Imagine That IP Law Blog

OSI Pharmaceuticals, LLC (OSI) discovered and patented a method of treating non-small cell lung cancer (NSCLC). See US Patent No. 6,900,221. The method comprises administering a therapeutically effective amount of erlotinib to an NSCLC patient. Apotex Inc. (Apotex) challenged OSI’s patent by filing a petition for Inter Partes Review (IPR) at the United States Patent and Trademark Office (USPTO). Three references were produced during the IPR, as follows:

(1) The reference “Schnur” disclosed erlotinib along with 104 other different compounds, as potent inhibitors of a class of oncogenic proteins, stating that compounds in this class can be used for the treatment of a variety of cancers, including “lung.” Schnur did not say “NSCLC.”

(2) The reference “Gibbs” reviewed various signaling mechanisms and how they relate to cancer. Two references cited by Gibbs discussed (A) a different but related compound, compound ZD-1839 and its effects on NSCLC, and (B) erlotinib and its effects on mouse liver tumors and a human head and neck tumor – but not NCLSC.

(3) OSI’s own SEC submission (a 10-K), filed before its patent application over NCLSC, disclosed erlotinib as a potent inhibitor of EGFR, which is a key oncogene in a variety of cancers, including NSCLC. The 10-K also indicated that Phase I FDA safety trials had been completed and that Phase II clinical trials were being initiated in cancer patients.

In short, the prior art of record established that (1) erlotinib was an old drug, known well before the OSI patent application, (2) erlotinib would be useful to treat cancer, and maybe even NSCLC, and (3) OSI had success in advancing erlotinib from Phase I to Phase II trials for NSCLS treatment. All of the parts were assembled for the USPTO to conclude that one of ordinary skill in the art would have expected that erlotinib would successfully combat NSCLS. On this basis, the USPTO held the claims to be obvious and invalid. OSI appealed, and won; i.e., the Court of Appeals for the Federal Circuit reversed the USPTO holding. See OSI Pharmaceuticals, LLC v. Apotex Inc., Fed. Cir. 2018-1925, October 4, 2019.

There were missing pieces in the case for obviousness. Phase I trials only establish that a drug has interesting biological activity in the lab, and is safe when administered people. And recently initiated Phase II trials do not yet prove a drug will work to treat a particular disease. Indeed, the record in the USPTO showed that between 1990 and 2005, there were 1630 new drug compounds that targeted NSCLC that made it to Phase II trials. Of those, 99.5% failed. Other evidence of record also raised doubts whether data pertaining to one cancer would apply to another. In sum, one or ordinary skill in the art would have expected that yet another candidate agent for NSCLC treatment would fail — again. But surprisingly, the use of erlotinib to treat NSCLS made it through FDA trials, erlotinib was used under the mark Tarceva® to treat NCLSC, and OSI’s patent was listed on the FDA “Orange Book” for Tarceva®. It worked.

The record at the USPTO included this evidence, but the USPTO did not consider it by the correct standard, according the Federal Circuit. A case establishing patent claim obviousness requires, among other thing, a showing that one of ordinary skill in the art would have had a reasonable expectation of success at arriving at the claimed invention. And the record must include “substantial evidence” of the reasonable expectation of success.

Here, with respect to Gibbs, the specific references relied upon in Gibbs did not teach what the Board alleged. The first reference referred to the compound ZD-1839 and its effects on NSCLC, and the second referenced referred to erlotinib and its effects on mouse liver tumors and a human head and neck tumor. The Federal Circuit explained that these two separate pieces of evidence cannot be broken and recombined to teach what the Board was looking for – support that erlotinib is used to treat NSCLC. This was also presented to the Board in a Declaration by Dr. Jackson B. Gibbs, the author of the article. He also declared that he was not aware of any published articles describing the clinical or preclinical response of a NSCLC to erlotinib at the time the review article was published.

The record —as a whole— had to be considered under a “clear and convincing” standard. The USPTO, instead, focused on isolated facts, and used a lower standard of review, “preponderance of the evidence.” Having examined the record as a whole and using the appropriate standard of proof to evaluate the evidence, the Federal Circuit determined that one of ordinary skill would not have had a reasonable expectation of success at arriving at the method of treating of NSCLC by administering erlotinib, as claimed in OSI’s patent.

A patent claim may be non-obvious and valid even when all of the parts of the invention are found in separate prior art references. To find to the contrary, it must be shown by substantial evidence that one of ordinary skill in the art would have and a reasonable expectation of success.

A claim to a new method of treatment for an old drug might pass the obviousness test even if there is a suggestion of the method in the art, but not yet any data demonstrating success.