Apple Inc v. AliveCor Inc

1. Case Identification

• Case #: IPR2021-00970
• Patent #: 9,572,499
• Filed: June 9, 2021
• Petitioner(s): Apple Inc.
• Patent Owner(s): AliveCor Inc.
• Challenged Claims: 1-20

2. Patent Overview

• Title: Methods and Systems for Arrhythmia Tracking and Scoring
• Brief Description: The ’499 patent discloses methods and systems for detecting the presence of arrhythmia. The claimed method involves sensing a user's heart rate with a sensor (e.g., photoplethysmography or PPG), determining heart rate variability (HRV), sensing the user's activity level with a motion sensor, comparing the HRV to the activity level, and alerting the user to sense an electrocardiogram (ECG) in response to an identified irregularity.

3. Grounds for Unpatentability

Ground 1: Obviousness over Shmueli and Osorio - Claims 1-6, 10-16, and 20 are obvious over Shmueli in view of Osorio.

• Prior Art Relied Upon: Shmueli (WO 2012/140559) and Osorio (Application # 2014/0275840).
• Core Argument for this Ground:
  • Prior Art Mapping: Petitioner argued that Shmueli taught a wrist-worn mobile computing device that continuously monitors a user via a PPG sensor to detect an "irregular heart condition" (arrhythmia). Upon detection, Shmueli’s device prompts the user to perform a confirmatory ECG measurement. Petitioner contended that Osorio remedied a deficiency in Shmueli by teaching the use of a motion sensor to determine a user's activity level. Osorio explicitly disclosed that comparing a user’s HRV to their concurrent activity level improves arrhythmia detection accuracy by establishing non-pathological HRV ranges for different activities (e.g., sleeping, walking, running) and reducing false alarms. The combination of Shmueli's PPG-to-ECG alert system with Osorio's activity-based HRV analysis allegedly rendered the limitations of the independent claims obvious.
  • Motivation to Combine: A POSITA would combine Osorio with Shmueli to improve the accuracy of Shmueli’s arrhythmia detection. Osorio expressly taught the benefits of using activity level data to avoid false diagnoses, a known problem in the field. Incorporating Osorio’s method into Shmueli’s device was presented as a predictable solution to increase reliability and reduce the number of unnecessary ECG alerts, thereby enhancing user experience.
  • Expectation of Success: A POSITA would have had a reasonable expectation of success because both references are directed to body-worn medical devices for detecting pathological heart conditions. Integrating Osorio's known technique of activity-level monitoring into Shmueli's device was a straightforward application of known sensor technology and processing techniques to yield the predictable result of improved accuracy.

Ground 2: Obviousness over Shmueli, Osorio, and Hu-1997 - Claims 7-9 and 17-19 are obvious over Shmueli in view of Osorio and Hu-1997.

• Prior Art Relied Upon: Shmueli (WO 2012/140559), Osorio (Application # 2014/0275840), and Hu-1997 (a 1997 IEEE journal article).
• Core Argument for this Ground:
  • Prior Art Mapping: This ground built upon the Shmueli-Osorio combination from Ground 1 and added Hu-1997 to address the limitations of claims 7-9 and 17-19, which required determining arrhythmia using a "machine learning algorithm." Hu-1997 disclosed a patient-adaptable ECG beat classifier using a "mixture of experts" machine learning approach. Petitioner argued this taught using machine learning trained on both user-specific historical data and general population data to detect arrhythmia. This machine learning algorithm would be applied to the multifactor data (PPG, motion, ECG) collected by the combined Shmueli-Osorio device.
  • Motivation to Combine: A POSITA would be motivated to incorporate a machine learning algorithm, as taught by Hu-1997, into the Shmueli-Osorio device to further enhance detection accuracy. By the patent's critical date, machine learning was a well-known and superior method for analyzing complex, multimodal biomedical data. Shmueli itself contemplated improving its detection algorithms by searching for correlations between sensor signals. Hu-1997’s method offered a sophisticated, proven way to perform this analysis and improve performance.
  • Expectation of Success: A POSITA would have reasonably expected success in applying Hu-1997’s machine learning techniques. Machine learning was a known and popular tool for arrhythmia detection, and its application to analyze sensor data from the Shmueli-Osorio device was a simple application of a known technique to a known system to achieve predictable improvements in detection accuracy.

4. Arguments Regarding Discretionary Denial

• Petitioner argued that discretionary denial under §314(a) based on the Fintiv factors was inappropriate. The petition was filed very early in a parallel ITC investigation (within seven weeks of the complaint), before any significant milestones and well before the one-year statutory bar. Petitioner asserted that institution would support a stay in the parallel proceedings. Crucially, Petitioner stipulated that if the IPR was instituted, it would not pursue in the parallel ITC or district court proceedings any invalidity grounds based on Shmueli, Osorio, or Hu-1997. Petitioner argued this stipulation created "zero overlap" between the IPR and the parallel proceedings, weighing heavily against discretionary denial.

5. Relief Requested

• Petitioner requested institution of an inter partes review and cancellation of claims 1-20 of the ’499 patent as unpatentable.