We can leave it at that. I think 3 is the better number but if you want it to be 5 to bolster your argument - fine. But 38 months is another egregious error. There are only 2 people in the entire study that, to date, have lived longer than 38 months. How you think this could be a median is beyond me? Maybe you meant 28, which would close. Never said 26 was hard and fast. Either way if you buy into the study that says pseudo median OS is 3t months the results look very bad and if you like the study that has it at 24 the results are - meh which is all I have been saying all along.
"There is simply no valid way to reclassify after the fact like you are trying to do."
Well of course there is. And I concede it is not definitive but it is certainly valid --that's why we have a word called "estimate" And we estimate every day. It helps if you use sound assumptions, (unlike most on this board) drawn from real life data (as I did) An argument can be made that the reported results are invalid because they don't include known mortality data from 5 trial participants. The fact that NWBO didn't classify doesn't mean you can't estimate what the effect of the known data would be if they were placed in each cohort. And it is sound to assume that the lower mortality are rapid because there is at least a 14 MONTH difference in OS between the 2 groups. Is it possible that they were pseudo ? Yes but that's not the way to bet (or model) it. Measuring the effect of those 5 on the median OS results based on sound estimates is more than valid. - it is essential if you are trying to truly understand the data.
And if you want to just lump everything together and look at the median OS for pseudo and rapid-- 18.3 months is not particularly exciting. In fact its right on the screws if you use 10 months OS for rapid and 24 for pseudo (indeterminate)
Your calks are off. there is only 1 survivor in the unclaasifieds If you cherry pick the two longest you get n-27 and the pseudo makes 28. There are only 11 living patients The median has been reached. Even if you take out my estimate of 3-5 of the pseudos as rapid, the median is still reached and becomes about 26 months as I stated earlier. Only if assume away more than 5 indeterminate do you get a an N that hasn't reached the median. Thats a fact.
huuuuuh? Yes you could have a new lesion in which case your "double" is nominally correct As long as you're imagining things why stop at one? Could be a quadruple. But of course it's all in your imagination since you have no breakdown of how many were second lesion vs 25% growth. But is it s a fact that 25% growth twice - is not a double. Depending on how the 25% growth was calculated It's either 50%.or 56.25% of the baseline scan.
And for a second time for the short bus crowd, you cannot determine, from the way NWBO wrote the poster whether the 25% increase on the second scan is measured from the baseline or the first post op scan. I know you don't understand the ambiguity but it is there.
should read OR the scan after the baseline (the first scan)
duuuuude Made me laugh out loud doubling down on schtoopid. How many mistake can you make where there is no number larger than 2? There is an ambiguity in NWBO's language0 you won't spot iy. It's not clear whether the second additional increase is compared to the baseline of the first scan each is equally plausible. If its compounded by referencing the first scan you make 2 mistake if compared to baseline scan you only make one that a fith grader would not make 25% and 25% i snot a double jesssssus man its 50% If its compounded that math is still simple
1x 1.25= 1,25 Ya with me so far?
1.25 x 1.25= 1.5625 or about a 56% increase either way donk its not a double.
I read your post and the mediana patient has been reached if you take out as many #$%$ from the from the indeterminate group an dive them to rapid. My estimate of remaining g rapid in the indeterminate cohort is 3-5 tops for all the (sound) reasons set forth elsewhere. How'd you like the study showing pseudo median OS is 37 months?
Once again keeping your perfect record of bad logic and erroneous statements. There is ZERO evidence that L is responsible for any increase or decrease in longevity. But you won't understand that. If they are responding better at all, they might be responding better to Temozolmide or radiation. It's a common mistake. The ONLY way to determine the effect of L is via double blinded trial of sufficient power
That is funny stuff. If you knew anything about the percentage of pseudo vs rapid you would'n look so much like a fool by believe that there is only 1 pseudo patient. OH and BTW 2, 25% pct increases is not a double. Now I know why some folks vote GOP--they just can't think.
Well sport the second group is "indeterminate" -- I suggest you look up the meaning. Under your "logic" the indeterminate group is 100% pseudo. Fine. Run with that. It's almost certainly not true. But so what? Whatever gets you through the night.
Didn't change the data. As a matter of empirical fact, it is NWBO who altered the data by excluding people from their results. Don't bother telling me that they couldn't include them, but that is the point. You CAN include them and make reasoned estimates about them. Well not you, but some folks.
Did it occur to you that you are trying to make a point by proposing inaccurate and irrelevant data? Magical screening methods . That's new one.
Sorry. It isn't sound to assume the REMAINING cohort is distributed along historical lines. The original N is 55. The indeterminate group has been culled, which has the effect of "enriching the pseudos in the remaining cohort. It is no longer a group you would find in a random sample. This fact seems to elude you. But feel free to estimate however you want.
As far as longevity, you make an argument that may be true but is very unlikely ( I could flip a coin 100 times and all come up heads!!!) given the known differences in median OS . So once again, feel free to make bad estimates.
You can ignore the 5 unclassified if you want but it is a fact that they exist and are either pseudo or rapid. So in the real world they are one or the other. The fact that NWBO didnt classify them doesn't change that fact. Making some reasoned estimates about them is what smart folks would do. Doesn't make them correct but is a worthwhile exercise if you have serious money at risk. And yes you can fiddle around with the indeterminates and take out 8 or 9. I don't think that is sound because, in the indeterminate group there would be a bias towards pseudo since you have already culled out the obvious rapid progrssors but hey make your own estimates. My numbers are sound and based on ALL the humans in the trial.
I also took a few illustrative tidies as comparators but there are studied out there that put OS in pseudo much higher--frinstanceconsecutively treated patients with pathologically diagnosed GBM.
Pseudoprogression was observed in 10 (12%) cases applying the stringent criteria, and in 18 (23%) patients when using the liberal criteria, in the cohort treated with RT/TMZ. Pseudoprogression was observed in only one patient treated with RT alone. The median time to pseudoprogression was 4 weeks after the end of RT. Patients with pseudoprogression had a median survival time of 28 months, compared with 12 months for patients without pseudo progression.
and this oneThe overall median survival was 19.9 months (95% CI 15.1-22.5). Median survival in 24 (35.3%) patients with pseudoprogression was 34.7 months (95% CI 20.3-54.1), significantly longer than the 13.4 months (95% CI 11.1-19.5) in 44 (64.7%) patients without pseudoprogression (P
"We" apparently does not include you my little friend.
PS you are never allowed to use the word "statistically" again since you are clueless on the subject
Believe what you want sport It's America and you are free to remain uninformed.. But the FACT is this---somewhere between 35 and 50% of those initially diagnosed as "rapid progressors turn out to be pseudo. Why a smart person would check this out and a really smart person like me would already know that.
Actually median OS for pseudo would improve to about 26 mos. My mistake. Better but not exciting.
you're wrong on the "double" see Patriot's post if you want to be informed. Or go to NWBO's website and READ. But like all here you don't really understand the numbers. If you did you would have some concerns. For instance, the 5 excluded patients didn't vanish, NWBO just didn't count them. But we know what happened to them. Of the five patients that were excluded, 3 died well before 10 months. If (as is likely) they were in fact "rapid progressors' and added to the N of the rapid progressors, the median would be somewhere around 13 months and if N was just one or two more and died in less than 10 months, the median would be 10 months. Just saying.
As for the "indeterminate".group, well there is no medical classification of "indeterminate" The fact that you didn't "determine" them is meaningless. They were, as a matter of empirical fact, either rapid or pseudo.. There were probably a few true rapid progressors but the majority would be Pseudo., since somewhere between 35 and 50% of rapid progressors turn out to be pseudo. If you assume the second groups is 100% pseudo, 21.5 months ain't good news. In fact it is lower than at least one study, of pseudo that had median OS at 24 months. A second study had OS at about 19 months.
If you really wanted to make a more accurate estimate, you would assume that a few of the 25 indeterminate are true rapid progress ors say 4 or 5. It would also be fair to assume they died pretty early. 8 indeterminates died before 14.6 months. If you assumed that most in not all of the 8 were true rapid progress ors ( a reasonable assumption given the disparity between the OS's of the 2 groups 8-10 for rapid and 19-24 for pseudo) and you removed them from indeterminate and placed them in "rapid", the OS for rapid plummets. The OS for the pseudo improves but only to about 21--right in line with historical OS for pseudo progressors
All in all the data is not compelling and is, in fact , troubling.