Those are cumulative sensitivities on the right. Try and work out the individual sensitivites for each range and compare that to Avacta's. Imagine a situation where you had all positive samples having a Ct for e.g. between 30 and 32 and the sensitivities will collapse.
CI are dependent on sample size sure and in the context of a headline sensitivity, but when broken down like mentioned, reveal no new information. Sample size aids both.
@tecbadger
You don't need confidence intervals if details about the true/false positives/negatives are broken down by specific Ct ranges like the case with Avacta's results, not simply cumulative sensitivities.
The majority of these companies validate their tests on one sampling method and then advertise the same performance for another - the ambiguous 'nasal swab'. Also, companies purify their antibody reagent to a quality beyond what it would typically be in a mass manufacturing setting (batch to batch consistency). And then you have your statistical magic ('pooled results' of 2 sampling methods) and PCR equipment variations that can in effect boost Ct values. And the obfuscation of certain details such as how many positive samples were in the higher Ct range that were correctly/incorrectly identified vs just a cumulative sensitivity (with a very small % of total positive samples being in the higher ranges). These are more applicable to Mologic's studies than Avacta's (but the Chinese ones ultimately). Avacta would be better off with more studies being published that's for sure. The difference with Avacta compared to many other testing developers though is that their performance marries with theory quite easily and this is why I personally am not too worried about inflated results. Affimers > antibodies so of course the test achieves those results. It's a lot more obscure with other companies who seemingly only upgrade to more expensive materials etc - they use gold instead of latex for e.g.
https://www.omegadiagnostics.com/Products/COVID-19/VISITECT-COVID-19-Antigen
Ask yourself why haven't ODX updated this page with the independent, (seemingly) much better FindDx study results? Why even quote 98% sensitive at Ct value of 20 when the 'standard' is 25? Probably the same reason why Avacta quote 98% sensitivity despite achieving 100% at Ct<27.
Pretty sure it was the La Paz Hospital in Madrid that did the validation but could be wrong about that.
Anyone's guess whether they've been producing lfts since before recent validation results, though probably not given Alistair said validation results can now make commercial commitments firm. After all its an inventory risk to the manufacturers as well as to Avacta. May have been stockpiling the Affimers though so reagent production bottleneck may not be a problem like others will have if that's the case.
That is a subtle but very significant boon for Affimers - batch to batch consistency is very real. Apart from validating their tests using nasopharyngeal or mid turbinate/oropharyngeal sampling despite advertising anterior, other lfts have almost certainly been using the purest of antibodies just for their validations. In a real world mass manufacturing setting, they wouldn't be able to replicate that quality without increased lead time or a financial hit. Govt have highlighted their preference for batch to batch consistency probably because they've evaluated many inconsistent lfts.
Yes, pls provide a little more detail if you can. If I remember correctly, phase 3b (or 3a) was to be done in lots of 2 tests at a time. Avacta/GAD already producing quantities for 'other validation studies' - perhaps the 4000 needed for phase 3a/b.
'Reading the test results EARLIER than the recommended 5 minute time can cause false negative results. Reading AFTER 5 minutes can cause false positive results.'' Quote from another Mologic test's IFU here https://www.google.com/url?q=https://mologic.co.uk/wp-content/uploads/2017/03/MOLOGIC-PERiPLEX_IFU-V2-Dec2016.pdf&sa=U&ved=2ahUKEwiJ476JpbLwAhXzSxUIHYpKCgMQFjAAegQIAxABusg=AOvVaw278i5rvyvIJmB40jEPGvr8
Beyond the speed of the sample/buffer solution "wicking" up the test strip, the speed to results should then be about minimising the false negatives and positives as stated above. Avacta illustrated the difference in read times 10 vs 20 in the investor presentation with that graph. A 10min read would be possible for high viral load samples (which was the original purpose of these lfts anyway), with lower viral load samples requiring longer simply because the test is immunochromatic which means, at 10 min for a low viral load sample, the line would be so faint you'd need to wait up to 20min for a discernible colour to form. This is why Excalibur's sensitivity was boosted through their partnership with Sensodyne, which enables even a faint colour to be read from the test strip. This is why lft readers are useful though impractical for home-use.
At the other end, waiting too long can cause false positives and given that Affimer's are high affinity molecules that inherently boost specificity, Avacta could theoretically "wait longer" without suffering false positives; the Mologic test would not. In the context of the 30min tests, those probably utilised different materials/buffer that "wick" slower. I am of the opinion Avacta's CV used a universal 20min read time regardless of viral load (and this perhaps explains the 1 false positive reading), explaining why their sensitivity ended up so much better than Mologic's despite them both having similar analytical LODs. Had they encouraged a 10 min read time, the false negative rate increases and sensitivity reduces. As Mologic don't have Affimers they can't wait as long to effectively boost sensitivity given the negative effect it'd have on their specificity. Given Mologic have been working with Avacta to enhance their test and govt interest in a standard scaleable design, I doubt materials/buffer that affect "wicking" speed is the cause for the difference in recommended read times. Overall point is Mologic have to be 10min give or take while Avacta can be anywhere between 10 and 20min give or take (or even less than 10 if the viral load is particularly high).
Some good points raised about the practicality of having a smaller range of time from which to read results in an at-home setting. Again, Avacta's test is seemingly more robust in this respect.
Been trying to work out why the Mologic CV results seemed pretty poor in the previously posted results here https://mologic.co.uk/wp-content/uploads/2021/03/COVID-19-rapid-antigen-test-February-16-2021-v1.pdf
And why they seemed to have improved a lot in the recent Finddx report here https://www.finddx.org/wp-content/uploads/2021/04/Mologic_Ag-Public-Report_v1-20210423.pdf
From the Avacta investor presentation (and originally from the Liverpool Covid SMART study), there’s a table that compares viral loads and the corresponding Ct values of the PCR equipment used by PH Porton Down and Glasgow. There’s an obvious discrepancy in that Glasgow’s Ct values are lower for the same viral load than Porton Down’s. Ct values across PCR equipment is not equivalent, this is well established in academia too as far as I’m aware.
If you go through the 1st set of Mologic data above you’ll see they used Cepheid’s Xpert Xpress, and in their 2nd set of results, they verified roughly half their samples with the Roche Cobas PCR kit. The Cobas kit is the same kit used in the PH LFT evaluation protocol here https://www.gov.uk/government/publications/assessment-and-procurement-of-coronavirus-covid-19-tests/protocol-for-evaluation-of-rapid-diagnostic-assays-for-specific-sars-cov-2-antigens-lateral-flow-devices
So it could be assumed the PCR kit/equipment used for the Porton Down Ct values in the above table is Cobas.
The investor presentation also indicates what PCR kit Avacta’s CV results used – Thermofisher’s TaqPath. A (very rough) comparison of TaqPath and Cobas can be seen here https://www.medrxiv.org/content/10.1101/2020.11.15.20231795v1.full
The median Ct value for Cobas was 36 and about 32 for TaqPath’s. Just to reiterate, this is all very rough and I’m sure there are details etc I’ve overlooked. So in other words, when Mologic claim 96.4% sens in Ct values <25, this actually translates to a Ct range of something like <22 (give or take) had they used the same equipment as Avacta. And is a lot more believable given the 1st Mologic results of 98.4% sen at Ct <20. Even more if you consider that ‘Ct values ranged from 30.1 to 37.9 (mean 36.7 ± 1.9) on the Roche cobas 6800 assay and 24.6 to 42.4 (mean 32.8±4.1) on the Cepheid assay’ here https://www.biorxiv.org/content/10.1101/2020.05.05.078501v1.full
This may mean that Cepheid is of a more similar standard to TaqPath and by extension the Glasgow PCR kit in the above table. Cobas requires more cycles to detect the same viral load than these alternatives and therefore, in the context of Ct comparison, is quite liberal.
For whatever reason, it appears that Avacta have been using a more conservative PCR kit (as far as these Ct values are concerned anyway) so anyone accusing Alistair of ramping needs to look elsewhere.
This is some very rough research and I don’t assume that any of this is correct, so I’d welcome anyone with expertise to provide their pov.
You mean the breath test by DeepVerge that uses Affimers? If saliva was so hard in being a consistent sample type I would imagine breath similarly so. Welcome to corrections on that assumption though.
LFTs are complementary to PCR according to Alistair.
Death throws for growth of the market but not the market entirely.
And the PCR competitors that rise to the top will probably do so my cutting prices given the huge expense that PCR is atm, not as much through technological innovation. So shrinking market with shrinking margins with shrinking forst mover advantage.
@allah4uk
Alistair did say during the investor presentation (as part of the explanation why BAMS commercial strategy is being revised) that PCR companies were losing contracts because the PCR market is oversaturated at the moment. Probably because LFTs are replacing a lot of the testing demand (and way too many PCR competitors).
Also worth not forgetting that if Avacta hadn't have strived for a UK govt contract, the partnerships we have developed along the way (which will no doubt help the future diagnostics business) would not have happened at least not to the extent they have.
You still want Avacta at the heart of UK govt testing, even with the excellent CV results, the endorsement would propel demand even further as we soak up competitor demand that would otherwise rely on potential buyers being competent at reading, understanding and comparing CV results (of which the majority are misleading). This is more relevant for longer term demand as the higher quality LFT (at least perceived) revenue will survive far longer. Not to mention the endorsement for the other therapeutic side of the business. I do get that demand > supply atm but still very important to maintain that demand for as long as possible. Besides I expect overseas manufacturing to dwarf UK manufacturing so we might very well be able to service the extra demand.
Maybe as a distributor if anything
Antigen, home use, anterior, 15min, detect all variants. Pretty standard to me if you're suggesting it is Avacta they have been working with.
Also worth pointing out that we were also given confirmation that Mologic had been working on our LFT architecture (outside of the S-N 2nd gen test). The originators of the pregnancy test have sacrificed not only manufacturing capacity and low mid income country sales (though at low margins if at all) but now their own test's commercial success by improving a competitor's product. Chris Malloy (key DHSC figure if I remember correctly) commented on a LinkedIn post a while back I believe about companies dropping commercial rivalry for the sake of the better good - this is surely what he was talking about.
So 5 mil low end and 30mil by end of year dependent on access to uk govt loaned equipment. Fits nicely with idea uk manufacturing capacity has been largely absorbed by the uk govt. Only after the uk govt have finished producing their sovereign test as part of the free lft scheme will Avacta get more than a sliver of uk manufacturing for private selling (after all govt paid for ODX's ramp up). The uk market will be dominated by the sovereign test that's clear of which Avacta seemingly are front runners in having a hand in. Focus on Europe etc therefore makes sense as the only thing Avacta can actually talk about outside of an NDA.
Was raised before but Avacta have been in talks with overseas manufacturing for months. Alistair said the CV had prevented any firm commitments. With ongoing talks with distributors being so extensive it seems assured that the overseas manufacturing will be similar in extent, dwarfing any uk capacity, probably at a discount too. Not to mention Medusa's supposed 50m/mo capacity. If Avacta is central to the sovereign test, regardless of the excellent CV results, our test will be favoured by all given the endorsement.
I am of the opinion that the top performance of antigen LFTs so far based solely on materials etc has been achieved, i.e. the best combinations have been refined. With the only bottleneck in performance left being the reagent.
What's (actually) funny is Novacyt holders berating antigen LFT stocks for having tests they denounced regularly as harmful and utterly ridiculous vs PCR and then need an antigen LFT themselves to get themselves out of a hole. Given Novacyt have no innate advantages of developing an antigen LFT themselves, knowing the partner is key to determining what sort of quality test it would be. Novacyt management does seem better than most (not including Alistair fyi) so I expect them to expect it to be at least be better than the current crop of antigen LFTs ex Avacta to even attempt to sell.
The application to the antigen LFT of the 'detect all variants' rhetoric Novacyt have adopted is topical but most LFTs would detect all variants so not exactly a USP. To what degree they are affected by variants is the real question. But I would have thought Novacyt would be boasting about having a very sensitive and specific LFT instead. Perhaps nothing spectacular in terms of quality, just another product to get shareholders off their back about what the **** they're going to do after Covid.