A better way to treat cancer?

Factual Questions
1

If the current approach is to find treatments that are effective for many people, we end up with one-size-fits-all chemotherapy with horrible side effects.

But if we change the assumption that people are all the same, maybe we can find more effective, personalized drug therapies. Just like Thomas Edison would keep a workshop full of possible filament materials and try every conceivable combination (the 99% perspiration part) or like Dr House might try a number of treatments if desperately pressed for a cure, maybe the shotgun approach is better?

So, concentrate on building automated testing equipment that can test a large number of drugs or therapies against the patient’s own cell samples to see which ones kill the cancer without harming the healthy cells.

Seems like such a simple, obvious approach. What’s wrong with the idea?

2

Nothing wrong with that idea at all, and in fact a lot of the cutting edge of cancer research is working on similar approaches. The term of art is “personalized medicine”. Now, I don’t know of any group that’s trying your idea of testing drugs against a patients tumor in tissue culture. In the clinic, that would be inordinately time consuming and expensive.

Now there are research groups that test various cancer cell lines against a library of drugs. In particular, they’re trying to find specific drugs that are useful against cancers with specific mutations. And some hospitals are now genotyping the cancers of patients that come in for treatment, in the hope that we can later pull out some correlations between particular genetic tumor markers and which drugs worked. The eventual hope is that a patient can come in for treatment, and doctors will be able to determine which treatment is most effective by sequencing the tumor*.

There’s one good example of how well this approach can work – but unfortunately I can’t remember any of the specifics, and my google-fu is weak this early in the morning. Anyways, the crude outline is that for one particular type of cancer, there’s a new drug that’s very effective in most of patients, and totally useless in the remaining minority. Turns out that the difference is due to a single gene – patients with tumors that have one form will respond to the treatment, and those with another form will not respond. Hopefully someone else can remember the actual story…

  • an aside: sequencing is getting cheaper and cheaper, and now it costs something like a few tens of thousands to sequence a complete human (or tumor) genome. Someday it might be less than a thousand, at which point it would be reasonably economical to do for all patients.
3

And a whole ton of medical research studies are asking for permission to, besides testing their new drug on you, also test your blood samples/whatever to see that very thing, if your body might be more/less receptive than others to work with their new drug, and why that might be.

So it is a great idea… we just don’t know yet for the vast majority what makes any given drug work better or worse on someone and what makes the side effects better or worse on someone. It’ll take time to work all of that out.

4

The premise of the OP is wrong.

Current treatments do their best to tailor treatment to the specific type of cancer and to the specific patient.

Next!

5

It’s trivially true that oncologists do the best they can, but there’s still a whole lot of room for improvement. Right now the best treatment involves identifying the type of tumor. Unfortunately tumor categorizations are pretty crude – all that can be identified are some pretty broad characteristics, like size, density, histology (what it looks like under a microscope), etc. The problem is that for a given type of cancer, as defined in the clinic, there are multitudes of different genetic causes that could cause it. And (as the current thinking goes) the genetics are what really determine how a tumor will respond to different treatment.

So, currently, an oncologist can identify a cancer type, but in a lot of cases there’s not much good data about which treatments work best. For a given cancer, we might know that treatment X works effectively in 40% of cases, Y works in 20%, Z works in 30%, and some proportion is effectively untreatable*. Which works best for a given patient is pretty much unpredictable, so treatment breaks down to “try all the options until one works”. We don’t know why one treatment will cure one patient, but exarcebate the tumor for another patient.

The hope is that with better understanding of genetic causes of cancer, and better genetic tests for patients, we can make better treatment decisions. We’ll instead be able to say that one particular treatment will be effective for nearly all cancers of a given genotype. And we’ll also be able to identify new targets for currently untreatable cancers.

This won’t happen overnight of course, but there are a lot of very smart people making good progress, a little bit at a time.

But the OP’s idea? Actually rather good idea, it just happens to be very very expensive – costing something like six or seven figures. Which is too expensive to be used in treatment, but within the budget of well-funded research groups, so it’s an approach they use to understand the genetics of cancer. Or, conversely, I know of many groups that take a converse sort of approach, and test a given experimental drug against libraries containing hundreds of different cancer cell lines.

*Totally made up but not unrealistic numbers. Some tumors are much more treatable (treatment Q works almost always), and others are untreatable (nothing works reliably, but one patient in a hundred might respond very well to something you try randomly).

Usual disclaimer: I’m not a cancer researcher, but most of the people in my department are, so I’ve talked to them a lot and seen a their work in talks and posters.

6

Mammalian cell culture is hard, and every cell line from each individual patient will require optimization of growing conditions. That’ll be expensive.

7

When my SO was dying of cancer, we researched this extensively and communicated with this doctor, who was very nice and personally accessible.

This all sounds very good in theory but there are some issues in practice.

The first is getting the actual cell samples. In order to do a full test panel a good sized tumor sample is required and must be surgically obtained. Thisis hardly ever easy and in some cases it requires major surgery, which is hard on people with cancer.

The patient must also be chemo free for a month before the samples are extracted and of course cannot do chemo while recovering from surgery. This 2 month delay in starting chemo can be devasting.

It is espcially difficult to get the samples on early stage cancers where the tumors are small and contained and the cancer can be spread via the required surgery. Fine needle biopsy is usually used for tumor typing and that is safe procedure but a chemosensitivity assay requires a MUCH larger sample than can be obtained by a fine needle biopsy and the surgery can spread the cancer cells to surrounding tissues

If you can get to a sample the testing can be a strong predictor of what chemo drugs WON’T work and can spare the patient from suffering through an ineffective chemo regimen. If the chemo doesn’t kill the cancer cells in vitro then it’s not likely to work.

The reverse isn’t really true though because the tumor samples aren’t a living growing tumor. While a drug may appear to be very effective in testing it can be ineffective in practice. This is simply because many cancers are just really aggressive and they just grow faster than the chemo can kill them.

My SO wanted to do the chemosensitivity testing but getting the sample was a problem, then as his pancreatic cancer progressed he had to have another major surgery anyway and the surgeon determined that they could get an effective sized sample from a liver metasis.

However, he did not survive the recovery from major surgery and died two days after getting the test results. However, the test result claimed that the chemo he was using was actually the most effective for him so his cancer was just overgrowing the chemo, there was no magic bullet that was missed.

Another thing is that most chemo is effective at first, that is why it is so commonly used to shrink tumors before surgery. The real issue is that it becomes ineffective quite rapidly… the cells that are susceptible to the chemo die, then the only cells reproducing are those that are resistant to the chemo so you end up with a resistant tumor really fast.