DCT
1:24-cv-10437
Progenics Pharma Inc v. MIM Software Inc
I. Executive Summary and Procedural Information
- Parties & Counsel:
- Plaintiff: Progenics Pharmaceuticals, Inc. (Delaware) and EXINI Diagnostics AB (Sweden)
- Defendant: MIM Software Inc. (Ohio)
- Plaintiff’s Counsel: Choate Hall & Stewart LLP
- Case Identification: 1:24-cv-10437, D. Mass., 02/23/2024
- Venue Allegations: Plaintiff alleges venue is proper in the District of Massachusetts because Defendant conducts business, derives revenue, and maintains systematic contacts in the district, including by selling the accused software products to Massachusetts-based hospitals and cancer centers.
- Core Dispute: Plaintiff alleges that Defendant’s medical imaging software products infringe patents related to automated, AI-driven analysis of medical scans for cancer diagnosis and treatment evaluation.
- Technical Context: The technology at issue involves using computer systems and machine learning to automatically segment anatomical structures and analyze radiopharmaceutical uptake in medical images (e.g., CT, PET, SPECT), aiming to provide more reliable and accurate clinical decision support.
- Key Procedural History: The complaint alleges that the parties engaged in partnership and licensing discussions from 2020 to 2022, including entering into a Confidential Disclosure Agreement in 2021. In October 2022, Plaintiffs allegedly sent Defendant a draft Collaboration Agreement that expressly listed the patents-in-suit, suggesting Defendant had pre-suit knowledge of the patents.
Case Timeline
| Date | Event |
|---|---|
| 2016-10-27 | Earliest Priority Date for '346 Patent |
| 2018-01-08 | Earliest Priority Date for '486 Patent |
| 2020-05-26 | '346 Patent Issued |
| 2020-06-01 | Partnership/Collaboration Discussions Allegedly Began (approx.) |
| 2021-04-13 | '486 Patent Issued |
| 2021-10-01 | Confidential Disclosure Agreement Signed (approx.) |
| 2022-10-01 | Plaintiffs Sent Draft Collaboration Agreement to Defendant (approx.) |
| 2023-06-01 | Plaintiffs Discovered Allegedly Infringing Features (approx.) |
| 2024-01-08 | GE HealthCare Announced Agreement to Acquire Defendant |
| 2024-02-23 | Complaint Filed |
II. Technology and Patent(s)-in-Suit Analysis
U.S. Patent No. 10,665,346 - "Network for Medical Image Analysis, Decision Support System, and Related Graphical User Interface (GUI) Applications," issued May 26, 2020
The Invention Explained
- Problem Addressed: The patent’s background describes the difficulty and potential for human error when physicians must interpret and combine different types of medical scans (e.g., anatomical CT scans and functional PET scans) to diagnose and treat cancer, a process that can lead to inconsistent or inaccurate outcomes (Compl. ¶¶ 18-21; ’346 Patent, col. 2:37-61).
- The Patented Solution: The invention is a network-based system that automates this process. It uses a machine learning algorithm to analyze medical images and generate a "risk map," which is a visual representation of tissue overlaid with graphical markings indicating regions of cancer risk. This risk map is then displayed to a user via a graphical user interface (GUI) to serve as a decision-making support tool (’346 Patent, Abstract; ’346 Patent, col. 3:46-4:24).
- Technical Importance: The technology aims to provide a more objective, automated, and reliable method for medical image interpretation, moving beyond subjective physician assessment to improve the accuracy of cancer diagnosis and treatment monitoring (Compl. ¶ 22, ¶ 28).
Key Claims at a Glance
- The complaint asserts infringement of at least independent claim 1 (Compl. ¶ 29).
- The essential elements of claim 1 are:
- A network-based system with a processor and memory.
- Accessing one or more medical images of a patient from a database.
- Performing an analysis on the images using a machine learning algorithm to generate a risk map, which comprises a visual representation of tissue overlaid with graphical denotations of cancer risk.
- Displaying the risk map to a user via a GUI.
- The analysis further comprises creating a 3D image of the risk regions overlaid on the medical images, with the 3D image including geographic identification of specific tissue regions.
U.S. Patent No. 10,973,486 - "Systems and Methods for Rapid Neural Network-based Image Segmentation and Radiopharmaceutical Uptake Determination," issued April 13, 2021
The Invention Explained
- Problem Addressed: The manual process of identifying specific 3D organs or tissues within complex 3D medical images is difficult and inefficient, limiting the ability to perform accurate quantitative analysis (Compl. ¶ 20; ’486 Patent, col. 2:37-52).
- The Patented Solution: The invention proposes an automated two-step segmentation method. A first module rapidly identifies a broad, initial volume of interest (VOI), such as the pelvic region. A second, more fine-grained module then operates only within this smaller VOI to accurately identify a specific target, like the prostate. This approach improves both the speed and accuracy of segmentation, enabling the determination of quantitative metrics like radiopharmaceutical uptake (’486 Patent, Abstract; ’486 Patent, col. 4:28-54).
- Technical Importance: By automating segmentation with a two-step localization and identification process, the invention enables more efficient and reliable analysis of medical scans, which is critical for quantifying disease states and making treatment decisions (Compl. ¶ 27).
Key Claims at a Glance
- The complaint asserts infringement of at least exemplary claim 42 (Compl. ¶ 30).
- The essential elements of method claim 42 are:
- Receiving a 3D anatomical image and a 3D functional image of a subject’s pelvic region.
- Using a "first module" to determine an initial volume of interest (VOI) within the anatomical image corresponding to the pelvic region.
- Using a "second module" to identify a prostate volume within that initial VOI.
- Determining one or more radiopharmaceutical uptake metrics using the functional image and the identified prostate volume.
- Displaying the images on an interactive GUI as selectable and superimposable layers.
III. The Accused Instrumentality
Product Identification
- The complaint identifies Defendant's software products Contour ProtégéAI® and MIM SurePlan™ MRT as the "Infringing Products" (Compl. ¶ 5).
Functionality and Market Context
- The accused products are described as software imaging tools for radiation oncology, radiology, and nuclear medicine (Compl. ¶ 5). The complaint alleges they provide "AI-based segmentation and contouring," "integration of diagnostic images from multiple modalities into treatment plans," and "quantitation and advanced processing in diagnostic imaging" (Compl. ¶ 46).
- Contour ProtégéAI is alleged to use a neural network to segment structures on CT images, while MIM SurePlan MRT is alleged to use "PET Edge" to create tumor contours overlaid on SPECT/CT images (Compl. ¶ 54).
- The complaint alleges that these functionalities were a key driver of GE HealthCare's decision to acquire Defendant MIM Software in January 2024 (Compl. ¶¶ 45-47).
IV. Analysis of Infringement Allegations
’346 Patent Infringement Allegations
| Claim Element (from Independent Claim 1) | Alleged Infringing Functionality | Complaint Citation | Patent Citation |
|---|---|---|---|
| A network-based system for generating a disease risk map for use as a decision-making support... | The accused products are described as a "network-based system" that can "Deliver AI on Premise or in the Cloud" and function as a "decision-support tool." | ¶52 | col. 35:46-49 |
| access one or more medical images associated with a particular patient from a database; | The accused products allegedly "access scan images for particular patients from an electronic database," and its MIMcloud service can "push and pull images." | ¶53 | col. 36:1-3 |
| perform an analysis...using a machine learning algorithm, to generate the risk map, wherein the risk map comprises a visual representation of tissue overlaid with graphical denotations marking one or more regions of risk of cancer... | MIM SurePlan MRT allegedly uses a neural network (Contour ProtégéAI) to segment structures and PET Edge to create a "contour of tumor regions that is overlaid on a SPECT/CT image." A brochure image shows a tumor segmentation overlaid on a medical image. | ¶54, ¶55 | col. 36:4-10 |
| cause display of risk map via a graphical user interface (GUI)... | A brochure image shows the accused product displaying a tumor segmentation via a GUI. This image shows a medical scan with a red color outlining a tumor. | ¶55 | col. 36:11-13 |
| wherein the analysis...comprises creation of a 3D image of the one or more regions of risk...overlaid on the...medical images, and wherein the 3D image comprises geographic identification of one or more specific tissue region(s)... | The complaint alleges that brochure images show the accused products displaying "normal organ contours and tumor contours overlaid on medical images." | ¶55 | col. 36:14-22 |
- Identified Points of Contention:
- Scope Questions: A central question may be whether the accused products' "contour of tumor regions" meets the claim limitation of a "risk map." The defense could argue that merely outlining a tumor is anatomical segmentation, not a "risk map," which may be construed to require a more explicit assessment or quantification of cancer risk.
- Technical Questions: The complaint alleges that a neural network for "segmenting normal structures" (Compl. ¶ 54) satisfies the "machine learning algorithm" element for generating a risk map. It raises the question of what evidence demonstrates this algorithm is used to generate a map of cancer risk as opposed to simply identifying non-cancerous anatomy.
’486 Patent Infringement Allegations
| Claim Element (from Independent Claim 42) | Alleged Infringing Functionality | Complaint Citation | Patent Citation |
|---|---|---|---|
| (a) receiving...a 3D anatomical image...at least a portion of which corresponds to a pelvic region... | The accused products allegedly use CT cameras to provide 3D anatomical images and have specific applications for the pelvic region, including the prostate. A marketing image shows a "Pelvis CT" menu. | ¶58, ¶17 | col. 74:10-17 |
| (b) receiving...a 3D functional image...wherein at least a portion...represent physical volumes within the pelvic region... | The accused products allegedly use SPECT imaging, described as a 3D functional image, for applications within the pelvic region. | ¶59 | col. 74:18-28 |
| (c) determining...using a first module, an initial volume of interest (VOI) within the 3D anatomical image... | The complaint alleges that when Contour ProtégéAI "detects [the] treatment site to apply correct AI model," it is performing the function of determining an initial VOI. This is supported by a workflow diagram from the Defendant's website. | ¶60, ¶19 | col. 74:29-34 |
| (d) identifying...using a second module, a prostate volume within the initial VOI... | The complaint alleges that Contour ProtégéAI's use of a "Pelvis CT" model to specifically segment the prostate after the initial site detection constitutes the claimed "second module." | ¶61 | col. 74:35-38 |
| (e) determining...the one or more uptake metrics using the 3D functional image and the prostate volume... | A brochure for the accused products allegedly shows the calculation of "dose accumulation statistics" and uptake measurements for tumors and organs. An image from the brochure displays a table of these metrics. | ¶62, ¶15 | col. 74:39-42 |
| (f) causing...display of an interactive graphical user interface (GUI) for presentation...of the 3D anatomical image and/or the 3D functional image; | The complaint points to brochure images that display SPECT/CT image overlays within a GUI. | ¶63 | col. 74:43-47 |
| (g) causing, by the processor, graphical rendering of...the 3D anatomical image and/or the 3D functional image as selectable and superimposable layers... | The complaint alleges the same brochure images show that the accused products display SPECT/CT overlays, which are alleged to be selectable and superimposable. | ¶64 | col. 74:48-55 |
- Identified Points of Contention:
- Scope Questions: The infringement theory relies on mapping the accused product's workflow to the patent's "first module" and "second module." A key dispute may arise over whether "module" requires structurally or functionally distinct software components, or if it can read on different logical steps within a single, integrated algorithm.
- Technical Questions: The allegation that "detects [the] treatment site to apply correct AI model" (Compl. ¶ 60) is equivalent to determining an "initial volume of interest...excluding tissue outside the pelvic region" (Compl. ¶ 30) will likely be contested. The defense may argue that detecting a general treatment area is not the same as defining a specific anatomical VOI as required by the claim.
V. Key Claim Terms for Construction
The Term: "risk map" (’346 Patent, claim 1)
- Context and Importance: This term's definition is critical, as the infringement case for the ’346 patent hinges on whether the accused products' "contouring" of tumors constitutes a "risk map." Practitioners may focus on this term because Defendant is likely to argue its feature is merely segmentation, while Plaintiff will argue any graphical identification of cancerous tissue on an image meets the definition.
- Intrinsic Evidence for a Broader Interpretation: The claim defines the term functionally as "a visual representation of tissue overlaid with graphical denotations marking one or more regions of risk of cancer" (’346 Patent, col. 36:7-10). This language does not explicitly require a quantitative score, potentially allowing any visual flagging of a tumor to qualify.
- Intrinsic Evidence for a Narrower Interpretation: The patent's abstract and detailed description repeatedly link the "risk map" to the computation of a "risk index," such as a Bone Scan Index (BSI) (’346 Patent, Abstract; col. 4:2-3). This context suggests the "risk map" is a visualization of a calculated, quantitative risk, not just an anatomical outline.
The Term: "first module" / "second module" (’486 Patent, claim 42)
- Context and Importance: The viability of the infringement allegation for the ’486 patent depends on successfully mapping these two distinct "modules" onto the accused software. The dispute will likely center on whether the accused product, which the complaint describes using an AI workflow diagram (Compl. ¶ 19), actually performs the claimed two-step process.
- Intrinsic Evidence for a Broader Interpretation: The patent does not define "module" in structural terms, instead describing the modules by their function: a first for determining an initial VOI and a second for identifying the prostate within it (’486 Patent, col. 74:29-38). This could support an interpretation where the terms refer to distinct logical operations within a single algorithm rather than separate software components.
- Intrinsic evidence for a Narrower Interpretation: The specification justifies the two-step approach as being "more computationally efficient" because the fine-grained segmentation is performed on a "smaller, more computationally intensive" basis only after the initial localization (’486 Patent, col. 4:48-54). This suggests a sequential process with distinct inputs and outputs for each step, which could support a narrower construction requiring functional separation.
VI. Other Allegations
- Indirect Infringement: The complaint alleges that Defendant induces infringement by "actively and knowingly inducing its customers to use the Infringing Products" in a way that practices the claimed inventions (Compl. ¶ 69, ¶ 76). The factual basis for this allegation includes Defendant’s marketing brochures, website materials, and user documentation, which allegedly instruct physicians and other users on how to operate the accused "AI-based segmentation," "contouring," and "integration" features (Compl. ¶¶ 49-64).
- Willful Infringement: Willfulness is alleged based on Defendant’s purported pre-suit knowledge of the Asserted Patents. The complaint asserts that the parties engaged in collaboration discussions for years, and that Plaintiffs sent Defendant a draft Collaboration Agreement in October 2022 that "expressly listed the Asserted Patents" (Compl. ¶ 65).
VII. Analyst’s Conclusion: Key Questions for the Case
- A core issue will be one of definitional scope: can the term "risk map" in the ’346 patent, which is tied in the specification to quantitative indices like BSI, be construed broadly enough to read on the accused product's function of drawing a "contour" around a tumor?
- A second central issue will be one of functional mapping: does the accused product's AI-driven workflow, as depicted in marketing materials, actually perform the distinct, two-step process of the ’486 patent—first determining a broad "initial volume of interest" and then identifying the prostate within that volume—or is it a single, integrated process that does not align with the claimed method?
- A key evidentiary question will concern knowledge and intent: given the alleged history of collaboration discussions and the explicit identification of the patents-in-suit in a draft agreement, the court will have to evaluate the strength of the evidence supporting Plaintiffs' claim that any infringement by Defendant was willful.