Delhi HC to Patent Office: "Skilled Artisans Would Combine Them" Is Not an Analysis
By Prashant Kotian, PhD Researcher (Chemistry), ICT Mumbai
Array BioPharma's application claimed a specific combination: encorafenib (a B-Raf inhibitor) paired with either cetuximab or erlotinib (EGFR inhibitors), and optionally a third agent, alpelisib (a PI3K-alpha inhibitor). The clinical data behind it was the kind that gets a pharmacologist's attention — the triple combination produced tumor regression of −2%, against monotherapy results ranging from 57% to 95% tumor progression on the same models. Whatever else is true about this application, the combination did something the individual agents didn't.
The Controller rejected it anyway, on three grounds stacked into one order: no inventive step, barred by Section 3(d) as a known combination, and improperly drafted as a method-of-treatment claim rather than a product claim because of language about simultaneous, separate, or sequential administration. The Delhi High Court set aside all three — and the reasoning is more interesting than the result, because the Court didn't say the Controller reached the wrong conclusions. It said the order never showed how it reached them.
On inventive step, the Controller had cited four prior art documents and asserted that a skilled person would have been motivated to combine these agents. The Court's problem wasn't the citation count, it was the missing middle step: nothing in the order identified which document was the closest prior art, articulated the specific technical problem the invention solved, or explained why a skilled artisan would have selected these three compounds in particular out of everything the prior art made available. "Skilled artisans would combine them" is a conclusion, not an obviousness analysis — and an order that skips from citation to conclusion doesn't survive review even when the citations themselves are real.
The Section 3(d) ground is the one worth sitting with longest, because it runs the opposite direction from where 3(d) usually shows up in this space. Most 3(d) fights are about whether a NEW FORM of an already-known compound — a salt, a polymorph, a deuterated analog — clears the enhanced-efficacy bar. Here the Controller used 3(d) to reject a combination of two or three already-approved, independently-known active agents. The Court drew a real boundary: Section 3(d) addresses derivatives and new forms of a known substance, not combinations of distinct active agents each with its own independent mechanism. A combination claim lives or dies on inventive step and synergy, not on 3(d) — conflating the two isn't a stricter reading of the statute, it's the wrong statute for the question.
The method-of-treatment ground came from the claim language itself describing "simultaneous, separate or sequential administration" — the Controller read that as claiming a method of treating a patient, barred under Section 3(i). The Court read the same language as describing dosing modes for a product, not narrowing the claim into a treatment method. Administration language describing how a combination product can be dosed doesn't, on its own, convert a product claim into a process claim.
Taken together, the three reversals share one structural theme: an examiner is entitled to reject a claim on any of these grounds, but the order has to show the reasoning, not just cite the outcome. Remanded, not granted — Array still has to make its case again. But it gets to make it against an order that actually engages with a triple combination that took tumors from progression to regression, instead of one that waved at four documents and called it obvious.
What we do
See if this applies to one of your matters
We run genus/species (Markush) claim coverage analysis, Section 3(d) compliance screening, and prior-art structural triage on real chemical structures — not keyword search. If a case like this one is live for you, we'll run a free sample on a compound of your choosing.