Ahjay – would like a second opinion?Its a good question that...

Ahjay – would like a second opinion?

Its a good question that arou011 has asked. Essentially if PFS is defined from CT scans 8 weeks doesn’t this muck up the calculation of PFS.

And the answer would be yes. But this is not how it is defined.

Lets assume that Can-004 follows the same basic outcome methodology to Can-003.

The interim data poster from Can-003 specifies the how progression of disease (PD) is determined.

Two serum values of CA-125 = 2× ULN performed at least 1 week apart, regardless of CT scan results, based on Gynecologic Cancer Intergroup (GCIG) criteria, or

Increasing clinical or radiological evidence of disease since study entry regardless of serum CA-125 per Response Evaluation Criteria in Solid Tumors (RECIST) criteria.

I am assuming that serum levels are tested weekly and in practice this is what will mainly determine PFS.

At the next level of detail you can go down to the GCIG and Recist criteria to see in turn how each of these are defined.

For the truly obsessed there is a FDA guidance related to outcome measures for oncology trials where some of the complexities around event timing and how it interacts with PFS are discussed.

With your Poisson distribution question arou011 – I dread crossing swords with someone with background in theoretical maths and applied statistics but aren’t you mixing up the idea of a count model here with essentially survival analysis. There are all sorts of distributional assumptions you can make or not make in a survival analysis.

But this is not where the complexity is – the statistics are fairly trivial. The hard bit is more related to pragmatic design issues - working out a plan for missing data (holes where appointments were missed - the disease progresses - when did it happen? - and particularly if this different between the two groups). And then sensitivity analyses which can then be performed testing different ways of modelling investigator decisions.

All this stuff is in the SPA – which the likes of us will never see – unless someone here could break into PRR HQ.

Apologies in advance if I have misunderstood your point here.

All the best Southoz