The results Grail presented as a talk at ESMO are fairly similar to those they presented in somewhat less detail at conferences earlier in the year (AACR and ASCO).
For those who are interested in a deeper dive on this technology, its potential limitations, and its potential benefits, I wrote up a three-part series earlier this year breaking down recent advances in liquid-biopsy-based cancer detection:
Part III: clinical performance and economics - integrating the sensitivity and specificity numbers with a simple model for important clinical parameters like adherence (how many people prescribed to take the test actually do), and cost: https://ihaque.org/posts/2019/08/18/early-detection-mid-2019...
Disclaimer: I'm the former CSO of a competing company, but I no longer work in this space.
From [0], the types of cancer they can screen make up at least 81.4% of new cancer diagnoses, which means that with their 99.4% true positive rate, this test should be able to correctly identify ~80% of new cancer incidences. If I'm doing my math right, that's really impressive.
It seems to have a great false positive rate, but a not-good false negative rate. It misses the bulk of stage one cancers. So its utility for early diagnosis (the important bit) is marginal at this time.
For those not in the lab business: a low false positive rate and a high false negative rate are useful for confirmatory testing, not for screening. Such tests are terrible for screening. So this might be able to insert itself between the screening test (e.g. mammo, psa, or CXR) and the biopsy, but it won't replace the biopsy because you still have to characterize the cancer for further treatment decisions.
From a market positioning standpoint, it's not bad, because you don't piss off the generalist physicians by stealing their screening role, and you don't piss off the pathologists or surgeons by stealing their biopsies. The question is, does it add sufficient value for the payer to pay?
> For those not in the lab business: a low false positive rate and a high false negative rate are useful for confirmatory testing, not for screening. Such tests are terrible for screening.
Newbie question: wouldn't a test with very few false positives, but frequent false negatives, still be good for screening if it was cheaper/easier than existing screening, so it could be given to more of the population based on a weaker suspicion threshold? Like, in the limit of it being free and having zero false-positive rate, you'd give it to everyone.
(Maybe given your expertise it's obvious to you that the test is sufficiently expensive that this argument doesn't apply?)
Frequent false negatives mean it will miss many people who do have the disease, leading them to believe they are healthy until the disease progresses to a point it may be more difficult to treat.
Many of these early cancers are asymptomatic and universal diagnostic imaging in the absence of symptoms is cost prohibitive. Better blood tests address that and other important gaps.
You’d follow an asymptomatic false positive blood test with at worst a whole body MRI and at best a targeted imaging study (if the blood markers are specific). You’d have a further false positive rate from that but it drops off quickly. The cost is manageable. And some of that cost is offset because you catch cancers earlier when they are less expensive to treat. A substantial cost savings will come through first line analysis of imaging by AI, to highlight images for radiologists to focus on.
But if the tests are cheap enough you can repeat them and compare. Bringing cost of blood diagnostics down is I think a good idea. I have fitness and health conscious friends who regularly get direct-to-consumer tests as a matter of course. In a few cases they have caught issues very early.
This is understanding there are some false positives, but being able to recognize them and learn to ignore them is better than not testing out of fear you overreact.
Repeating a test doesnt help. The blood result likely wont be different the second time. A false negative is a person for whome the test just doesnt work, no matter how many times you repeat it. (In reality, these sorts of important tests are run multiple times against the single sample before they print the results.)
How so they determine false negatives? Are there other less broad and more expensive test?
That said, isn't it also possible at Stage 1 for the body to cure itself? That is, __in theory__ is it possible Stage 1 defecation is too early? That med intervention could be premature?
In addition to the press release someone posted, if you’re interested in some of the (high level) science behind this, I’d encourage you to read the ESMO presentation slides this news comes from: https://grail.com/wp-content/uploads/ESMO_2019_Oxnard_CCGA2_...
I don’t think the article mentions it, but the study is funded by Grail, which is commercializing a variety of related technologies.
A pathologist friend told me “there’s a good chance you already have the cancer that might eventually kill you.” The danger of early detection of cancer is that we’ll end up treating fat more people than we would otherwise, and cancer treatment is often brutal. I’ll be interested to see what better detection actually does for mortality rates.
One day I hope that I will have a device that can digitize my blood sample at home so that remote cloud processing can warn me of issues. Life can sometimes get in the way of preventative checks. Social anxiety might be a part of that too.
I was thinking about it in terms of cataloguing the foods you eat for later analysis, but could work for this too.
Would it be possible to make portable device that could freeze dry a drop of liquid? Heat pumps and pressure vessels both get much simpler with smaller sizes.
This is very promising. The performance will definitely get better, this is like the first smoky pictures coming out of an MRI.
The real challenge is going to be dealing with the result... it isn’t simply a matter of continuing to use the same treatments we do now but applied to earlier stage disease.
...Actually, it is. Survival when using the same treatments is much higher in earlier stage patents is much better than in later stage. Full stop. Of course not everyone will be cured, but this will make a big dent.
By the birthday paradox, that means 11% of healthy people will be misdiagnosed for one form of cancer on the 20 test panel.
That's going to require some in person discussion and review... maybe they can handle that with the NHS, but I think the US patient would end up spending an extra $1000+ due to required human contact, beyond any test cost.
It isnt 20 tests. It is one test regime that detects the presence of 20 diseases. It doesnt diagnose any one disease, just red-flags patients for more accurate testing.
It seems quite obvious to me that the 0.6% is already accounting for there being multiple tests, and even if not, confounding factors are likely not completely independent per test.
For those who are interested in a deeper dive on this technology, its potential limitations, and its potential benefits, I wrote up a three-part series earlier this year breaking down recent advances in liquid-biopsy-based cancer detection:
Part I: regarding tumor fraction (this goes into depth about the stage I limitations noted by other commenters) https://ihaque.org/posts/2019/07/02/early-detection-mid-2019...
Part II: technical advances from Grail (the authors here) and Guardant (one of their main commercial competitors) https://ihaque.org/posts/2019/07/08/early-detection-mid-2019...
Part III: clinical performance and economics - integrating the sensitivity and specificity numbers with a simple model for important clinical parameters like adherence (how many people prescribed to take the test actually do), and cost: https://ihaque.org/posts/2019/08/18/early-detection-mid-2019...
Disclaimer: I'm the former CSO of a competing company, but I no longer work in this space.