Avid Bioservices Inc (CDMO)

[Protector]

SUNRISE Enrolment & Eventing Simulation 1st/2nd Look-in

JUN 16th, 2015 SUNRISE day #533, 1 year, 5 months and 16 days into the clinical trial.

ATTENTION: This post holds simulation data parameters which are NOT EXACT PREDICTIONS and depending on the precision of some assumptions the real results may differ.

NOTE: This simulation has been update to date and the basis for eventing has been changed. Before the 33% and 50% events needed for 1st and 2nd look-in have been calculated on the total of the patients (582). Now they are calculated on the X-value (non-disclosed by PPHM) that we fill in below 582 and for which the data below holds an ASSUMED value (guestimate). This lowering of the 582 to X has a BIG impact on the 1st and 2nd look-in readiness. Actually for BOTH there should have been sufficient events at this moment instead of the previous simulation where second look-in was more for Sept/Oct 2015.

 
  FACTUAL DATA                                                              SOURCE  
  Start Date                             : 31 Dec 2013                      (PPHM PR 30 DEC 2013)  
  Start enrolment                        : JAN 2014                         (PPHM PR 30 DEC 2013)  
  Planned End of enrolment               : DEC 2015                         (clinicaltrails.gov)  
  Planned End of trial                   : DEC 2016                         (clinicaltrails.gov)  
     
  Announced # treatment centres          : more then 100.                   (Shan on a PPHM quarterly/CC)  
  Current # of treatment centres open    : 159                              (clinicaltrails.gov)  
  Most were open on                      : Q3/2014                          (clinicaltrails.gov history files)  
  Simulation Dates Used                  : Real Opening Dates per centre    (clinicaltrails.gov history files)  
  Start recruitment                      : recruiting as of opening day     (clinicaltrails.gov history files)  
  Last ON TRACK report by PPHM           : June 2015                       (Company UPDATE PR)  
    
  Total # patients in trial              : 582                              (clinicaltrials.gov)  
  Arm A                                  : Docetaxel+Placebo Control arm  
  Arm B                                  : Docetaxel+Bavituximab 3mg/Kg arm  
  Randomizing Placebo Ctrl/Bavi          : 1/1                              (clinicaltrails.gov)  
    
  # of planned look-ins                  : 2                                (Quarterly/Annual CC's)  
   
  
    
  ESTIMATES/ASUPTIONS FOR ENROLMENT SIMULATION  
  # patients available at centre open    : 0     
  Days needed to enrol 1 patient/centre  : 112 (about 3.5 month)            (IHub posters educated consensus)  
  Max Patients per Centre                : 4                                (Average)  
  Placebo Eventing days                  : 210 (7 months)                   (Golfho's assumption)  
  Bavi Eventing days                     : 365 (12 months)                  (Golfho's assumption)  
  PPHM PII Placebo Eventing days         : 145 (5.6 months)                 (Chicago Spet 7th 2012 conference)  
  PPHM PII Bavi Eventing days            : 365 (12 months)                  (idem, between 1mg 11.5 and 3mg 13.5)  
  Historical Placebo Eventing days       : 110 (3.5 months)                 (for comparable ECOG groups)  
  
  ASSUMPTIONS FOR EVENTING 
  X-value for end of trial               : 500 of 582 
  1st Look In # Events                   : 165                              (since SK's presentation @ ROTH 33% of X)  
  2nd Look In # Events                   : 250                              (since SK's presentation @ ROTH 50% of X)    
    
  ESTIMATES FOR EVENTING  
  Patients days mid-of-period            :  56 (assumed enrolled mid of 112 days)  
  No eventing in first period            : 112 (patients in 1st 112 days excluded from eventing)  
    
  SIMULATION RESULTS per 16th JUN 2015  
  Patients Enrolled                      : 464  
  Evented Placebo Patients               : 180  
  Evented Bavi Patients                  : 100  
  Total Evented                          : 280  
  Remaining Events Needed for 1st Look-in:   0 (READY)  
  Remaining Events Needed for 2nd Look-in:   0 (READY)   
    
  Placebo Patients Alive                 :  52  
  Bavi Patients Alive                    : 132  
  Patients Alive                         : 184 
   
 

The quick progression, between this and previous simulation, in the patients numbers and alive vs evented is related to the fact that since this month the bulk of the centres that were opened mid of last years are all open since 12 months or longer. This speeds up the control arm events that are at 7 Months but now also starts to have more impact on the Bavi arm that is at 12 months. It is the non-linearity of the hockey stick enrolment that now kicks in into the linear progression model of patients as of their enrolment.

Conclusion
The 1st look-in should have taken place (as well with the new X as on the old 582), may still be in processing, or the patients didn't event as fast as assumed (then for the trial lets hope it are the Bavi patients). There still is the possibility of PPHM not announcing 1st look-in or keeping the announcement for another event such as the JUL quarterly as the 1st look-in (if a DMC CONTINUE advice) is probably not considered a material event since nothing that was projected to shareholders (EOT DEC 2016) changes.

The number of events for the 2nd look in should have taken place with the new X while that was rather for SEP/OCT with the old 582. However, if the second look-in is a DMC STOP advice or STOP order for ethical reasons then PPHM will receive way more data to process (they can unblind) and so results would in any case be for Q4/2015 or Q1/2016.

While I feel quite confident about the 1st look-in, I am less about this early 2nd look-in event reaching. Intuitively I'dd keep 2nd look-in events reached as achieved end Q3/2015 but the simulation says otherwise.
Then again I may have been focussing to much on the European/World treatment centres and forget that the US centres are open much longer, some up to 18 months now. So surprises are not impossible. Since this is a competitive enrolled clinical trial and since the maximum # patients per centres set to 4 is an ASSUMPTION, surprises are not impossible.

This emphasizes again how many parameters are involved and how changing a single one can impact the end result tremendously. Keep this in mind : It is a simulation with many assumptions - not FACT.