Hi Fosco,
Thanks for putting this together such efforts are much appreciated. I'll admit I'll need to review your calculations properly, but at the moment, the short answer is no I haven't taken account for drop-outs.
However I would think drop-outs to be low and definitely lower than 10% of the sample.
My reasoning for this would admittedly be based on the following logic and assumptions.
1. While small in sample there were no observed severe adverse events in the Phase II that would have warranted drop-out.
2. The IDMC did a safety review and recommended to continue the trial, therefore one would assume that on the safety aspect where concerning drop-outs would originate from, was not an apparent issues for them.
3. The patients are getting a series of injections during the few weeks between diagnosis and radio/chemo. Therefore they are just beginning their treatment and would I think have little reason to drop out during the course of the injections, especially when considering how patients routinely put up with the almost ubiquitous horrid physical and/or physiological side-effects of chemo in particular.
4. Following the 2-3 weeks of injections, there is no additional intervention. All that is needed is to keep in touch with the patients and see if there are any SAE that may be MK treatment or non-treatment related and if they have passed away or not. However again, no SAE's have yet been reported with MK. Thus I would not think requiring particularly burdensome requirements of the patients.
From these standpoints I would only realistically think drop-outs to be non-medical treatment related which can of course happen to any patient at anytime, e.g. they just don't consent to follow-up for whatever reason or have entered new circumstances that privy them from continuing with the follow-up. However 10% would seem quite high to me for non-medical treatment related drop-outs. Again if I were made to hazard a guess, just for arguments sake I would go with a high of 5%.
AI
Market Data 
Markets 
