Top Story Today
Roche Walked Into a Two-Horse KRAS Race and Beat Both Horses at Once
Roche's divarasib beat both Amgen's sotorasib and BMS's adagrasib in a direct, randomized Phase 3 trial in previously treated KRAS G12C-mutant lung cancer. The drug hit its primary endpoint (progression-free survival) and showed a statistically significant improvement in overall survival at interim analysis. That's both survival measures, against both approved competitors, in one shot. Specific numbers are being saved for a medical conference, but the company confirmed the results were "clinically meaningful." Preclinical data had suggested divarasib was 5 to 20 times more potent than the first-gen drugs; now the gold-standard trial backs it up.
Why it matters: A head-to-head Phase 3 win on both PFS and OS against two approved competitors could rewrite treatment guidelines for KRAS G12C lung cancer and position divarasib as best-in-class, with the real prize (the multi-billion-dollar first-line market) still ahead.
Read more →Clinical Breakthroughs
The Cancer Drug Where Every Single Patient's Tumor Vanished
GSK's dostarlimab hit its primary endpoint in the AZUR-1 rectal cancer trial, reinforcing a streak that started at Memorial Sloan Kettering: all 42 patients who completed treatment saw their tumors disappear entirely, with no surgery, no chemo, no radiation. The catch is that only dMMR/MSI-H tumors (a small subset) respond this way, but for those patients, immunotherapy is replacing the scalpel at major cancer centers.
Read more →The Hair-Loss Drug That Came Back from the Dead
Q32 Bio's stock surged after its alopecia drug bempikibart posted dramatically better results at 36 weeks than the underwhelming 24-week data that tanked the stock last year. The kicker: zero serious adverse events and early signs that patients kept their hair even after stopping treatment. In a market dominated by JAK inhibitors carrying black box warnings, that safety profile could be a game-changer.
Read more →Regulatory Watch
The FDA Went Silent for Three Months, Then 14 Rejection Letters Dropped at Once
After an unnamed pharma company challenged the FDA's transparency policy through a citizen petition, the agency largely stopped publishing drug rejection letters for approximately two to three months. Then 14 complete response letters appeared on the public database. The legal fight over whether the FDA can publish these letters at all has an October deadline, and the outcome will determine whether investors get to read the FDA's actual words or go back to relying on corporate spin.
Read more →Leqembi Just Escaped the Infusion Center
The FDA approved an at-home starting dose for Leqembi, letting Alzheimer's patients skip the infusion center from day one. It's not a minor convenience upgrade: the infusion bottleneck was the single biggest reason only about 13,500 patients had cumulatively received the drug through late 2025 despite 7 million Americans living with Alzheimer's. Revenue impact won't fully kick in until 2027 when Medicare reimbursement catches up.
Read more →Global & Pandemic Response
Big Pharma Spent Billions Shopping in China Before Lunch
Chinese biotech out-licensing deals surged in the first half of 2026, with Q1 alone exceeding $60 billion and the pace on track to surpass last year's full total of roughly $137 billion. AstraZeneca's $18.5 billion CSPC deal, BMS's $15.2 billion Hengrui pact, Pfizer's $10.5 billion Innovent haul: every major Western pharma company now has a serious China strategy. One-third of new compounds in U.S. pipelines now originate from Chinese companies, up from roughly 10–12% in the early 2020s.
Read more →Ebola Has No Vaccine for This Strain. Oxford Just Started Fixing That.
Congo's Ebola outbreak (1,926 cases, 702 deaths) just spread to two new provinces, and the approved vaccines don't work against the Bundibugyo strain. Oxford launched the first-ever human trial of a Bundibugyo-specific vaccine, built on the same platform as its COVID shot. The Serum Institute of India already has 620,000 doses manufactured and ready if the data holds up.
Read more →