Wednesday, September 3, 2014

Predicting outcome to therapy

There is an interesting article out today via pubmed that talks about the ability to predict response to chemotherapy in CLL (summary editorial here, actual article here).

As I've been talking about quite a bit, it involves knowing the detailed molecular genetics associated with an individual patient which you can now test for.

I wanted to briefly point out the difference between "prognostic" and "predictive" biomarkers.  First of all, a biomarker is any sort of test such as a blood test, type of scan, etc that correlates with a biologic behavior.  Ideally those biomarkers can help influence treatment decisions to make more personalized treatment decisions.

"Prognostic biomarkers" give a general sense of how a patient is likely going to do with their disease.  In CLL, you might look at things like ZAP-70 or CD38 to estimate how a patients disease is going to behave over time.  You might say that they are "high risk patients" or "low risk patients" for progression / survival etc. but it doesn't necessarily tell you if a specific therapy is going to be useful.

"Predictive biomarkers" on the other hand help with specific therapeutic decisions.  They can be either positive or negative predictors.  Patients with 17P deletion for instance do not experience durable benefit from FCR therapy so in essence it "predicts" an inadequate response to this treatment.  Patients with TP53 or BIRC3 mutations have similar negative outcomes therefore these are considered "negative predictive biomarkers."

On the other hand, some biomarkers may be "positive predictive biomarkers."  I am VERY interested to find out if CLL patients with NOTCH1 mutations derive unique benefit from new research medications that target that mutation specifically.  In other cancers like melanoma or lung cancer, the presence of BRAF or EGFR mutations are required before you can even use certain meds that have very high levels of activity in those specific settings but are useless or even harmful if a patient lacks the mutation.

This particular CLL paper looks at the German CLL8 study which compared FC to FCR and was the first paper to establish a survival benefit in CLL.  It offers a wealth of new insight into the newer prognostic and predictive biomarkers that have emerged in CLL.  We find out that patients with SF3B1 mutations respond well, but experience more rapid relapse.  Patients with NOTCH1 mutations don't derive benefit from the addition of rituximab to FC.  Patients with 17P deletion / TP53 mutation and IgVH unmutated CLL have shorter survival following FCR therapy compared to patients lacking these abnormalities.

Both the technical article and the associated editorial are worth the read.