- the drug is for "ALK positive Non–small cell lung cancer (NSCLC)"
- NSCLC is 85% of all lung cancers, of which 3% to 5% are ALK+.
so it would appear to apply to at most 4% of all lung cancers. Are the numbers correct? Even so, huge deal for tens of thousands of people each year.
People complain about expensive drugs "only extending life by a few months". But they forget that each advance adds on top of that.
So over the last few decades survival has gone from a few months to several years.
Great, so this is a targeted approach that works 60% of the time (in the case of Lorlatinib) in one highly specific area of the highly diverse space of possible cancers, many of which have never been observed before and for which you could never enroll enough patients to have a robust clinical trial.
How can this approach be generalized and scaled, meaning that for a given patient with a given cancer, you can run an algorithm and synthesize molecules that will work?
In CS the paradigm involves knowing how to solve an unseen problem instance, not just remembering solutions to specific, well-studied problem instances.
The paradigm is called rational drug design (+ personalised medicine) and its scope is quite limited at the current level of technology. To achieve what you describe, we need to solve all of the following:
- a way to recognise the molecular mechanism of a particular cancer and identify potential drug targets
- a way to characterise the targets - in the previous step you identified a protein, now you need to know its shape (X-ray crystallography (potentially years), AlphaFold). We need to know how the target should be modified to disrupt the disease.
- a way to design a molecule that binds the target specifically and in the desired way
- a way to synthesise the molecule quickly & efficiently
- a way to predict the molecule's interactions at all stages of metabolism, taking into account the patient's individual phenotype (different people have significant variations in drug-digesting enzymes)
Significant (decades at the current rate) progress in all of these areas is needed before this sort personalised medicine has a chance of becoming feasible.