Intel's Collaborative Cancer Cloud, an Open Platform For Genome-Based Treatments

Lucas123 writes: Intel and the Knight Cancer Institute have announced what will be an open-source service platform, called the Collaborative Cancer Cloud. The platform will enable healthcare facilities to securely share patient genomic data, radiological imagery and other healthcare-related information for precision treatment analysis. Key to averting HIPAA privacy issues will be Intel's Trusted Execution Technology, its embedded server encryption hardware that tests the authenticity of a platform and its operating system before sharing data. Intel said it will be opening that technology up for use by any clinic that want to take part in the Collaborative Cancer Cloud or to build its own data-sharing network with healthcare partners. Dr. Brian Druker, director of the Knight Cancer Institute, said the Trusted Execution Technology will allow healthcare centers to maintain control of patient data, while also allowing clinics around the world to use it for vastly faster genomic analysis.

Comments

fu /.

[anapsix]

wtf slashdot? just lost my entire comment, when data connection was lost as train was going though a tunnel. da faq you internet connection for when I'm filling out a comment form? fuuuuu!

Re:fu /.

[Anonymous Coward]

wtf slashdot?
just lost my entire comment, when data connection was lost as train was going though a tunnel.
da faq you internet connection for when I'm filling out a comment form?
fuuuuu!

Put down the crack pipe.

Re:fu /.

[dinfinity]

I think (or hope) he meant that /. sometimes auto-refreshes the page, even when you are still writing a comment. I've had this happen to me recently as well.

Prior art!

[Anonymous Coward]

I believe Beijing's got the patent on collaborative cancer clouds.

open source software, yet proprietary hardware

[anapsix]

According to an engineer giving talk on collaboration of companies involved with Intel's CCC, they are going to open-source it, yet it's all running on proprietary hardware. He made a point that Intel is not making any money on this and doing it for the sake of humanity.. Except you'd need proprietary hardware manufactured by Intel.. some of which has no drivers yet. hmm.. Also, I wonder how they deal with exporting health data across international bordets, since some compan(y|ies) are Canada-based.

Translation

[markdavis]

>"its embedded server encryption hardware that tests the authenticity of a platform and its operating system before sharing data"

Translation: "Use our proprietary hardware and software and forget about using anything open-source like Linux".

Reminds me a lot of that horrible, crappy "Trusteer" junk that some banks are trying to force on people, especially corporate customers. https://www.trusteer.com/Prote...

"Cancer Cloud"

[tsotha]

They really could have come up with a better name. This one makes it sound like something you'd really like to avoid.

Re:"Cancer Cloud"

[dinfinity]

So it's the perfect name for it.

Intel should call it Trusted Insecurity Tech

[Anonymous Coward]

It's one big fraud. Intel knows it's systems aren't secure or it would reveal the code. It's just that simple. You can't trust what nobody other than Intel can analyse.

I don't understand

[behrooz0az]

I don't understand how a hardware feature is going to make shit secure.
How does it stop shoulder surfing, MITM, software bugs, cloning, 100 more things????????//

Most cancers cured within 50 years

[Anonymous Coward]

My fearless prediction is that, within about 50 years, cures will have been found for most cancers. It will take about 15 years for full genome sequencing of cancer patients to become routine. And then it will take another 15 years to understand the full biological basis of most cancers (using the genome sequences of everyone in developed counties who has had cancer in the last 15 years). And then it will take another 15 years to design effective therapies based on understanding the full biological basis of most cancers. And, finally, it will take another 5 years to roll the cures out to the clinics.

Clinical/personal genome sequencing is one of the biggest revolutions in the history of biology and medicine. With our grandchildren's lifetime, cancer will go from being a death sentence - to an "Oh, you've got some cancer - just take a couple of these pills and it should clear up in a couple weeks". Of course, we can get there faster if governments put more money into to finding the cures - there are far more people trained up to do biomedical research than available jobs. But, unless we manage to completely destroy our civilization in the next few decades then we will definitely get there.

Exciting times!

Re:Most cancers cured within 50 years

[Anonymous Coward]

I doubt it. They are still stuck on the somatic mutation idea. Whether or not it is the cause, cancers are almost always aneuploid (have wrong number of chromosomes) meaning that the expression levels of hundreds or thousands of genes are messed up and they are genomically unstable. Somatic mutations also appear to happen very rarely, too rarely to account for the age-specific incidence of most cancers. The mutations they target may work as markers for a certain type of cell, which will help in killing them, but it also selects for resistant cells so it will come back. Those mutations are never required for the cancer cell to survive, nor form.

If you want to throw money into cancer research, we need basic information like how many times healthy cells are supposed to be dividing in each tissue by age. Then this can be compared to various error rates to figure out what is driving the formation and survival of these cancer cells, and also to provide a way to dose drugs in a way that distinguishes between healthy and cancerous tissue. Currently, we don't even know the total number of cells to within 8 orders of magnitude:

The aim of this work is to discuss the theoretical issue of the total number of cells that compose the standard human adult organism.

First, we noticed that these data were typically mentioned in the literature without citing a reference; second, we observed wide ranges among data reported by different sources, ranging from 10^12–10^20.

http://www.ncbi.nlm.nih.gov/pubmed/23829164

In reality, the state of affairs is very sad. You've been fooled by press release propaganda. Now, if tomorrow the course was immediately changed to focus on collecting data required to test quantitative theories regarding cancer and cancer researchers stopped believing that if the null hypothesis is false it means their theory is true, I think in 20 years we will be able to figure it out. That is, if current funding levels are sustained (adjusted for inflation).

I do agree that personal genome sequencing is cool, but it will not achieve what you are thinking.

Re:Most cancers cured within 50 years

[Anonymous Coward]

...cancers are almost always aneuploid (have wrong number of chromosomes) meaning that the expression levels of hundreds or thousands of genes are messed up and they are genomically unstable.

The are quite a few genetic conditions caused by duplications and even deletions of large portions of various chromosomes - where every cell in a person's body has the wrong copy number for a large set of genes. But many of these genetic conditions have no additional risk of cancer. So major structural variations, per se, aren't necessarily the key driver of cancer.

Somatic mutations also appear to happen very rarely, too rarely to account for the age-specific incidence of most cancers.

Ah, but certain specific mutations can increase the mutations rate and also cause general genomic instability.

The mutations they target ... are never required for the cancer cell to survive, nor form.

Well, the general model is that you start off with a mutation or two that increases the mutation rate and general genomic instability which then causes you to get mutations in certain key genes that cause the cells to proliferate out of control. So targeting the proteins of the genes for cell proliferation is actually very effective in slowing down and even killing the cancer cells.

In reality, the state of affairs is very sad. You've been fooled by press release propaganda.

It's definitely sad that people are dying of cancer. But, even though every cancer is different, it's overwhelmingly likely that most cancers are going to fall into common patterns - starting off with a particular set of mutations and then progressing in a typical way for that type of cancer.

I'm not saying it's an easy problem but it would be difficult to understate the amount of insight that will be gained when full genome sequencing of people with cancer becomes routine.

Re:Most cancers cured within 50 years

[Anonymous Coward]

Well, the general model is that you start off with a mutation or two that increases the mutation rate and general genomic instability which then causes you to get mutations in certain key genes that cause the cells to proliferate out of control.

Do the math for the multistage model for yourself. What is the waiting time for x mutations to occur at rates r1,r2,...rx needed to explain peaks in age specific incidence = around 80 yrs old (tongue, tonsil, hypopharynx, liver, esophagus, lung/bronchus, larynx, corpus/cervix uteri, prostate, testis kidney/renal, brain, thyroid, lymphomas, myelomas, breast)? What is the somatic mutation rate? Something is wrong.

Data-Sharing

[Etherwalk]

"by any clinic that want to take part in the Collaborative Cancer Cloud or to build its own data-sharing network with healthcare partners."

Oh, yeah, because my local clinic's tech guy is TOTALLY going to do that with good data security practices.

WOW!!! crazy

[Anonymous Coward]

Look at how many comments are posted here, and how many are positive?
this is ground-breaking stuff and yet no one realy cares..
Lets compare this to um .. the Ashley Madison article, which has 100's more..
What a damn shame..