Prototype Scanner Detects Cancer In Under 1 Hour

Ian Lamont writes "Researchers at Stanford say they have developed a blood scanner that can search for cancer-associated proteins in a blood sample and returns results in less than an hour. The device looks in a blood sample for cancerous proteins, and attempts to match them up with complementary proteins using chips based on magnetic nanotechnology. One of the researchers says the device could potentially help doctors identify lung cancer, ovarian cancer and pancreatic cancer at an early stage. The device still has to undergo clinical testing and trials before it can win regulatory approval."

Comments

yawn

[jDeepbeep]

The device still has to undergo clinical testing and trials before it can win regulatory approval."

Meanwhile.... 14 years later...

Re:yawn

[the4thdimension]

Indeed. This thing is a LONG way off. By the time they get this out the door to hospitals for use, someone will have an instant test coming out and we should just be fast-tracking that.

I work for a company that makes such devices and clinical trials and testing are not even close to the last step. Clinical trials are the beta test, so to speak, and often mean you have months and months of bug fixing and documentation to do. Take a device intended to diagnose patients, and you can multiple that by years. Fourteen years might seem funny, but its actually somewhat accurate. My company has been working on a product for nine years now seeking US approval.

Re:yawn

[Invicta{HOG}]

I agree - studies like this are quite common in the medical literature and, while exciting to think about, have a long way to go before they find their way into the clinic. For instance, this chip is looking for many different kinds of proteins. Each protein will have a specific false positive and false negative rate of detection. Because the chip has so many proteins it looks for, the total false positive/negative rate for cancer detection of the chip will have to be determined and, likewise, a decision made as to whether this is an acceptable rate for clinical practice. For instance, it might do well for each individual cancer/protein, but when you are looking for so many different cancers, you might find that an unacceptably high number of chips return an answer of "CANCER." Since this might necessitate costly evaluations with their own inherent risks, you will need to insure before this comes to market that the results are clinically relevant and have an acceptable positive/negative predictive value.

Re:yawn

[Invicta{HOG}]

There is no guarantee that this will save lives. That's what studies are for. You are confusing cancer detection (or in this case cancer related protein detection) with cancer therapy.

Take two examples now available to detect cancer/cancer related proteins. The first, prostate specific antigen (PSA) is elevated in most cases of prostate cancer. This simple blood test was recommended for all men above a certain age to screen for cancer. However, it does its job too well. It finds cancer in so many men that people started wondering whether finding all of this cancer is a good thing. A lot of men are old enough (and prostate cancer grows slowly enough) that they would certainly have died from other causes without ever having a single symptom of prostate cancer. Additionally, the number of biopsies done to find early cancer (or to find that there was no cancer after all!) combined with the amount of morbidity from current modalities of prostate cancer therapy have led many organizations to conclude that screening for prostate cancer does more harm than good.

A second example comes from the detection of thyroid cancer. Above a certain age, a simple thyroid ultrasound has a fairly good chance of detecting a thyroid nodule. There's a fairly good chance that this nodule will be cancer. However, the risk from dying from this thyroid cancer end up being very low - low enough that the risks from detecting and treating it are higher than the risks of just leaving it be.

When this has been proven to save lives or improve the quality of life of people with asymptomatic cancer, I will join you in criticizing an admittedly slow and often overly burdensome system. But at this point you can't criticize when the device is so far from proven.

Re:yawn

[jcr]

You are confusing cancer detection (or in this case cancer related protein detection) with cancer therapy.

I'm doing nothing of the kind. I'm taking note of the fact that the earlier a cancer is detected, the better the chances of survival.

But at this point you can't criticize when the device is so far from proven

The hell I can't. The FDA kills people every day.

-jcr

Cancer Joke

[amclay]

Doctor: Well, we better discuss treatment now for your testicular cancer. I recommend hormone therapy. Man: Are there any side-effects? Doctor: A few. You will have a loss of potency. You might get some hot flashes. And when lost, you will have an inexplicable urge to ask for directions. http://www.phoenix5.org/humor/HumorRVYjokes.html

Re:Usefulness?

[MozeeToby]

Well, according to the summary the scanner is looking for protiens that are produced by cancerous cells, not the cells themselves. And even if it were the case that it could only detect the cancer cells in the blood, it would still have it's uses. If it could be made cheap enough, it could become a standard test, everytime you visit the doctor. It would still allow us to catch cancer cases earlier than they would have been otherwise even if we couldn't rely on it to detect 100% of all cancer cases.

Re:Usefulness?

[CorporateSuit]

I don't see this being useful for detecting breast... tumors

That's what hands are for. (Preferrably my hands)

Bah, I had that beat years ago!

[Chris Burke]

Not only did I invent a device which would correctly diagnose cancer 99.999% of the time, not only did it work in only half an hour, it also didn't involve any of this expensive magnetic nanotechnology la-dee-dah. Plus the device was so ridiculously simple anybody could use it, which you'll see once I describe the device itself.

Basically it's a big box, kinda like a front-loading washing machine. In front of the box is the scanning aperture. On top of the box is a single button labeled "Detect Cancer". You stand in front of the scanning aperture, and you hit the button. Over the next half hour, the box scans you with very high levels of x-rays. Once the scanning was done the only other feature on the top of the box, a green LED with a label that says "Cancer Detected", would light up.

You see, so simple a child could use it! I should know, too, because I had some try it out. But those bastards at the FDA brushed me off, even threatening me if I continued performing clinical trials! Even after I showed them it had the same accuracy detecting radiation burns and radiation sickness! But what do you expect from bureaucrats? More concerned with their "rules" and "regulations" than helping people. I wouldn't be surprised if this new one gets a pass because being hard to build and complicated to use they can regulate the hell out of it, even if it is inferior. :(

One hour compared to what?

[Unknown Relic]

How does this compare to traditional tests? One hour is great and all, but how long to today's tests take to return results?

Re:One hour compared to what?

[RingDev]

No idea, but if it means I don't need the doc to stick a finger up my bumm, I'm all for it!

-Rick

Re:One hour compared to what?

[Invicta{HOG}]

This technology is not currently available in the marketplace. There are blood tests that look for tumor markers such as PSA, CEA, CA19-9, etc. and they generally are sent to a large reference laboratory for analysis. This can take up to a week. Traditionally cancer is diagnosed pathologically by looking at a tissue sample underneath a microscope. Aside from the obvious need to undergo a biopsy, this can often be done quickly (pathologist standing in the OR, the surgeon hands the sample over, they read it then and there). However, the hour time frame is not the real story here - it's the ability to combine all of this screening in the first place.

Re:One hour compared to what?

[jandoedel]

Well, since it tests the blood of the patient, and doesn't need to scan patient per patient, that probably means that you can put a lot of blood samples together in one big pot, and then test that mixture (and do this a couple of times, to reduce the error margin, but that can be done at the same time.) If your batch is cancer free, you have just tested X people in one hour. If not, then you split the batch in two, and test both batches to know which of them contains the cancerous blood. Continue a couple of times until you found the patient. Testing 256 people at the same time takes 1 test if everyone is healty, or 8 tests if one has cancer. (or a bit more if more than one person has cancer) So that means that the average time per person can be very low. As opposed to PET scans, MRI, CT, SPECT,.. where you can only cram one (or two if you squeeze) person in the machine at the same time. If a PET scan takes 15 minutes, then you can only scan 4 people per hour with one machine.

wrong kind of dog

[slew]

You can use your Deterministic Oncological Generating box ;^)

or perhaps take advantage of another type of dog...

http://news.nationalgeographic.com/news/2006/01/0112_060112_dog_cancer.html

Apparently this more common type of dog can be trained to smell certain types of existing cancer (instead of deteriministically generating them) ;^)

Re:Usefulness?

[coolsnowmen]

The key to the gap in your understanding is that cancer proteins can be found in the blood with out there being any cancer cells that have actually metastasized to the blood. When any cells replicate proteins slip in to the blood for various reasons. Looking at presence and the relative increase of these proteins is the focus for early detection of cancers.

tNOX (tumor-associated NADH oxidase) is a protein some research was looking at.

serum amyloid A elevates for lung cancer

Doctors in india found a protein to indicate the precursor to colon cancer

early detection of ovarian cancer based on four proteins: leptin, prolactin, osteopontin and insulin-like growth factor-II.

All this research is from the last couple years, so it appears that measuring the correlation of these proteins with cancer has been an area of hot research.

Re:Usefulness?

[Walpurgiss]

While that may seem like fun, I bet it would be pretty awful when you do find cancer.

Re:Usefulness?

[ColdWetDog]

It would still allow us to catch cancer cases earlier than they would have been otherwise even if we couldn't rely on it to detect 100% of all cancer cases.

Careful what you ask for, you might get it. A large number of cancers, perhaps the majority of them, are 'cured' by the immune system at very early stages. Even some Breast Cancers seen on mammograms will involute. (A Google search is in order if you're curious). If you are not very careful to understand the biology of the cancer in question you will end up creating a) a lot of angst on the patient's part b) a lot of angst on the doctor's part c) extra costs for what amount to unnecessary tests d) the real possibility that those tests might HARM the patient rather than help.

Other cancers grow so slowly that detection of small numbers of cells very early on will create enormous clinical controversies - how do you treat a $_random_cancer that is seen only in a micro array test given that standard therapy for clinically apparent $_random_cancer might include radiation, surgery or a host of chemicals that would give even Saddam Hussein the willies?

These will be interesting lab devices, but I don't see picking one up in your local Lucky Dragon anytime soon.