The White House recently announced the Cancer Moonshot Task Force, an effort to “focus on making the most of Federal investments, targeted incentives, private sector efforts from industry and philanthropy, patient engagement initiatives, and other mechanisms to support cancer research and enable progress in treatment and care.” Yesterday, in support of this project, the White House and the USPTO launched “Patents 4 Patients,” an initiative to “free fast-track review for cancer treatment-related patents.” More specifically, the objective of the program “is to complete the examination of an application containing a claim(s) to a method of treating a cancer using immunotherapy within twelve months of special status being granted.”
In theory this sounds like a great idea. In practice, there’s a significant problem: Section 101 and the scientifically incoherent definition of a ‘law of nature’ imposed by the Supreme Court’s decision in Mayo Collaborative Services v. Prometheus Laboratories, Inc. In Mayo the Court essentially ruled that any natural relationship or correlation that “exists in principle apart from any human action” sets forth a law of nature, and thus is not patent eligible without the addition of "significantly more."
This definition is not scientifically correct, as mere correlations of naturally occurring conditions do not rise to the level of “laws of nature” just by dint of the statistical relationship. First, as any student of statistics knows, correlation does not imply causation: that A is correlated with B does not mean that A is the cause of B (or vice versa). It is often the case that there is a third and yet unknown agent that is the cause of both: C causes both A and B. Concluding that the relationship between A and B based on the correlation alone is a logical fallacy. By coincidence, just yesterday the Federal Circuit itself latched on to a similar fallacy, rejecting an argument made by the USPTO in support of a decision by PTAB: the USPTO’s “suggestion is fallacious: It employs a version of the fallacy of the undistributed middle, under which the two statements, ‘a dog is a four-legged pet’ and ‘a cat is a four-legged pet,’ are asserted to give rise to the inference that a dog is a cat. A common characteristic [between two things] does not allow a conclusion that one can be the other.”
Second, in the field of biology specifically, there is considerable debate about whether there are any laws of nature at all. The Evolutionary Contingency Thesis states that there are no laws of biology, since biological relationships are the result of evolutionary forces—random mutations, environmental variables and selection pressures, and multiple functionally equivalent adaptions. As Stephen Gould put it vividly: “evolution is like a videotape that, if replayed over and over, would have a different ending every time.” That a particular sequence of amino acids codes for a specific protein is just as much a result of contingent evolutionary forces as the ability of aspirin to stop headaches: neither was a necessary outcome of the universe. Or to put in the terms overused by the Supreme Court, neither of these are on the level of Einstein’s E=mc2 or Newton’s universal law of gravitation.
The district courts and Federal Circuit have adopted the Supreme Court's naked statement, sometimes yielding absurd results. The court in Genetic Technologies Ltd., v Laboratory Corp. of America, et. al, (CV 12-17360LPS-CJB) held that "the link between a particular genetic variation and the potential for elite athletic performance embodied in Claim 1 is "a natural process, an eternal truth that 'exists in principle apart from any human action." I seriously doubt that elite athletic performance is the kind of thing Newton, Einstein or Richard Feynman would put the same plane as the fundamental laws of physics.
The USPTO has also followed this 'definition' literally and expansively, and as a result has rejected thousands of patent applications as claiming laws of nature. Particularly relevant to the Patents 4 Patients program is that the USPTO has issued Section 101 rejections in hundreds patents related to the identification and treatment of cancer —and many of these have been abandoned (often without any further continuations). Here is a small sample of cancer-related applications that have been rejected and abandoned:
|Publication Number||Title of Invention||Assignee||Foreign Patents?|
|US20100196907||MARKERS TO PREDICT AND MONITOR RESPONSE TO AURORA KINASE B INHIBITOR THERAPY||ABBVIE||CN, EP, JP|
|US2014020023 3||USE OF MAD2L2 AS A STRATIFICATION MARKER IN THE TREATMENT OF BREAST TUMOURS WITH NOVEL PAN-CDK INHIBITORS||BAYER INTELLECTUAL PROPERTY GMBH||CN, EP|
|US20120115151||METHOD FOR TESTING A SUBJECT THOUGHT TO BE PREDISPOSED TO HAVING METASTATIC CANCER USING DELTA133P53BETA||CENTRE NATIONAL DE LA RECHERCHE SCIENTIFIQUE||CN|
|US20140030715||METHOD OF USING BOTH MIR-196A AND MIR-196B AS BIOMARKERS FOR DETECTING ORAL CANCER||CHANG GUNG UNIVERSITY|
|US20120183518||ASSAY FOR DETERMINING HEALTH OF CD8+ T CELLS||COLUMBIA UNIVERSITY IN THE CITY OF NEW YORK||CN, EP|
|US20100129799||CANCER SUSCEPTIBILITY VARIANTS ON CHR8Q24.21||DECODE GENETICS EHF||JP|
|US20110269143||GENETIC VARIANTS AS MARKERS FOR USE IN URINARY BLADDER CANCER RISK ASSESSMENT, DIAGNOSIS, PROGNOSIS AND TREATMENT||DECODE GENETICS EHF|
|US20110117545||GENETIC VARIANTS ON CHR2 AND CHR16 AS MARKERS FOR USE IN BREAST CANCER RISK ASSESSMENT, DIAGNOSIS, PROGNOSIS AND TREATMENT||DECODE GENETICS EHF||CN, JP|
|US20130029319||MEANS AND METHODS FOR PREDICTING THE RISK OF MORTALITY OF PATIENTS WITH HPV POSITIVE OROPHARYNGEAL SQUAMOUS CELL CANCER||DEUTSCHES KREBSFORSCHUNGSZENTRUM RUPRECHT-KARLS-UNIVERSITAT HEIDELBERG|
|US20130143237||COMPOSITIONS AND METHODS FOR PROSTATE CANCER ANALYSIS||F HOFFMANN-LA ROCHE AG|
|US20110159492||DIFFUSE LARGE B-CELL LYMPHOMA MARKERS AND USES THEREFORE||FRED HUTCHINSON CANCER RESEARCH CENTER|
|US20100092965||GENETIC VARIATIONS ASSOCIATED WITH TUMORS||GENENTECH||EP|
|US20130224193||METHODS AND COMPOSITIONS FOR DIAGNOSTIC USE IN CANCER PATIENTS||GENENTECH INC.||CN, EP|
|US20130224194||METHODS AND COMPOSITIONS FOR DIAGNOSTIC USE IN CANCER PATIENTS||GENENTECH INC.||CN, EP|
|US20120028264||METHOD FOR USING GENE EXPRESSION TO DETERMINE PROGNOSIS OF PROSTATE CANCER||GENOMIC HEALTH||EP|
|US20130084564||ASSESSMENT OF CANCER RISK BASED ON RNU2 CNV AND INTERPLAY BETWEEN RNU2 CNV AND BRCA1||GENOMIC VISION|
|US20130243758||METHOD TO IDENTIFY A PATIENT WITH AN INCREASED LIKELIHOOD OF RESPONDING TO AN ANTI-CANCER THERAPY||HOFFMANN-LA ROCHE INC.||CN, EP|
|US20130217712||PREDICTIVE MARKER FOR EGFR INHIBITOR TREATMENT||HOFFMANN-LA ROCHE INC.||CN, EP|
|US20110281284||NOVEL LIVER CANCER MARKER||HYOGO COLLEGE OF MEDICINE OTSUKA PHARMACEUTICAL CO||CN, EP|
|US20100317001||METHOD OF USING TUMOUR RNA INTEGRITY TO MEASURE RESPONSE TO CHEMOTHERAPY IN CANCER PATIENTS||LAURENTIAN UNIVERSITY||EP, JP|
|US20110306049||METHOD FOR DETECTING GYNECOLOGIC CANCER||LSIP||CN, EP|
|US20110117551||DETECTION AND PROGNOSIS OF LUNG CANCER||MDXHEALTH||EP|
|US20120004129||BIOMARKERS FOR INHIBITORS WITH ANTI-ANGIOGENIC ACTIVITY||MERCK PATENT GMBH||CN, EP|
|US20090181393||METHODS FOR THE IDENTIFICATION, ASSESSMENT, AND TREATMENT OF PATIENTS WITH PROTEASOME INHIBITION THERAPY||MILLENNIUM PHARM||EP, JP|
|US20140065619||MARKER SET OF HIF-1ALPHA, HDAC3 AND WDR5 FOR PREDICTING METASTASIS AND SURVIVAL OF CANCERS||NATIONAL YANG-MING UNIVERSITY|
|US20130344510||IDENTIFICATION OF CANCER PATIENTS LIKELY TO BENEFIT FROM RADIATION THERAPY||NOVARTIS AG|
|US20130224209||METHODS FOR DIAGNOSING AND TREATING PROSTATE AND LUNG CANCER||PICOBELLA LLC||CN, EP|
|US20150072355||FLAP ENDONUCLEASE-1 AS A MARKER FOR CANCER||ROCHE DIAGNOSTICS GMBH||CN, EP, JP|
|US20110244497||METHOD FOR JUDGING SUSCEPTIBILITY OF CANCER CELLS TO ANTHRACYCLINE ANTICANCER AGENT AND COMPUTER PROGRAM||SYSMEX CORPORATION||EP, JP|
|US20090246789||GENE MUTATION PROFILING OF CSMD1||UNITED STATES NATIONAL INSTITUTES OF HEALTH|
|US20070092893||METHODS AND COMPOSITIONS FOR IDENTIFYING CANCER-RELATED BIOMARKERS||UNITED STATES NATIONAL INSTITUTES OF HEALTH||EP|
|US20100221722||METHODS FOR EVALUATING BREAST CANCER PROGNOSIS||UNITED STATES NATIONAL INSTITUTES OF HEALTH|
|US20120028260||RAD9 AS A DIAGNOSTIC, PROGNOSTIC AND THERAPEUTIC TOOL FOR PROSTATE CANCER||UNITED STATES NATIONAL INSTITUTES OF HEALTH|
|US20100035240||METHODS AND KIT FOR THE PROGNOSIS OF BREAST CANCER||UNIVERSITY COLLEGE CARDIFF CONSULTANTS LIMITED||CN, EP, JP|
|US20130137086||HPV DNA METHYLATION PATTERNS OF DIAGNOSTIC OR PROGNOSTIC SIGNIFICANCE IN CERVICAL CANCER SCREENING||YALE UNIVERSITY|
|US20140004521||MICRORNA-BASED METHODS FOR PROGNOSIS OF HEPATOCELLULAR CARCINOMA||YEH CHAU-TING HUANG YA-HUI CHEN HUA-CHIEN LIN KWANG-HUEI|
The far right column indicates whether the patentee was able to obtain patent protection for these inventions in China, the EPO, or Japan. In other words, these countries accepted these cancer-related inventions as eligible subject matter even though the US did not, and thus granted the patents. That hardly seems encouraging for research in the US or good for patients.
The assignees of these patent applications are listed in the third column. As you can see, the National Institute of Health is the assignee of four of these. That the NIH was denied patents here is a bit troubling: US taxpayers funded the research and then were denied the patent protection, in some cases only on the basis of a Section 101 rejection, not over any prior art.
Let’s take a look at US20090246789, for GENE MUTATION PROFILING OF CSMD1. This application relates to a method of identifying the risk of colon cancer based on a mutation in the CSMD1 gene. Here is the claim:
1. A method for detecting a mutation in the CSMD1 gene, the method comprising:
purifying genomic DNA from epithelial cells of a tumor sample from a human subject;the genomic DNA including the CSMD1 gene;
amplifying exons of the CSMD1 gene to form CSMD1 amplicons;
sequencing the CSMD1 amplicons to an average depth of 1800 fold, such that 90% of the amplicons are represented by greater than 300 sequencing reads per sample;
detecting, in the sequenced CSMD1 amplicons, a mutation in the CSMD1 gene, the mutation comprising an alteration selected from the group consisting of:
a Tat genomic position 1588709, encoding a G733>STOP CSMD1 protein mutation;
a Tat genomic position 1595339, encoding a G777>V CSMD1 protein mutation;
an A at genomic position 1687052, encoding a P1298>T CSMD1 protein mutation;
a Tat genomic position 1907660, encoding an R2476>STOP CSMD1 protein mutation; and
an A at genomic position 1994688, encoding a G2682>8 CSMD1 protein mutation.
The examiner’s final rejection is typical:
With regard to the claims as the judicial exception may be considered to be a natural principle the rejected claims are drawn to methods for detecting a mutation in the CSMD1 gene. The use of such a method, as asserted in the specification, is in assessing the risk of node positive colorectal cancer.
In so far as the instant claims are based on a law of nature-namely relationships between gene mutation and a clinical phenotype, as asserted by the Applicants, the relationship is a consequence of the way in which gene mutations affect subject biological processes – a entirely natural relationship. And so a patent that simply describes that relation sets forth a natural law.
As detailed earlier in this rejection, the judicial exception of the rejected claims may also be considered to be the naturally occurring nucleic acids in which a mutation is detected.
This is simply wrong scientifically. That these specific genomic mutations in CSDM1 are correlated with a risk of colon cancer is not, scientifically speaking, a law of nature. First, it could be that there is some other factor, as yet unidentified, that causes both the CSDM1 mutations and the increased risk of cancer. Second, the correlations (which are not even claimed here) may in fact be entirely spurious: the invention was based on a DNA analysis of just 28 specimens (16 with colon cancer, and 12 without). While the correlations may be statistically significant, there’s simply not enough data at this point to definitely establish a causal relationship; history teaches us that many so-called medical discoveries of just this type turn out to be incorrect in the long run, and thus not “laws of nature” at all. Indeed, the Examiner issued a Section 112 rejection on essentially the same reasoning:
[T]he specification does not provide any validation or replication of any particular mutations in specific cancer stages, nor is there any statistical analysis of the identification of mutations in particular stages…While the state of the art and level of skill in the art with regard to identification of known mutations in human samples is high, the unpredictability in associating any particular mutation with a specific phenotype or risk is even higher. The unpredictability is demonstrated by the related art and the instant specification.
If the Examiner is correct here, then this contradicts the rejection that the claims are directed to an actual law of nature: how the claims be a law of nature, if the unpredictability of associating any specific mutation with the appearance of cancer is so high? Laws of nature allow us to predict with great accuracy exactly what future outcomes will occur under particular circumstances (even probabilistic laws operate with significant accuracy). If that cannot be done here, then there is no law of nature being claimed.
More fundamentally, medical researchers do not wait–indeed cannot wait–until their discoveries are definitively validated by the scientific community: they must file for patent protection well before that. Moreover, the scientific method precludes proving something is true: it can only falsify a hypothesis, never prove it. Together this means examiners are very likely to identify as laws of nature relationships that turn out to be at best incomplete or at worst simply incorrect. A patent on non-law of nature does not block any fundamental research, but the absence of such a patent may undermine the research efforts of the patentee.
Finally, even if it is shown that the mutation in question here in fact causes the cancer, that is not enough since that relationship was the result of evolutionary forces–it could have turned out that the videotape (or more accurately, game) of evolution resulted in an entirely different gene mutation causing colon cancer–or none at all. If we intend to grant special "law of nature" status to biological relationships, then that is something that Congress can do in a more precise and appropriate manner, since as the Mayo Court acknowledged "[c]ourts and judges are not institutionally well suited to making the kinds of judgments needed to distinguish among different laws of nature."
The listed applications are only a sample of cancer-related patent applications which are now abandoned, meaning that the Section 101 rejections were received more than six months ago. There are several implications from this simple fact. First, there are approximately 900 cancer-related applications that have received Section 101 rejections (both final and non-final) since January 1, 2016, many which are likely to be abandoned. Second, even if not abandoned, these applications are facing additional costs and delays—they are under-going continue examination and or appeal. In my review of the impact of Alice two years after the decision, I found that in the Microbiology and Immunology art units in Tech Center 1600, examiners issued final Section 101 rejections in 27.3% and 16.9% of their office actions after July 30, 2015. Finally, all of my figures are for applications that have published; published applications account for about 60% of all applications. Assuming that cancer-related applications are published at about the same rate as other applications, then the actual number of rejected and abandoned applications is higher still by about 66%.
Thus, if the USPTO continues in this manner, what the Patents 4 Patients program offers for a large number of cancer researchers is a fast track to a Section 101 rejection, and possibly abandonment or years of waiting on appeal. And on appeal, the likelihood of reversing a Section 101 rejection is less than 10%.
If the USPTO is serious about its Patents 4 Patients efforts, if it is "standing up and doing its part to help bring potentially life-saving treatments to patients, faster," then it needs to recalibrate how it applies Section 101 to patents in this field. Yes, the USPTO is bound by the Supreme Court, but that does not preclude it from interpreting the Court in a way that is scientifically coherent. The USPTO should focus on the Supreme Court’s stated concern—avoiding preemption of actual laws of nature, "the basic tools of scientific and technological work"—and adopt a more nuanced and scientifically accurate definition of what counts as a law of nature. What the Court implicitly assumed was that all natural relationships are universal, necessary, and causal relationships (again, think of Einstein and Newton, the Court’s go-to examples). The USPTO should make express this assumption of the Court in its test for a law of nature. Second, the burden is and remains on the examiner to demonstrate that the claimed subject matter rises to this level. With these two relatively simple rules in place, the unnecessary Section 101 rejections of cancer-related inventions can be substantially eliminated, and the patents start issuing for the patients.