Graduate student takes a problem-solving approach to his brain cancer | Hacker News Reader

Graduate student takes a problem-solving approach to his brain cancer

newsoffice.mit.edu

69 pointsuser_23571139 comments

Not the best place to vent this but having a conversation with my mother the other day about cancer sent my mind racing. I don't think cancer will cured for a very long time.

She had watched some evening news show talking about injecting polio into cancer cells, which is some really cool research [1], but in explaining it to me it brought my attention that she held some serious misunderstandings of what cancer is. She said something like, "They would inject polio into the cells so that the body would recognize the cancer as foreign, because for some reason our bodies don't recognize cancer as foreign."

I just wonder how many people have absolutely no idea that cancer is just a damaged cell behaving naturally. It's not foreign. And with that in mind, it's kind of ridiculous to think we can do some kind of checksum of DNA at a cellular level, for every cell, for everyone. Cancer is here to stay, because it's as fundamental to life as aging and growing.

[1]: http://www.cancer.duke.edu/btc/modules/Research3/index.php?i...

actually, tumors are being immunosurveilled constantly, and when the immune system identifies those cells, it tries very hard to eliminate them. The immune system doesn't know or care that the cells are "foreign", just that the cells have properties, or tags that correlate with properties, that make those cells dangerous.

While I generally agree there will be no single silver bullet technology by which we "cure all cancer" in a general way with little to no side effects, I think it's not inconcievable that improving our ability to identify potentially damaging our cells will contribute a lot to keeping people cancer-free.

And while I also generally agree that cancer is inextricably intertwined with tissue growth, as an existence "proof", cancer has never been observed in naked mole rats; this observation alone is significant.

The significance of the Naked Mole Rat finding is that not only does it not develop cancer, it also has a very long life-span for a rodent (decades).

Wow thanks for commenting, the bit about the naked mole rats is really interesting.

Can you explain the significance of injecting polio into cancer cells? Is it just to make the immune system definitely see that the cell is bad?

The theory is that cancer cells develop a protein barrier that hides their anomaly from the immune system. The polio virus is thought to damage that and the immune system thus starts to recognize them as mutated cells.

For the patients in the 60 minutes special who experienced remission (if you will), they received one injection, then it took over a year for the tumors to go away.

That's not the problem: the problem is "cancer" is not one illness, with one cause. It's a grouping term for literally thousands, and likely millions, of specific variations which result in unique cells which evade normal anti-cancer immunoresponses.

Even if you developed a vaccine which sensitized the immune system to one type of cancer, chances are the "type" would be limited to pretty much just the 1 person you tested it on.

The MIT article is heartening, but Steven's speech about his experience is even better. If you have 12 minutes to watch it -- it's worth it: https://www.youtube.com/watch?v=-L-WFukOARU

The talk was both profoundly touching, informative and surprisingly cheerfully optimistic. He really did convince me that MIT and Harvard are some of the best places to be at if you have cancer in terms of resources.

A highlight from that talk was the state of the art robotic surgery facility at Brigham and Women's neurosurgery dept.

http://www.brighamandwomens.org/departments_and_services/neu...

He got operated on by one of the best neurosurgeons specializing in glioma (his condition):

http://physiciandirectory.brighamandwomens.org/Details/886

He 3d printed his brain tumor because his doctors gave him access to his health data. Such problem solving!

You jest, but this is not the first time a 3D printed tumor has allowed for doctors to better solve the problem:

"Man Saves Wife’s Sight by 3D Printing Her Tumor.

When his wife was misdiagnosed, Michael Balzer used 3D printing and imaging to get her well"

http://makezine.com/magazine/hands-on-health-care/

With your own health data and access becoming easier to revolutionary medical advancements, you should expect to see more stories like this. You are your own best health advocate, and your time is worth more than a single doctor's experience. You can dig longer and further than someone who needs to get you in and out in 15-30 minutes.

EDIT: A good example is my wife's endometriosis. There is no cure, just treatment. I'm working on a non-profit program to have genetic testing done for volunteers to isolate the parts of the genome responsible for endometriosis (statistical analysis; just like 23andme does) so that research can move to removing the responsible genome section using the recently discovered CRISPR protocol: http://phys.org/news/2015-03-crispr-technology-precise-genet...

My girlfriend is basing her entire PhD around spatial modeling of tumors. It's not as simple as you would expect. She's actually working in a physics lab despite being in a bioinformatics program because the PI is focused on material packing research. Turns out physics does impact cancer!

Understanding how cells and populations of cells grow, from a 3 dimensional and spatial perspective, is a fundamental step in understanding our biology because it's research rooted in math and the real world. What do cells actually, physically look like, and how do they interact? These are questions biologists barely have the answers to, if they have them at all, and only recently have we unlocked the ability to try to answer them. Cancer research is a main beneficiary of super computing because it allows us to build simulations much closer to reality than we previously could.

Don't take the ability to 3d print a tumor for granted!

Tumors are also, sadly, the less scary subcase of cancer. Single tumors can be targeted, shrunk, and often taken out. It's the undetectable "lucky" cells that might survive chemo and radiation that make the disease so hard to fight (because, as with antibiotic resistance, they've now been selected for resistance to the drugs/radiation).

If you are pointing out how university news departments are hype machines, you are right. I've seen that this is particularly true of MIT. I don't track these things, but I've seen too many over-hyped research stories coming out of MIT news and Tech Review. Granted, they're not the same organizations, but somehow they both have this common trait.

yes - this is the kind of BS that I alluded to because he is working with Neri Oxman.

Neri is an architect who says her work is 'biologically inspired'. BS Central.

The problem here is that he is joining Neri's group. Neri is an architect who tries to design articles inspired by biology. It's all wishy washy with her (at-least from the science perspective).

So all that enthusiasm is going to Zilch, unless he wants to be a brain skeumorphia artist.

Again, this might inspire some awesome things, but maybe not traditional science research. I think most of HN can buy that.

One symptom, phantosmia, is the only information he needed to suspect seizures or cancer.

Is there any reason why we don't do full-body scans on a regular basis? Is it too much radiation? It seems to me that (a) this is the only way to deal with silent cancers like many pancreatic kinds that are discovered at Stage IV (at which point, the prognosis is terrible) and (b) it would be cheaper to do the scans than to deal with the (expensive, and often ineffective) late stage cancers.

I'd imagine that there was a time when this was impractical because of the amount of radiation involved, but is that still the case? Don't we have extremely low-radiation X-ray devices that can make regular full-body scans a reality?

There are multiple reasons, not only the direct harm from radiation. The latest book from Dr. Welch "Less medicine, more health : 7 assumptions that drive too much medical care" explains some of them. The table of contents provides a good summary of his message:

ASSUMPTION #1: ALL RISKS CAN BE LOWERED

Disturbing truth: Risks can’t always be lowered—and trying creates risks of its own

ASSUMPTION #2: IT’S ALWAYS BETTER TO FIX THE PROBLEM

Disturbing truth: Trying to eliminate a problem can be more dangerous than managing one

ASSUMPTION #3: SOONER IS ALWAYS BETTER

Disturbing truth: Early diagnosis can needlessly turn people into patients

ASSUMPTION #4: IT NEVER HURTS TO GET MORE INFORMATION

Disturbing truth: Data overload can scare patients and distract your doctor from what’s important

ASSUMPTION #5: ACTION IS ALWAYS BETTER THAN INACTION

Disturbing truth: Action is not reliably the “right” choice

ASSUMPTION #6: NEWER IS ALWAYS BETTER

Disturbing truth: New interventions are typically not well tested and oft en wind up being judged ineffective (even harmful)

ASSUMPTION #7: IT’S ALL ABOUT AVOIDING DEATH

Disturbing truth: A fixation on preventing death diminishes life

This seems like a repetition of the same point seven times.

Let me try to rephrase your seven assumptions into a single, strong argument: "Doctors and patients treat intervention as the default response to medical issues, but there are situations where intervention is (on average) neutral or harmful."

This is true. It is not a strong argument against medical scans however[1]. For instance, the result of the scans could be analyzed by a third party who isn't the patient or the doctor. Issues could only be passed along to the doctor if it will lead to a statistically better outcome for the patient.

[1] It might be a strong argument given a sufficiently weak culture, for instance one that is unable to provide legal immunity to the statistics-running third party. If your argument is, "We suck so bad we can't do 'statistics', so putting our heads in the sand is actually the best outcome we can get", you can just say that directly.

EDIT: Changed "can't" to "can". Thanks kgwgk.

> If your argument is, "We suck so bad we can't do 'statistics', so putting our heads in the sand is actually the best outcome we can't (?) get", you can just say that directly.

As far as I know, the statistics we have show that it is better to avoid doing full-body CT scans to asymptomatic people in the first place. Maybe you have different statistics?

ASSUMPTION #7: IT’S ALL ABOUT AVOIDING DEATH

Disturbing truth: A fixation on preventing death diminishes life

I don't like the wording of that last point. Yes, there are more things to life than avoiding death. And yes, fixation on preventing death can, in some cases, be counterproductive. I much prefer: as long as life is worth living, death is worth avoiding.

There's an opinion that his wording is exactly correct, and concentration on avoiding death makes life unworthy by definition. Although that opinion might be incomprehensible for some cultures. I can't think of anything that would explain that opinion clearly enough, but you could read Hagakure anyway.

>Disturbing truth: A fixation on preventing death diminishes life

Define 'fixation'. Total prevention of death obviously does not diminish life.

If there ever comes a day when we truly conquer aging, it is going to be one of sadder days for humanity.

Scanning is still fairly invasive (w.r.t. radiation/dosages). And in general, the better the resolution, the more damaging.

The bigger reason we don't continually scan people is that everyone has all sorts of little cancerous growths that are entirely benign. They would show up on these scans and freak people out. And often the overreaction of invasively removing those tiny little benign cancers can cause more problems. Don't misuse or abuse this information, but in general the 'abnormal growths' in a person are just that. They do very little and are just little bugs in the system. But the overall system is capable of dealing with them. It's the rare kind that is able to get around the entire body's defenses. But unfortunately it's difficult to tell which/when that happens.

But again, if you took whole body scans all the time, and tried to root out all those little growths, you'd cause a lot more harm than good.

>Scanning is still fairly invasive (w.r.t. radiation/dosages). And in general, the better the resolution, the more damaging.

This guy had MRI scans which don't rely on ionising radiation and have no known adverse effects.

>everyone has all sorts of little cancerous growths that are entirely benign

When discussing tumours, cancerous and benign mean different things. You cannot have a benign cancerous growth.

I think the main reason people aren't scanned routinely is due to the enormous cost of the scanning and the data analysis that currently relies on physicians for interpretation. My personal view is that with ever decreasing costs of MRI and improvements in computer vision and diagnostic algorithms, this could become a viable process in the future. As for not over-treating, that is a physician led decision - just because there is an anomaly on the scan does not mean that it will be removed or even necessarily treated at all.

While you are technically correct, it appears you've missed the point: being able to detect tumors better doesn't improve lifespan if we end up aggressively treating benign ones. This isn't a theoretical problem, rather one we are seeing right now.

Look at thyroid cancer: detection and treatment rates are much higher than they used to be but the death rate has remained steady.

Source: http://www.nytimes.com/2014/11/06/health/study-warns-against...

Actually, I addressed this point in my original comment:

> As for not over-treating, that is a physician led decision - just because there is an anomaly on the scan does not mean that it will be removed or even necessarily treated at all.

Perhaps I wasn't clear. I am saying that having cheaper, better scans with computer aided analysis is likely to be beneficial. I am not saying that every anomaly detected requires treatment. Treatment decisions are influenced by imaging but are not enforced by them.

> This guy had MRI scans which don't rely on ionising radiation and have no known adverse effects.

http://www.ncbi.nlm.nih.gov/pubmed/21412810

Also, the price of a single scan is measured in thousands of dollars.

I skimmed that study, but it looks like all the samples they tested came from the same guy, and they didn't actually test putting people in MRIs.

Yes, the authors did base their paper on in vitro tests from a single individual.

Another group attempted to reproduce the results of Lee, et al by performing a similar in vitro experiment, but failed to detect any DNA damage.[1]

Nevertheless, in an article commentary published by the European Society of Cardiology [2], DNA damage was detected in an in vivo study of Cardiac MR (CMR) imaging, and a precautionary approach to MR imaging is recommended. They admit, though, that the study could not differentiate between the effects of the contrast agent used in CMR, and the effects MR imaging alone.

[1] http://www.ncbi.nlm.nih.gov/pubmed/23679232

[2] http://eurheartj.oxfordjournals.org/content/34/30/2337

> I think the main reason people aren't scanned routinely is due to the enormous cost of the scanning and the data analysis that currently relies on physicians for interpretation.

The main reason is that doctors and regulators warn against it because, as patent says, people are chock full of scars and lumps and stuff that shows on scans but that is not in anyway harmful.

http://news.bbc.co.uk/1/hi/health/7150583.stm

> "If 100,000 people undergo a CT scan every five years from age 40 to 70 years...then the estimated impact is approximately 240 excess fatalities."

Although that's talking about CT, not MRI, scans.

Yes, my comments are limited to MRI. A full-body CT scan is a very large dose of radiation.

If this technology came with a feature to easily biopsy all the growths and tumors, and send them to a lab, do you think it would be beneficial then?

There's a significant school of thought that the structure of biological material helps ensure it works properly, and that an otherwise benign feature might actually become more harmful when their local environment is disturbed (possibly, by a biopsy). Biopsies are a very useful tool, but you don't want to do lots and lots of them if you can help it.

I understand, but I think medical technology is inevitably moving in the direction of some degree of consistent self-testing and monitoring, and I expect small biopsies of suspicious masses will eventually be implemented in a way that causes almost no structural changes.

I think this is how average lifespans are going to start rising dramatically in the next few centuries, but it's going to be a long and slow process.

Even urine screening would be excessive on a regular basis. The pretest probability is too low in the general population for mosts tests that even with incredibly high sensitivity and specificity the positive predictive value sucks due to all the false positives. This adds up into undue stress and unnecessary economic burden.

Edit: Just to add, low dose CT was recently approved for lung cancer screening from a study which showed increased detection and survival for patients 55-80 yrs old with at least a 30-pack year history (20 cigarettes per day for 30 years).

All that said, screenings will ultimately be grounded in genetics more than lifestyle as we become more preventative focused. This also adds granularity to pre-test probabilities of screening tests to ensure they are useful for the population tested.

In addition to the radiation, which can also induce cancer, you don't want to go scanning for things without a good reason. The false positive rates would increase a lot and you'd end up worrying a bunch of people for no good reason. Not to mention, you'd then have to have secondary biopsies or scans for these people, and those are not without risk.

Let's say that you had a low-radiation source imaging device that only had a 0.01% chance of causing cancer (I don't know the actual rates, so this is just a hypothetical). If you only get the scans when needed, you might only have 10,000 people getting the scan in a year. That means you'd have roughly 1 radiation induced cancer, and that would be in a patient that was already undergoing some form of treatment. Now let's say that you open up the criteria for who can get a scan. So instead of the 10,000, you now have 1,000,000 people. Now you have 100 people that will get a radiation induced cancer. While each of them has a low risk, the risk for the population is still too high to warrant unrestricted testing.

MRI's are a different beast, but you still have the risks of worrying a bunch of people that may not have anything wrong with them. And then there are the costs involved - both in terms of $$$ and time. For each healthy person getting a prophylactic MRI, a sick person would have to wait.

CT scans expose you to about 100-1000x the radiation of a single x-ray but that isn't the main reason frequent tests aren't done. The main reason based on my understanding is two-fold: First, cost would be prohibitive, and second the absolute number of false positive would do more harm than good overall.

I wonder if the relative rate and effect of false alarms is constant or if it increases as a function of how frequently scans are done. This is related to what studies are finding about cancer screenings. The more screenings you have, the time at which potential tumors are spotted is pushed back, and more non-dangerous tumors are caught. These benign tumors are mistakenly treated as malignant tumors because they were detected too early. I would imagine something similar would happen if people went for all out CT-scans.

Reply on news.ycombinator.com