Monthly Archives: April 2021

Patient selection bias of evolutionary based adaptive scheduling in cancer – benefit to public health?

In this blog-post we shall discuss an interesting approach to cancer treatment but also pose the question as to who is likely to benefit from such an approach.  Let’s first begin with the idea…

What is evolutionary based adaptive dosing/scheduling? The idea revolves around the concept that we have two or more types of cancer cells competing with each other, but only one population is sensitive to the treatment we give. (It must be noted that clinically competition hasn’t actually been observed, we will discuss this in a later blog-post.)  This drug sensitive population is in excess of the drug resistant cells and “out-competes” the resistant cells. The idea is you apply the drug to kill sensitive cells but then remove the drug once you see a certain drop in tumour burden.  This allows the sensitive cells to grow back and suppress the resistant cells. You then re-apply the drug and then stop again and continue to cycle. Note, that we are not changing the dose but simply deciding when to start-stop the dose. Let’s take a step back and discuss what the dose cycling represents before moving onto the main question in the title of this post…

Consider for example a hypothetical drug that has just been approved for a new indication and its schedule is continuous daily dosing. That’s the schedule but what about the dose? There are two choices: you could either use the maximum tolerated dose, or a dose based on exposure-response relationships if one is found. Now the latter does happen more than people realise, especially when there is an interest in combining drugs, in which case using the maximum tolerated dose may not be appropriate nor needed.

The adaptive therapy community has somewhat confused the choice of dose with the choice of schedule, this was apparent at their inaugural meeting CATMO 2020 (http://catmo2020.org/). As highlighted above their interest has been in the choice of schedule not choice of dose, they still use the maximum tolerated dose or the dose based on exposure-response relationships depending on the drug.  The maximum tolerated dose may well be different for a continuous schedule versus an adaptive schedule, that is the patient may be able to tolerate an even higher dose when there are dosing holidays introduced than when they are not. Let’s now move away from this long aside and get back to the main question…

What is the selection bias mentioned in the title? Well, in order to apply adaptive scheduling, you first need patients who actually respond to treatment, in fact you will want to ensure they have a reasonable depth of response i.e. reasonable number of drug sensitive cells. The response rate for most Oncology drugs is around 20-30%, this can increase in certain areas but let us consider this value as the “average”. It is only these patients, who respond to the original schedule, for whom adaptive scheduling can be applied. Patients who respond will have better prognosis (live longer) than those that don’t. Therefore, adaptive scheduling is only applicable to patients who gain the most benefit from the existing schedule. What about the patients who don’t have a strong depth of response and therefore have poor prognosis?  Should research funds and research time be spent on those that do best on treatment or those that don’t benefit at all?  What is better for public health: helping those that need it most, i.e. those that don’t respond to treatment, or those that need it least, i.e. those that do respond to treatment?