That's a fair point. The rejection of the null hypothesis here doesn't necessarily have to do with an increase in concentration with all other things equal in this case. However, in random samplings from the same heavily pre-treated population, it does seem like the recommended dose of Cabo is less well tolerated than the recommended dose of Everolimus judging from the relative degree of dose reduction. Anecdotal evidence from COMET-1 experimenters describing results seemed to indicate that relatively healthier patients could tolerate more exposure to Cabo, whereas healthier patients would presumably take roughly the same dose of Everolimus as sicker patients since there wasn't much dose reduction for them so it's possible HR for a frontline Cabo vs Everolimus trial could be better. We'll see when they publish CABOSUN topline results and the dose adjustment they used if this theory holds water.
("Conclusions: Based on the ER analysis, C was efficacious at all doses evaluated in METEOR, including those resulting from dose reductions, with higher exposures predicted to correlate with improved PFS."
Also, in that same abstract:
"The best Emax model (p less than 0.001) was achieved with EC50 = 100 ng/mL, a value below the typical individual predicted steady-state average concentration at 60 (1125 ng/mL) or 20 (375 ng/mL) mg/day. "
I take it this means that less than full titration gave them the clearest model...)
That's set to finish in 2019 but I think it's already finished enrolling so results may drop sooner than expected. There could be a lull in the immune-checkpoint storm when Cabo has a chance at the firstline therapy spot. Toxicity and durable response look very encouraging for the Nivo3 + Ipi1 combo; it's unclear if this is better than Cabo's AE profile, ORR rate, and OS firstline, but I think PFS is almost certainly worse if Sunitinib is getting ~11 months. Also, 8 months of the combo therapy will be something like $220k unless BMY cuts a deal for the first 4 months as they seem to do for Melanoma.
It still isn't curative for the majority of patients. Cabo has a totally different mechanism of action and will find plenty use second line as the king of the VEGFR TKIs as compared to other PD-(L)1 agents.
Here's an abstract from ASCO to support that a higher concentration of Cabo is correlated with disease progression rate:
"Population pharmacokinetic (PopPK) and exposure-response (ER) modeling of cabozantinib (C) in patients (pts) with renal cell carcinoma (RCC) in the phase 3 METEOR study."
"A statistically significant relationship was identified between the rate of progressive disease (PD) or death and the time-varying average daily C concentration. Increases in C concentration were predicted to decrease the rate of PD or death in a nonlinear manner."
The median everolimus dose from METEOR was 9mg (from 10mg starting)for Everolimus versus 44mg (from 60mg starting) for Cabo. A healthier population like that in CABOSUN could probably tolerate something closer to 60mg for longer and derive more benefit as evidenced from this abstract.
Presumably many practitioners already have Everolimus as it's a common second-line therapy and they call up Eisai for the lenvatinib to add to the mix. They'll make a better profit selling both drugs at retail prices.
Sumanta Pal weighed in again on Onclive today stressing caution over safety concerns and ultimately concluding once again that Cabo is preferred in second-line:
"Where does lenvatinib plus everolimus fit into the treatment paradigm, in terms of sequencing?
The level of evidence is higher with cabozantinib and nivolumab, as both of those agents were approved on the basis of randomized phase III trials that included many more patients.
In my opinion, the standard second-line treatment should be cabozantinib, because we know the agent led to improvement, not just with PFS and response rate, but also in overall survival (OS), as well. The third-line treatment of choice would be nivolumab, given that this agent has been shown to improve both response rate and OS. Beyond that, perhaps lenvatinib plus everolimus represents a reasonable option. Certainly, the combination is preferred to everolimus alone now."
It's especially galling that they excluded patients with liver mets in COMET-1. The ongoing Posadas trial is more focused on prostate cancer patients with visceral mets and this may guide further development in prostate. (I'm not going to go so far as to suggest approval based on a good result here as the standard of care in prostate is rather different after abiraterone and enzalutamide demonstrated OS benefits with modest toxicity.)
Google "A Closer Look at METEOR Results Confirms Benefits of Cabozantinib Across RCC Subgroups". I pm'd you a more detailed message.
If we assume assume the cox model for hazard and proportional hazards along with the same relative performance of Cabo vs Everolimus in first-line as in METEOR (PFS HR = 0.52), an estimate using RECORD-3 Sunitinib vs Everolimus (PFS HR = 1/1.4) would give Cabo a hazard ratio of roughly 0.52/(1/1.4) ~ 0.728. (Using the earlier METEOR HR would give ~0.8)
This is extremely rough but well within the realm of possibilities for what the results of a smaller trial like CABOSUN could show.
A update in January showed a revised PFS HR of 0.52 for METEOR and a nicer separation at the tail end of the KM curves for PFS. The same nicer separation happened to the OS KM curve as quantified by the better HR of 0.66 on the FDA label than the 0.67 w/ a higher p value achieved on the first interim.
In heavily pretreated METEOR indicated a lot of dose modifications for the Cabo arm despite having the almost the same discontinuation rate as Everolimus; pretreated patient ability to tolerate extended Cabo therapy had also been cited as a major reason for COMET-1 failure. I hypothesize that Cabo would be relatively better than everolimus in first line because the same full dose everolimus could be given in both situations to patients whereas a relatively larger dose might be tolerated in first line with Cabo than in second line so they might be exposed to a greater effective dose in CABOSUN than in METEOR.
Hbomb, the down playing of CABOSUN was done by Morrisey at the Cowen conference earlier this year. He said that it had been the initial intent to try to move to front-line therapy but it sounded like plans had changed once PD-(L)1 inhibition came into vogue. In all the quarterly transcripts, it is mentioned that this is not designed to be registrational; but as was mentioned in the other thread, an independent radiological review was permitted for the L+E trial (the FDA review docs attest to this) and if the results are indeed stat-sig in the traditional sense, we might be in the door w/ CABOSUN. The safety threshold for adjuvant therapy is more rigorous but the FDA did make a show of leniency for L+E.
"They actually first released the HR with the July 2015 press release along with the PFS HR and p values. The only stats they did not release were the medians, which were okay, but not inspiring. They cherry picked from what was available and told the best story they could. I don't intend it as a criticism. It is what it is. Even back in 2011 with the EXAM trial they released all the topline PFS stats with no mention of the interim OS analysis that at the time had HR=1.0. The E1512 press releases also highlighted the best stats. The rational thing to do reading their press releases is to assume they are they are painting the best picture possible. It's not dishonest, its good PR, but from an investor's viewpoint you should be careful to not assume more than they are telling you."
I think it's entirely possible that the effect size is big enough for Cabo to achieve statsig PFS against sunitinib whose HR in frontline clear cell RCC vs everolimus appears to be 0.70 from the RECORD-3 trial. The biggest clue is simply the PFS 0.52 HR against everolimus from METEOR, and I would also expect frontline patients on Cabo as their first VEGFR based TKI to do much better than patients who have had any number of prior therapies, and have mets (including bone and pretreated CNS)
One more potential thing to add to your list is EXEL retiring some debt early provided we hover around $7 for long enough to trigger the necessary conditions. Is that a better return (prepaying while share prices are still lowish) than investing in a new trial? I don't know but it would be a vote of confidence from the management in future sales and dispel the debt gloom overhanging the stock since 2012.
Hind sight is 20/20, and I think given what they knew at that time, they weren't as bass-ackward as they seemed.
At that time, AVEO and BMY seemed likely to sew up the second line space with their drugs Tivozanib and Nivolumab respectively. Tivozanib was a TKI and the early PFS numbers vs Sorafenib seemed really solid, so adding another TKI with slightly better PFS than the standard of care to the mix wouldn't have seemed exceptionally compelling. You'll note they started METEOR after the ODAC rejected Tivozanib due to its rather questionable trial design allowing for crossover from the control arm to tivozanib but not the other way around. (Presumably they wanted to save a little money by not having to buy the extra Sorafenib, but this resulted in a slightly negative trend in OS since sequencing TKIs seems to generally confer survival benefit.)
You can also imagine the prostest they would have drawn from investors if they didn't pursue the unprecedented bone scan responses in prostate cancer. We would have revolted over something which ended up not being a very durable response, and not improving OS by much in the population they selected.
Nice reading between the lines. I think the embargo on trial data has more to do w/ meeting submission deadlines than hiding bad data.
This onclive article: "Frontline Cabozantinib Improves PFS in Renal Cell Carcinoma" adds the following details:
"Exelixis plans to submit the full CABOSUN results for presentation at an upcoming medical meeting and communicate with regulatory authorities about a potential first-line cabozantinib indication in RCC."
"Attaining p=.05 or less requires an HR of approximately .70. That would imply a 2-3 month PFS improvement over Sunitinib."
HR of Cabo in second line vs Everolimus in METEOR was 0.52 in January. HR of Sunitinib vs Everolimus in first line RCC from RECORD-3 was 0.71 (=1/1.4). Just fudging around cross trial but ratio of the hazards of each versus everolimus would be around 0.52/0.71 ~ 0.73. In a small trial like Cabosun, I think
I think it could mean another swing at prostate cancer as well. The sentiment of the medical community seems to be that anti vegfr post anti AR are distinct and the AEs stack. COMET-1 was also on an especially advanced population who all had received anti-AR therapy and wouldn't be able to handle this agent very well; despite this there were benefits to PFS and bone related metrics like time to first SREs. Also, patients with visceral mets and those who had received prior Cabazitaxel did better (as the patients who didn't, especially in the placebo arm more often went onto Cabazitaxelthan in the Cabo arm).
All these can inform the design of a trial with a better chance of success in a more appropriate population, perhaps in combination with something like abiraterone.
Take a look at:
"Hypertension Induced by VEGF Signaling Pathway Inhibition: Mechanisms and Potential Use as a Biomarker"
"Unlike traditional “off-target” side effects, hypertension is a mechanism-dependent, “on-target” toxicity – reflecting effective inhibition of the VEGF signaling pathway rather than non-specific effects on unrelated signaling pathways."
So that there is a lot of hypertension in Cabo patients means it's working. VEGFR inhibitors block abnormal blood vessel growth and normalize vasculature; this can allow for more efficient delivery of various other drugs to the tumor.
Here is a very on point article about this topic as it relates to PD-1 checkpoint inhibition:
"A Surprising Match: Cancer Immunotherapy and Mismatch Repair"
The immune system preferentially targets cells with many DNA errors as there are more unusual proteins being produced from such a cell. There's also alkylating and crosslinking agents whose mechanism of action is to induce these types of errors in tumor cell DNA.
Nice point. Do you know if this strategy of using a MEK inhibitor to sensitize checkpoint inhibitors to mismatch repair proficient tumors is more generally applicable than in CRC?
If Cobi were expanded into another indication with a Roche drug, what would prevent them from further lowering the price of Cobi and raising the price of their wholly owned drug and sweeping more of the profit into their coffers?
Looks like a Phase I. BRAF+ Melanoma will probably be dominated by Nivolumab + Ipi since cost and durable efficacy may be comparable w/ a triplet. Not to mention: time to market may well be early 2020's if the trial succeeds.