Blood Test That Finds 50 Types of Cancer Is Accurate Enough To Be Rolled Out (theguardian.com) 32
A simple blood test that can detect more than 50 types of cancer before any clinical signs or symptoms of the disease emerge in a person is accurate enough to be rolled out as a screening test, according to scientists. The Guardian reports: The test, which is also being piloted by NHS England in the autumn, is aimed at people at higher risk of the disease including patients aged 50 or older. It is able to identify many types of the disease that are difficult to diagnose in the early stages such as head and neck, ovarian, pancreatic, esophageal and some blood cancers. Scientists said their findings, published in the journal Annals of Oncology, show that the test accurately detects cancer often before any signs or symptoms appear, while having a very low false positive rate.
The test, developed by US-based company Grail, looks for chemical changes in fragments of genetic code "cell-free DNA (cfDNA)" that leak from tumors into the bloodstream. The Guardian first reported on the test last year and how it had been developed using a machine learning algorithm a type of artificial intelligence. It works by examining the DNA that is shed by tumors and found circulating in the blood. More specifically, it focuses on chemical changes to this DNA, known as methylation patterns. Now the latest study has revealed the test has an impressively high level of accuracy. Scientists analyzed the performance of the test in 2,823 people with the disease and 1,254 people without. It correctly identified when cancer was present in 51.5% of cases, across all stages of the disease, and wrongly detected cancer in only 0.5% of cases.
In solid tumors that do not have any screening options "such as esophageal, liver and pancreatic cancers" the ability to generate a positive test result was twice as high (65.6%) as that for solid tumors that do have screening options such as breast, bowel, cervical and prostate cancers. Meanwhile, the overall ability to generate a positive test result in cancers of the blood, such as lymphoma and myeloma, was 55.1%. The test correctly also identified the tissue in which the cancer was located in the body in 88.7% of cases.
The test, developed by US-based company Grail, looks for chemical changes in fragments of genetic code "cell-free DNA (cfDNA)" that leak from tumors into the bloodstream. The Guardian first reported on the test last year and how it had been developed using a machine learning algorithm a type of artificial intelligence. It works by examining the DNA that is shed by tumors and found circulating in the blood. More specifically, it focuses on chemical changes to this DNA, known as methylation patterns. Now the latest study has revealed the test has an impressively high level of accuracy. Scientists analyzed the performance of the test in 2,823 people with the disease and 1,254 people without. It correctly identified when cancer was present in 51.5% of cases, across all stages of the disease, and wrongly detected cancer in only 0.5% of cases.
In solid tumors that do not have any screening options "such as esophageal, liver and pancreatic cancers" the ability to generate a positive test result was twice as high (65.6%) as that for solid tumors that do have screening options such as breast, bowel, cervical and prostate cancers. Meanwhile, the overall ability to generate a positive test result in cancers of the blood, such as lymphoma and myeloma, was 55.1%. The test correctly also identified the tissue in which the cancer was located in the body in 88.7% of cases.
The punch line is... (Score:2, Interesting)
that we frequently have cancer cells floating around in our bodies that do not develop into "cancer", which must be treated
I can see a lot of good coming from this, but I also see a lot of people having unnecessary surgery and getting worse outcomes than if they knew nothing at all
Re:The punch line is... (Score:4, Informative)
Cancer doesn't "develop into" cancer. It's just that some cancers can be effectively fought by the immune system. A positive test wouldn't lead straight to treatment. It's certainly a way to narrow down the search.
Re: (Score:1)
"We got a call saying a test detected cancer cells in your body. Please step out of the car."
"I don't give you the right to search my person."
"Then we'll call for the K-9 unit [the-scientist.com] to confirm."
Re:The punch line is... (Score:5, Insightful)
I think perhaps there's another issue at play here. In most developed countries that have universal healthcare these new screening tests would be overwhelmingly welcome news because they'll save lives. However, do you think that in a country which has a for-profit, private healthcare system, people would be anxious about yet another battery of tests to be subjected to & charged for & then the worry of whether their health insurance will pay for it? For example, in the USA, the leading cause of bankruptcy is medical bills. Getting inflated bills every time you go to see the doctor because they recommend batteries of tests just because you've reached some age-group can make life really difficult if you're poor &/or don't have health insurance. Just going to the dentist is out of reach for a large number of Americans. Have a look on crowd-funding websites to find a lot of Americans begging for help with their medical bills.
tl;dr - This is fantastic news for people living in developed countries with universal healthcare. Bittersweet/mixed feelings for countries with predatory, privatised healthcare.
Re: (Score:2)
Absolutely. Because, just like national health plans, for-profit insurance companies have an incentive to get you treated early, rather than after a disease that's just starting develops into something that will be more costly for them to pay for treating.
Th
Re: (Score:3)
Re: (Score:3)
Re:The punch line is... (Score:4, Informative)
but I also see a lot of people having unnecessary surgery and getting worse outcomes than if they knew nothing at all
That doesn't seem likely at all. Nobody does surgery to remove a tumor if they can't see it from imaging first. The only time they ever perform surgery to remove a tumor they can't see is if they were able to see it previously, i.e. before chemotherapy shrunk it to the point of being undetectable by imaging.
Re: (Score:2)
I think he's referring to things like prostate cancer, which sometimes is better off without treatment [wikipedia.org].
Still, it seems like knowing about the cancer is better than not knowing, even if it doesn't need treatment immediately.
Re: (Score:2)
A long time ago, in a galaxy far, far away, I was working in a hospital that had just installed one of the first MRI units. As it was super high tech, quite a lot of staff volunteered to be test subjects while the machine was being calibrated before being put into service, so they could say they had had an MRI scan.
One of the junior doctors saw a small lump in his arm, so, fearing the worst, arranged for some investigative surgery. By the time that happened, nothing could be found and this lump had disappea
Re: (Score:2)
> I think that shows that we did not really know at the time what normal was when it comes to a Gods Eye view of our bodies, and probably still are missing a great deal of knowledge in that area...
We‘ll then acquire this missing knowlegde.
I‘m from a curious family of physicians, my father told me this story:
When my mother was pregnant with my sister they received a long time ecg (may even have been acoustic). It was quite experimental and they mainly used it because it was interesting and new
Re: (Score:2)
I can see a lot of good coming from this, but I also see a lot of people having unnecessary surgery and getting worse outcomes than if they knew nothing at all
This is a simple dragnet type diagnostic test. You don't take the results and head for treatment, you investigate further and identify the trigger. It would be like if you told your mechanic you keep hearing a "rar rar raa" sound and he immediately replaced a part in your car without any testing. It just doesn't work that way.
Elementary, my dear Holmes (Score:1)
Elizabeth should feel vindicated now.
Cost (Score:2, Insightful)
Re:Cost (Score:5, Informative)
It's $949. Insurance will not cover it yet but they are working on it.
https://www.galleri.com/ [galleri.com]
Re: (Score:2)
That does not include administering the test and the $200 band aid applied after they prick you.
Nothing of worth in the US healthcare system costs $949. I just paid $224 for a tele-doc visit WITH INSURANCE!
Re: (Score:2)
Because of Covid-19 some insurance companies are waiving telehealth co-pays [ahip.org] if the issue relates to Covid-19.
--
Re: (Score:2)
I'm so happy I live in a country with universal healthcare, AKA "socialized medicine", with "death panels" & everything. Seeing your doctor is free. Getting treatment is free. There's a small, fixed prescription charge (same price for all medicines) which is waived if you're retired, unemployed, etc.. Nobody talks about health insurance.
I lived in Canada for a few years. Healthcare there is semi-privatised, i.e. NOT universal. I met people who had to decide between buying food & paying for their pre
Re: (Score:2, Interesting)
It's $949. Insurance will not cover it yet but they are working on it.
https://www.galleri.com/ [galleri.com]
When a blown tire is enough to trigger bankruptcy for a lot of people, this might as well be $949,000. And it likely will be, by the time this little test drums up the rest of the business associated with treatments.
Sorry, but the Cancer Industrial Complex IS corrupt enough to create tests like this, and demand that insurance companies force you to start treatment right away. Others are correct regarding immune systems beating cancer. TBH, I'm sure how accurate this next statement is, I heard a statistic
Blech (Score:2)
> You simply get old and weak enough to be consumed by it eventually but that doesn't mean you're not naturally strong enough to beat it.
I just buried another friend who died of cancer (pancreas) and he was not even 45. He survived about a year.
I am old enough to remember 3 people who would likely be alive today because the cancer they had back then is treatable today. It is very likely would have been healed and it‘s likely I forgot someone because a lot has changed within the last 40 years.
I
Re: (Score:1)
Re: (Score:2)
Yes, they will. It's cheaper to treat cancer early than it is to deal with the long expensive final decline of someone who wasn't discovered to have cancer till way too late.
May not be sensitive enough (Score:3)
From the article:
"“The study was done in patients whose cancer was already diagnosed based on other tests and this screening technology still needs to be tested in actual screening trials before routine use.
“But it already allows a glance at early cancer detection in the future which will almost certainly be built around liquid biopsy tests, which detect cancer DNA in the bloodstream.”
Meanwhile, the results of the NHS pilot of the test, which will include 140,000 participants, are expected by 2023."
This test was developed using patients who have diagnosed cancers. That means those patient's had tumors advanced enough to become symptomatic or be caught be existing screening. However, I suspect that trying to detect smaller cancers in much earlier stages may be far more difficult. And it sounds like it is not sensitive enough to replace good screening tests that we already have like the colonoscopy. Once we see how it performs in the mentioned pilot study and how much it costs, then we can start the process of figuring out what it will be good for.
Is predictive power good enough? (Score:2)
I'm just wondering if the predictive power of this test is good enough?
To actually answer the question we'd need data for cancer incidence for the population for which this test would be used.
But I'll do some rough estimation. From [1] I'm estimating cancer prevalence to about 5% (Note: I should be using incidence, so this will be a very conservative over estimate):
cancer_prevalence = 16e6; # Based on 10 most common types of cancers in the United States
for a population of:
Re: (Score:2)
[High rate of false positives if undetected cancer is rare may be a problem]
But note that this test also says WHICH of 50+ cancer types it's saying "maybe, watch out" for. The next step would be to confirm the suspicion by taking one of the specific tests, if one is available, or examine imagery or the like if one is not. Doing that for (typically) just ONE type of cancer is a LOT cheaper than doing screening by doing a WHOLE BUNCH of tests and imagery for those 50+ cancer types.
What seems even more impor
Re: (Score:2)
"I'm just wondering if the predictive power of this test is good enough?"
It's good enough for a: "You have pre-existing conditions, Sir, I'm sorry"
This Test Only Proves... (Score:2)
...that the more you look around the human body for problems...the more likely it is that you will find problems.
The human body is an incredibly imperfect machine that works pretty well inspite of all of it's flaws. Some of those flaws are due to it's imperfect gene replication process of the human body.
IMHO...programming in C with compiler errors only to guide you will probably uncover fewer problems that poking around the human body.