Friday, 16 December 2016

Optune OS for rGBM (at first recurrence)

I initially dismissed the idea of using the Optune device because the OS for rGBM in the EF-11 trial (as shown on the Optune website) was only 6.6 months vs. 6 months using chemotherapy.

However, since reading the PRiDe (Patient Registry Data) analysis (and Stephen's summary on of Optune users I'm becomming more optimistic. The PRiDe analysis is based on a separate database of 457 patients using Optune outside of clinical trials. The PRiDe analysys parses patient data by the number of recurrences and indicates an OS for patients at first recurrence  of 20 months (from the time of recurrence).

The source of this comes directly from the  PRiDe study:

The table indicates that Optune works substantially better for 1st recurrence patients than for patients with 2 or more recurrences.

It appears that the reason the OS of 6.6 months from the EF-11 study is so much shorter (6.6 vs. 20 months) is because only 9% of the patients in the EF-11 study were at first recurrence (i.e.  if 91% of the patients were at 2 or more recurrences and if these patients do much worse, then the OS of the entire group is heavily weighted towards patients with more than one recurrence - resulting in a much lower overall surviva).

The source of this is also shown within the PRiDe study:

It was pointed out to me that there are some flaws with the PRiDe data:

1. The research was based on patients outside of a clinical trial, so these patients may have used additional treatments.

2. The OS of 20 months comes from limited follow through data in that the 20 month OS is really based on the average of a range of 14 months to 26 months from the Kaplan-Meier data. I’m a little ignorant when it comes to Kaplan-Meier data, but somehow they use averages when there is insufficient follow-up data . It was explained to me that the 20 month OS is likely overstating the true OS, but it can still be concluded that the OS for first recurrence using the Optune (and possibly additional treatments) was at least 14 months.

So my understanding of the data is that patients using the Optune at first recurrence would have an expectecd OS (from recurrence) of closer to 14 months than the 6.6 months from the EF-11 study.

The PriDe paper is in the library that Stephen W put together.

Mike B


  1. Thanks for the reminder Mike, I really need to update the PRiDe section on the Optune page on Astrocytoma Options and will do that now.

  2. Kaplan-Meier methodology "censors" patients at the time of last follow-up and assumes those patients will have the same risk as the patients still being followed up. When there are little marks on a Kaplan-Meier curve, those marks mean that a patient was not followed up beyond that point in time. That's why the "stair steps" of a Kaplan-Meier curve get much larger on the further out time points - because of censored data due to lack of follow-up. The accuracy of the survival curve is also less at those farther time points for the same reason - there are more assumptions being made about the survival of patients who were not followed up long enough. This is why they're called Kaplan-Meier "estimates".

    It's not surprising that patients at first recurrence in this study did better than those at later recurrences. I think you'd see the same trend in any clinical trial. As far as trials for recurrent GBM, the only somewhat fair comparisons are between trials where all the patients are at first recurrence.

  3. More Optune data:

    Mike B

  4. How can we access the Patient Registry data? Empire Blue Cross/Blue Shield is denying coverage and we need to send off final documents by Wednesday. It would be great to include this. Thanks

    1. You can download the study here:

      But this was a retrospective non-randomized study so doesn't carry the same weight as randomized phase 3 trials (EF-14).

      You might try bringing this up also at the Novocure yahoo discussion group run by Al Musella, as there might be others in that group that have had similar issues.

  5. Dear friends! Please advice the contact details of professor Ben Williams! Thanks a lot.