Authored by: Jeremy J. Gustrowsky
In a closely watched case, the Federal Circuit has affirmed the Patent Trial and Appeal Board’s (PTAB) decision that all claims of three AliveCor patents related to wearable heart monitoring technology are unpatentable as obvious over prior art. The patents at issue—U.S. Patent Nos. 9,572,499, 10,595,731, and 10,638,941—cover systems and methods for detecting cardiac arrhythmias using smartwatches and related devices. Apple challenged these patents through inter partes review (IPR), arguing that earlier publications already taught the claimed inventions, particularly the use of machine learning and confirmation of arrhythmias with ECG data.
The Federal Circuit found that substantial evidence supported the PTAB’s conclusions. The court agreed that prior art references, including Hu 1997 and Li 2012, disclosed the use of machine learning for analyzing heart data, and that it would have been obvious for someone skilled in the field to apply these techniques to arrhythmia detection. Importantly, the court noted that the AliveCor patent claims did not require any specific type of machine learning algorithm, making Apple’s burden easier to meet. The court also found that the prior art reference Shmueli adequately taught the step of confirming arrhythmias using ECG data after a potential issue is detected by a photoplethysmography (PPG) sensor.
A notable aspect of the case was AliveCor’s attempt to argue that Apple failed to meet its discovery obligations by not producing certain evidence from a parallel International Trade Commission (ITC) proceeding, where AliveCor had successfully defended its patents. However, the Federal Circuit ruled that AliveCor forfeited this argument by not raising it with the PTAB during the IPR process. The court emphasized that parties must preserve such issues for appeal by first presenting them to the agency.
This decision highlights the importance of presenting all relevant arguments and evidence during the PTAB proceedings, especially when parallel litigation is ongoing. It also reinforces that broad, functional patent claims—such as those simply requiring “machine learning”—may be more vulnerable to obviousness challenges if the general concept is already known in the field.