Sherlock, the basket trial arithmetic you've walked through deserves more attention than it has received, and your bevacizumab correction is on point.
On the vascular normalization point: the AI characterization holds up against the published literature. Bevacizumab recalibrates abnormal tumor blood vessels, causing a brief normalization period during which previously tortuous and highly permeable tumor blood vessels become more organized and leak less, improving tissue perfusion and drug delivery. The decreased permeability seen with bevacizumab creates a pressure gradient across blood vessel walls and tumor interstitium conducive to drug penetration into tumor cells. So bevacizumab's contribution to the triplet is not primarily cytotoxic; it is architectural. It cleans up the vascular plumbing such that TAS-102 can actually reach the tumor cells which it is designed to kill.
Leronlimab opens the immune gate. Bevacizumab normalizes the delivery infrastructure. TAS-102 does the killing. Three mechanisms, three distinct failure points addressed simultaneously. That is the triplet's logic and you have described it correctly.
https://www.ajmc.com/view/fda-panel-votes-to-...gs-to-come
https://www.merck.com/news/fda-approves-merck...cer-whose/
On the monotherapy patient as the de facto control: this framing is analytically important and underappreciated. Four patients with a functional chemotherapy partner produced a 100 percent DCR at a dose the team itself considers suboptimal for liver-metastatic disease. The one patient without a chemotherapy partner did not. The CLOVER trial's design, ensuring no patient receives leronlimab without TAS-102, is the direct institutional lesson from that single data point. The basket trial is a five-patient experiment with an accidental control arm, and the control arm answered the question.
On the 350mg ctDNA clearance observation: three of the four patients who cleared ctDNA entirely were in the 350mg cohort. That could reflect receptor occupancy saturation at lower doses in patients with lower metastatic burden, or patient selection differences between dose cohorts, or a genuine dose-response inversion worth understanding before the dose-response data due this summer resolves it. Any of those explanations would be meaningful. The right move is to hold that observation carefully rather than extrapolate from it and let the summer dose-response data clarify the clearer architecture.
On the 60 percent plus ORR prediction: the basket trial extrapolation supports genuine optimism and the historical comparator of 6.3 percent from SUNLIGHT makes the bar look manageable. The honest framing is that the basket trial patients received FOLFOX as the chemotherapy backbone rather than TAS-102, and TAS-102's trifluridine mechanism of DNA incorporation and strand break induction is distinct from fluoropyrimidine-based replication inhibition. Whether TAS-102 plus leronlimab produces more RECIST-qualifying tumor cell death than FOLFOX plus leronlimab is a question the CLOVER data answers and the basket trial cannot directly predict. The direction your argument points is sound. The specific number awaits January 2027 at ASCO GI.
The FDA drooling scenario is not implausible in a population where 6.3 percent is the current ceiling. The boy's dream has better published support than most dreams in this space.
On the vascular normalization point: the AI characterization holds up against the published literature. Bevacizumab recalibrates abnormal tumor blood vessels, causing a brief normalization period during which previously tortuous and highly permeable tumor blood vessels become more organized and leak less, improving tissue perfusion and drug delivery. The decreased permeability seen with bevacizumab creates a pressure gradient across blood vessel walls and tumor interstitium conducive to drug penetration into tumor cells. So bevacizumab's contribution to the triplet is not primarily cytotoxic; it is architectural. It cleans up the vascular plumbing such that TAS-102 can actually reach the tumor cells which it is designed to kill.
Leronlimab opens the immune gate. Bevacizumab normalizes the delivery infrastructure. TAS-102 does the killing. Three mechanisms, three distinct failure points addressed simultaneously. That is the triplet's logic and you have described it correctly.
https://www.ajmc.com/view/fda-panel-votes-to-...gs-to-come
https://www.merck.com/news/fda-approves-merck...cer-whose/
On the monotherapy patient as the de facto control: this framing is analytically important and underappreciated. Four patients with a functional chemotherapy partner produced a 100 percent DCR at a dose the team itself considers suboptimal for liver-metastatic disease. The one patient without a chemotherapy partner did not. The CLOVER trial's design, ensuring no patient receives leronlimab without TAS-102, is the direct institutional lesson from that single data point. The basket trial is a five-patient experiment with an accidental control arm, and the control arm answered the question.
On the 350mg ctDNA clearance observation: three of the four patients who cleared ctDNA entirely were in the 350mg cohort. That could reflect receptor occupancy saturation at lower doses in patients with lower metastatic burden, or patient selection differences between dose cohorts, or a genuine dose-response inversion worth understanding before the dose-response data due this summer resolves it. Any of those explanations would be meaningful. The right move is to hold that observation carefully rather than extrapolate from it and let the summer dose-response data clarify the clearer architecture.
On the 60 percent plus ORR prediction: the basket trial extrapolation supports genuine optimism and the historical comparator of 6.3 percent from SUNLIGHT makes the bar look manageable. The honest framing is that the basket trial patients received FOLFOX as the chemotherapy backbone rather than TAS-102, and TAS-102's trifluridine mechanism of DNA incorporation and strand break induction is distinct from fluoropyrimidine-based replication inhibition. Whether TAS-102 plus leronlimab produces more RECIST-qualifying tumor cell death than FOLFOX plus leronlimab is a question the CLOVER data answers and the basket trial cannot directly predict. The direction your argument points is sound. The specific number awaits January 2027 at ASCO GI.
The FDA drooling scenario is not implausible in a population where 6.3 percent is the current ceiling. The boy's dream has better published support than most dreams in this space.
12
