Monday, February 15, 2016

Surrogate Biomarkers, Diagnostic Biomarkers, Prognostic Biomarkers, and Predictive Biomarkers

A biomarker is a biologic molecule, such as a protein or gene, that is measureable in tissue, blood, or other body fluids, and is an indicator of some clinically significant condition. Biomarkers can be diagnostic, surrogate, prognostic, or predictive. Biomarkers can be very useful in clinical trials. They can be used as the inclusion criteria to identify the right patient population for the clinical study. They can be used as the efficacy endpoint (specifically the predictive biomarkers).

Surrogate Biomarkers: Biomarkers are easier to measure and can be used as screening or surrogate measures for more sophisticated, more accurate, but more cumbersome measures. For example, while the gold standard of diagnosis in oncology is a pathologic tissue review, a highly elevated prostate-specific antigen (PSA) level in the right clinical setting can be diagnostic of prostate cancer. Although the value of PSA level as a diagnostic biomarker is limited by its sensitivity and specificity, it can be an excellent surrogate biomarker for monitoring prostate cancer response to treatment. Surrogate biomarkers can be diagnostic, prognostic, or predictive. In clinical trials, the term 'surrogate endpoints' are also used. Surrogate biomarkers can be surrogate endpoints, but not all surrogate endpoints are surrogate biomarkers. For example, the imaging endpoints (e.g. MRI, CT Scan) may be used as surrogate endpoints, but they are not surrogate biomarkers.

Diagnostic biomarkers indicate if a disease already exists. They are often used for screening for diseases such as cancer. If diagnostic biomarkers are used for screening, they must have good sensitivity and specificity and also must be sufficiently noninvasive and inexpensive to allow widespread applicability.

Prognostic biomarkers indicate how a disease may develop in an individual case regardless of the type of treatment and show the progression of disease with or without treatment. In other words, prognostic biomarkers refer to markers that correlate with the natural progression or aggressiveness of a disease. In oncology, prognostic biomarkers are useful for informing patients about the risk of recurrence or median survival for their particular type of malignancy and for minimizing confounding factors when analyzing clinical trial cohorts or when prospectively stratifying patients in randomized clinical trials.

Predictive biomarkers are defined by their role in predicting a response to a given treatment. Therefore, these are most useful if they can be assessed before the initiation of treatment. Predictive biomarkers help to assess the most likely response to a particular treatment type. If we are looking surrogate endpoints for efficacy measure in clinical trials, predictive biomarkers are most useful.

When we discuss the biomarkers, it is necessary to distinguish whether or not they are diagnostic biomarkers, prognostic biomarkers, or predictive biomarkers. 

Medscape has an article by Tezak, Kondratovich, and Mansfield "US FDA and Personalized Medicine: In vitro Diagnostic Regulatory Perspective". The article included the following diagram to distinguish the differences between prognostic biomarkers and predictive biomarkers.


Predictive versus prognostic biomarkers.
Marker-positive population is marked in red, and marker-negative population is marked in blue. The figures only illustrate a few simple ways in which biomarker–therapy–outcome interactions might occur. Other factors (such as risk:benefit ratio, safety concerns, availability of other treatment and so on) that may affect assessment of the biomarker and therapeutic effect are not taken into account. (A) No biomarker effect tested. The effect of T versus S is assessed. T shows improved outcome (green arrow) compared with the S in all comers. (B) Prognostic biomarker. Only S is used to assess the effect of biomarker; the effect of therapy is not assessed. When the same type of care is used (regardless of whether there is treatment or no treatment), marker-positive population (dashed red line) shows better outcome than the marker-negative population (dashed blue line). Biomarker shows prognostic effect (yellow arrow). (C) Prognostic biomarker. The effect of S versus T is assessed in both biomarker-positive (red) and biomarker-negative population (blue). Similar therapy versus standard-of-care effect size is observed (green arrows), regardless of biomarker status. For the purposes of the point described, the therapeutic effect is the same, for example, in an 'absolute' survival sense (the green arrows are the same length). Biomarker-positive population has better outcome than biomarker-negative population (yellow arrows) regardless of whether the S or T is used. The biomarker shows prognostic effect, and there is no predictive biomarker effect (i.e., treatment effect is independent of marker status). (D) Predictive biomarker. The effect of S versus T is assessed, in both biomarker-positive (red) and biomarker-negative population (blue). T does not appear to improve patient outcomes over S in the marker-negative population (circled green arrow between blue lines T and S). T shows large improvement in patient outcomes when compared with S in marker-positive population (green arrow between T and S red lines). Biomarker shows predictive effect. (E) No biomarker effect. The effect of S versus T is assessed, in both biomarker-positive (red) and biomarker-negative population (blue). Similar therapy versus standard-of-care effect size is observed (green arrow), regardless of biomarker status, and T shows improved patient outcomes when compared with S. There appears to be no biomarker effect on patient outcomes in either S or T arm (marked by yellow circles). There is no predictive or prognostic biomarker effect.
Figures are simplified illustrations of the relevant points, and not depictions of biological data.
S: Standard of care; T: New therapy.
One of the slides from Roche is also a good summary for the differences among three type of biomarkers.

References:


Monday, February 01, 2016

Estimating the average treatment effects at two different visits and its implementation using MMRM model

Orkambi is a combination drug including Ivacaftor and Lumacaftor and is approved by FDA for the treatment of cystic fibrosis in patients less than 12 years with homozygous F508del mutation. The approval was based on two large pivotal studies with identical study design. The treatment effect is trivial, but statistically significant. One thing that is interesting to me is that the primary efficacy endpoint and the key secondary efficacy endpoint are based on the average of two different visits (visits 16 and 24).
  • The primary endpoint was absolute change in ppFEV1 from baseline at week 24 (assessed as the average treatment effects at week 16 and 24).
  • Average relative change from baseline in ppFEV1 at Week 16 and at Week 24
Typically, for a clinical trial with fixed treatment duration, the treatment effect will be assessed at one specified time point. In Orkambi pivotal studies, the primary efficacy endpoint was measured at post-baseline visits Day 15, Weeks 4, 8, 16, and 24. The treatment effect would usually be estimated at Week 16 or at Week 24.

In Vertex’s briefing book for FDA advisory committee meeting on May 12, 2015, the rationale of using the average of two visits was indicated as below:
Change in ppFEV1 at Week 24 was assessed as the average of the treatment effects at Week 16 and at Week 24 to provide a more precise estimate of the treatment effect at the end of the treatment period, given the inherent variability in ppFEV1. 
The average treatment effects at Weeks 16 and 24 was obtained from the mixed model with repeated measures (MMRM). It would be naïve if one thinks that the average value of week 16 and week 24 is calculated for each individual subject. According to FDA’s statistical review, the statistical methods were described as below:
For the primary efficacy endpoint, absolute change from baseline in ppFEV1 at Week 24, the primary analysis was to test the difference between each active combination treatment group versus placebo using a mixed model with repeated measures (MMRM). Both on-treatment measurements and measurements after treatment discontinuation (for subjects who discontinued dosing early) were included in primary analyses. The MMRM analysis included subject as a random effect, treatment, visit, and treatment-by-visit interaction as fixed effects, with adjustment for sex, age group at baseline, and ppFEV1 severity at screening. An unstructured covariance structure was assumed to model the within-subject errors. A Kenward-Roger approximation was used for the denominator degrees of freedom. The primary result obtained from the model was the average of the treatment effects at Week 16 and at Week 24.
The summary results were depicted at the following graph.


Based on the description above, the following SAS codes can be used to calculate the average treatment effect at weeks 16 and 24. Notice that in MMRM model, all observations for the change from baseline to all visits (day 15, weeks 4, 8, 16, 24) are used. The last two estimate statements are to calculate the treatment effect at week 16 or at week 24 separately. 

proc mixed data=FEV1;
     class subject treat visit sex agegrp ppFEV1_severity;
      model Chg_ppFEV1 = treat visit treat*visit sex 
                         agegrp ppFEV1_severity/ddfm=kr;
       repeated window / sub = subject type = un;
       estimate 'Average Treatment Effect at Weeks 16 and 24'
                             treat -1 1
                             treat*visit 0 0 0 -0.5 -0.5
                                         0 0 0  0.5  0.5/cl;
       estimate 'Treatment Difference at Week 16'
                             treat -1 1
                             treat*visit 0 0 0 -1 0
                                         0 0 0  1 0/cl;
       estimate 'Treatment Difference at Week 24'
                             treat -1 1
                             treat*visit 0 0 0 0 -1
                                         0 0 0 0  1 /cl;
run;