PTAB

IPR2025-00630

MIM Software Inc v. Progenics Pharmaceuticals Inc

Log In
Key Events
Petition
petition

1. Case Identification

2. Patent Overview

  • Title: NETWORK FOR MEDICAL IMAGE ANALYSIS, DESIGN SUPPORT SYSTEM, AND RELATED GRAPHICAL USER INTERFACE (GUI) APPLICATIONS
  • Brief Description: The ’346 patent describes a network-based system for automatically analyzing medical images using a machine learning algorithm. The system generates a "disease risk map" to support the evaluation of cancer risk by creating a 3D image of risk regions overlaid on medical images, which includes the geographic identification of specific tissue regions.

3. Grounds for Unpatentability

Ground 1: Claims 1-13 are Anticipated by Renisch

  • Prior Art Relied Upon: Renisch (Application # 2012/0123253). An alternative obviousness ground was asserted for claims 1-12 over Renisch in view of Wu (Application # 2016/0335395).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner argued that Renisch discloses every limitation of the challenged claims. Renisch teaches a diagnostic system for automatically detecting cancerous lesions ("hotspots") from 3D anatomical (e.g., CT) and functional (e.g., PET) images. The system uses a segmentation unit, which can employ machine learning algorithms like a "neural network," to identify specific anatomical structures (e.g., organs). Renisch then generates and displays on a GUI a 3D image where graphical denotations of risk regions (hotspots) and specific tissue regions (segmented organs) are overlaid on a medical image. This, Petitioner asserted, constitutes the claimed "risk map." The combination with Wu was presented as a fallback to explicitly teach the "network-based" system limitation if Renisch were found deficient on that point.
    • Motivation to Combine (for §103 grounds): A person of ordinary skill in the art (POSITA) would combine Renisch with Wu to provide the medical image processing services via a network server rather than a standalone workstation, enabling remote and simultaneous access for multiple users.
    • Expectation of Success (for §103 grounds): A POSITA would have an expectation of success implementing Renisch's system on a network server using Wu's methods, as it combines known elements for a predictable result.

Ground 2: Claims 1-10 and 13 are Anticipated by Sprencz

  • Prior Art Relied Upon: Sprencz (Application # 2015/0063667).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner contended that Sprencz anticipates the claims by disclosing a network-based system that automatically segments, displays, and quantifies cancerous lesions from 3D multi-modality imaging data. Sprencz uses machine learning algorithms, such as "clustering," to automatically segment a patient's 3D bone volume from an anatomical image. It then identifies lesions and displays images of the "segmented skeletal volume overlaid with contours of lesion candidates." This visual representation of a specific tissue region (the skeleton) overlaid with graphical denotations of risk regions (lesion contours) on a medical image meets the limitations of the claimed "risk map" generated by a machine learning algorithm.

Ground 3: Claims 1-10 and 13 are Anticipated by Suehling

  • Prior Art Relied Upon: Suehling (Application # 2011/0007954).

  • Core Argument for this Ground:

    • Prior Art Mapping: Petitioner argued that Suehling discloses a network-based system for automatic lesion detection in 3D medical images that anticipates the claims. Suehling uses trained machine learning algorithms ("Marginal Space Learning") to segment an image into search regions corresponding to particular organs and bones. It then detects lesions within these specific regions. Suehling's system displays the results as a graphical overlay of segmented lesions on the segmented anatomical structures, which constitutes the creation of a 3D image of risk regions overlaid with the geographic identification of specific tissue regions, as required by the claims.

  • Additional Grounds: Petitioner asserted additional obviousness challenges for specific dependent claims. These grounds combined the primary references (Renisch, Sprencz, Suehling) with secondary art to teach limitations such as prostate cancer specificity (in view of D’Souza or Applicant's Admitted Prior Art), generation of a radiologist report that includes reference images (in view of Maier, Application # 2016/0203263), and updating the machine learning algorithm with accumulated data (in view of Menon, Application # 2016/0180042).

4. Key Claim Construction Positions

  • "Risk Map" (claims 1, 10, 11, and 13): Petitioner argued this term lacks a well-established meaning and should be construed based on the patent's intrinsic evidence and the Patent Owner's statements in parallel litigation. Petitioner asserted that the Patent Owner, in its infringement allegations, equated "risk map" with "tumor segmentation." Based on this, Petitioner contended that a "risk map" is simply "a visual representation of tissue overlaid with segmented lesions," a feature Petitioner argued is plainly disclosed in the prior art.

5. Arguments Regarding Discretionary Denial

  • Petitioner argued that discretionary denial under Fintiv would be inappropriate because the parallel district court litigation is in its infancy.
    • Factor 1 (Stay): While no stay is in place, the court has previously stayed discovery, indicating a stay is likely if inter partes review (IPR) is instituted.
    • Factor 2 (Trial Date): The court has not adopted a case schedule or set a trial date.
    • Factor 3 (Investment): Investment in the parallel case is minimal, as the complaint was recently answered and no discovery has occurred.
    • Factor 4 (Overlap): The extent of overlap is unknown because infringement and invalidity contentions have not been served.
    • Factor 5 (Parties): The parties are the same, but this factor weighs against denial given the nascent stage of the litigation.

6. Relief Requested

  • Petitioner requests institution of an IPR and cancellation of claims 1-13 of the ’346 patent as unpatentable.