Why does T-DXd stop working? Paired biopsies before and after ADCs, presented by @PTarantinoMD at ESMO 2026, give a clear first answer: the tumor rarely loses HER2. Instead, several mechanisms coexist in the same tumor:
Drug pumps up: ABCB1 rose in 5 of 7 T-DXd-treated tumors, ABCC1 in 4 of 7.
Uptake down: MAN2A2, a glycosylation enzyme, fell in 5 of 7. In cell models, silencing it reduced T-DXd binding and internalization.
After sacituzumab, no recurrent pattern at all. For clinic: a repeat biopsy that only re-checks HER2 will usually come back "still positive" and explain nothing. The resistance happens after the drug reaches the cell. Seven patients, hypothesis-generating. But it tells us what the next biopsy should measure. #bcsm #ESMO26 @myESMO

In ~2.5 years, MASH has gone from its first FDA approval to a ~$1.5B annualized drug launch, $8B+ in upfront FGF21 acquisitions and a crowded 500+ asset landscape. But what's fascinating is the commercial / clinical split is happening much earlier than usual. By all accounts, the commercial market is barely penetrated. Madrigal (while biased) estimates 1% penetration - the prescribing and diagnostic infrastructure needed to identify the ~410K untreated patients in the US is in its infancy. Against this backdrop, we historically used to see a number of fast followers with near-identical trial designs and target patient populations, even in mid-stage development. Instead, even later-stage R&D is already planning for a more fragmented, multi-drug treatment environment canonically associated with more mature markets. Many trials are segmented by fibrosis stage, metabolic comorbidities and, importantly, response to emerging frontline treatments, including GLP1s (Wegovy was approved in MASH in August 2025). Planning for future treatment sequencing has always been part of the playbook, but the degree of competition in every corner of drug development these days has pushed forward the need to have a differentiated patient angle. Every shift in the SOC creates a new patient phenotype that did not exist before. In this case, GLP-1 exposed MASH patients will differ histologically from the untreated population and the bet is that the broad reach of GLP1s will flip these ratios over time (Novo indicated 1/3 of patients in efruxifermin's Ph2 study were already on GLP1s), as it may be a more natural discovery point for undiagnosed patients. And that means the next drug will need different comparators, patient selection criteria and, potentially, endpoints to answer emerging treatment questions, like: what additional benefit should a liver-directed drug deliver? When do you add it? An asset's competitive position isn't fixed the moment it enters clinical development, so staying nimble to skate where the puck is going is critical. Going back to MASH, we pulled together an interactive landscape covering mechanisms, fibrosis stages, efficacy, deals, China and upcoming readouts: sleuthinsights.com/analysis/… Some takeaways from that work: • F4 is the next expansion opportunity - current approvals cover F2/F3, but ~250K additional US patients have compensated cirrhosis • The next 12 months are packed with catalysts, including Inventiva's pivotal lanifibranor readout, Phase 3 data from Novo and Roche's FGF21s, and the expected Rezdiffra cirrhosis outcomes readout • GLP1s haven't yet boxed out liver-directed drugs. Rezdiffra is still adding ~6-7K net patients per quarter despite Wegovy's entry I've included a snapshot below but it's definitely worth playing with yourself! For those working in MASH, what are we missing? Particularly interested in which patient segments and treatment combinations you think will matter most over the next few years.

Where did 2025’s 30 best-selling medicines actually come from? Industry made more of the molecules than the “pharma just buys innovation” story suggests, but 28 of the 30 rest on biology first found in academic labs. Where a molecule starts tells you surprisingly little. Maybe an interesting set-up for AI and the next 15 years... asymmetriclearning.substack.…

FDA new drug approvals in Q3 2026 nature.com/articles/d41573-0… The third quarter of 2026 was the busiest of the year so far in terms of novel FDA drug approvals, with a total of 22, including anticipated blockbusters for pancreatic cancer, cholesterol lowering and narcolepsy

2024 to present, went down a rabbit hole of proxy documents from all the mergers and acquisitions in bioland after @daphnezohar tweet. To try to see if I can find any insights into what the turnaround period could be or what the outcomes may look like. I&I , Onco , rare disease and neuro been the hottest areas, will that continue? Nephrology have the most parties interested 16 per deal n=2. (Also why $vera m&a will have legs) Actually read below with the graphics in the subtweets that will be easier and better. Findings: The first mover usually loses once a second bidder appears. In the 18 competitive deals the first party to name a price won only 44% of the time. $nvs took $trml, $rgls and $mor from earlier bidders; $pfe took $mtsr from $nvo; $mrk took $cdtx from Company D and Terns from Party C. •Competitive processes run on a clock the company sets. Process letters with bid deadlines ($dcph, $akro, $cdtx, $trml, $mrus) and a best and final round are the standard tools. $cdtx committee added a rule that the winner had to clear the next bid by 5% or the top two went another round. •Single bidder deals price the gap with CVRs instead.CVRs appear in a larger share of single bidder deals; with no second bidder, the contingent payment is how the board closes a valuation gap (Poseida, $etnb , $verv, $bpmc, $aclx, $apls). •Exclusivity is a prize, not a default. It is granted in a similar share of both groups, and when it is granted it is usually short and bought with a price step ($dawn got $1 more for 8 days; $mor got $57 for exclusivity to Jul 11; $swtx got $50 instead of $49). Deals that refused exclusivity entirely did not give up any bump. •Competition does not make deals slower from first price. Median first price to signing is similar in both groups. What stretches competitive deals is the longer run up: more parties approached and more diligence before the board lets prices crystallise. More insights below.. Deals announced in Nov would have most likely had their 1st offer late Sept or Oct.