As MedWhat scores an initial win in its courtroom battle with Stanford, Paige.ai finds itself in the middle of allegations that its licensing deal with Sloan Kettering was improper.
It can be easy to forget that, while university venturing is on an upward trajectory with more money being invested in spinouts and more funds being launched, we are not living in a perfect world. But the past couple of weeks have brought a range of headlines that underline just how difficult a sector this remains and how easily things can go awry.
First came news that household appliances manufacturer Dyson – which had acquired University of Michigan’s solid-state battery spinout Sakti3 for $90m in 2015 – decided to write off the investment.
The move was not entirely out of the blue. Dyson abandoned Sakti3’s patents in April 2017, giving up on claims made at the time of the acquisition that the spinout would prove transformational. Dyson’s founder James Dyson went as far as to claim that Sakti3’s technology could one day be worth more than the multibillion-dollar appliances maker.
He also acknowledged that it might require total investments of up to a $1bn and it looks like that requirement eventually became too much for the corporate, which has multiple competing battery technologies making its way through the R&D pipeline and may have realised that there is a superior – read cheaper – technology.
Solid-state batteries have long been seen as something of a holy grail, as they are less flammable, more compact and enable faster charging cycles. They are also considered a fundamental step to making electric cars more viable. But so far companies have struggled to bring the technology to a standard where it can be widely deployed, and it looks like Sakti3’s resources, despite the underlying research at Michigan backing it up, were not enough.
Second came the announcement that MedWhat, a US-based virtual medical assistant developer, had scored an initial win in its legal battle with Stanford University and its Stanford-StartX Fund.
The conflict began in April, when the Stanford-StartX Fund, Caixa Capital, Regent Capital Ventures and Startcaps Ventures sued MedWhat for fraud.
The court mandated mediation in June, but this failed shortly thereafter and led to MedWhat filing a countersuit within weeks, alleging a wide range of issues from a conflict of interest – StartX and Magic Stone Investments hold shares in MedWhat and its direct competitor Sensely – to fraud and blackmail.
The Stanford-StartX Fund tried to have MedWhat’s bank accounts frozen through a court filing in July, but on Thursday, in a first indication that MedWhat’s arguments may have some swing, the judge at the Supreme Court of California in San Francisco dismissed the claim, though admittedly without prejudice – meaning StartX could bring the case again if it finds additional arguments. The judge did not specify whether the alleged conflict of interest played any role in the decision to dismiss.
MedWhat remains one of the most intriguing stories being tracked by Global University Venturing – it is impossible to predict which way the case will go and it could prove a powerful moment either for Stanford-StartX or spinouts.
Finally, the New York Times and ProPublica published an in-depth look into an allegedly improper licensing deal between Memorial Sloan Kettering Cancer Centre (MSK) and its artificial intelligence spinout Paige.ai, which emerged with $25m in series A funding in February 2018.
Paige.ai aims to tackle the inefficiencies of clinical diagnosis in oncology, which currently relies on pathologists manually going through and interpreting as many as 60 slides of a biopsy, though in many cases only a handful turn out to be relevant.
The ability to digitise the slides has existed for decades but is hardly used as it does not improve workflow – simply moving the slides from under a microscope to a screen.
Paige.ai’s technology combines AI with clinician-generated annotations and incorporates anonymised clinical, treatment, genomic and survival data. The approach means pathologists can interpret data without the need to compile it first manually, speeding up work while reducing cost.
The technology is based on research by Thomas Fuchs, director of computational pathology in the Warren Alpert Centre for Digital and Computational Pathology at MSK and a professor at Cornell University’s Weill Cornell Graduate School of Medical Sciences.
Jim Breyer, co-founder and chief executive of VC firm Breyer Capital, co-led the series A round with Julian Robertson, founder of investment firm Tiger Management. But more importantly, MSK obtained a 9% equity stake in the spinout in return for providing Paige.ai with exclusive access to its library of 25 million pathology slides.
The exclusivity of that deal has come back to haunt MSK, as the research institute’s scientists and physicians have voiced their reservations about the terms at a staff meeting last month – a single company stands to benefit significantly from more than 60 years of research – and patients have come forward with concerns that their health data is being commercialised by a for-profit venture without their consent.
What has not helped matters is that the New York Times and ProPublica uncovered failures by Paige.ai’s chief medical officer, José Baselga, to disclose some financial ties to pharmaceutical firms in dozens of research articles.
Baselga resigned in mid-September and Craig Thompson, chief executive of MSK, announced a taskforce to review the institute’s conflict-of-interest policies.
There are also concerns that MSK, a charitable organisation, may have broken state and federal law by undervaluing the commercial worth of its assets – legally such institutions must prove they did not provide access for less than a fair market value.
Apart from MSK itself, a member of its executive board, the chairman of the pathology department and the head of a research lab also each hold shares in Paige.ai, leading Marcus Owens, the former head of US tax office the Internal Revenue Service’s division in charge of tax-exempt outfits, to tell the Times: “It just seems awfully coincidental that the individuals involved happen to be people in control and influence of that asset, and they ended up with an exclusive use of it.
“It seems to create a cascading series of conflicts for the operation of Sloan Kettering.”
MSK has defended the deal, saying some investors helped value the assets, with guidance from hedge fund managers on its board and a law firm. Patients that did not consent to the use of their data would have all personal information removed from slides and notes.
Paige.ai will also not receive exclusive access to any slides that are the result of federally funded research.
MSK clearly did not anticipate the internal and public backlash to the deal, with the centre’s chief operating officer, Kathryn Martin, noting: “I think we could have done a better job communicating it.”
Of course, Paige.ai is an early-stage company and it is unclear when – if ever – it will find commercial success. Much of the discussion could therefore be considered moot, but it should not. As artificial intelligence makes its way through every sector, discussions around the ethics of giving a private company all this research – particularly without informing patients and pathologists first – are needed sooner rather than later.
Charities, research institutes and universities might be expected to come out on the side of their researchers and patients but it will become increasingly difficult to figure out where that line is – after all, improving healthcare is for the greater good of all of humanity. Let us hope it will not take another court battle or a public relations nightmare to make it happen.