Anavex Life Sciences Corp (AVXL)

[Investor2014]

Let's dive in and separate CGI-Anchoring and EOT vs. AUC, two very different concepts.

Dr.M still stumbled over explaining RSBQ, RSBQ-AUC, CGI, and CGI anchored scores.

Starting with the FDA guidance: Patient-Reported Outcome Measures: Use in Medical Product Development to Support Labeling Claims, which is what Anavex are referring to on slide 8 of the Top-line AVATAR readout slide deck, as well as on below slide from the obligatory Anavex readout rescue / explainer deck follow up.



There is guidance on Anchoring, but not EOT vs. AUC. When Missling says the endpoint for Rett is harder to achieve (than for tronifinetide), it is due to the Responder Analysis Threshold as discussed in slide 9 of same deck. I will get to AUC vs. EOT below.

E. Planning for Clinical Trial Interpretation Using a Responder Definition
Regardless of whether the primary endpoint for the clinical trial is based on individual responses to treatment or the group response, it is usually useful to display individual responses, often using an a priori responder definition (i.e., the individual patient PRO score change over a predetermined time period that should be interpreted as a treatment benefit). The responder definition is determined empirically and may vary by target population or other clinical trial design characteristics. Therefore, we will evaluate an instrument’s responder definition in the context of each specific clinical trial.

The empiric evidence for any responder definition is derived using anchor-based methods. Anchor-based methods explore the associations between the targeted concept of the PRO instrument and the concept measured by the anchors. To be useful, the anchors chosen should be easier to interpret than the PRO measure itself. For example, the number of incontinence episodes collected in incontinence diaries has been used to determine a responder definition for PRO instruments assessing the annoyance of incontinence. A 50 percent reduction in incontinence episodes might be proposed as the anchor for defining a responder on the PRO instrument. Confirmation of this anchor approach in early clinical trials can provide the basis for the proposed responder definition in the confirmatory trials.

Another anchor-based approach to defining responders makes use of patient ratings of change administered at different periods of time or upon exit from a clinical trial. These numerical ratings range from worse to the same and better. The difference in the PRO score for persons who rate their condition the same and better or worse can be used to define responders to treatment. Patient ratings of change are less useful as anchors when patients are not blinded to treatment assignment.

Another set of approaches to defining a responder are distribution-based methods that use, for example, the between-person standard deviation or the standard error of measurement to define a meaningful change on a scale. Distribution-based methods can be used to categorize these changes as small, moderate, and large and often can be combined with anchor-based estimates to provide confidence in the responder definition. Distribution-based methods for determining clinical significance of particular score changes should be considered as supportive and are not appropriate as the sole basis for determining a responder definition.

Alternatively, it is possible to present the entire distribution of responses for treatment and control group, avoiding the need to pick a responder criterion. Whether the individual responses are meaningful represents a judgment, but that problem is present with almost all endpoints except survival. Such cumulative distribution displays show a continuous plot of the percent change from baseline on the X-axis and the percent of patients experiencing that change on the Y-axis. This display type may be preferable to attempting to provide categorical definitions of responders. A variety of responder definitions can be identified along the cumulative distribution of response curve.

Guidance on interpretation considerations for a clinical trial’s SAP is found in section V.E., Interpretation of Clinical Trial Results.
25

Now for AUC vs. EOT, starting with the Q1 cc Q&A.

Yun Zhong

Okay. And then switching to the Rett syndrome study. I believe though press release announcing over enrollment had the language that with the FDA's input, you are using the primary endpoint. So I wanted to confirm that the primary endpoint is RSBQ and you see similar to -- or the same to the one used in the EBITDAR study? And so has the FDA agreed that the AUC, the modified RSBQ scale can be an appropriate endpoint for Rett syndrome study?

Christopher Missling

Yeah. We have it described in clinicaltrial.gov, and it was also never change in clinicaltrial.gov for the EXCELLENCE study. It is the RSBQ primary endpoint, and the CGI is key secondary endpoint over the course of the trial.

Yun Zhong

Is that the same endpoint that was used in the AVATAR study?

Christopher Missling

Slightly different. So it's actually the measurement over time from beginning to end of trial. AUC.

Yun Zhong

Not AUC?

Christopher Missling

Not AUC.

Yun Zhong

Not AUC?

Christopher Missling

Exactly, yes. Because the study is large enough that it can carry the signal by itself without AUC.

So I think we got that cleared up in this exchange. There will be NO resorting to flicking between EOT or AUC in the EXCELLENCE trial, like Missling fancied for the AVATAR readout.

The reason of course for that is that AUC makes zero sense in a short trial with only one mid-point data sampling and one EOT., which would be subject to a binary outcome of either AUC or EOT being stat sig positive. That is of course unless both AUC and EOT shows success or failure. Jin Kun has already been useful teaching Missling a lesson in stats. That will be especially important after EXCELLENCE as the time comes to present the collective data from the three Rett trials to regulators without obvious statistical nonsense.

Actually no even sure if the EXCELLENCE 12 week trial has a mid-point data sampling event, can't find any slide with the design. Of course if there isn't AUC is not even possible.

In any event AS WITH THE AVATAR and US Rett trials, EXCELLENCE endpoints are all "Change from baseline to End of Treatment (EOT)". No AUC hanky panky!