Life Sciences Patent Strategy

Life sciences is not a monolithic IP landscape — pharmaceutical; biologic; medical device; diagnostics; and digital health companies face fundamentally different patent term economics; regulatory interactions; and patent eligibility challenges: PHARMACEUTICAL PATENTS: PRIMARY REVENUE PROTECTION: composition of matter patents (the molecule itself) + method of treatment patents + formulation patents + dosing regimen patents; PATENT TERM EXTENSION: 35 U.S.C. § 156: up to 5 years extension for FDA regulatory approval delay; maximum 14 years post-approval; one extension per drug; apply within 60 days of NDA/BLA approval; FDA DATA EXCLUSIVITY: separate from patent protection; NCE 5-year exclusivity (new chemical entity); 3-year exclusivity for new clinical study; prevents ANDA filing regardless of patent status; GENERIC DRUGS (HATCH-WAXMAN): Orange Book listed patents; paragraph IV challenge; 30-month stay; 180-day first filer exclusivity for para IV ANDA filer; key strategy: thick patent estate around each drug (primary + secondary + tertiary patents); BIOLOGICS: LONGER DEVELOPMENT TIMELINE: 10-15 years; more complex manufacturing = harder to biosimilar; 12-YEAR DATA EXCLUSIVITY under BPCIA (Biologics Price Competition and Innovation Act): 4-year safe harbor; 12-year data exclusivity prevents most biosimilar approvals regardless of patent status; PATENT DANCE: BPCIA optional but strategic; provides information for license negotiations or litigation; post-Amgen v. Sandoz (S.Ct. 2017): patent dance steps are largely optional for biosimilar applicant; ANTIBODY PATENTS: Amgen v. Sanofi (S.Ct. 2023): functional antibody genus claims require enabling examples across the full scope; MEDICAL DEVICE PATENTS: FDA REGULATORY PATHWAY INTERACTION: 510(k) predicate device finding + patent: device with predicate = clears FDA faster + may have related patents to leverage vs. design-around competitors; PMA (premarket approval): de novo patents + FDA clinical data investment create compound barriers; DESIGN + UTILITY: medical devices benefit from both design patents (ornamental appearance) + utility patents (functional aspects); design patents fast (14 months; no maintenance); DIGITAL HEALTH (SaMD): SOFTWARE AS A MEDICAL DEVICE (SaMD): FDA regulates AI/ML-based medical software; § 101 CHALLENGE: diagnostic + therapeutic decision support software faces Mayo/Alice; must clear abstract idea + natural phenomenon; DTCA: direct-to-consumer devices (Apple Watch ECG; blood oxygen) raise regulatory + patent questions simultaneously.

Mayo v. Prometheus (S.Ct. 2012) is one of the most consequential Supreme Court decisions for life sciences patent law — it dramatically restricted diagnostic method patents by holding that laws of nature cannot be patented even when a human-engineered step is added: MAYO v. PROMETHEUS — THE CASE: PATENTS AT ISSUE: US6,355,623 and US6,680,302: methods for optimizing therapeutic efficacy for treatment of immune-mediated gastrointestinal disorders; CLAIMED PROCESS: (1) administering thiopurine drug to patient; (2) determining the level of 6-thioguanine metabolite in blood; (3) correlating the level to optimal dosing; PRIOR KNOWLEDGE: the correlation between metabolite level and drug effectiveness was a natural phenomenon (a law of nature — the body naturally metabolizes the drug and the metabolite level correlates with outcome); HOLDING (9-0): the correlation between metabolite level and drug efficacy is a law of nature; adding 'administering the drug' is conventional; adding 'determining the metabolite level' is conventional laboratory technique; adding 'wherein' statements correlating to outcomes adds nothing beyond the natural law; result: patent claims held invalid under § 101; BREADTH OF IMPACT: most diagnostic method patents now face serious § 101 challenges; any claim that says 'measure biomarker X; if X > threshold, diagnose or treat condition' is vulnerable to Mayo challenge; WHAT SURVIVES MAYO: SPECIFIC TECHNICAL NOVELTY BEYOND THE CORRELATION: if the diagnostic method includes a specific novel assay step (not conventional laboratory techniques); if the detection method itself is novel and non-obvious (not merely applying a known technique to detect a known correlation); SPECIFIC NOVEL SAMPLE PREPARATION; SPECIFIC NOVEL DETECTION CHEMISTRY: claims covering a genuinely novel assay platform; a new antibody with novel binding characteristics; novel sample preparation for liquid biopsy; COMPANION DIAGNOSTICS + PATENTS: CDx (companion diagnostic) + drug combination: the drug patent + CDx method may together be patentable if the CDx method is technically novel; Roche; Abbott; Foundation Medicine; LIQUID BIOPSY: Grail (Galleri multi-cancer early detection test); Guardant Health; Foundation Medicine; PERSONAL GENOME DIAGNOSTICS; key innovations: cfDNA (cell-free DNA) fragment analysis; methylation patterns; tumor mutation burden calculations; these may survive Mayo if the detection method — not just the correlation — is novel; ASSOCIATION FOR MOLECULAR PATHOLOGY v. MYRIAD GENETICS (S.Ct. 2013): isolated genomic DNA is a product of nature = not patentable; cDNA (complementary DNA; not naturally occurring) may be patentable; MYRIAD claimed BRCA1/BRCA2 gene sequences; lost on isolated genomic DNA claim; won on some cDNA claims.

The interaction between FDA regulatory timelines and patent protection is the most important IP strategy consideration in life sciences — because the commercial value of a patent depends critically on how much exclusivity remains when FDA approval is obtained: PHARMACEUTICAL FDA-PATENT TIMELINE COORDINATION: IDEAL TIMELINE: File patent application → enter clinical trials → complete Phase III → file NDA/BLA → receive FDA approval while 10+ years of patent remaining → sell under patent exclusivity → generic entry triggers 6-12 months after patent expiry or upon successful paragraph IV challenge; TYPICAL REALITY: basic compound patent filed during discovery (preclinical); clinical trials take 8-12 years; FDA approval process adds 1-2 years; total: patent may be 10-15 years old at first approval; remaining patent term: 5-10 years; patent term extension (§ 156): up to 5 years back for regulatory review time; PATENT TERM EXTENSION CALCULATION: regulatory review period = IND date to NDA approval date; extension = 1/2 of IND-to-Phase III start period + Phase III + FDA review; maximum extension: 5 years; maximum post-approval term: 14 years; PEDIATRIC EXCLUSIVITY: 6 additional months added to ALL unexpired patent terms for a drug if sponsor conducts pediatric studies per FDA written request; significant value; ORANGE BOOK PATENT LISTING STRATEGY: list ALL patents that claim the drug or method of use; each listed patent gets its own 30-month stay upon paragraph IV challenge; strategy: stagger patent expiration dates to extend total exclusivity period; MEDICAL DEVICE FDA-PATENT INTERACTION: FDA CLEARANCE TIMING vs. PATENT LIFE: 510(k) clearance: 3-12 months; minimal patent consumption; PMA: 3-7 years; significant patent consumption; device patent term extension: NOT available (§ 156 only covers drugs and biologics); DESIGN SPACE AND 510(k): 510(k) requires 'substantial equivalence' to predicate device; if you patent specific design features and a competitor changes those features to achieve substantial equivalence to a different predicate, your patent may be avoided; consider patenting functional performance outcomes (X% improvement in flow rate; Y% reduction in leakage) in addition to specific design features; EU MDR (Medical Device Regulation): new EU rules increased clinical evidence requirements for device CE marking; patent strategy must account for longer EU clearance timelines; IPR AND MEDICAL DEVICES: medical device patents face IPR challenges when competitors use Hatch-Waxman analog strategies; consider pre-issuance submissions on competitor applications and early monitoring.

Digital health represents the fastest-growing and most IP-uncertain area of life sciences — FDA regulatory frameworks for AI/ML-based software as a medical device (SaMD) are still evolving, and § 101 patent eligibility challenges are the norm rather than the exception: DIGITAL HEALTH IP LANDSCAPE: SOFTWARE AS A MEDICAL DEVICE (SaMD) DEFINITION: FDA guidance: software that performs medical device function; risk classification based on the significance of the clinical information provided; FDA AI/ML ACTION PLAN (2021): recognized that AI/ML-based SaMD adapts over time ('locked' vs. 'adaptive' algorithms); predetermined change control plan (PCCP) framework to allow algorithm updates without full resubmission; MAJOR DIGITAL HEALTH PATENT HOLDERS: APPLE HEALTH: ECG algorithm (AliveCor licensed); atrial fibrillation detection; SpO2 sensing; fall detection; ABBOTT DIABETES: Libre CGM (continuous glucose monitoring) sensor + algorithm; DEXCOM: CGM accuracy algorithm; predictive alerts; MEDTRONIC: closed-loop insulin delivery (MiniMed 780G): basal insulin adjustment algorithm; MDT Connect cloud connectivity; INTUITY MEDICAL: lancing device; CLINICAL AI COMPANIES: Viz.ai: AI detection of large vessel occlusion stroke from CT images; FDA De Novo clearance 2018; Aidoc: radiology AI triage; Butterfly Network: AI ultrasound interpretation; PathAI: digital pathology AI; Tempus AI: genomic + clinical data + AI for oncology decisions; PATENT ELIGIBILITY FOR DIGITAL HEALTH: § 101 STRATEGY: anchor ML claims in specific hardware (specific sensor type; specific imaging modality; CT scanner characteristics); specific neural network architecture (ViT; ResNet variant; specific layer configurations) with measured clinical performance (sensitivity; specificity; AUC improvement); provide § 1.132 declarations with clinical validation data; NATURAL PHENOMENON vs. MACHINE LEARNING CLAIMS: 'detecting atrial fibrillation from ECG data' is a natural phenomenon if the correlation between the ECG pattern and AF is known; but 'a specific convolutional neural network trained on X+ labeled ECG segments achieving sensitivity Y and specificity Z' is a specific technical implementation; TRADE SECRETS IN DIGITAL HEALTH: FDA-cleared training datasets (labeled medical images; EHR data with ground truth); model weights achieving clinical performance; relationship with health system data partners; REGULATORY-PATENT COORDINATION: FDA De Novo clearance creates a new regulatory pathway predicate — patent the cleared algorithm specifications so competitors cannot use your predicate without licensing.