Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
Spidey, I couldn't wait
And I computed your SOC survival in TAB 4 of my spreasheet,
So now I have
1- Cel SCI initial assessement (which was proven wrong)
2- my assessment
3- UK data (same as lightrock)
4- Spidey
Do you agree with the log regression based on your figure : showing 62% at year 5 and 78% at year 1 ?
Here is my critical analysis
Feb 2018 with 208 events leads to 13,2% survival improvement
Your Feb 2017 events miss heavily the 133 mark (you get 162), as you only get 46 events in the 1 year period
Your endpoint is reached in November 2021. I don't know how you got Feb 2019 as you state 23 events per 3months : it's too much according to your model.
In conclusion : Unless my chart proven wrong, your SOC data shows way too much survival as it leads to end of 2021 for completion of trial and insufficient deahts in Feb 2017. Survival in SOC has to be lower, hence MK % survival improvement has to be better than 13,2%
Hi Spidey,
I am very glad that you computed your model with those figures. We need triple check.
Whether lightrock could try to compute something too this would be great.
While your model is in line with the 208 figure, is it also in line with the 133 figure of Feb 2017 ? This is important as it indicates the further trends and can indicate MK efficacy superiority (%). You could make any good SOC matching the 208, it's harder to make it match the 133 plus a endpoint in 2019.
What is needed is to match 133, 208 and endpoint in 2019
May I suggest that you to tweak your data (including SOC Stats) so that SOC + MK % effect make those three milestones match. (feb 2017, feb 2018 + end point in 2019)
in any case, in every model I made, endpoint is reached beyond feb 2019, earliest in April 2019, therefore I am quite confident we will not have news before, which is very good news in all our models
Re : "definitely believe you are incorrect on the understanding of the type of cancer we are looking at treating. "
please have a look at p 18, bottom
https://www.entnet.org/sites/default/files/NeckDissection_QuickRefGuide_highresFINAL.pdf
Or
https://www.cancercenter.com/head-and-neck-cancer/stages/tab/stage-4/
then, please come back to me...
Kind regards
Fosco
To get 208 events in Feb 3rd 2018, as figure is pretty low, I need 25% efficacy from MK group otherwise SOC is not good enough even very good SOC stats (and too good SOC leave us to 2020).
With 25% I am getting 131 events in Feb 4 2017 VS 133 (the real figure from CEl SCI) as there had been 75 events between those two dates !!!
Now 90 deaths are required from this date of 3 Fev 2018 to get to 298. I guess Geert did a linear regression and 90/75 + feb 2018 = April 2019 = H1 2019. I don't think it'll be linear but W&S !
Another good news !!!
Something that passed totally unnoticed. Thanks Andy for the tip !
In Feb 2018 208 events had occurred and in feb 2017 133 as per a safety brochure published by CEl SCI?
(All links are in my spreadsheet).
And that's very good because 208 is not as much as expected for this date !
Therefore, I have been applying 03 Feb 2018 as a date and need a 25.2% MK benefit in order to get this 208 figure in my "realistic model 2".
Then when I computed 04 Feb 2017 I got 131 deaths on my simulation which is very close to the real figures !!!!! (77 events between those dates instead of 75)....
The only thing a bit in contradiction with CEl SCI is that it leads us towards October 2019 to reach 298. Might be the log model not 100% accurate, or a trick to conceal us that they 'll need to finance again in 2019 at low pps. However I can now understand the enthusiasm from Cel SCI management, why they bought so many shares and why they got the go from FDA and the IDMC !!!! Survival is really good and can only be explained by MK, otherwise it could be SOC survival but then the trial would end in 2020 or 2021 as too many people from the 795 pool would survive to be able to finish the trial in 2019 !!!!
Staticmirror
I have made long DD on this one based on inclusion criteria in CT site (N&T)
It is important to know because otherwise you might be biaised by survival data from Stage I and II cancers (small tumors, non metastatic, no nodes)
My conclusion is that the inclusion criteria aims at Stage III and IVa H&N SCC
Hence valid statistical data has to include stage IVa and stage III and exclude stage I and stage II
I suspect that too optimistic survival models might include stage I and II
lightrock & spidey
Wow, I am not getting headaches with those spreadsheets !
@Lightrock Ok so I understand you took UK data for your calculations, so I built a third sheet with the UK data and the earliest date computed (+10% efficacy) for end point falls March 20th 2019. It was quick to do as I already had the UK log regressions.
@Spidey : ok I see that depending on MK efficacy you are getting the 298 to Feb (+10%) or May 2019 (+15%). I am surprised you got February as your survival model is even more optimistic than the UK one.
Therefore, given that I consider UK data as very "optimistic survival" as most of patients won't come from the UK (and US by the way)but from countries with less performant healthcare system, I think there are reasons to be optimistic by now.
Of course the savvy investor will wait till end of February... if nothing happens till then.
Cheers
Fosco
sure, it's in post #17698 below
Hello There !
It's me, Fosco
- to my "supporters", I ammended my spreadsheet and added an "introduction" tab, I have put more explanatory notes and rounded survival figures to avoid figures such as "3,7 survived" as someone didn't like this
My spreadsheet is more flexible than the others I have seen so far (hence the reason I made it) in the sense that it allows to compute a MK survival VS SOC rate and an End Date dynamicaly as survival is calculated from a formula (log curve) and not using a value table. Therefore one can ammend it everyday that passes and whatch number improve.
- @lightrock : i 'd love to see your spreadsheet and assumptions to reach Feb, (I love spreadsheets . More seriously I 'd like to see your source of survival data, eventually to compute it in my spreadsheet (please feel free to do it as well if you wish)
- @ Spideyboy : their are 3 arms, two receive LI (4/7), one of the two (1/7) do not receive CIZ in order to assess CIZ efficacy but is not counted to assess LI efficacy
- @ On TREND : 298 is the number of the expected deaths, not the number of tested patients
Nice day to all, hopefully this week brings an uptick
Fosco
lol TB ... was wondering whether my english was so poor that I could not understand what he meant . Thanks for saving my day
Fosco
hahahaha Sab...
Concerning the number of events there are at least two times I believe they knew this number : first time when they mentioned in July 2016 : "Last year we reached the recruitment target of 880 patients, but, based on pertinent data evaluated at that time, we observed that the death rate of the enrolled patients was lower than had been expected ". Overall survival was much better than expected (how could they know otherwise ?) and second time when mentionned early 2019 as the date when endpoint should be reached. There can't be other way to state this without knowing the number of events. That might be unverified events though as the fact that a patient does not attend a follow up medical meeting does not necessarily mean he is dead (hence double check needed in the final count).
Ok, so 270 is just speculation. No official communication on this number. Nevertheless I don't understand why CEl SCI does not disclose this figure.
May be because it needs triple checking that those people are really dead prior to any annoucement ?
Nostradamus : Michel de NostreDame, famous astrologer from the 16th century
"we had over 270 events" : where did you get this info Sab ?
Very Interested to know as well as which date this number was reached, as I get exactly this number in one of my models for October 2018 which reaches endpoint by April 2019
sab :
"I read here that the trial has concluded already (June?), "
* last patients were recruited in 2016 (september I think)
"now we need to WAIT until the total 298 events have occurred "
* Yes
"with a "potential" current count of 290)"
* I don't know where this figure comes from (or this is speculation from someone from the M/B). My understanding is that Geert does not know how much events have occurred yet.
"....so we have already about 4 months past that date (PLEASE: correct me if I am wrong). "
* in the video dated 10/24 I think I heard that he stated between 6 or 9 months
Hi Sab63090
I meant in my google sheet simulation (with the assumptions I mentionned),
The endpoint not being reached yet (and within 3 to 9 months according to Kersteen) this means that probably overall survival is better than expected
Fosco
Thanks Spidey
I wouldn't expect many drops-out, as well, before all are treated, but afterwards.
After the last SOC treatment and many years many things could happen in this fragile population which can include elderlies: such as the guy dying of another cause or the guys just doesn't showing up to say "hello I'm alive !" (for any reason either good or bad for him). I remember reading from Taïwan study that they had to search for death certificates database to confirm death and cause of death of patient (of course they didn't show up to inform the doctors).
So I have no clue of how much it could be, some website mention enormous attrition and others figures like a few %. I'd be glad for it to be less than 5% of course as this would confirm survival is currently doing well
Fosco
RobotDroid,
One can compute the sheet by downloading it and changing the figures in red (as such it is read-only) and try to reach the 298 events
For instance I ran it with a 10% improvement of survival with MK
12% rounded drop-out rate
Cel Sci SOC expected survival rates from CT(which they openly said it was too pessimistic in a 2017 press release)
End point (298 events) was supposed to be reached by July 2018
Which means, as obviously we have passed that date, one and/or the others :
- SOC Survival is better than expected by Cel SCI (Please have a look at the other SOC survival models in the other folders)
- MK survival is better than 10%
- Drops-out are worse than 12%
Hope it helps
Fosco
Spideyboy
Did you include Indeed, patients who dropped out ?
Something not to forget if you want study to be realistic
Feel free to play with my model here and give me the output. I am particularily interested in cancer survival figures you might compute.
https://docs.google.com/spreadsheets/d/1hhOKr9DXRWKJbhaSjYDIojfCE8bPR6z84Eg2FB3jbwo/edit#gid=1489531853
(first folder : 1 have put a 10% attrition, figures based on low SOC survival)
Fosco