I was reviewing the bar graph last night that shows the dramatic ctDNA decreases from Sept 2026; the 28 patients with 2 week data. I started fooling around with the numbers, trying to come up with some rough ORR calculations.
The assumption I made was that anyone with an 85% decline would go on to shrink tumors 30% or more, and would be classified as a Partial Responder. Using this metric, 7 patients in the 350mg arm qualified, and 8 in the 700mg arm, for a total of 15 out of 28. I don’t own a calculator, so I did the multiplication with pen and paper (literally on the back of an envelope!). To make is easy I multiplied the 15 patients by 2.25 (28 patients x 2.25 = 63, a tad under the 66 enrolled in the trial). Using this back-of-the-envelope algorithm I got an estimated ORR of 33.75%. Not bad considering these are 3rd-line mCRC patients.
I plugged in some other cut-off numbers:
75% decline as the cut-off figure (18 total, 11 at 700mg, 7 at 350mg)--Estimated ORR of 40.50%
A 70% decline as the cut-off figure (22 total; 12 at 700mg, 10 at 350mg)--Estimated ORR of 49.50%
A 90% decline as the cut-off metric (11 total; 8 at 700mg, 3 at 350mg)--Estimated ORR of 25.75%
At the 90% cut-off it looks like you start to see the dose-response effect in the 700mg arm, with 8 patients reaching that metric and only 3 in the 350mg arm. But you also see 3 additional patients in the 350 arm with declines of 89%, 88%, and 87%. And the 350 arm includes 1 patient with a 100% decline, and 2 others with 99% declines. (Best in the 700 arm are 2 patients at 97%, and 1 at 96%). I personally hope we see a little more separation between the arms; as several have noted here, the FDA likes a clear dose-response benefit.
The above calculations make me wonder if Hoffman was kind of low-balling it when he mentioned the 16% figure. He did say something to the effect that the number might be better; I forget the exact quote. Tactful, and appropriate, on his part... given Cytodyn's history. Best to let the data do the talking. People in the Pharma Press would probably laugh out loud if he said what he really expects.
Well, I haven’t done a deep dive into what those kind of ctDNA declines really mean in terms of ORR or OS outcomes, just taking Dr Kasi’s comments seriously, and responding to his work. And his enthusiasm. It’s getting late or I'd run these numbers by AI. Perhaps someone else could.
Let me end with a saying I heard just yesterday. ‘Figures Lie, and Liars Figure.” Please do not take this as investment advice, or that the trial is going to be a slam dunk! These patients are very sick. The above is more a thought experiment than any kind of hard knowledge you can take to the bank. But I did want to bring some figures into the discussion. And I tried to explore a range of cut-off numbers—from 70% to 90%—reflecting those excellent ctDNA declines from the data available in Sept 2026. We will know more in just over a month.
The assumption I made was that anyone with an 85% decline would go on to shrink tumors 30% or more, and would be classified as a Partial Responder. Using this metric, 7 patients in the 350mg arm qualified, and 8 in the 700mg arm, for a total of 15 out of 28. I don’t own a calculator, so I did the multiplication with pen and paper (literally on the back of an envelope!). To make is easy I multiplied the 15 patients by 2.25 (28 patients x 2.25 = 63, a tad under the 66 enrolled in the trial). Using this back-of-the-envelope algorithm I got an estimated ORR of 33.75%. Not bad considering these are 3rd-line mCRC patients.
I plugged in some other cut-off numbers:
75% decline as the cut-off figure (18 total, 11 at 700mg, 7 at 350mg)--Estimated ORR of 40.50%
A 70% decline as the cut-off figure (22 total; 12 at 700mg, 10 at 350mg)--Estimated ORR of 49.50%
A 90% decline as the cut-off metric (11 total; 8 at 700mg, 3 at 350mg)--Estimated ORR of 25.75%
At the 90% cut-off it looks like you start to see the dose-response effect in the 700mg arm, with 8 patients reaching that metric and only 3 in the 350mg arm. But you also see 3 additional patients in the 350 arm with declines of 89%, 88%, and 87%. And the 350 arm includes 1 patient with a 100% decline, and 2 others with 99% declines. (Best in the 700 arm are 2 patients at 97%, and 1 at 96%). I personally hope we see a little more separation between the arms; as several have noted here, the FDA likes a clear dose-response benefit.
The above calculations make me wonder if Hoffman was kind of low-balling it when he mentioned the 16% figure. He did say something to the effect that the number might be better; I forget the exact quote. Tactful, and appropriate, on his part... given Cytodyn's history. Best to let the data do the talking. People in the Pharma Press would probably laugh out loud if he said what he really expects.
Well, I haven’t done a deep dive into what those kind of ctDNA declines really mean in terms of ORR or OS outcomes, just taking Dr Kasi’s comments seriously, and responding to his work. And his enthusiasm. It’s getting late or I'd run these numbers by AI. Perhaps someone else could.
Let me end with a saying I heard just yesterday. ‘Figures Lie, and Liars Figure.” Please do not take this as investment advice, or that the trial is going to be a slam dunk! These patients are very sick. The above is more a thought experiment than any kind of hard knowledge you can take to the bank. But I did want to bring some figures into the discussion. And I tried to explore a range of cut-off numbers—from 70% to 90%—reflecting those excellent ctDNA declines from the data available in Sept 2026. We will know more in just over a month.
7
