IP Law Daily, PATENT—D. Mass.: Medical image analysis patents using machine learning are not directed to unpatentable abstract idea, (Jan 15, 2025)

By Kevin M. Finson

Claims related to the generation of composite images with machine learning for the diagnosis and monitoring of cancer were plausibly alleged to contain an inventive concept because they contained more information than was used by doctors in the prior art.

A seller of medical analysis software plausibly alleged that its patents directed to improving medical imaging analysis recited an inventive concept, even assuming that the claims were directed to an abstract idea, the U.S. District Court in Boston has held. The claims addressed not just having computers do what doctors themselves did, but also claimed the creation of overlaid maps created with machine learning ((Progenics Pharmaceuticals, Inc. v. MIM Software Inc., No. 1:24-cv-10437-PBS (D. Mass. Jan. 14, 2025)).

Progenics Pharmaceuticals, Inc. and Exini Diagnostics AB (collectively, Progenics) were the owners of U.S. Patents Nos. 10,665,346 “the ’346 patent); 10,973,486 (the ’486 patent); U.S. 11,424,035 (the ’035 patent); 11,894,141 (the ’141 patent); 11,657,508 (the ’508 patent); 11,721,428 (the ’428 patent); and No. 11,941,817 (the ’817 patent), all of which claimed computer systems designed to improve the analysis of medical imaging for diagnosing and monitoring cancer and other diseases. Progenics brought suit for infringement against MIM Software, Inc. (MIM), alleging that MIM’s Contour ProtégéAI, MIM SurePlan MRT, LesionID, and LesionID Pro products infringed each patent. MIM moved to dismiss the claims based on the ’346, ’035, ’141, and ’817 patents on the ground that they were invalid for being directed at an abstract idea, and moved to dismiss the claims based on the remaining three patents for failure to state a claim for infringement.

Abstract idea. The parties agreed the ’346 patent was representative for invalidity purposes. For purposes of the present ruling, the court assumed at Alice step one that the claims were directed to the abstract ideas of “accessing, analyzing, and displaying medical imaging data to identify regions of cancer risk,” as MIM contended.

At step two of the Alice test the court considered whether the ’346 patent contained an inventive concept that saved its patentability. Progenics argued that the ’346 patent’s description of composite images of a risk map showing risk of recurrence generated with machine learning overlaid on one or more medical images was inventive, while MIM argued that this was simply an automation of work traditionally done by doctors. According to MIM, the requirements of “machine learning” and a “risk map” do not render the claims non-abstract because the claims recite use of a generic machine learning algorithm and the concept of a “risk map” is not inventive.

The court found that Progenics had at least plausibly alleged that the composite images produced by its software were unlike anything in the prior art and that they contained additional, useful information. Accordingly, the court denied the motion to dismiss as to the four patents challenged on invalidity grounds.

Infringement. As to the ’486 patent, the court found that the complaint lacked information that one of the accused products practiced the asserted patent. As to the ’508 patent, the court found that the complaint contained conflicting allegations about whether the accused products practiced the claimed method. Accordingly, these claims were dismissed with leave to amend.

The court denied the motion to dismiss as to invalidity and granted it, with leave to amend, as to non-infringement of the ’486 and ’508 patents.

The Case is No. 1:24-cv-10437-PBS.

Judge: Saris, P.

Attorneys: Michael H. Bunis (Hall & Stewart LLP) for Progenics Pharmaceuticals, Inc. Alexis Cohen (Wilmer Cutler Pickering Hale and Dorr LLP) for MIM Software Inc.

Companies: Progenics Pharmaceuticals, Inc.; MIM Software Inc.

Cases: Patent TechnologyInternet MassachusettsNews