Stealing Fire: A Retrospective Survey of Biotechnology Patent Claims in the Wake of the U.S. Supreme Court s Mayo v. Prometheus Decision

Size: px
Start display at page:

Download "Stealing Fire: A Retrospective Survey of Biotechnology Patent Claims in the Wake of the U.S. Supreme Court s Mayo v. Prometheus Decision"

Transcription

1 Stealing Fire: A Retrospective Survey of Biotechnology Patent Claims in the Wake of the U.S. Supreme Court s Mayo v. Prometheus Decision Elizabeth J. Haanes, Ph.D.,J.D., and Jaume M. Cànaves, Ph.D., J.D. Sterne, Kessler, Goldstein & Fox, P.L.L.C. SUPPLEMENTARY MATERIALS Patent Database Search and Analysis I. Search Design Patent searches were conducted using Thomson Innovation (Thomson Reuters), a commercial database search tool (available at Searches for granted patent documents were conducted against the Thomson Innovation s U.S. Granted Patent Database (U.S. Grant), which as of April 19, 2012 contained 8,651,180 records (1836- present). The selection of terms for the search string was based on the language used in relevant claims in the Prometheus case and other relevant case law. In particular, we selected terms that are commonly found in method or process patent claims in medicine, biomedicine, or biotechnology, as well as terms related to some type of mental process, such as, identifying, categorizing, selecting, assessing, etc. The original data set used in the present study was the result of applying the following search string: CL=(((((A ADJ method) or (A ADJ process)) ADJ15 (therap* OR diseas* OR disorder* OR prognos* OR diagnos* OR immuniz* OR immunis* OR treat* OR patient* OR subject* OR human* OR individual*))) ADJ75 (determining OR assessing OR comparing OR screening OR concluding OR identifying OR diagnosing OR analyzing OR analysing OR detecting OR categorizing OR examining OR monitoring OR categorising OR predicting OR obtaining OR correlating OR inferring OR selecting OR observing)); CL indicates that the search was conducted in the claims field of each patent record. The proximity operator ADJ between two terms indicates that the two terms are contiguous and in the specified order. The proximity operator ADJn between two terms indicates that the terms are within n words of each other and in the specified order. The wildcard character * can correspond to any number of characters, including zero.

2 The search string comprised two components: (i) (ii) The substring "(A ADJ method) or (A ADJ process)) ADJ15 (therap* OR diseas* OR disorder* OR prognos* OR diagnos* OR immuniz* OR immunis* OR treat* OR patient* OR subject* OR human* OR individual*)," which was aimed at detecting the occurrence of keywords within the claim preamble; and The substring "(determining OR assessing OR comparing OR screening OR concluding OR identifying OR diagnosing OR analyzing OR analysing OR detecting OR categorizing OR examining OR monitoring OR categorising OR predicting OR obtaining OR correlating OR inferring OR selecting OR observing)" which was aimed at detecting the occurrence of keywords in the body of the claim, preferably steps in a method or process. The election of the proximity operator ADJ15 within the first substring and the proximity operator ADJ75 between the first and second substrings was the product of empirical observations indicating that the terms "A method" or "A process" and certain keywords occurred frequently within 15 words in the preamble, and the keywords describing steps in a method or process tended to occur within 75 words from the preamble. The use of the wildcard character * in the preamble substring was due to the fact that terms in the preamble can have different endings, e.g., a claim can use indistinctly language such as "A method of treating" or "A method for treatment." In contrast, the present participle form was used for the terms in the claim body substring because such verbal form is conventionally used in steps in method or process claims, e.g., "a method of comprising: (a) assessing (b) comparing. (c) diagnosing." Each patent record included the following fields: patent publication number, publication date, patent application date, title, first claim, abstract, IPC classification, U.S. classification, and INPADOC family ID. This patent data was exported as a Microsoft Office Excel spreadsheet for further analysis. Patent publication number: A unique alphanumeric identifier composed of a country code and a number which distinguishes the granted patent application from any other existing patent document, e.g., U.S. 7,654,321. Publication date / Publication year: The date and year when a patent was officially published. Patent application date / Patent application year: The date and year when a patent application was filed. Title: The title shown on the patent document. A short description of the invention, not always sufficiently informative. Abstract: A brief summary or description of the invention. First claim: The first of the claims/assertions made in the patent about the subject matter and scope of the invention. Claims are a particularly important part of the patent document that specifically defines the boundaries of what is to be protected. IPC class / IPC class group / IPC section / IPC subclass / IPC subgroup / IPC current full: International Patent Classification (IPC) code to which this patent is assigned. This - ii -

3 classification system is administered by the World Intellectual Property Organization (WIPO) (see The classification represents the whole body of knowledge which may be regarded as proper to the field of patents for invention, divided into eight sections. Sections are the highest level of hierarchy of the classification. Each section is subdivided into classes. Each class comprises one or more subclasses. Each subclass is broken down into subdivisions referred to as "groups," which are either main groups or subgroups. IPC record format, e.g., for entry A01K 1/00 is as follows: A (section), 01 (class), K (subclass), 1 (main group), 00 (subgroup). U.S. class / U.S. class-main / U.S. class-original: The U.S. Patent Classification code which this patent is assigned. The classification system is administered by the U.S. Patent & Trademark Office (USPTO) (see The original class (U.S. class-original) is the class that was assigned when a patent was originally published. The current class (U.S. class-main) is the class to which a patent is now assigned. INPADOC family ID: International Patent Documentation Centre (INPADOC) family identifier. INPADOC is part of the European Patent Office. An INPADOC family consists of an earlier patent application ("parent application") and subsequent applications or publications anywhere in the world that claim priority from the parent application ("child applications") (see II. Search Results and Data Set Curation A search using the above described search string against the U.S. Grant database returned 10,293 patent records, which was the data set used in all further analysis. We curated the 10,293- patent record data set as follows. To reduce bias due to the presence of large patent families, we filtered the data set to remove multiple records corresponding to the same INPADOC family, leaving only a representative member of each INPADOC family in the data set. Patents that issued before 1992 were filtered out since all patents which granted prior to 1992 are likely expired by now. The Thomson Innovation search was conducted on the entire claim set for each patent. Relevant claims for the purpose of this study might be in any part of the claim set, which would require individual review of each claim in each patent to identify relevant claims. To expedite analysis, we limited our further analysis to only the first claim in each patent in the data set, and we filtered-out patents in which the first claim was not a method or process claim (e.g., patents with first claim directed to a composition of matter, a device, an apparatus, etc.). The curated data set was further limited to include only those patents in which the first claim was a method and process claim and the patent was classified in IPC biotechnology-related classes A61 or C12. The method and process claims in the data set were also filtered to remove those records corresponding to methods or processes with clearly transformative steps (e.g., methods of synthesis of therapeutic products, methods to perform surgical procedures, methods of treatment with novel drugs where the only mental step is selecting patient, methods directed to complex medical or technological procedures, etc.). The final curated data set comprised 2,218 patent records. - iii -

4 III. Curated Data Set Analysis The records in the curated data set corresponded to 38 different U.S. classes and 428 subclasses. The top five U.S. classes, comprising 85% of the total number of patents, were classes 435, 600, 607, 424 and 514. Although it has a different class number, Class 514 is considered to be an integral part of Class 424. See Supplementary Table 1 and Supplementary Figure 1. The trend in number of patents filed and issued in the curated data set is shown in Supplementary Figure 2. The number of issued patents in the patent classes represented in the curated data set has grown explosively in the past four years ( period) compared to the historical trend. The distance between the peaks in the filed applications and issued patent distributions is shown in Supplementary Figure 3 and indicates a substantial delay (on average, about 4 to 5 years) between patent application and patent issuance. This delay is shown in further detail in Supplementary Table 2. For our detailed claim analysis, we selected Class 435, which comprised 1,180 patent records (54.15% of the curated data set). These Class 435 patents corresponded to a total of 59 subclasses. The most common Class 435 subclasses are shown in Supplementary Table 3 and Supplementary Figure 4. The ten subclasses included in Supplementary Table 3 comprise a total of 947 records (80% of the Class 435 patents in the data set, 43% of the patents in the curated data set). IV. Data Set Limitations Our data set is exemplary and not comprehensive. The final curated data set is limited by the choice of terms and operators used to construct the database search string. For example, different terms and less stringent proximity operators could have resulted in a substantially larger number of hits in the search, although probably at the cost of higher number of false positives. Another caveat regarding the curated data set relates to the fact that curation was limited to the first claim of each patent in the original data set. Thus, the exclusion of a patent record from the curated data set does not imply that such patent does not contain claims that may be affected under Prometheus, only that those claims may be elsewhere in the claim set. Conversely, the presence of a first claim which might be patent-ineligible under Prometheus does not imply that the rest of claims in the patent would likewise be affected. The 1,180 record subset of the curated data set selected for claims analysis in the present study is limited to Class 435 biotechnology-related patents, which fall within the general field of expertise of the authors of the present study. Patents corresponding to other areas of science and engineering identified in the present search, or patents that may be identified through alternative search strategies may also be affected by Prometheus. Thus, the curated and analyzed data set represents a non-comprehensive cross-section of biotechnology-related patents that may be affected by Prometheus. - iv -

5 V. Claim Analysis and Categorization To quantitate the extent of potential vulnerabilities related to Prometheus in the Class 435 subset of the curated data set, we manually analyzed and classified the claims into four categories (see Supplementary Table 4): (1) Claims reciting no, or generalized "pre-solution activity" and no "postsolution activity"; (2) Claims reciting specific "pre-solution activity" and no "post-solution activity"; (3) Claims with generalized, non-specific "post-solution activity;" and, (4) Claims with specific "post-solution activity" A fifth category included claims which could not be clearly classified in view of Prometheus. These claims were, for example, directed to methods of screening therapeutic compounds, methods to perform in vitro assays, etc. The classification of these claims may be reassessed in the future as subsequent cases, e.g., Myriad, are decided. - v -

6 Supplementary Table 1 Top 5 U.S. Classes in Curated Data Set Class Top 5 Total % Description CHEMISTRY: MOLECULAR BIOLOGY AND MICROBIOLOGY SURGERY SURGERY: LIGHT, THERMAL, AND ELECTRICAL APPLICATION DRUG, BIO-AFFECTING AND BODY TREATING COMPOSITIONS DRUG, BIO-AFFECTING AND BODY TREATING COMPOSITIONS - vi -

7 Supplementary Table 2 Time from Patent Application Filing to Patent Grant in the Curated Data Set Filing to Grant (days) # Patents % Patents vii -

8 Supplementary Table 3 Top 10 U.S. Subclasses in Class 435 Patents in Curated Data Set Class 435 Top 10 Subclasses # of records % of records in curated set DETECTING CANCER Description TEST STRIP involving a nucleic acid encoding a protein related to the nervous system, (e.g., nerve related factors, brain-derived cytokines, nerve cell biomarker, etc.) TEST STRIP nucleic acid based assay involving a hybridization step with a nucleic acid probe, involving a single nucleotide polymorphism (SNP), involving pharmacogenetics, involving genotyping, involving haplotyping, or involving detection of DNA methylation gene expression TEST STRIP involving antigen-antibody binding, specific binding protein assay or specific ligand-receptor binding assay TEST STRIP TEST STRIP with significant amplification step (e.g., polymerase chain reaction (PCR), etc.) TEST STRIP tumor cell or cancer cell TEST STRIP involving virus or bacteriophage TEST STRIP involving nucleic acid TEST STRIP involving a nucleic acid encoding an enzyme - viii -

9 Supplementary Table 4 Claim Classification in Class 435 Subset Claims Reviewed Classification Group % relative to full 1180 records in Class 435 subset % relative to records in Classification Groups 1-4 No Post-Solution Activity Post-Solution Activity Methods of drug No or Nonscreening, in specific Pre- Specific Non-specific Specific vitro assays, Solution Pre-Solution Post-Solution Post-Solution etc. Activity Activity Activity Activity % 18.97% 0.59% 0.68% 5.67% 78.64% 20.11% 0.63% 0.72% -- - ix -

10 Supplementary Figure 1 U.S. Class distribution in curated data set. The most common classes in the curated data set are shown in the X axis. Percentages are shown in the Y axes and on top of the bars. The most common class is Class 435, comprising 54.15% of the records in the curated data set x -

11 Supplementary Figure 2 Trends in application filings per year and issued patents per year in the curated dataa set. The values are represented as percentages with respect to the total number of records in the curated data set Filed (%) Issued (%) xi -

12 Supplementary Figure 3 Days from application filing to patent issue. The X axis shows day ranges, whereas the Y axis shows percentages with respect to the total number of records in the curated data set. Most patents issued between 1,201 and 1,400 days after filing xii -

13 Supplementary Figure 4 Most common Class 435 subclass in the curated data set. U.S. Class and Subclass identifiers are shown in the X axis. The percentage of records in the curated data set corresponding to each subclass is shown in the Y axis xiii -