Novartis's Ribociclib Patent Survived Four Oppositions - The Chemistry Argument That Actually Held Up

Indian Patent 594997 was filed in 2011 and granted in 2026 - fifteen years, four pre-grant oppositions, and a 198-page Controller's order to get there. The compound is ribociclib, sold as Kisqali, a CDK4/6 inhibitor used in HR+/HER2- breast cancer. Natco Pharma and three individual opponents fought the grant. We looked at the actual chemistry each side argued, not just who won.

The bioisosterism argument, and why it didn't hold up

Opponents pointed to two prior-art compounds from an earlier Novartis application (WO'222) and said ribociclib was an obvious tweak of each: Example 338 differs by a phenyl-for-pyridyl swap, Example 392 differs by the absence of one C5-methyl group. Structurally, both are true. Chemically, "isosteric" doesn't mean "interchangeable" - swapping phenyl for pyridyl changes electronics, hydrogen-bonding capacity, and binding-pocket fit in ways that aren't reliably predictable without actually testing the analog. The Controller adopted the same standard the EPO uses: bioisosteric replacement is an empirical rule that needs case-by-case experimental verification, not a shortcut to "obvious." On the chemistry, that's the right call. A same-atom-count substitution is not the same claim as a same-activity substitution, and treating them as equivalent is exactly the kind of hindsight reasoning patent law is supposed to guard against.

Novartis's own earlier patent, IN 283133 (priority 2006), carried a broad Markush genus claim wide enough to cover ribociclib's structure. Opponents argued this meant Novartis had already claimed the compound once. The Controller disagreed: a genus claim and an individually exemplified species aren't "literally identical," so no prior-claiming. That's settled law, and narrowly correct here. But it's worth naming the pattern plainly: file a broad genus first, then patent a specific, better-performing species from inside that same genus years later, and you get two bites at exclusivity from one underlying discovery. Whether ribociclib's own facts cross that line is a separate question from whether the pattern itself deserves more scrutiny across the industry, and we think it does.

The Section 3(d) question, where we land

That distinction matters chemically, not just legally. CDK1 is essential to cell division in every dividing cell in the body, not just tumor cells. Poor selectivity against it is a direct route to the toxicity (myelosuppression, in practice) that limits how much drug a patient can actually tolerate. A compound that hits its target roughly 30 times more selectively, even at somewhat lower raw potency, can have a meaningfully wider therapeutic window in real patients. That's a different claim than the one Section 3(d) was written to stop. The Novartis v. Union of India Glivec case rejected "enhanced efficacy" built on bioavailability alone, with no shown clinical benefit, a real distinction from a novel drug candidate. Selectivity-driven reduction in off-target toxicity is a mechanistically grounded, clinically meaningful property, not a formulation trick. On the chemistry, we think the Controller's efficacy finding holds up.

What we're not weighing in on