Discloser: I am long this stock; I have no inside information; I am a clinical researcher.
Assumption 1: Recently the company disclosed the trial entry data and indicated that as of 7-15-2017 there were 98 patients out of the 331 randomized, still alive. Utilizing this new data, I endeavored to model a worst-case DCVax-L GBM survival curve for all patients. Since 90% of the patients in this trial have received the experimental treatment DCVax-L, this trial is really a randomization of upfront early use DCVax-L vs Delayed used of DCVAX-L given after the first progression. For the purpose of creating this curve, one assumes that only the 98 most recently randomized in the trial are still alive. Hence this will provide us with the least amount of survival time as all their other patients were randomized at an earlier date. Furthermore, in these remaining patients, we know for a fact the amount of time for follow-up as we have been given there entry randomization date and we have been told that there will be 98 patients still alive by 7-15-2017.
Assumption 2: If you, therefore, assume that all remaining patients randomized early survived by an amount equal to the quoted standard of care (SOC) median survival of 15-17 months, one can generate a Kaplan-Meyer Survival curve for the entire 331 patients which will be the worse case scenario especially for the tail of the survival curve.
When one does this the 36 month tail survival for all patients is equal to nearly 30% . This tail long-term survival is essentially equivalent to what has been reported and published by the Optune dataset.
Please remember that this analysis is intended to return the worse case survival curve because we already know that many of the earlier randomized patients have lived longer than SOC.
THEREFORE this analysis would suggest that the total group of patients in the DCVax-L Trial will have overall survival curves much better than the one shown that has a tail equivalent to what has been published with the OPTUNE dataset.
I wish I could model this more precisely but the limited data points makes that impossible.
It does look like the survival will be quite good in this trial as a whole.
Summary statistics:
Total observed Total failed Total censored
331 233 98
Kaplan-Meier table:
Months At risk Failed Censored Proportion failed
15 331 78 0 0.236
16 253 77 0 0.304
17 176 78 0 0.443
20.26666667 98 0 1
20.507 97 0 1
20.618 96 0 1
20.729 95 0 1
20.73333333 94 0 1
20.84 93 0 1
20.951 92 0 1
21.062 91 0 1
21.173 90 0 1
21.284 89 0 1
21.395 88 0 1
21.506 87 0 1
21.617 86 0 1
21.728 85 0 1
21.839 84 0 1
21.95 83 0 1
22.061 82 0 1
22.172 81 0 1
22.283 80 0 1
22.394 79 0 1
22.505 78 0 1
22.616 77 0 1
22.727 76 0 1
22.838 75 0 1
22.949 74 0 1
23.06 73 0 1
23.171 72 0 1
23.282 71 0 1
23.393 70 0 1
23.504 69 0 1
23.615 68 0 1
23.726 67 0 1
23.837 66 0 1
23.948 65 0 1
24.059 64 0 1
24.17 63 0 1
24.281 62 0 1
24.392 61 0 1
24.503 60 0 1
24.614 59 0 1
24.725 58 0 1
24.836 57 0 1
24.947 56 0 1
25.058 55 0 1
25.169 54 0 1
25.28 53 0 1
25.391 52 0 1
25.502 51 0 1
25.613 50 0 1
25.724 49 0 1
25.835 48 0 1
25.946 47 0 1
26.057 46 0 1
26.168 45 0 1
26.279 44 0 1
26.39 43 0 1
26.501 42 0 1
26.612 41 0 1
26.723 40 0 1
26.834 39 0 1
26.945 38 0 1
27.056 37 0 1
27.167 36 0 1
27.278 35 0 1
27.389 34 0 1
27.5 33 0 1
27.611 32 0 1
27.722 31 0 1
27.833 30 0 1
27.944 29 0 1
28.055 28 0 1
28.166 27 0 1
28.277 26 0 1
28.388 25 0 1
28.499 24 0 1
28.61 23 0 1
28.721 22 0 1
28.832 21 0 1
28.943 20 0 1
29.054 19 0 1
29.165 18 0 1
29.276 17 0 1
29.387 16 0 1
29.498 15 0 1
29.609 14 0 1
29.72 13 0 1
29.831 12 0 1
29.942 11 0 1
30.053 10 0 1
30.164 9 0 1
30.275 8 0 1
30.386 7 0 1
30.497 6 0 1
30.608 5 0 1
30.719 4 0 1
30.83 3 0 1
30.941 2 0 1
31.052 1 0 1
25.6717551
Mean survival time:
Mean survival time (Months<30.940999999996)
Mean survival time (Months<30.940999999996) Standard deviation Lower bound (95%) Upper bound (95%)
20.391 0.375 19.656 21.125
Quantiles estimation:
Quantile Estimate Lower bound (95%) Upper bound (95%)
75%
50% 17.000 16.000 17.000
25% 16.000 15.000
AI
