In the Information Age, the marriage of software and hardware enables the use of computer system platforms capable of solving very complex data problems. Computers wouldn’t exist without hardware, the physically tangible equipment capable of storing and processing data. However, that hardware would be pretty meaningless without software providing instructions on how the data is to be processed and for what purpose.
Medical software has been developed to benefit both patients and medical practitioners by providing better diagnostics, which ultimately lead to new and better treatments. This was explained perhaps best in an amicus brief filed by Medtronic, Inc. in Bilski v. Kappos. Medtronic makes the following extraordinarily compelling argument:
In the context of medical technology, the proper evaluation and effective treatment of patients depend upon complex correlations assessed over prescribed times. This, in turn, relies upon the generation of predictive models from a comparison of an individual patient’s signs and symptoms against a database of studied human wellness parameters, which contain patterns of diagnosis, chosen treatment, and outcome. These efforts are far from trivial.
Medtronic brings home this point — if you cannot understand, monitor and diagnose the problem you cannot treat the problem. “[T]he development of a diagnostic test almost always precedes the ability to treat the disease and is often a distinct research enterprise separated by years, if not decades,” the Medtronic brief argued.
Those who make the argument that medical software is abstract, or trivial, are just wrong.
Alice Runs Amok
There are those who have tried to make the case that software doesn’t deserve the same type of intellectual property rights as other innovations because the algorithms undertake processes that could be performed by a person. This kind of reasoning has led to results like the one seen in the U.S. Supreme Court’s decision in Alice Corporation v. CLS Bank International, which held that a software method of exchanging financial obligations was unpatentable under Section 101 as an abstract idea.
Despite the fact that the Supreme Court only reviewed financial software patents and methods in Bilski v. Kappos and Alice, Alice has had impacts to software innovations in many fields, including medical fields. For example, software expert Robert Sachs who writes for The Bilski Blog has identified dozens of patent applications relating to cancer treatments that have gone abandoned after the patent applicant received an Alice rejection. Such broad application of Alice, which dealt only with computerized process for electronically managing a checkbook register, seems wholly misplaced if not completely illogical. How can it be possible that Alice sheds any light on cancer treatments? Still, that is how the Patent Office is interpreting and applying Alice.
While anti-patent advocates may rejoice, those without an ideological dog in the fight should think twice about celebrating with those who would prefer that all software be patent ineligible. Without the protection provided by an issued patent there is no incentive for companies to pursue these diagnostic protocols, which means new treatments and cures will not be realized. In truth, the lack of patent protection stands in the way of a cure. Investors and Boards of Directors simply cannot justify spending the time, money and energy to create these monumentally difficult innovations without patent protection.
Still, many will choose to believe that software is abstract, as if a piece of hardware would innately be able to perform the computerized process on its own if only given enough time. But when software results in the saving of a human life — a life that may otherwise have been lost through conventional treatments — the result certainly doesn’t seem abstract. Neither should the concrete and identifiable innovation represented in the computerized process be considered abstract.
A Brief Survey of Recent Events
In the middle of August, Israeli imaging analytics firm Zebra Medical Vision announced that it had developed two machine learning algorithms, which can quantify both the amount of calcified plaque in coronary arteries as well as the presence of fatty liver from medical images. The system can help doctors determine patients who are at a higher risk of experiencing heart attack or stroke early enough to take therapeutic intervention.
Digital analysis of medical images can also aid in the fight against cancer. Research conducted at Stanford University has led to the recent development of a machine learning algorithm, which can provide better analysis of hematoxylin and eosin (H&E) stains from lung cancer patients than trained pathologists. The algorithm analyzes 10,000 image features from histopathological images to predict long-term and short-term prognoses with greater than 85 percent accuracy. According to quotes from the research team which were published by Medscape Medical News, experienced pathologists looking at H&E slides agree about 60 percent of the time on what the slide shows, so the algorithm brings a great deal of objectivity to this area of medicine.
Problems for Medical Software Patent Applicants
Increasingly the United States Patent Office (USPTO) is operating in ways that treat applicants with similar technologies very differently. For example, the Technology Center and Art Unit where your application is assigned makes all the difference in the world when it comes to software. If you are unlucky enough to have your software related application assigned to Technology Center 3600 you have about a 10% chance of success based on 2016 data. Elsewhere within the USPTO your odds are much higher. Thus, it is probably best to think of the USPTO as a parent corporation with a number of wholly owned subsidiaries (i.e., Technology Centers) that are run by very different individuals with very different views of the patent system.
Luckily for those with medical software, Technology Center 3600 deals with e-commerce patents (i.e., business methods and financial patents primarily). Unfortunately, for those with medical software, particularly those with diagnostic software, Alice is not the only problematic Supreme Court decision. Perhaps the worst Supreme Court patent decision over the last generation (perhaps longer) was in Mayo v. Prometheus, which purposefully and intentionally conflated patent eligibility with novelty and obviousness. Thus, for those with medical software you not only have to be aware of the tendency of many to want to believe software is all just an abstract idea, but you also have to contend with the language of Mayo, as elaborated upon by the Federal Circuit in Ariosa v. Sequenom. Thanks to Mayo and Ariosa v. Sequenom, the patent eligibility of diagnostics of various sorts is substantially in question.
Medical Software Patents
Not all medical software patents are denied. Indeed, a search for medical software patents using Innography’s patent portfolio analysis tools shows us that Irish medical device developer Medtronic (NYSE:MDT) owns 11.1 percent of the medical software patent market in the U.S. Following close behind is Fairfield, CT-based conglomerate General Electric (NYSE:GE), which holds 10.4 percent of the market, and then German engineering giant Siemens AG (ETR:SIE), which has 9.4 percent of the total medical software share in the U.S.
Recently, we took a look at advances in the realm of artificial pancreas systems and found that computer algorithms were being used to dramatically improve the delivery of insulin to patients suffering from diabetes. U.S. Patent No. 9107623, titled Signal Processing for Continuous Analyte Sensor, claims a glucose sensor system which utilizes a pattern recognition algorithm to recognize a time of day associated with a physiological pattern of the host’s glucose concentration over a past time period and to detect an approaching clinical risk associated with the recognized pattern. In response to detecting an approaching clinical risk, the system displays a recommendation to aid the host in proactively avoiding the clinical risk. This system provides an improvement in data processing of continuous glucose sensors to prevent inconsistencies in data output when using reference glucose values for monitoring. This patent was issued last August by San Diego, CA-based continuous glucose monitor developer DexCom Inc. (NASDAQ:DXCM)
Software techniques for improved monitoring of cancer treatments are protected within U.S. Patent No. 9289140, titled Systems and Methods for Imaging Changes in Tissue. This patent discloses a computer-based system for generating a parametric response map by obtaining a parametric measurement data set for a tissue region with an imaging device, administering treatment, obtaining subsequent sets of parametric measurement data and identifying whether voxels within the tissue region have increased, decreased or remain unchanged. Claim 10 of the patent specifically claims software including the method to generate the parametric response map. This system can help medical professionals find whether or not a cancer treatment is having its intended effect in a much shorter timeframe, down from three to six months to less than a month in some cases. The patent was issued this March to the University of Michigan. During the prosecution of this patent, the patent application faced a non-final rejection in December 2014 as the examiner found that multiple claims were directed to non-statutory subject matter. The examiner cited the Alice case, finding that the claim elements in combination “do not amount to significantly more than an abstract idea.” The prosecution team was able to overcome this rejection by arguing that both the collection of parametric measurement data as well as the voxel-by-voxel comparison and analysis are specific limitations to the method claimed in claim 1 which would not limit others from practicing the abstract idea of an algorithm for determining the disease progression of a tissue sample.
Swiss pharmaceutical company Novartis AG (NYSE:NVS) has also gotten into the medical software sector as is evidenced by the issue of U.S. Patent No. 9412161, titled Systems and Methods for Medical Use of Motion Imaging and Capture and issued in early August. It claims a touchless system for capturing and evaluating motion data of a subject having a neurological condition with the use of a pattern-recognition algorithm to analyze the motion data. This innovation provides a system that can accurately determine fine motor activity and impairment by evaluating the physical ability of subjects. This technology improves upon the use of physical tests which are not suitable for measuring fine motor movements to determine the appropriate treatment for neurological diseases.
Signal processing techniques to support minimally invasive surgeries have been developed by the University of South Florida and are protected through U.S. Patent No. 9402530, titled Minimally Invasive Networked Surgical System and Method. It protects a system for performing networked medical procedures on a subject by a plurality of in vivo medical devices communicating wirelessly across a network that permits the receipt and transmission of control signals between the medical devices which have been generated by an ex vivo control unit to control the behavior of the medical devices. This invention allows the use of in vivo medical devices which can communicate to perform complex surgical tasks with less clutter while providing reliable communications.