Can someone explain to me - assuming they have enough data - why not train different models explicitly for each group / subgroup you want to model? You even could then just take the top N% of each group by score, effectively guaranteeing equal treatment for each group. Why would this not work?
Regeneron is among the most science and data focused biopharma and has a long, long history in genetics. They know how critical privacy is and will ensure the data are used to advance human health.
Potentially an even better home for the data than in the company, since now will not have pressure of quarterly reports.
Genetic genealogists are not so optimistic. 23andMe had already made their family discovery options poorer, and now they're bought up by a company with no interest in that.
This was the same sentiment about 23andme many years ago.
If regeneron cared about user privacy, then they would work to make sure consent to use of every piece of data is obtained.
It should be the default to delete any data that has not received a consent upon transfer and users must opt in for regeneron to use it for research or otherwise.
Telling the slave how lucky he is to have been bought by a good slave owner.
Buying and selling humans has never been a good thing. Trading their DNA in the marketplace without consent is just the new evil our generation is too blind to see.
Not mentioned in the article is the fact that lorlatinib, original studied in the phase 3 CROWN trial, was used in frontline management of advanced ALK-mutated non-small cell lung cancer (NSCLC) [0]. Five-year data indicate a significant progression-free survival benefit, with 60% of patients free of progression. Even after 60.2 months of follow-up, the median progression-free survival has not yet been reached [1]. It is the durability of progression-free survival that are most impressive and surprising here. That said, only 4% to 8% of patients with NSCLC will have ALK mutations/rearrangements, so while amazing, this finding is not applicable to all patients [2].
I believe the median progression-free survival is the length of time where 50% of people are still free from progression. The numbers you mentioned would mean that we don't know how long that is for this treatment but it's definitely longer than 5 years. That sounds pretty good depending on the type of cancer and stage when treatment started.
> it is fully approved for ALK+ NSCLC and has been since 2015.
Not exactly. FDA granted it an orphan drug designation in 2014, and then full approval for second- or third-line therapies in 2018 (on the basis of a Phase I/II study). The Phase III study (NCT03052608) is the gift that keeps on giving - it ran 2017-2020, and the progression-free survival in the lorlatinib arm is 60% vs 8% for the control.
