the growing number of pilots, point products and AI agents entering the market.

Glover sees the clearest traction in operational work with a defined job to be done, such as intake, referrals, prior authorization, and value-based care workflows, where reliability and return on investment are easier to measure. Gordon added a practical test for scalability: the strongest tools often spread because they save time without requiring clinicians to change behavior or receive extensive training.

Chen drew lessons from more than a decade of clinical AI in imaging. Accuracy alone is not clinical value. AI becomes useful when it is embedded deeply enough in the workflow to influence how data is acquired, interpreted and acted on. Podradchik made a similar point from the health system side, noting that revenue cycle applications are producing some of the clearest financial returns and helping fund other clinical AI priorities.

The conversation ultimately came back to defensibility. Solutions closest to the creation of clinical data, clinical action or a hard-to-replicate workflow may be best positioned to survive as EHR companies, hyperscalers and startups increasingly compete on the same terrain.

Healthcare is not the first industry trying to figure out how artificial intelligence changes work at enterprise scale. Niyum Gandhi brought together Jeff Busconi, Athina Kanioura PhD, Margaret Pierson PhD and Scott Sperling to compare what banking, consumer goods, e-commerce and private equity have learned from deploying AI inside large organizations.

The examples ranged widely. Bank of America is using AI to help junior bankers research companies, build presentations and work at a higher level. PepsiCo is combining digital twins with physical AI across operations. Wayfair is using generative AI to improve tasks that previously depended on narrower predictive models, from finding errors in product data to qualifying sales leads. THL has made decades of investment and offering memoranda searchable so teams can interrogate prior decisions and comparable companies in new ways.

Across very different businesses, a common lesson emerged: replacing one task with AI rarely transforms the whole process by itself. Better models can expose bottlenecks elsewhere, change job responsibilities and force organizations to rethink how decisions are made. Kanioura stressed linking technology investments to business strategy and measurable value, while the broader discussion emphasized governance, workforce adaptation and the need to scale what works rather than accumulate disconnected pilots.

For healthcare, the takeaway was less about copying individual tools and more about learning how other industries redesign operations around technology once it proves useful.

Carl Berke PhD opened a conversation with Clay Johnston MD, PhD, Mihir Kamdar MD, and John Voith around a deceptively simple question: why did care models developed around academic medicine need new companies to reach more patients?

For Johnston, the answer was incentives. Harbor Health combines insurance and care delivery so it can reward prevention, primary care, and better outcomes rather than treating them as financial losses. Kamdar described a similar mismatch in palliative care. Research at Mass General showed that earlier palliative support could improve symptoms and reduce unnecessary hospital use, but fee-for-service economics and a severe workforce shortage made scaling difficult. Tuesday Health uses technology to identify patients, extend specialist support into the home and intervene earlier when symptoms begin to worsen.

Voith told the same story from pediatric mental health. InStride Health grew out of an evidence-based McLean Hospital model for children and young adults with complex anxiety and OCD, but long wait lists, geographic limits and self-pay economics kept it out of reach for many families. Technology and insurance partnerships created a path to scale the clinical model without stripping away the involvement of families, schools and care teams.

The common thread was that tech-enabled care is not simply digital care. Technology works best when it helps proven clinical models reach more people, aligns payment with outcomes, and preserves the human relationships that made the care effective in the first place.

Curative cell and gene therapies can create extraordinary value for patients while creating equally difficult questions about who pays, when they pay and who carries the risk. Steven Pearson MD led a discussion with Angela Shen MD, Sanjay Srivastava PhD, Eric Herbek and Will Shrank that followed the problem from early development through manufacturing, hospital delivery and insurance coverage.

Shen described why academic centers can sometimes move early therapies into patients at far lower cost than large companies, but also why that approach often postpones harder questions about manufacturing scale and commercialization. Srivastava pointed to automation, robotics and closed manufacturing systems as part of the path toward lower development and production costs. Herbek brought the health system perspective, where hospitals may need to invest in specialized infrastructure without knowing how many patients they will treat or when they will be reimbursed.

Shrank focused on the financing architecture around those clinical realities. High-cost, one-time therapies create volatility that individual employers, payers and providers are poorly equipped to absorb on their own. Risk pooling, outcomes-based payment and intermediaries that can align manufacturers, payers and providers may all have a role.

No one on the panel offered a single solution. The stronger conclusion was that affordability is an ecosystem problem. Development incentives, manufacturing, provider economics, insurance design, and patient access must be addressed together if curative therapies are to remain both investable and broadly available.

Diagnostics is no longer just about confirming a disease after symptoms appear. David Walt PhD joined John Iafrate MD, PhD, Gianluca Pettiti and Paul Ridker MD to examine how AI, molecular testing, multi-omics and longitudinal data are expanding diagnostics into prediction, treatment selection and continuous disease monitoring.

Pettiti argued that high-quality, well-annotated data is the foundation for useful healthcare AI. Iafrate added an important caution from pathology: AI is already well suited to labor-intensive tasks such as counting, scoring and pattern recognition, but the performance expected of a definitive clinical diagnostic is far higher than the threshold for a productivity tool. The field should take advantage of what works now without assuming that every foundation model is ready to replace a clinical gold standard.

Ridker brought the discussion into cardiovascular medicine, where companion diagnostics are much less developed than in oncology. Biomarkers can identify risk and reveal disease biology beyond traditional measures such as LDL cholesterol, but recent trial results also show that a promising biomarker does not automatically guarantee a successful therapeutic target.

The panel was optimistic about multi-omics and AI, particularly when built on carefully designed longitudinal cohorts. The message was equally clear about the prerequisites: better data, prospective validation, and a disciplined distinction between tools that assist clinicians and tests trusted to make high-stakes clinical decisions.

RNA interference has moved from a Nobel Prize-winning discovery to a validated drug platform, and John Harris MD, PhD used that progress to ask Anastasia Khvorova PhD, Francesco De Rubertis PhD and Craig Shepherd what it will take to move siRNA beyond the liver and into a much wider range of diseases.

Khvorova described the chemistry and biology that made liver delivery possible, including chemical stabilization and receptor-mediated uptake, and pointed to the unusual durability now achievable with some siRNA medicines. That raises the possibility of replacing daily pills with treatments given only once or twice a year. The next frontier is extrahepatic delivery, with work advancing in tissues including skin, muscle, heart and the central nervous system.

The financing discussion showed how different investors approach the same scientific risk at different points in development. Shepherd reflected on Blackstone’s large, structured investment in Alnylam after the platform had survived major setbacks and demonstrated clinical validation. De Rubertis explained Medicxi’s asset-centric model, which generally favors focused programs with a clearer path to proof of concept rather than financing broad platforms for decades.

Across science and capital, the conversation came back to de-risking. RNAi is no longer being asked to prove that gene silencing can work in humans. The challenge now is delivering the right molecule to the right tissue, showing meaningful clinical benefit and matching the financing model to the time required to get there.

Healthcare has no shortage of data. The harder problem is turning fragmented clinical and biological information into a coherent patient representation that supports better research and care. Faisal Mahmood PhD explored that challenge with Tiffany Chen MD, Rowland Pettit MD, PhD, Jorge Reis-Filho MD, PhD and Celestine Schnugg.

Chen focused on tissue, where spatial relationships between cells can reveal biology that disappears when cells are studied in isolation. Combining pathology with spatial transcriptomics and proteomics may help researchers understand disease mechanisms, predict therapeutic response, and design more precise clinical trials. Pettit moved to the other end of the scale, describing foundation models that can bring together structured records, clinical text, imaging and other longitudinal data into reusable patient-level representations.

Reis-Filho connected those approaches to drug development, describing how AI can identify biomarkers that traditional methods miss and how foundation models can dramatically reduce the manual annotation required to build new tools. Schnugg brought the investor and biobanking perspective, where disease-specific translational datasets can help connect emerging models to questions that matter in real clinical development.

The larger idea was not simply to build bigger models. It was to make medicine more computable across scales, from tissue architecture to an entire patient trajectory, while preserving the ability to search, predict and explain why a model concluded.

The next phase of obesity treatment may be less about maximizing the number on the scale and more about improving the quality of weight loss. Moderated by Jason Gerberry and Alec Stranahan, the discussion with Caroline Apovian MD and JoAnn Manson MD, DrPH focused on what patients are losing along with fat as GLP-1-based therapies produce increasingly substantial reductions in body weight.

Muscle mass, bone health, strength, physical function and quality of life were recurring concerns. The panel emphasized that changes in lean mass should not be interpreted in isolation. Researchers still need to determine when muscle loss translates into meaningful functional decline, which patients are most vulnerable and whether therapies designed to preserve muscle actually improve strength, mobility and daily life.

The conversation also kept lifestyle in the picture. Resistance training, adequate protein intake and other behavioral strategies may remain important even as pharmacologic treatment becomes more effective. Older adults, frail patients and other high-risk groups will need particular attention as the evidence base expands.

New combinations and mechanisms may eventually help preserve lean tissue while maintaining weight-loss efficacy, but the panel was careful about how much is known today. Phase 2 signals are not enough. Larger trials will need to show that preserving muscle changes outcomes patients can actually feel.

American academic medicine’s global role is changing. Instead of primarily bringing international patients to the United States for highly specialized care, major health systems are increasingly helping countries build that expertise closer to home.

David Duncan led Aaron Gerber MD, Heitham Hassoun MD and Bishoy Mikhail through that evolution. Gerber described Mass General Brigham’s growing emphasis on long-term partnerships, including work in the UAE that combines facility development, clinical operating models and a willingness to put the MGB brand behind care delivered locally. Hassoun emphasized that no single international business model fits all. Cedars-Sinai’s approach ranges from partnerships in Mexico City to a wholly owned London clinic, depending on what the market and relationship require. Mikhail traced Cleveland Clinic’s progression from international patient care and physician training to advisory relationships, managed facilities and major operations such as Cleveland Clinic Abu Dhabi.

Affordability, access and geopolitics remain major constraints, particularly as visa challenges and regional instability affect patient travel. But the panel saw those pressures as another reason to bring care closer to patients. International healthcare partnerships are becoming less about exporting an American hospital model and more about building sustainable local capability, with lessons that may eventually help US health systems deliver high-quality care more efficiently at home.

The first generation of cell and gene therapies proved that genetic disease can be treated at its source. The next generation focuses on making those therapies easier to deliver, safer, less expensive, and available to many more patients.

Roger Hajjar MD brought together Hanny Al-Samkari MD, Florian Eichler MD, Samarth Kulkarni PhD and Patty Musolino MD, PhD to examine what the field has learned. Kulkarni pointed to the curative potential of CRISPR-based therapy in sickle cell disease and beta thalassemia, while acknowledging the cost, hospitalization and conditioning required by current autologous approaches. Al-Samkari emphasized that the risk-benefit calculation changes dramatically by disease. Patients with severe sickle cell disease may accept risks that make far less sense in hemophilia, where highly effective alternatives already exist.

Mousolino focused on delivery, from lipid nanoparticles and engineered viral capsids to more localized approaches that could reduce dose and toxicity. Eichler added another essential variable: timing. In neurological disease, correcting a mutation may come too late once irreversible tissue injury has occurred, making newborn screening, biomarkers and earlier diagnosis increasingly important.

The discussion also looked beyond rare disease. Gene editing is beginning to move into cardiovascular disease, hypertension and autoimmune conditions, while manufacturing improvements and allogeneic approaches could make treatment much more scalable. But the panel cautioned that better technology alone will not determine success. Clinical trial infrastructure, regulatory pathways, patient selection, meaningful biomarkers and true unmet need will shape which therapies ultimately reach patients.

Getting an innovation into the community takes more than making the technology portable. It requires trust, people and a care model designed around where patients actually live their lives.

Elsie Tavares MD led Allison Bryant, Valerie Dao, Rian Hasson MD and Priya Sarin Gupta MD through four very different examples of that principle. Dao described how Oben Health turns barbershops and salons into healthcare access points by training trusted community members as health workers and using technology to make screening, documentation and connections back to clinical care easier. The technology supports the relationship rather than replacing it.

Bryant described the same balance in maternal health, where community health workers, doulas, remote blood pressure monitoring and mobile postpartum care are being combined to support patients throughout pregnancy and after delivery. Hasson is applying the model to lung cancer screening by bringing low-dose CT into communities, using data to identify areas with high rates of late-stage disease and then building navigation around the screening itself.

Sarin Gupta brought the discussion to substance use disorder, where mobile clinics can provide naloxone, wound care, testing, harm reduction or simply a conversation without requiring someone to enter a hospital first. Digital tools can help with follow-up and navigation, but she drew a clear boundary: technology should not become a substitute for trust.

Across maternal health, hypertension, cancer screening and addiction care, the same idea kept resurfacing. Access improves when healthcare stops expecting patients to find the right front door and starts redesigning the front door around the community.

AI is beginning to move beyond documentation and administrative work into clinical decision-making, raising harder questions about trust, accountability and how much autonomy healthcare is ready to accept. Bret Bostwick MD moderated the discussion with Muthu Alagappan MD, Allon Bloch, Rebecca Mishuris MD and Haider Warraich MD.

Mishuris said truly autonomous clinical care remains limited, with humans still involved in evaluating risk and making consequential decisions. Bloch argued that patients are already using AI extensively and that health systems need to engage rather than wait. Alagappan described how messaging and AI can gather patient histories over time, creating a different way to understand symptoms outside a single office visit. Warraich focused on the safeguards needed as agentic systems become more capable, including post-deployment supervision and monitoring.

The panel saw the larger opportunity in continuous, more proactive care. Getting there will require real-world evidence, new payment models and clearer approaches to responsibility and disclosure as AI becomes a more active part of the care team.

Cardiovascular innovation is moving well beyond the traditional focus on lowering cholesterol. Jim Tornatore and Patrick Ellinor MD, PhD brought together Paul Bolno MD, Martin Cowie ChB, MD, Clive Meanwell and Alvin Shih to discuss new therapeutic pathways across cardiovascular and metabolic disease.

Cowie described a future that combines lipid lowering with new metabolic and genetic approaches rather than relying on a single mechanism. Shih focused on cardiac metabolism and the potential to improve how diseased heart muscle generates energy. Bolno highlighted how human genetics and RNA medicines could lead to more durable treatments, including therapies designed for infrequent dosing and combinations that address multiple targets at once.

Meanwell pushed the discussion toward scale and access. Better science will matter little if patients cannot receive or remain on treatment. That brought adherence, manufacturing, consumer engagement and distribution into the same conversation as genetics, biomarkers and clinical trials.

New Takeda CEO Julie Kim joined Merit Cudkowicz MD to discuss the company’s next chapter, the long path from scientific discovery to a successful medicine and how artificial intelligence is changing drug development.

Kim described a two-horizon strategy. The immediate focus is to successfully launch late-stage medicines, advance the pipeline, and transform the organization. Longer term, the goal is sustainable growth and scientific leadership across Takeda’s therapeutic areas.

The development of Takeda’s orexin therapy for narcolepsy type 1 illustrated the persistence required. Earlier molecules showed the biology could work but faced formulation and safety challenges before the program ultimately produced an approved medicine.

AI is now being used across Takeda’s research and development process, from identifying candidates and designing clinical trial protocols to improving operational efficiency. Kim was equally clear that technology does not eliminate the need for human scientists, clinical trials or partnerships with academic researchers, biotech companies and patients.

After several difficult years for biotech financing, investors see signs that stronger science, more mature companies and renewed M&A activity are bringing capital back into the sector. Glenn Silverstein and Anthony Yu led Michael Gladstone, Roger Kitterman, Nilesh Kumar and Laura Stoppel through the market from academic company creation to public investing.

The panel described a more disciplined environment than the market of 2020 and 2021. Investors are looking not only for compelling science, but for clear development paths, efficient use of capital and multiple potential routes to value creation. Kitterman noted that the downturn affected even very early academic ventures, although appetite for earlier-stage opportunities has begun to return.

China is also becoming an increasingly important source of innovation and partnering opportunities. Across private investment, IPOs, reverse mergers and M&A, the discussion suggested that capital is available, but companies face a higher bar for differentiation and execution.

The 2026 Kraft Prize for Excellence in Innovation in Community Health was awarded to Oben Health, an AI-enabled platform that transforms trusted community settings such as barbershops and beauty salons into healthcare access points.

The award highlighted Oben Health’s approach to reducing cardiovascular risk by combining technology with community relationships. The model trains trusted community members to support screening and connect people back to clinical care, creating a scalable way to reach populations that may face barriers to traditional healthcare settings.

The presentation also emphasized the Kraft Prize’s broader purpose: identifying innovations that can move beyond promising ideas to measurable impact in community health.

Scaling artificial intelligence across healthcare requires more than deploying new tools. Najat Khan PhD joined Kyle Bryant, David Rhew MD and Ron Walls MD to examine how organizations move from isolated pilots to changes in workflows, workforce roles and enterprise operations.

Bryant distinguished between personal productivity, process transformation and entirely new AI-enabled products, cautioning that adding AI to a poorly designed process can increase cost rather than create value. Rhew pointed to the rapid adoption of ambient documentation as an example of technology spreading quickly when it solves a clear problem for clinicians.

Walls described MGB’s use of an AI-enabled virtual primary care pathway that can gather patient histories, triage needs and prepare information for clinicians before the visit. As AI becomes more capable, the panel emphasized that culture change must be accompanied by governance, human oversight and greater AI literacy throughout the workforce.

Two investigators received the 2026 First Look Awards following presentations from 12 Mass General Brigham innovators whose work was evaluated for scientific, translational and commercial potential.

Lloyd Bod PhD was recognized for Cytagonists: Restoring Immune Function by Neutralizing Disease-Promoting Autoantibodies, while Korkut Uygun PhD was recognized for Organs Should Wait for Patients: Creating the Organ Logistics Infrastructure of the Future.

Miceal Chamberlain presented the awards, describing the selected projects as addressing significant unmet needs while showing potential for collaboration and future impact on patient care.

China’s biotechnology sector has moved rapidly from generics and biosimilars toward novel therapies and global drug development. Aamir Mecklai and Timothy Ng explored that transformation with Hui Feng, Judith Li, Jianguo Michael Yin and Weian Zhao.

The panel traced the growth of an ecosystem that now combines scientific talent, venture capital, hospitals, CROs, CDMOs and increasingly sophisticated regulatory infrastructure. Speed, cost and access to patients remain advantages, but the discussion also emphasized the growing quality and originality of Chinese science.

Li argued that the relationship between East and West should not be viewed only as competition. Western companies increasingly look to China for assets and development capabilities, while global clinical expertise remains important for translating those programs across populations and markets. New licensing and partnership structures are bringing the two ecosystems together much earlier in development.

Rare disease drug development has benefited from new genetic technologies, more flexible regulatory thinking and a much stronger patient voice, but the smallest patient populations remain difficult to serve sustainably. Jason Zemansky and Patty Musolino MD, PhD led Carole Ben-Maimon, Charlie Gayer, David Meeker MD and Kasper Roet PhD through the scientific and economic challenges.

The panel highlighted RNA medicines, gene editing, gene therapy and patient-derived cell models as technologies that are changing what can be treated. At the same time, finding patients remains a major obstacle. Gayer described using AI and electronic health record data to identify patients who might otherwise remain undiagnosed, while Meeker emphasized that the diagnostic odyssey is still very real.

For ultra-rare conditions, science alone is not enough. Organized patient communities, natural-history data, academic partnerships, regulatory flexibility and business models that group multiple small indications together may all be needed to make development viable.

Solid tumor treatment has progressed through successive waves of chemotherapy, targeted therapy, immunotherapy and increasingly precise approaches to previously difficult targets. Aamir Mecklai and Rebecca Heist MD joined Roy Baynes MD, PhD, Ryan Cohlhepp PharmD and Jonathan Lim MD to explore what is changing now.

Baynes described a new phase of precision oncology enabled by better molecules, diagnostics, patient segmentation and trial design. Lim used RAS as an example of a target once considered undruggable that has opened through new chemistry, while emphasizing that the next gains will likely depend heavily on combination therapy.

Cohlhepp described a similar effort to revisit TGF-beta by targeting tumors more precisely and reducing the toxicities that limited earlier approaches. Across the panel, the focus was not simply on discovering more targets, but on improving first-in-class innovation, moving promising therapies into earlier treatment settings and designing trials that make better use of every patient who participates.

Bank of America Chair and CEO Brian Moynihan joined CNBC’s Leslie Picker for a wide-ranging conversation spanning the economy, artificial intelligence, healthcare innovation and employee health.

Moynihan discussed inflation, interest rates and consumer spending before turning to AI, which he described as both an opportunity and a disruptive force across banking and healthcare. In healthcare specifically, he emphasized the importance of sustaining the academic research base that feeds future innovation while the industry adapts to rapidly changing AI capabilities.

The conversation also addressed Bank of America’s decision to cover GLP-1 therapies for employees and families. Moynihan framed that spending within a broader commitment to workforce health, arguing that the value of improved health, satisfaction and retention can extend beyond the immediate medical cost.

Asked about financing innovation, he returned to the importance of clearly connecting the science to its eventual use, market and business opportunity so investors can understand the path from discovery to impact.

Cardiovascular medicine is one of modern medicine’s great success stories, and cardiovascular disease remains a leading cause of death worldwide. The panel took up that contradiction.

Emily Lau MD argued the field treats established disease well and intervenes far too late. Her specific case concerns women, where evidence now identifies menarche, pregnancy and menopause as inflection points in risk decades before overt disease. The field has not married preventive strategies to those transitions, which she called the biggest opportunity of the coming decade.

Christian Ruff MD divided the unmet need differently. One half is implementation, since excellent therapies are delivered ineffectively, particularly in primary care. The other is phenotyping residual risk, so that foundational therapy goes to everyone while specific pathways are targeted in the patients active in them.

Jorge Plutzky MD, reflecting on thirty years in preventive cardiology, called GLP-1 access offering cardiovascular benefit independent of diabetes a seismic change. It answered a question that had frustrated the field for decades: why treating obesity had not reduced cardiovascular events.

Two disease-modifying therapies now carry full FDA approval and are in routine use across Mass General Brigham. John Dickson MD, PhD was careful about what they do. They remove amyloid and appear to slow cognitive decline in at least some patients. They do not stop or reverse it. For a devastating illness with no previous disease-modifying option, that is still a paradigm shift.

The limiting factor is not efficacy but logistics. Both launched as infusion-only with extensive MRI monitoring for ARIA, the swelling or bleeding that worries patients most in clinic. In practice, with careful patient selection and appropriate monitoring, Dickson reported the therapies have proven quite safe.

Subcutaneous administration now removes the infusion center bottleneck. What he wants next is easier monitoring, ideally a blood test rather than MRI, and risk stratification good enough to identify which patients can safely skip routine imaging.

On the next generation, he ranked improved safety above improved efficacy as both more realistic and more desirable, since safety is what reduces the burden patients carry.

The search in lung cancer is for a combination capable of unseating the current standards of care, and two components dominate: PD-1 VEGF inhibitors and antibody-drug conjugates.

Justin Gainor MD described the debate shifting from progression-free survival to overall survival, with recent readouts beginning to deliver the latter. Three things stood out to him. VEGF appears safer in squamous disease than the old teaching about fatal hemoptysis risk suggested.

Activity looks somewhat better in squamous than non-squamous. And hazard ratios between Chinese and Western populations looked comparable.

Zofia Piotrowska MD supplied the qualifier. EGFR-mutated populations are relatively homogeneous across ethnic groups, so comparability there is less informative than it appears. Non-genomically defined populations carry far more variation in demographics and smoking history, which matters particularly for immunotherapy.

Rebecca Heist MD explained what made the squamous results notable. Despite exclusion criteria, those trials included patients with hemoptysis, cavitation and large central tumors, where delivering this drug class safely has not previously been possible.

Personalized cancer vaccines moved quickly from concept to excitement, and the panel’s question was where the approach goes next.

Ryan Sullivan MD explained why melanoma was the right starting point, and much of the reasoning was practical rather than biological. Vaccine manufacturing takes time, so the setting had to be one where patients would not have progressed before receiving treatment. An adjuvant setting after surgery provides that window, and an approved comparator already existed.

He contrasted that with the audacious decision to pursue pancreatic cancer, where a small subset of patients who generated documented immunization did better than those who did not. Given results at both ends of the immune spectrum, his conclusion was that everything is now fair game.

Xin Gao MD addressed cold tumors, where prostate sits at the far end. His point was that unresponsiveness to checkpoint inhibitors does not mean unresponsiveness to immunotherapy. The right combination partner may simply be a different immunotherapy approach entirely.

Psoriasis took roughly thirty years to go from methotrexate to targeted therapy, and the path ran backwards from the usual order. Anti-TNF worked first, and curiosity about why drove the translational research that identified IL-17 and then IL-23 upstream of it.

John Harris MD, PhD explained what makes IL-23 inhibition unusual. The target sits far enough upstream to eliminate the memory cells driving psoriasis, so stopping treatment does not bring the disease back.

He was also candid about the cost of that success. Companies cared about nothing else for years. He would bring a new target and be told to try it in psoriasis first, even when it had nothing to do with psoriasis. Only once IL-23 inhibitors became good enough that the field concluded it could not do better did attention turn to atopic dermatitis, vitiligo, alopecia areata and the rest.

Avery LaChance MD described a newer oral matching biologic efficacy with a clean safety profile and no lab monitoring, which for the first time lets clinicians genuinely ask a patient which they would prefer. She cautioned against assuming the answer, noting that many patients choose four injections a year.

Japan is the world’s third largest pharmaceutical market and runs on universal health insurance, which means a drug approved by the PMDA is almost immediately reimbursed and reaches every patient who needs it. Despite that, foreign companies routinely treat Japan as a late-stage afterthought, which costs the industry its most valuable resource.

Wakako Toga PhD described a government-funded initiative built to fix that, covering all disease areas and providing the strategic planning, regulatory clarity and partner introductions a company with no presence in Japan lacks. She was notably direct that the initiative supports a no-go decision as readily as a go.

Her worked example was patent term extension, which allows the patent clock to stop during Japanese clinical trials for up to five years, producing up to twenty five years of exclusivity.

Liz Lewis explained what changes structurally when Japan is planned for from day one rather than addressed through bridging studies after US and European approval. Asking early what data Japan requires opens access to pathways and pricing mechanisms that a company planning retrospectively has already forgone.

Reza Dana MD went back 120 years to make his point about ophthalmology’s role in medicine. The first successful allotransplants came in 1905, half a century before the first successful kidney transplant. The field led on lasers for shaping tissue, on stem cell transplantation, and on topical biologics.

The change of the last twenty years is a shift toward regenerative approaches that remove the indication for surgery. Keratoconus was a surgical condition until roughly a decade ago, when corneal thinning past a threshold meant a transplant. Patients now receive cross-linking.

Rishi Singh MD traced the retina arc from anti-VEGF’s arrival in 2004, a single drug class addressing four leading causes of blindness, through Ang-2 biology, first-in-class treatment for geographic atrophy, and gene therapy for inherited retinal degeneration.

Asked what remains unmet, he was unambiguous. It is not visual acuity. It is durability. Patients come in five to six times a year, spend four hours in the office and cannot do anything for twenty four hours afterward, and many of them are working-age adults.

Inflammation and immunology diseases are classified by tissue of origin or clinical specialty, which conceals enormous diversity in mechanism and presentation within any one category. Niranjana Nagarajan MSc, PhD identified the consequence: targets pursued broadly tend not to address enough of the problem for anyone.

Samantha Singer, speaking publicly about her company for the first time, described an approach inferred from B cell depletion’s results. Rather than depleting an entire cell lineage, the aim is removing a small set of bad actors, a minority cell population responsible for most inflammation across many T cell mediated disorders.

She also objected to the field’s vocabulary on substantive grounds. Reset implies wiping the slate clean, which she considers a bad idea for an immune system that has learned a great deal worth keeping. Her preferred term is rebalancing.

Adrian Ray PhD made the practical case that delivery can disqualify an effective therapy. Giving a child an injection every two days does not fit a pediatric food allergy population, regardless of how well it works.

Ireland is the fifth largest source of foreign direct investment into the United States, with 700 to 800 companies employing more than 250,000 Americans across all 50 states. Jennie Lynch noted, with acknowledged bias, that this is reasonable for a country smaller in population than Massachusetts.

Brian Shields described inverting the conventional sequence. Medical device companies traditionally sought European validation first, then brought products to the US. Neurent went US-first, partly because regulatory barriers are now lower there than in Europe, and partly because the commercial model benefits patient, surgeon and system simultaneously in a way Europe does not offer.

Sharon Cunningham was candid about why most of her company’s original ideas died. The problem was not science but economics. Products better for patients were discontinued because they would not be reimbursed or prescribed, since the economic benefit to the prescriber was insufficient.

Frank Keane offered the sharpest strategic lesson. His company acquired a Massachusetts business for its technology, but what it really bought was their address book and the thousand hospitals it opened. US revenue has risen fivefold since.