Ken, my AI buddy, really likes you!
Your reasoning is directionally sound, but I would temper the expectation that the poster will necessarily contain complete ctDNA and PD-L1 results for all 66 patients.
How many patients were probably enrolled by the end of March?
The strongest public clue is CytoDyn’s April 30 presentation:
22 patients already had Week-8 RECIST data available.
To reach an eight-week scan by April 30, those patients generally had to begin treatment by approximately early March.
The company also reported ctDNA results from the first 19 City of Hope patients and stated that PD-L1 had increased numerically in a majority of the patients evaluated.
Therefore, we can say with reasonable confidence that at least 22 patients had enrolled by early March, and probably more because some patients enrolled during that period would not yet have reached their Week-8 scan by April 30.
My best estimate would be:
Approximately 35–45 patients enrolled by March 31, with a reasonable broader range of perhaps 30–50.
I would not claim a more precise number without a site-by-site enrollment graph. Enrollment probably accelerated substantially in March and April, because CytoDyn reported target enrollment complete by April 30 and later described the total as just over 60 participants.
What might have been available for the May 12 abstract?
Using a 45-day PD-L1 observation point, only patients treated by approximately March 28 could theoretically have generated that follow-up by May 12.
That suggests the abstract might have included PD-L1 results from roughly:
30–45 patients, before accounting for missing or unevaluable samples.
If the relevant sampling window was 60 days, the cutoff moves back to approximately March 13, perhaps leaving only 22–35 potentially evaluable patients. At 90 days, the number would be considerably smaller.
One important qualification: the earlier mTNBC analysis discussed PD-L1 changes occurring over a 30–90-day period, but that does not necessarily mean CLOVER has only one fixed 45-, 60-, or 90-day PD-L1 measurement. CLOVER may collect serial blood samples, and the April 30 presentation already said that numeric PD-L1 increases had been observed in a majority of patients.
Could the eventual poster contain all 66 patients?
For Week-2 ctDNA: potentially close to all 66, assuming:
every patient had measurable baseline ctDNA;
the Week-2 sample was obtained;
the patient remained on study long enough;
and the laboratory analysis passed quality control.
Thus, the poster could reasonably report something such as “60 of 66 evaluable”, rather than literally 66 of 66.
For PD-L1: the eventual poster could contain substantially more mature data than the abstract, especially because ESMO is not until October. However, “all 66” still is not assured. Some participants may lack usable paired samples, discontinue early, die, or have assay failures. In clinical posters, the relevant denominator is usually the number of evaluable paired samples, not total enrollment.
My overall interpretation
The Investors Hangout post is making a reasonable and potentially exciting inference, particularly concerning the Week-2 ctDNA dataset. The likely May 12 abstract population was probably materially smaller than 66, while the October poster could be far more mature.
The potentially powerful poster would be one showing:
consistent early ctDNA reductions across most evaluable patients;
a dose relationship between 350 mg and 700 mg;
PD-L1 increases across a substantial paired-sample population;
correlation between ctDNA reduction, PD-L1 induction, Week-8 scans, and clinical outcomes;
encouraging durability rather than merely an early biomarker change.
That truly could be a doozy. But the most important issue will not be whether the denominator is exactly 66. It will be whether the biomarker changes consistently predict objective tumor control and whether results compare favorably with the expected performance of Lonsurf plus bevacizumab alone.
Your reasoning is directionally sound, but I would temper the expectation that the poster will necessarily contain complete ctDNA and PD-L1 results for all 66 patients.
How many patients were probably enrolled by the end of March?
The strongest public clue is CytoDyn’s April 30 presentation:
22 patients already had Week-8 RECIST data available.
To reach an eight-week scan by April 30, those patients generally had to begin treatment by approximately early March.
The company also reported ctDNA results from the first 19 City of Hope patients and stated that PD-L1 had increased numerically in a majority of the patients evaluated.
Therefore, we can say with reasonable confidence that at least 22 patients had enrolled by early March, and probably more because some patients enrolled during that period would not yet have reached their Week-8 scan by April 30.
My best estimate would be:
Approximately 35–45 patients enrolled by March 31, with a reasonable broader range of perhaps 30–50.
I would not claim a more precise number without a site-by-site enrollment graph. Enrollment probably accelerated substantially in March and April, because CytoDyn reported target enrollment complete by April 30 and later described the total as just over 60 participants.
What might have been available for the May 12 abstract?
Using a 45-day PD-L1 observation point, only patients treated by approximately March 28 could theoretically have generated that follow-up by May 12.
That suggests the abstract might have included PD-L1 results from roughly:
30–45 patients, before accounting for missing or unevaluable samples.
If the relevant sampling window was 60 days, the cutoff moves back to approximately March 13, perhaps leaving only 22–35 potentially evaluable patients. At 90 days, the number would be considerably smaller.
One important qualification: the earlier mTNBC analysis discussed PD-L1 changes occurring over a 30–90-day period, but that does not necessarily mean CLOVER has only one fixed 45-, 60-, or 90-day PD-L1 measurement. CLOVER may collect serial blood samples, and the April 30 presentation already said that numeric PD-L1 increases had been observed in a majority of patients.
Could the eventual poster contain all 66 patients?
For Week-2 ctDNA: potentially close to all 66, assuming:
every patient had measurable baseline ctDNA;
the Week-2 sample was obtained;
the patient remained on study long enough;
and the laboratory analysis passed quality control.
Thus, the poster could reasonably report something such as “60 of 66 evaluable”, rather than literally 66 of 66.
For PD-L1: the eventual poster could contain substantially more mature data than the abstract, especially because ESMO is not until October. However, “all 66” still is not assured. Some participants may lack usable paired samples, discontinue early, die, or have assay failures. In clinical posters, the relevant denominator is usually the number of evaluable paired samples, not total enrollment.
My overall interpretation
The Investors Hangout post is making a reasonable and potentially exciting inference, particularly concerning the Week-2 ctDNA dataset. The likely May 12 abstract population was probably materially smaller than 66, while the October poster could be far more mature.
The potentially powerful poster would be one showing:
consistent early ctDNA reductions across most evaluable patients;
a dose relationship between 350 mg and 700 mg;
PD-L1 increases across a substantial paired-sample population;
correlation between ctDNA reduction, PD-L1 induction, Week-8 scans, and clinical outcomes;
encouraging durability rather than merely an early biomarker change.
That truly could be a doozy. But the most important issue will not be whether the denominator is exactly 66. It will be whether the biomarker changes consistently predict objective tumor control and whether results compare favorably with the expected performance of Lonsurf plus bevacizumab alone.
11
