Lucentis/Avastin redux
Over the past decade, first-line standard of care for multiple sclerosis (MS) has evolved toward defaulting to anti-CD20 monoclonal antibody therapy such as Ocrevus (ocrelizumab, Roche) or Kesimpta (ofatumumab, Novartis). These are effective but prodigiously expensive agents. There is however another anti-CD20 monoclonal: rituximab, with biosimilars available at a fraction (<10%) of the price of Ocrevus. It’s been shown to work in MS in a phase 2 by Roche, before Roche decided to develop Ocrevus instead, and it was never prospectively directly compared with Ocrevus. Given that substitution of Ocrevus would result in very large savings, Norway has undertaken a head-to-head RCT of rituximab vs. ocrelizumab in newly diagnosed relapsing MS with a primary goal of checking non-inferiority for rituximab. The goal was met – Ocrevus and rituximab (the Sandoz biosimilar Rixathon was employed) performed very similarly. I have two comments.
First, although rigorously conducted, this was a comparatively small study of 216 patients. A similar study, but with 600 Danish patients (DanNORMS) is about to read out. If it shows the same thing, the publicly funded healthcare systems of Norway and Denmark will likely switch to using rituximab for MS as a way to spare the public purse and the wallets of the taxpayers (that’s why they are funding these studies). Will US payers do the same and try to save some money for their members getting crushed by premium increases and deductibles? Can they push this with providers even though the label of rituximab does not include MS, and the original manufacturer of reference – Roche! – is not about to ask for an expansion?
Second, it seems obvious that payers should sponsor more of these studies. Almost two decades ago, a small VA study showed that $50 of Avastin did the job of 40x more expensive Lucentis, and this was confirmed by a larger publicly funded trial. Couldn’t a consortium of payers fund these types of studies? By my back-of-the-envelope calculation, it’ll roughly take the therapeutic conversion of 5,000 MS patient years to defray the cost of the validating studies, perhaps triple that with discount and risk adjustments. At 10,000 new patients a year in the US, that seems like a good potential ROI! Rituximab versus Ocrelizumab in Newly Diagnosed Relapsing Multiple Sclerosis
Letters from China: in vivo CAR-T for autoimmune disease
A bit more than two years ago, a study from Germany reported in the Journal and these pages (Hitting reset for the immune system) showed the potential of CAR-T therapy directed against CD19 (a B-cell target that casts a wider net than CD20 mentioned above) in unresponsive severe autoimmune disease including lupus and scleroderma. However, the cumbersome nature of ex-vivo CAR-T therapy (collect patient cells, culture cells, transfect cells, grow cells, test cells, sort cells, condition patient, reinject cells), has led to interest in methods to create the appropriate CAR-T in vivo (i.e. in the patient directly), either via mRNA, or via gene therapy administration. Both have now been done and published as correspondence (rather than Brief Reports as they might have been – perhaps for speed to print?) by two different Chinese groups, with what appears to be minimal toxicity and early evidence of clinical effect. The key distinction between the two is that using mRNA should only create a transient CAR-T population, but using a lentivirus could well result in a permanent CAR-T population if stem cells are transfected. Would that be good or bad? Only long-term studies will tell us. But it does feel like in this cutting-edge modality/mechanism space, groups out of China have the lead, a state of affairs that would have seemed very unlikely just a decade ago. Lentiviral In Vivo CD19 CAR T-Cell Therapy in Neurologic Autoimmune Disorders; In Vivo CD19 CAR T-Cell Therapy for Refractory Systemic Lupus Erythematosus
A broadly active biologic for COPD
Until the smoking generations of yesteryear pass on, the burden of COPD will remain high, severely degrading the life of millions, and causing hundreds of thousands of hospital admissions for COPD exacerbations in the US each year. Though multiple biologics have had success in asthma, COPD has been a tougher area, and just a couple of biologics are approved for a specific sub-population with eosinophilia, a minority of COPD. This is despite attempts to target a broad number of pathways, and this is why the latest attempt with IL-33 blockade was not at all thought to be a slam dunk. However, two phase 3 studies of tozorakimab (AstraZeneca) have now successfully read out showing a decreased rate (by about 30%) of exacerbations in a broad COPD population of current and ex-smokers without regard to eosinophil count. Concretely, looking at the numbers carefully it appears that treatment shifts the hospitalization rate from 0.33/patient year to 0.23/patient year, so roughly 1 hospital admission avoided per 10 patients treated for a year. No doubt this is a case that AZ will soon be taking to US payers. Tozorakimab to Prevent COPD Exacerbations
Biohacking cholesterol with base editing
PCSK9 inhibition has been back in the Journal recently with the emergence of oral therapy to compete with injectables. Now comes the radical alternative: disable the gene using base editing. The mechanism is interesting: what’s edited is actually an intron to force its incorporation into the mRNA. That intron happens to encode a stop codon, which would otherwise never be read, but now that it is part of the mRNA, blocks translation into the PCSK9 protein. Delivery is with GalNAc lipid particles (no AAV here) that are highly trophic to the liver (via ApoE plasma binding). This was an ascending single dose study with 35 participants with heterozygous familial hypercholesterolemia or premature coronary disease and LDL > 70 (actual mean was 129) and on maximum statin +/- ezetimibe. At the highest dose, the LDL fell about 60%, which is what one expects with optimal PCSK9 inhibition, and the therapy was well tolerated.
One key question now is long-term outcomes: both in terms of durability (is the LDL decrease maintained?) and safety (did base editing hit anything that might lead to a malignancy)? Efficacy is less of a question to me, given the fact that this treatment recapitulates existing natural experiments with kindreds who have loss of function mutations of PCSK9 and have a decreased CV event rate. It will be interesting to see what the asset strategy becomes, now that Verve, the originator, is part of Lilly.
The other question is who will use this when it is possible to get the same level of PCSK9 inhibition via occasional injections or soon, a daily pill in a way that does not involve irreversible changes to one’s genome. Perhaps one should think about it like this: even in the era of GLP-1 receptor agonists, there are still some folks who opt for bariatric surgery. In Vivo Base Editing of PCSK9 with VERVE-102 for Hypercholesterolemia
Pan-RAS inhibition on a tear
After the remarkable results of the pan-RAS inhibitor daraxonrasib (Revolution Medicine) in pancreatic cancer comes a follow-on in non-small cell lung cancer (NSCLC). In about 30% of NSCLC, KRAS is mutated and a major advance was the approval of inhibitors sotorasib and adagrasib (Lumakras from Amgen, Krazati from Mirati/BMS). However, these agents are specific to the G12C which only represents 40% of the mutations leaving other mutations unaddressed. For the others, it was a natural next step for Revolution Medicine to try their pan-RAS inhibitors, and this has paid off. In a phase 1 study treating KRAS (non G12C) NSCLC progressing after 1st line chemotherapy, 40% of patients had a response – a result similar to or slightly better than what was seen in the studies with the G12C agents. A pivotal controlled study vs 2nd line standard of care (docetaxel) is on-going, but it’s one of those situations where one wonders if there is actual equipoise between the two arms – I wouldn’t be pleased being assigned to the taxane arm. Daraxonrasib for Previously Treated RAS-Mutant Non–Small-Cell Lung Cancer
Excitement in small cell lung cancer (SCLC)
Typically, SCLC initially responds well to 1st line therapy but the response duration is not long, and options in 2nd line are really sub-par. This is why there is real excitement about a number of ADC therapies in development targeting B7-H3, a common marker of SCLC with various topo-I directed payloads. The current leader is tambotatug pelitecan (MediLink Therapeutics) which reported a Chinese phase 3 study of the ADC vs. a “bare” topo-I agent in the form of topotecan. Median survival was 13.3 months vs. 9.4 months, supported by a 59% response rate in the tambotatug arm vs. only 10% in the topotecan arm. Interestingly, the enrollment criteria did not include measuring the presence of B7-H3 in the tumor, likely because phase 1 studies have not shown it to be predictive of response. There is a mystery to solve here. Tambotatug Pelitecan in Small-Cell Lung Cancer after Platinum-Based Therapy
Testing the test – the GRAIL story
Just out online in the Journal (print version will be in Q4), we have results of the NHS-Galleri study though the synchronous article in Nature Medicine (link below) is at least as interesting. In brief, this was a UK study involving a socioeconomically diverse set of more than 140,000 participants randomized 1:1 to control or to yearly testing with the Galleri multicancer early-detection test. Given that the goal of testing is earlier cancer detection, the primary end-point was the relative incidence of stage III and IV cancers in the tested vs. control group after 3 years. The difference did not reach statistical significance, but there was a notable trend toward fewer stage IV cancers at the third screening round. The data are so rich that one could write several long essays on the findings. I will highlight two things that struck me. First, that the timeline to see this kind of effect was in my opinion overly ambitious: for hard outcomes in prevention, 5+ years is more like it. Second, that I am struck by the remarkable specificity of the test: the false positive rate (i.e. # tested who screen positive / # tested (positive or negative) without cancer) is about 0.5%. Compare that with lung LDCT or mammography where the false positive rate is ~10 times that. Though the test is only moderately sensitive (~10-60% depending on the cancer and stage), when it’s positive, about half of the time, a cancer is found. What this means is that relative to lung and breast cancer screening, useless follow-up investigations are much fewer.
So, what does this mean? There are purists who think that because the NHS-Galleri trial did not meet its pre-established primary end-point the test should not be approved by FDA, and this is reflected in split votes from the FDA AdComm. Unusually for me, I find myself in the non-purist camp for several reasons. First, I think the standard for approval of a test ought to be mostly a clear evidence base of what it does and here GRAIL has provided the Rolls-Royce of data packages, the like of which one rarely sees in healthcare outside phase 3 drug trials. Second, it’s incontrovertible that the test detects some cancers that would have been detected much later without it (3,637 participants were diagnosed with cancer in the intervention group; in 937 of those, the cancer was found after a positive multicancer test). What the failure to get the expected end-point reflects is that at a population level the benefits are small. If payers cover this test (which should not be a given), they should only do so at a price commensurate with the benefit, as good stewards of the premiums paid by their members. But for individuals who are doing full body scans on their own dime – I’d be hard pressed to say that this is a worse investment. Finally, my PCP orders a complete blood count + differential (CBCdiff) on me every year… where is the evidence for this? There is none, even though that is a routine thing done in the US. Should regulatory authorities demand that purveyors of CBCdiff tests prove that they are valuable for the general population? – in my view that assessment is the job of the buyers (either payers, or the actual patient / provider). Effect of Screening with Multicancer Early-Detection Test on Late-Stage Cancer Diagnosis (NEJM); Performance of a multi-cancer early detection test in the randomized controlled NHS-Galleri trial (Nature Medicine)
The New England Journal of Medicine is a premier weekly medical journal covering many topics of interest to the health sector. In this series we offer an opinionated perspective on selected highlights that might be of interest to our clients and others.