Thread by @doctorhammond

#SIRT for advanced HCC has been recommended by @NICEComms. There may be some reasons for doing this but #costeffectiveness is not one of them. A short Tweetorial. @BSIR_News @cirsesociety @BASLedu nice.org.uk/news/article/h…

The relevant population here is people with advanced HCC. BCLC C, CP A, PS 1-2, unsuitable for conventional transarterial therapy (eg. portal vein thrombosis). These people have a dismal prognosis: most are dead within a year.

Treatment options are limited. Sorafenib is a protein kinase inhibitor and prolongs life slightly (vs best supportive care: SHARP trial) but at significant expense. It was originally rejected by NICE on this basis so Bayer applied via the Cancer Drugs fund (ie: reduced the cost).

The cost of sorafenib is commercially sensitive so we don't know the exact ICER. NICE says it is 'most plausibly' below £50k/QALY gained when compared with best supportive care (placebo). This is much more than the conventional 'willingness to pay' threshold (£20-30k/QALY)

The reason for approval was that Sorafenib qualifies for consideration as 'end of life' treatment. Essentially this means NICE values QALYs at the end of life more highly and is prepared to pay more for them. nice.org.uk/guidance/GID-T…

There is not much data on SIRT vs systemic treatment and no RCT data on it vs best supportive care. The SARAH found no difference in overall mortality between @SirtexMedical SIRSpeheres and sorafenib.

Note that patients in SARAH did particularly poorly compared with patients in the SHARP trial. A recent NIHR Health Technology Assessment concluded there were no meaningful differences in survival between SIRT and sorafenib. journalslibrary.nihr.ac.uk/hta/hta24480/#…

However, NICE's analysis is based on SIRT being advantageous over sorafenib in its side effect profile so may provide more QALYs (NICE suggests 0.047). Overall NICE's analysis suggests SIRT provides 0.029 fewer QALYs than sorafenib.

BUT, after discounts are applied to reduce the cost of SIRT, (Patient Access Scheme) SIRT comes in cheaper than sorafenib. Again, these discounts are commercially sensitive but NICE states the savings made per QALY lost were >£30k/QALY.

In other words: NICE says SIRT is not as effective but this is offset by the saving we make investing in SIRT instead of sorafenib.

(Aside: the uncertainties in these estimates of ICER are a problem, due to the commercial sensitivity of the actual cost, but they must be reasonably accurate as NICE would have used £40k/QALY or £50k/QALY if the ICERS were more in these ballparks)

The first cost-effectiveness problem arises here: in approving sorafenib, NICE decided it met end-of-life criteria and therefore the QALYS gained by sorafenib over BSC are valued at £50k each. So this is the relevant threshold by which to judge the value of QALYs lost using SIRT

We should GAIN £50k per QALY saved not using sorafenib. Even in NICE's calculations, this threshold is not met. So when judged by the relevant criteria, SIRT doesn't save enough money to suggest disinvestment in sorafenib.

The other cost effectiveness consideration is how SIRT compares with (cheap) BSC. I will be networking trials together here (SARAH and SHARP) which comes with problems but bear with me...

Because SIRT comes in at £30k/QALY vs sorafenib, that means it MUST come in at greater than sorafenib's ICER vs BSC. So SIRT vs BSC must cost at least £50k/QALY and may cost much more.

For those familiar with cost-effectiveness planes, see figure. SIRT vs BSC will be one of the green vectors, dependent on the relative number of QALYs gained by sorafenib vs BSC, and lost by SIRT vs sorafenib (the place of SIRT on the £30k/QALY line, C-B)

The only way SIRT can be <50k/QALY vs BSC is if the broad ICERS quoted by NICE were out by £20k/QALY which seems unlikely. So SIRT vs BSC is not cost effective at standard willingness to pay thresholds (£30k/QALY).

And it's not suitable for the end of life uplift as it does not prolong life vs standard care (sorafenib or BSC).

SIRT is cheaper but worse then sorafenib. If we are going to disinvest in sorafenib, we'd (in cost-effectiveness terms) be better to disinvest to BSC than to SIRT. There may be arguments for recommending SIRT for intermediate stage HCC, but cost effectiveness isn't one of them.

[Ends]