🧬 The week surgeons removed a man's lungs and said "trust the process"

Another week in biotech and apparently the theme is "what if we just... took the whole thing out?" Surgeons kept a man alive for two days with zero lungs (on purpose), Spanish researchers obliterated pancreatic tumors in mice by blocking every escape route at once, and Nike tried to convince neuroscientists that foam bumps on a shoe sole can rewire your brain. Meanwhile, a disgusting plant bacterium invented an entirely new category of biological goop, and AI just told the MS community it's been treating one disease when there are actually two. Science is undefeated at making us feel behind on the news.

RESEARCH

There's a point in every medical case where the textbook runs out of pages. For a 33-year-old Missouri man whose influenza B spiraled into necrotizing pneumonia, that point came when his lungs started, in his surgeon's words, "liquifying." His heart had already stopped once. The team at Northwestern Memorial Hospital got it beating again, looked at the CT scan, and made a decision that sounds like it belongs in a screenplay: remove both lungs entirely.

Dr. Ankit Bharat and roughly 15 staff performed a 12-hour bilateral pneumonectomy, connecting the patient to a custom-built Total Artificial Lung system that works nothing like standard ECMO. Where ECMO supports lungs that are still in the chest, Bharat's TAL drains blood from the right heart, oxygenates it externally, then returns it to the left atrium through two separate conduits. A self-regulating shunt prevents blood flow backup (basically a highway exit ramp for your circulatory system). They even stuffed saline-filled tissue expanders into the empty chest cavity to keep the heart from flopping around. Resourceful.

Within 48 hours, something remarkable happened. With the infected lungs gone, the sepsis source disappeared. Blood pressure stabilized. Kidneys recovered. The team performed a successful double-lung transplant, and nearly three years later, the patient is alive, independent, and breathing on someone else's lungs.

The innovation grew directly from COVID. Bharat performed the first US double-lung transplant on a COVID patient in June 2020 and discovered through molecular analysis that some ARDS lungs sustain irreversible damage identical to end-stage fibrosis, which no amount of waiting on ECMO will fix. The TAL gives those patients a shot they previously didn't have. To his credit, Bharat published every technical detail openly: "There is nothing proprietary about this. Anybody can replicate it."

The caveats are real, though. This is n=1. It requires a world-class surgical team, round-the-clock ICU monitoring, and donor lungs arriving on time. Mortality in this patient population exceeds 80%. But as Bharat put it: "In my practice, young patients die almost every week because no one realized that transplantation was an option." Sometimes the radical move is the only move left.

RESEARCH

Pancreatic cancer has a five-year survival rate of roughly 13% and a decades-long habit of humiliating every drug thrown at it. So when Dr. Mariano Barbacid's team at Spain's National Cancer Research Centre announced they'd achieved complete, lasting tumor elimination in mice with no resistance developing for over 200 days, the biotech world collectively leaned forward. Then, hopefully, read the fine print.

The approach is strategically elegant. KRAS mutations drive over 90% of pancreatic cancers but were considered "undruggable" for 40 years because the protein's surface offered nowhere for drugs to grab hold. Barbacid's insight: blocking one pathway just sends tumors scrambling to another exit. So his team blocked three at once. Daraxonrasib (a next-gen KRAS inhibitor already in Phase 3 trials with FDA Breakthrough Therapy status) hits the core driver. Afatinib (an FDA-approved EGFR blocker) seals the upstream escape hatch. And SD-36, an experimental STAT3 PROTAC degrader that doesn't just inhibit its target but destroys the entire protein, eliminates the backup survival system. Barbacid's analogy: a beam fixed at three points is far harder to break than one fixed at one.

The results across three mouse model types were genuinely unprecedented in pancreatic cancer research. Complete regression, no detectable tumor remnants, minimal toxicity. Any two-drug combination eventually failed. Only the triple worked.

Now for the asterisk. SD-36 has never been tested in humans, and no PROTAC drug has ever been approved for anything. Over 92% of drugs that work in mice fail in people. Chemist Simon Maechling put it well: "Mice are a filter, not a finish line." He estimates 8-10 years before meaningful clinical data. Barbacid, who isolated the first human oncogene back in 1982 and has some credibility on the subject, says 2-3 years to trial initiation with adequate funding. A public fundraising campaign has raised €2.5 million from over 50,000 donors, impressive but a fraction of the estimated €30 million needed.

The real significance isn't the headline ("cancer cured in mice" is a genre at this point). It's the framework: rational, simultaneous blockade of three independent survival pathways preventing the adaptive resistance that has defeated every prior KRAS strategy. That idea will shape clinical trial design for a decade, regardless of which specific drug combination crosses the finish line first.

NEWS & RESEARCH

Nike's Mind 001 ($95) and Mind 002 ($145) sold out instantly at their January launch, which proves that marketing works even when peer review doesn't exist. The shoes feature 22 independent foam nodes per sole that Nike claims activate the sensory cortex, suppress the Default Mode Network (the brain region responsible for mind-wandering), and induce "calm, focused" mental states. They were developed over a decade by Nike's Mind Science Department, staffed with neuroscientists and perception researchers, and tested with EEG equipment.

Sounds impressive. One problem: Nike has published zero peer-reviewed papers supporting any of these claims. No methodology. No sample sizes. No statistical analyses. Just internal EEG data that nobody outside Nike has seen.

Dr. Atom Sarkar, a neurosurgeon at Drexel University, published the most thorough independent takedown. Yes, the soles have thousands of mechanoreceptors that signal the somatosensory cortex. But "influencing movement is not the same thing as enhancing cognition," which depends on distributed networks across the prefrontal cortex, parietal lobe, and thalamus. His most cutting observation: the increased sensory input might actually increase cognitive load, drawing attention toward your feet rather than freeing mental resources. So the shoes might literally do the opposite of what Nike claims.

Technical critics on the MOVEMENT Movement podcast piled on. Functional podiatrist Dr. Emily Splichal noted the nodes are far too large and widely spaced for genuine two-point discrimination stimulation (which requires roughly 1mm precision through the SA1 Merkel disc at the bottom of the feet). The design also ignores the toes entirely, despite toes having the highest density of neural representation in the somatosensory homunculus, except for your fingertips, lips, and mouth. Harvard's Dr. Irene Davis was blunt: "Show me the data."

One of the most intellectually honest takes came from Psyche magazine, which compared Nike's approach to placebo construction. The elaborate narrative (Mind Science Department, neuroscientist staff, EEG language, Erling Haaland endorsement) functions exactly like the therapeutic ritual that magnifies placebo effects. And here's the twist: research on open-label placebos shows they work even when people know they're placebos. So the shoes might "work" not despite being scientifically unsupported, but because the story around them is so well-constructed. Consumer reviews confirm the sensation is real ("you feel it"), though nobody seems confident it's actually doing anything cognitive. This is fine.

RESEARCH

Here's a love story for the science nerds. Two UC Davis professors met at a new faculty training session before the pandemic, bonded over their "mutual love of goop," and five years later discovered an entirely new category of biofilm that's been quietly devastating global agriculture to the tune of over $1 billion annually.

Ralstonia solanacearum is ranked the #2 most important plant pathogen worldwide, affecting more than 200 crop species, including tomatoes, potatoes, bananas, peppers, and tobacco. It causes bacterial wilt, a disease where plants just... collapse and die within days. The mechanism has puzzled scientists for years: how does this bacterium spread so devastatingly fast through plant vasculature?

Turns out, the answer is goop. Plant pathologist Tiffany Lowe-Power and chemical engineer Harishankar Manikantan discovered that Ralstonia produces the first microbial biofilm ever characterized as a viscoelastic fluid rather than an elastic solid. Every previously known biofilm in microbiology behaves like a structured solid. This one flows. Its EPS-I exopolysaccharide creates a shear-thinning material that, at the pressures naturally present inside plant xylem vessels, moves readily through the plumbing while simultaneously clogging it. "They cause the plant version of a heart attack," Lowe-Power explained, "they clog up the vessels and cause plants to wilt and die."

Manikantan, whose NSF CAREER Award bio cheerfully states, "I love goop of all forms, including saliva, foams, lung surfactants, and tears," used Silly Putty to explain the physics. Bounce it, and it's solid. Leave it on a table, and it flows. The question is what timescale matters, and for Ralstonia's goop inside xylem, the answer is: it flows.

Graduate student Matthew Cope-Arguello developed a beautifully simple diagnostic. Grow the bacteria on a plate, tilt it. If the biofilm drips, you've got the pathogenic strain. Non-pathogenic relatives produce normal, solid biofilms that stay put. No fancy equipment required.

The implications are significant. With the pathogen surviving in water for up to 40 years, no effective chemical treatments, and climate change expanding its range into previously unaffected zones, new control strategies are desperately needed. This discovery suggests researchers could target the physical properties of EPS-I itself (disrupting the viscosity-elasticity balance) rather than trying to kill the bacterium outright, which hasn't worked. Bacteriophage therapy targeting Ralstonia is already under development. Understanding the goop might finally give us an edge against one of agriculture's most persistent villains.

RESEARCH

For decades, neurologists have classified MS into relapsing-remitting, secondary progressive, and primary progressive categories. Turns out those labels describe symptoms, not biology, and a UCL-led study published in Brain just found that underneath those clinical labels sit two biologically distinct disease trajectories that predict dramatically different futures and treatment responses.

Lead researcher Dr. Arman Eshaghi used an AI model called SuStaIn (Subtype and Stage Inference) to analyze 634 MS patients by combining brain MRI data with blood levels of serum neurofilament light chain (sNfL), a protein released when nerve fibers are damaged. The model identified two subtypes that cut across traditional categories. The "early-sNfL" subtype shows elevated nerve damage markers early, develops corpus callosum damage, and has a 144% increased risk of new brain lesions compared to the other group. It's more aggressive but, crucially, also more responsive to treatment. The "late-sNfL" subtype starts with grey matter volume loss before the blood marker rises, suggesting a slower, more insidious neurodegenerative process that may need neuroprotective drugs rather than anti-inflammatory ones.

The combined MRI-plus-blood-test model significantly outperformed MRI alone at predicting disability, with the correlation improving from 0.231 to 0.420. Both relapsing-remitting and secondary progressive patients could belong to either biological subtype, which means two people with the same clinical diagnosis might have fundamentally different diseases requiring fundamentally different treatments.

"MS is not one disease," Eshaghi stated, "and current subtypes fail to describe the underlying tissue changes, which we need to know to treat it." The MS Society called this part of "growing evidence supporting a move away from existing descriptors toward terms that reflect the underlying biology." More than 2.8 million people worldwide live with MS. If validated in larger populations, this subtyping could reshape how treatments are selected and how clinical trials are designed.

The caveats: the study excluded primary progressive MS entirely, the sample size is modest at 634 patients (a prior study used 9,390), and the trial data came from Merck KGaA, which funded the research. But the study builds on a decade of AI-driven disease subtyping that has already identified distinct biological trajectories in Alzheimer's and Parkinson's, and the direction of travel seems clear. We're moving from "what symptoms do you have?" to "what's actually happening in your nervous system?" which feels like progress, if overdue.

So there you have it. A man survived two days without lungs and walked out of the hospital, a 40-year "undruggable" cancer target finally cracked (in mice, with caveats), Nike invented the world's most expensive placebo, bacteria have been weaponizing goop against our salads this whole time, and MS might need to update its LinkedIn profile to say "two diseases."

The wildest part? This is just February.

Got opinions on foam nodes and brain waves? Want to argue about mouse models vs. human trials? Reply to this email. We read every response while nervously checking whether our houseplants are being attacked by viscoelastic fluid.

If you know someone who thinks science is boring, forward this. And if you’re still here, it means either you genuinely love biotech or our subject lines are unreasonably good. Probably both.

Keep questioning everything (especially shoe-based neuroscience),

Prateek & Jere

P.S. If your shoes start whispering about the Default Mode Network, that's not science. That's haunted footwear. Seek help