In theory this NCI trial could address Perioperative vs. Neoadj, but seems like it's still a few years out and won't answer the question of how they compare to only Adj

https://clinicaltrials.gov/ct2/show/NCT04267848

Study was initiated well in advance of any pivotal trial readouts in favor of neoadjuvant or adjuvant IO, so control arm is unsurprising in that respect. Merck, Roche and AstraZeneca all have very similar studies too

https://clinicaltrials.gov/ct2/show/NCT03425643

https://clinicaltrials.gov/ct2/show/NCT03456063

https://clinicaltrials.gov/ct2/show/NCT03800134

Drugs like Sovaldi and Gleevec are the outliers in terms of their miraculous, curative efficacy, and for the former an outlier in terms of seeing something resembling actual price competition in the US.

Advances in drug development are ostensibly a means to improve health outcomes, but it's no secret that health outcomes in the US are atrocious compared to other wealthy countries, despite the US spending much more per capita on healthcare. The hypothetical and debatable risk of slowing pharmaceutical innovation is much more palatable than allowing the status quo to persist, even if drug spending is only one, small part of the ridiculous state of healthcare in the US today.

The idea of forestalling drug pricing reform so that we can wait for bigger, more comprehensive healthcare reform is a nice idealistic thought, but completely insane given the reality of US politics in 2021. Maybe drug spending is only 15%, but it's a 15% where there's a glimmer of actual bipartisan will to do something, and that absolutely can't be ignored.

And for what it's worth, I work in pharma and believe all of the above despite the fact that my personal financial situation will undoubtedly suffer if any drug pricing reform passes

Sad to see so much unscientific disinformation on this board.

Those mortality numbers are estimates based on CDC modeling, based on the 0-17 year age group. https://www.cdc.gov/flu/about/burden/2019-2020.html

2018-2019: 136 reported pediatric deaths (480 estimated)
2019-2020: 199 reported (434 estimated)
2020-2021: 1 reported (??? estimated) - https://www.cdc.gov/flu/weekly/index.htm

In the same 0-17 age group, there were 198 reported pediatric (0-17) deaths from March to December 2020 (source). Of course this doesn't take into account any upward adjustment modeling like the numbers you quoted for flu deaths. It's extremely misleading to compare estimated flu deaths from 2018-2020 when there wasn't widespread social distancing & community masking, to reported COVID deaths in 2020.

And now it seems like severe outcomes among pediatrics are accelerating with the spread of Delta!

BIIB - Is there any reason to believe a CMS decision will be made sooner than 9 months? Any reasonable way to predict this or is it pure speculation?

RCUS - Agreed. Spending all that time & effort to say nothing meaningful through the PR and CC seems like a pretty big red flag.

BMY: my $0.02, I don’t think CM-816 will get approval without at least a trend to show an EFS benefit, given the lack of validation and FDA guidance supporting pCR as a surrogate for survival in lung (unlike in breast). Not sure how long they’d need to wait for that data, but if relatively immature EFS is sufficient then perhaps they’ll get the data relatively quickly?

Re: neo (816) vs. adjuvant (010). DFS is a survival based endpoint, but it’s still not OS, so I don’t know if it’ll be considered THAT much stronger than pCR, unless 010 hits a DFS result that’s like ADAURA. 816 also has the advantage of a short fixed course vs. longer duration with 010, which could be a competitive advantage for BMY but also limit the revenue opportunity. For comparison’s sake in metastatic NSCLC a patient on a Q4W regimen treated with Opdivo for a year would get 13 doses.

I’m curious to see what happens with MRK’s studies here, I suspect if they are positive any time soon then it’ll be their market to lose given their dominance in the metastatic setting.

PFE/BNTX PR clarification: the press release states that this was actually the first interim analysis, as they had decided to drop the previous first interim analysis at 32 events based on FDA feedback, and make the interim at 62 events the first interim analysis. However, it seems that by the time they had made this amendment to the protocol, they already had accrued 94 events.

At least that's how I read it, but I admittedly haven't been following the vaccine race with a ton of precision, so please let me know if I'm off base! Not sure if this had any material impact on the statistics/ability to meet the stopping threshold at the 1st interim, or if it's just trivia.

Pure speculation but perhaps blowback from GILD remdesivir pricing combined with expectation that Trump won’t be re-elected -> renewed fears over US gov’t drug price reforms?

BMY - minor correction. The CheckMate-9LA data were already released in the ASCO abstract since it's not a LBA, and BMS's press release for the abstract also included additional data, including the OS HRs by PD-L1 status. I don't think there was any data released as part of the FDA approval that hadn't already been disclosed in the abstract and/or PR.

I do think the 9LA data are good in a vacuum but underwhelming in the context of the various other IO options for 1L NSCLC. Hard pressed to see a clear advantage for the 9LA regimen over Keytruda-based regimens.

The comments in the original tweet by @Ananyo with the BMJ article do a good job poking holes in the optimism about higher than expected asymptomatic rates.

Boils down to the distinction between 1) asymptomatic at time of viral detection vs. 2) asymptomatic over the course of the infection & viral clearance. The China data would be indicative of 1), but this in itself doesn't suggest that "the cure [lockdowns, economic disruption] is worse than the disease", as the larger body of data suggest that 2) happens only in a minority of cases, i.e. most patients who are asymptomatic at the time of detection become symptomatic with more follow-up.

A couple references:
[url]https://www.cdc.gov/mmwr/volumes/69/wr/mm6913e1.htm?s_cid=mm6913e1_w
[/url][tag]Kirkland SNF CDC report.[/tag] Shows that 57% of +ve cases were asymptomatic at diagnosis (n=13/23) but of these 13 cases, only 3 remained asymptomatic 1 week later. So, 3/13 (23%) who were asymptomatic at diagnosis remained so, and only 3/23 (13%) of all positive cases were "true" asymptomatics (though presumably could have become symptomatic with >1 week of follow up). Given this is a SNF with a high-risk population, it may not be generalizable

Chinese study of 55 asymptomatic patients admitted to a hospital due to contact with an infected family member. 40% of these cases were >50 years old, 31% were 30-49 years old, and 27% were <18 years old. Only 2% were 18-29 years old, which is perhaps surprising and maybe also undercuts the theory of widespread asymptomatic infections... The paper is a bit hard for me to follow, but I think it looks like n=16/55 (29%) remained asymptomatic during hospitalization, which is actually pretty consistent with the 23% result from the Kirkland SNF study.

I believe the Diamond Princess experience had similar results but don't have that handy right now. There are probably other studies as well, but from what I've seen it doesn't seem like we should expect widespread asymptomatic, undetected infection to 'save us' or build herd immunity, as these cases will most likely become symptomatic given a week or two.

BMY. Presumably PDL1 <1% subgroup accrued RFS events much faster relative to >1% given that Opdivo is so remarkably effective for PDL1+ melanoma. Not sure the proportion of patients that fall into each group, but if >1% is sizable then it could slow down ITT.

BMY: not that I expect it to be positive, but technically the study hasn't failed... yet.

The PFS primary endpoint was negative, but there must be at least a trend to OS benefit for the IDMC to allow the study to continue. If the OS primary endpoint turns out to be positive, the lack of PFS will be inconsequential. While it's not all that unusual to see a OS benefit but no PFS with IO drugs, my guess is that the OS will wind up being negative here too.

NWBO: this guy was an MSL at Merck. That’s tantamount to hiring some random sales rep and appointing them as the Head of Commercial Operations. Not exactly a sign of being able to attract top talent for senior management.

ADXS: I tend to agree with jq1234t's tweet that this PR headline is a step beyond inevitable overly optimistic PR headline from small biotech. There is zero evidence that ADXS-PSA has any impact on overall survival, with or without Keytruda.

OS wasn't even cited as an exploratory endpoint on the poster, let alone a primary or secondary endpoint. PFS was listed as a secondary endpoint along with antitumor activity which presumably includes traditional RECIST response (notwithstanding lack of measurable disease barrier as mentioned by Dew) as well as PSA reduction. Guessing that they're not showing PFS because it doesn't look good compared to historical controls.

IMO any kind of OS data from a study like this are effectively worthless due to lack of control and small sample size. How many failed cancer vaccines have we seen over the years that have trumpeted their incredible mOS benefits from single-arm Ph1/2 studies that turn out to be expensive placebos?

What is anyone seeing here that makes it look like it won't be just another really expensive placebo? I'm not a deep expert in the prostate landscape, but the PSA waterfall plot doesn't look meaningfully different to me for the ADXS combo vs. pembro mono. And comparing the same data sets, the RECIST responses look better for pembro mono compared to ADXS-PSA + pembro.

It's a huge red flag to me for a biotech company to highlight the mOS so heavily from a trial like this, major blow to credibility and makes me immediately suspicious that they're trying to polish a turd (i.e. trying to pass the 'sight' test because they know they can't pass the 'sniff' test!)

MRK: this is the first time I've ever seen a study fail despite the confidence interval not crossing 1 for either primary endpoint. I know that with alpha splitting for multiple endpoints you can see p-values that are less than 0.05 but still not technically successful, but I always had thought that if the CI doesn't cross 1 then it's a positive study...

HALO/OT: power and alpha.

I’m no stats expert, but I have a basic understanding and love learning more. Here’s my guess as to why the increase in alpha from 0.04 to 0.05 didn’t result in a larger increase in power.

Take this graphical schematic showing the relation between alpha and beta/power as a starting point. https://www.animatedsoftware.com/statglos/sgrelabp.htm

This shows a hypothetical study designed with a power of 0.26. Increasing the power of the study would effectively be sliding the bottom distribution even further to the right, such that the dark green area under the curve to the right of the critical value was e.g. 0.92 for 92% power. Then, imagine changing the alpha on the top graph from 0.04 to 0.05 - the critical value would move very slightly to the left, which would increase the area under the curve that represents power. But the power would only increase a tiny amount, as it’s all coming in the tail of the curve which represents very little area

Good info on SCLC staging can be found here: https://www.cancer.org/cancer/small-cell-lung-cancer/detection-diagnosis-staging/staging.html

The long story short, extensive stage includes metastatic disease as well as non-metastatic disease that is too widespread in the lung(s) to be treated with a single field of radiation.

My understanding is that the staging is only used for treatment-naive patients; once a patient has relapsed, they're generally all bucketed together or sometimes split into platinum-sensitive vs. -refractory based on the amount of time it took them to relapse.

The Opdivo accelerated approval was for third-line SCLC based on a Phase 1/2 study. Since this approval happened, Opdivo has failed a Phase 3 study in second-line SCLC (Checkmate 331) as well as today's failure in a first-line maintenance Phase 3 study. As far as I know, there are no ongoing studies that BMS could use to convert the accelerated approval to a full approval. Seems like a case where the accelerated approval should be rescinded, but I have a feeling that won't happen.

Technically the RTOR approval expanded the Keytruda label from just carboplatin combination to carboplatin or cisplatin, along with the conversation from accelerated to full :-P

Correction/easy trivia re: FDA Real-Time Oncology Review

Adcetris was actually the third expedited label expansion under RTOR. The second label expansion happened earlier this year and received little fanfare as the approval didn't end up coming until relatively shortly before the PDUFA date. What was the 2nd RTOR approval?

NWBO: I don’t follow except for loose interest in seeing how biotech scam companies continually sink to new lows. Recent press release is a great example, I did a triple take when I saw first data cut being from the “top 100” patients in the study and then realized that and all subsequent data were still unblinded aggregate data from experimental and control arms.

Would be worth a laugh if it wasn’t effectively exploiting an extremely sick and desperate population.

https://www.nwbio.com/updated-interim-data-phase-3-trial-dcvax-l-glioblastoma/

AZN - agreed that MYSTIC is a negative trial more so due to the statistical design rather than the activity or lack thereof of Imfinzi monotherapy. That said, I highly doubt there will be any meaningful off-label use of Imfinzi in 1L NSCLC across any PD-L1 expression level. The mental leaps a clinician would have to make to justify such a decision are too great in the face of Keytruda's blanket approvals across the market, let alone any logistical leaps with regard to access and reimbursement.

As Dew mentioned, in trials that have multiple primary endpoints, the alpha (i.e. p-value) typically gets split amongst them. If 0.05 was used for each individual endpoint, the overall Type I error/false positive rate for the study would be >0.05.

Here’s link to interim 067 manuscript: https://www.nejm.org/doi/full/10.1056/NEJMoa1504030

sorry for double post, on mobile and not sure how to edit.

NKTR: the interim report of CheckMate 067 had 8.9% CR rate for Opdivo monotherapy. Opdivo + Yervoy CR rate was 11.5%. Opdivo monotherapy ORR in PD-L1 “negative” was 41%, granted using a 5% cutoff.

SITC data look good, but spectre of epac still looms large. Are the data so good that it looks clearly better than what you could expect from Opdivo monotherapy in a small study at (I’m assuming) top US academic centers? I don’t think so.

NKTR: good question, and I’m not sure why either. Obvious explanation would be if 214 was given on different dosing schedule than Opdivo, but looks like they’re both every 3 weeks.

BMY: I would bet that FDA doesn’t want to approve for TMB until they see OS results. So they give standard review to buy themselves a little more time for OS to mature and see if TMB is real deal or not. The kicker is that OS in TMB is not actually a powered endpoint per what BMY has disclosed, but I have a hard time believing FDA will approved Opdivo + Yervoy in TMB-H without at least a decent trend for OS even if it’s a secondary endpoint.

I'm not a biostats expert (though I do quite enjoy trying to educate myself on the subject!) but I'll share my thoughts:

- I've followed a number of the major IO studies over time since they were first presented through to the 2-year, 3-year, etc. updates, and while the OS HRs sometimes do improve, I've only ever noticed it on the magnitude of a couple of points -- not something that meaningfully changes the conclusion of the initial report of study results. If you have other more extreme examples, please share as I do find this fascinating

- Out of curiosity, I checked the protocols for a couple Keytruda studies pubbed in NEJM (2L bladder and 1L PD-L1 >=50% NSCLC), and they state "In case the proportional hazards assumption doesn’t hold, Fleming and Harrington’s weighted log-rank test or other methods, as appropriate, will be conducted". I don't exactly know what this means in technical terms, but I interpret it as the study sponsors knew that there could be non-proportional hazards and had a plan to test this

- In your hypothetical example, I would suggest that the issue isn't the non-proportional hazards, the issue is poor study planning in terms of not capturing enough follow-up before performing the analysis. As you say, you might only need to wait an extra 6 months to see a big difference, and it is incumbent on the study sponsor to account for this in their analysis plan

- No comment on FGEN, I haven't looked at this drug closely

NKTR: you need a controlled trial to have proof that this combo can turn pts from PD-L1 neg to pos. It is well known that PD-L1 is a dynamic marker that can change over time, so how does NKTR data prove that this change is due to combo, especially in highly immunogenic tumors like bladder, melanoma and renal? (Rhetorical question.)

I know I’m broken record but I still find NKTR data no more convincing than INCY epacadostat, and we all know how that turned out. No single agent activity (as far as I can find, if someone has evidence please share), no toxicity, and then PD-1 combos with a handful of pts in tumors that are highly responsive to single agent PD-1. That combined with what feels like undue arrogance from mgmt gives me a bad feeling.

MRK: the most important data which was presented but not included in the press release, was the OS in the 1-49% group. It was HR=0.92. I.e. the >=50% group drives all the benefit. With the strength of KEYNOTE-189 and -407, looks like chemoIO combos will be the standard for anyone <50%.

That is an extraordinary claim that requires extraordinary evidence to back up. What data have you seen that suggests LAG-3 could be anything close to PD-1 which as a class has generated billions upon billions in revenue already?

NKTR: I don’t closely follow NKTR but my understanding is that ORR in abstract is now lower across tumors vs previous reports with smaller sample sizes.

Yes, those are fair points, though that RCC abstract looks like a very small amount of single agent activity. I still would look at the Opdivo combo data with a healthy dose of skepticism.

MRK: this is the first time, to my knowledge, that they’ve disclosed a sBLA *filing* as opposed to the FDA *filing acceptance* which comes about 2 months later. Can’t really draw a conclusion about the FDA playing favorites with Keytruda insofar as they haven’t even accepted the filing yet; it does indicate how bold MRK is willing to be with its regulatory approach, which is impressive IMO. My guess is that they’ll be able to shoehorn in the PFS and/or OS when they look at it in the next few months.

BMY/MRK: for what it’s worth I still tend to think the drugs are practically identical, but MRK had a lot of luck and some smarts in their favor with regard to frontline NSCLC, which could be big enough to overshadow just about everything else.

I still don’t understand the reason to put much faith in NKTR after the INCY blowup, but at same time I think maybe it’ll be BMY’s turn for a little luck and -214 pans out!

BMY:

“One minor correction to your post, OS secondary endpoints for nivolumab + ipilimumab vs chemo is in TMB>=10 mutations/Mb population, only nivolumab vs chemo is in PD-L1+ and TMB>=13 mutations/Mb population.”

Yes that’s a secondary endpoint, but the Part 1 has two co-primary endpoints for O+Y: PFS in TMB-H (which was at AACR), and OS in PD-L1+ (which is still immature). Check NEJM if you want more details. This means that they could be in situation where they show primary endpoint PFS in one biomarker group (TMB-H) but OS in another biomarker group (PD-L1+). Extremely bizarre. Best case for BMY is all these OS primary and secondary endpoints are positive, but not sure how realistic that is...

BMY: I think the logic behind that tweet is essentially the best possible case for Opdivo + Yervoy, but the 227 data are a mess* and I have serious doubts they’ll be accepted by regulators, let alone clinicians.

*issues include extremely late protocol amendment to add TMB resulting in very low % of pts included in primary endpoint analysis, lack of OS in TMB-H, lack of benefit for Opdivo monotherapy in TMB-H, underwhelming PFS for O+Y in TMB-H, OS primary endpoint being PD-L1+ instead of TMB-H, commercial barriers to TMB testing, availability of an easier chemo-free option with Keytruda monotherapy for PD-L1 50% expressors and soon to be 1% expressors due to KEYNOTE-042.

Even if FDA approves these 227 data, I think BMY is going to have a pretty hard time convincing doctors that it’s the right regimen to use in the face of KEYNOTE-189, -024 and -042. Curious to see what the Opdivo + chemo data from 227 look like...

MRK: pretty remarkable that the most important trial in the history of Keytruda winds up being the most impressive, too. What incredible results. BMY’s prospects in frontline NSCLC are all but disappeared at this point IMO.