Doing things out of proper sequence can require a do-over.  But in law, as elsewhere, the do-over may come too late. 

Just ask Abraxis, a drug maker that today lost a lawsuit because the company it bought patents from didn't own them at the time.

Abraxis signed an Asset Purchase Agreement with AstraZeneca in April 2006.  The APA provided that AstraZeneca "shall or shall cause one or more of its Affiliates to, Transfer to the Purchaser [Abraxis], and the Purchaser shall purchase and accept from Seller [AstraZeneca] or its Affiliates, as applicable, all right, title and interests of the Seller and its Affiliates in" patents relating to the active ingredient in a pain medication.  A month later, AstraZeneca and Abraxis entered into an Intellectual Property Assignment Agreement by which AstraZeneca purported to assign the patents to Abraxis.

But AstraZeneca didn't own the patents, which instead belonged to a couple of its subsidiaries.

In March 2007, Abraxis sued Navinta for infringement of the patents.  On the same day, the subsidiaries assigned the patents to AstraZeneca.  But AstraZeneca didn't get around to inking an assignment of the patents to Abraxis until almost eight months later.

The district court refused to dismiss the case on the ground that Abraxis, at the time it filed the case, lacked legal standing.  A seven-day bench trial produced a judgment in Abraxis's favor. 

But today the Federal Circuit reversed.  The 2-1 majority held that Abraxis couldn't cure the jurisdictional defect retroactively.  More specifically, the subsidiaries' assignments to AstraZeneca on the day Abraxis filed suit didn't confer standing because AstraZeneca didn't make an effective assignment to Abraxis until months later.  Abraxis Bioscience Inc. v. Navinta LLC, No. 09-1539 (Fed. Cir. Nov. 9, 2010).

The Abraxis majority relied on the court's decision last year in Bd. of Trustees of Leland Stanford Junior Univ. v. Roche Molecular Sys., Inc., 583 F.3d 832 (Fed. Cir. 2009) (post here), which also turned on a promise to assign that the assignor didn't fulfill until too late. 

The Supreme Court granted review of Roche last month on a similar but different question — whether the federal Bayh-Dole Act entitles a university that conducted research with taxpayer money to trump one of its researcher's assignment of his patent rights. 

The Federal Circuit held that the inventor's assignment to Roche before Stanford asserted its rights under Bayh-Dole had the effect of ending Stanford's entitlement to ownership of the patent vis-a-vis the inventor's assignee.  The court likely would've reached the opposite conclusion if the inventor's assignment happened after Stanford stirred itself.  But, again, sequence seems all.
 
If the Court reverses, the ruling would probably allow university contractors to trump assignments by inventors no matter when they occurred.  That would mean, among other things:

• potential assignees like Roche would need to get assignments not only from the inventors but also from the universities.
• universities would need to become more assertive about claiming ownership rights under Bayh-Dole.
• universities and inventors should pay more attention to documenting their respective rights to inventions that result from collaborative research.

A reversal could have a big impact on industries that often use or wish to use inventions resulting from university research.  The difficulty and expense of obtaining patent rights would increase.  And that could cost many millions to medical device/diagnostics and pharmaceutical industries.