New Research Suggests Cancer Can't Be Eradicated
npr.org
npr.org
Part of the blame for this sensationalism seems to fall at the feet of the original scientist himself (Bosch). He makes an unfounded leap from the biological result that cancer seems to be an intrinsic property of genetic processes in cells (and the statements "You have to interfere with fundamental pathways", "It's a web of interactions" and "It's very difficult to do") to the assertion "will probably never be completely eradicated". Difficult is not the same as impossible.
He follows with more sensationalism: "You carry a time bomb in your body when you're born," he said. "It can explode early in life, or middle age or later."
I suppose it's understandable; he's probably just trying to get some publicity for his team's long-running research on hydra. And even he admits "that doesn't mean that, with a patient who develops cancer, there's nothing you can do."
But then the hedging of "probably" and "completely" was discarded by the article writers for a more sensationalist article and headline.
The claim that cancer probably never will be eradicated seems pretty defensible. Notice how it's the people who study it who are the least sanguine in their predictions. It's perfectly possible for us to have a complete understanding of cancer and find that we still can't eliminate it.
I love this quote, and I'm totally going to steal it.
There's just something about the futurists and singularity folks that mimics the Dunning-Kruger effect. The more you learn, the more you recognize that we're still operating on Cro-Magnon brains. All of our magnificent toys are merely low-hanging fruit in deference to the problems that seem to really matter.
I consider that extraordinary, and the exact opposite of low hanging fruit. It's hundreds of years of incremental progress across dozens of fields, resulting in reaching very high hanging fruit in the form of an iPhone and everything that goes with it.
I consider such amazing communication technology to be every bit as valuable to social beings such as humans, and to our general societal structures, as curing cancer is.
Search and traversal become increasingly difficult as any graph grows in size. It's not hard to generalize this to the graph of scientific and engineering knowledge. To maintain the connectedness of our ideas, we must either leave amazing literature and documentation or we must constantly seek to foster connections between those of separate, heterogenous fields and disciplines. Unless there is movement of ideas through collective human consciousness, key insights cannot percolate through to the point where they have the greatest impact. (For instance, not long ago biologists were still using Perl. :P) At some point, the time it takes for certain information to spread across the graph grows substantial enough that discoveries are simply repeated over.
Science and technology require ever-increasing levels of specialization. The network topology isn't going to get easier for the human mind to traverse. Imagine what kind of information lies just beyond your reach--the kind of insights that might truly propel you ahead, to reach your Eureka. If only you knew what you were missing out on. On the flip side, do you ever think about the places where your contributions would be the most impactful?
These problems are somewhat symptomatic of the fact that we haven't breached the complexity threshold of what our brains are capable of performing. We may have computer-aided tools to get us beyond some classes of problem--search and mathematical modeling come to mind. Still, our mastering of breadth at depth, our powers of abstraction and of reasoning, all have a hard limit. Everything we've ever done falls neatly within defined lines. Nothing being done today is substantially harder than what the previous generation did with its much more limited set of knowledge. (Have we discovered our generation's calculus or quantum mechanics? Advancements like that are special, truly rare, set apart by hundreds of years.)
Take your smartphone, for instance. You can't look at the following block diagrams and tell me they're too difficult to reason about:
* http://eda360insider.files.wordpress.com/2012/08/samsung-exy...
* http://www.arm.com/images/Cortex-A15-chip-diagram-LG.png
* http://www.wired.com/wp-content/uploads/blogs/insights/wp-co...
(Yes, I know the implementation details go much deeper and that these diagrams are meant for ease of conveying knowledge. You still understand them. And given time, a diligent student can learn to contribute to processor design using the pre-existing domain knowledge.)
Do you have a sense or appreciation for how complex the physical and chemical reality is by contrast? We can only partially understand minute samples taken from real systems. It's not especially owing to our primitive equipment or tooling; it's the sheer complexity that baffles us. There are things like helium atoms--we don't understand them. There's Hartree-Fock, but that's an approximation.
I find it hard to be beholden to the majesty of a phone when no one can stand up and explain a helium atom. I mean, don't you find it a little bit amusing we're still caking onto SGML, shoveling it onto handsets, and yet we don't understand the interstitial gas between us and the cell tower? The stuff we're breathing.
Anecdotes from the early days of AI research suggest that we didn't think we'd still be writing code imperatively, yet here we are writing Javascript module frameworks and the like. We have a stunningly powerful ability to underestimate how long and difficult things are. We are also often satisfied in rediscovery.
We're concerned with how to perform delivery in San Francisco, yet... can you imagine the vast number of computations the universe is performing every time a protein isomerizes? I wager it outweighs all of the computing power we have available to us through modern technology today. The modern smartphone could never compete in this arena; we would be comparing something not even at the level of a bacteria to the most omnipotent of gods. I'm not saying that one day we'll design computers out of polymers in solution. I don't even know if we'll survive as a species long enough to care. I'm just asking you to wonder.
Look to nature. Can you sense how incredibly infinitesimal and primitive we are as a species when you look up at the night sky? We've been feeding off of the cosmologically low hanging fruit, and we're still in the nursery. We spend on average 20 years learning what we need to begin making our cornerstone contributions to society. At that point, how much longer do we remain productive? It's a limited window with which we have to collapse all the reverse salients. Damn shortness of human productive lifetimes, and all of that. Oh well, there's always a new generation with which to start fresh. Rinse, repeat. Stand on a few more shoulders, see a little more of the picture.
Fill your mind with what might be possible. Let it be dissatisfied when you cannot reach a conclusion to a problem. Don't ever put a bound on the type of problems you're curious about. You'll scoff at press releases. Everything in our generation is transient, scaffolding for the next one gestating.
(Sorry for rambling. And also--I still hold that my flappy bird machine not even in the same galaxy when it comes to being as cool as science. It's a toy for which there will always be one better. How could it ever be so magnificent as the universe I get a few short years to experience firsthand?)
BUt it is, proof is that even after we invested tens of billions in funding basic research in cancer and other areas of biology we still know so little of practical value. Even the simplest organism is infinitely more complex than the most intricate human made object (which might very well be the microprocessor). We solved the fast computing and global communication problems first because they're way easier, and there's no denying it.
5-year survival rates has gone from 50% to 68%, other statistics are improving as well, although not particularly quickly.
I really hate 5 year survival rates as a measure of success as it is almost totally meaningless. We know most cancers have lead times from initial mutation to symptomatic diagnose of 15 to 30 years. This means that if I was able to make a test that could detected cancer 10 years after initiation I could in one swoop make the 5 year survival rate 100% for all cancers without changing age at death by one day.
"Treatment for this disease has become so effective that 80% of patients with metastatic testicular cancer can now be cured. Thirty-five years ago, 95% of these patients died, usually within 1 year of diagnosis."
From idlewords' post it might sound like we've made no improvement, but instead it is more like slow and patchy improvement.
We probably can't genetically or medically engineer humans never to generate cancers. That's what this guy seems to be saying: cancer is a consequence of normal operation, not a flaw or a disease.
We probably can get good enough at cleaning up the cancers, that our continual generation of them is no longer an issue.
There's the known linkage.
This says nothing about the runtime dynamics or higher order interactions, which involve heavy calculus.
This says nothing about the biochemistry under anything but isolated laboratory conditions.
This also doesn't deal with anything above the cellular level. In fact, these are just chemicals in solution--none of the intracellular systems with their own local chemistries are modeled here, and those are maddeningly complicated in their own right. If you think our biological understanding has progressed to the point where gleaning simple information about single edges in this graph are elementary, think again.
Most people have heard about the genome, for which we now have punchcard-level technology to deal with. The methylome sits on top of that, annotating the genome with important dynamic state. Then there's the transcriptome--utilizing nucleic acids, it's just one instance of intracellular messaging. The RNA field is extremely complicated due to its chemistry, which makes it annoyingly difficult to study. (Also, RNA isn't just messaging.)
Want to get at some systems-level interactions? There are the metabolome, reactome, interactome... And we're still at the sub-cellular level. I didn't even mention proteins, which seem to be somewhat important to the study of cancer. (Actually, entire immunological systems are. Yet we're still discovering new immune cell types and guessing at roles...)
If it wasn't something you already had an appreciation for, you can now see that the biology state machine is beyond your wildest dreams complicated. And really, this is glossing over things. You have to take into account a metric ton of chemistry. Unlike programming, nothing ever gets abstracted away in biology.
Welcome to the defining problem of the human race--solving it, if we ever do, will likely be our grandest moment.
If I could make the bet, I'd wager we won't have a cure to cancer in 200 years.
Yeah, because a scientific study can summarized from a couple of lines from a NPR article (and even those strawman-quoted in your comment).
>This only points out that cancer has been around for a long while - which I don't think is a contentious issue anyway - not that cancer can't be eradicated
At a basic level, this points out that cancer can emerge even with the very barest forms of life/DNA -- so it's something more inherent than something we acquired after several degress of evolution.
If you couldn't tell, I don't think the article is well-written and is probably an incomplete summary of the research being done. However, I think the statement I made there is an accurate summary of what the article - if not the research - is trying to claim. I.e. it is not a strawman.
1. Hydra are apparently immortal. They do not appear to age. It is likely that there is some biological trade off between immortality and cancer predisposition. Therefore they are not a good model for concluding that cancer everywhere in everything is incurable
2. Hydra don't have an adaptive immune system. This is a huge flaw in an experimental system for studying the 'roots' of cancer development.
The main point of the journal article is that maybe we should use metazoans to study cancer. The journalist here is way off the mark.
Since cancer is the consequence of the loss of the regulation of cell division it is axiomatic that in any cellular system given enough time and mutations the regulation will be lost in at least one cell of an individual. We don’t need to prevent disruption of the regulatory system, although this is good, we just need to be better able to identify and destroy those cells that are not regulated.
Typically there is a lot of work involved in "spontaneous" remission.
But yes, it's likely we can't eliminate cells' mistakes, just prevent, detect and fix them better in the future.
http://money.cnn.com/2014/02/04/news/economy/budget-outlook-...
There are only so many biologists, labs, etc. You can certainly entice more into the field with money, but I expect costs increase non-linearly with the number of scientists.
(I am not saying we can't do it. I'm saying it would cost much more than we expect from a linear extrapolation, especially in the near term. Imagine trying to use money to double, say, world food production – it couldn't be done, at least not in the short term.)
Labs are not hard to scale horizontally, and likely aren't even at capacity now due to lack of funded researchers.
Thank god I started programming when I was ~10 and was able to fall back on that after I decided against going to medical school, or I really have no idea what I would have done with my life.
In a $15 trillion economy, you'd think it would be possible to spend 10x on research. More private funding?
How many top notch leaders can you cultivate to lead big research teams. People that have a comprehensive understanding of what targets to have for the research, people able to recruit the best brains to the cause.
This seems to not be primarily a monetary bottleneck, but rather the inherently limited resource of how many really great people you can find to work on something at one time.
It strikes me that humans also can only move so quickly, and that perhaps automated / machine learning, testing and solving is the long-term solution to speed of progress. I know a lot of progress is actively being made on this, I expect it'll drastically accelerate the pace of breakthroughs in curing illness. I think of it this way in premise: the current approach is the NIH trying to sequence the human genome; the future approach is Venter'ized.
Medicine typically only knocks on the door with extracellular signaling. The cell remains a black box. Even if it weren't, with cancer we'd have to select certain cell populations based on internal cues that may never be exposed. Then there's the structure and pharmacology problems--there's a lot of chemistry and calculus along the path from drug introduction.
I suppose you can do this for a cell in a terminally differentiated state of your choosing. (But as you know, pre-cancerous cells are typically in fuzzy states that break our tidy little metabolic models.)
http://en.wikipedia.org/wiki/Budding
Even with normal binary fission, there is never an equal distribution of products between cells. Metabolite concentration, organelle distribution, etc. will vary. (If you think about it, it's kind of an interesting period of rough perturbation made to an extremely high-order dynamical system.)
At a kind of base level, though, the genome itself regularly experiences abnormal partitioning that leads to certain interesting states.
It's not strictly binary...
But back to the previous topic. Epithelia and endothelia are regularly dividing cell types. (It's interesting that most cancers arise from these lineages, right?) Blood cells are another lineage that undergo pretty frequent proliferation.
(On an entirely unrelated note, I can't wait for an HN discussion pertaining to quantum effects on biological systems--it's a topic I really want to learn more about. The absorption spectra discussions in the retinal thread has awoken my curiosity. For instance: though it's kind of a controversial, maybe even a pseudo-scientific topic, the quantum effects on microtubules are fun to think about and are still actively researched.)
It's hard to measure progress in something like cancer research, but in general, it hasn't been great. But the same might be said of the invention of flight. Flight had clearly been a goal across a variety of cultures for many years.
I suspect that a "cure for cancer" will not come at once, it will be more gradual (owing to the fact that cancer is actually many different diseases). Our knowledge of biology is improving at an exponential rate though, I wouldn't discount that entirely.
Cancer/dna-mutations happen all the time. They're triggered by the energy from cosmic rays, free radicals etc.
A lot of those reactions involve the oxygen we inhale, so one of the reasons for cancer and death is actually that we're slowly being burned in our hazardous atmosphere.
That said, there is obviously ways to prevent and treat but that's a different topic.