Novartis posts landmark CAR-T data as head-to-head rivalry with Kite Pharma looms
Vasant Narasimhan, chief medical officer of Novartis AG, looks on during a January 2017 news conference in Basel Michele Limina/Bloomberg
Novartis’ pioneering CAR-T drug CTL019 (tisagenlecleucel) scored high in treating diffuse large B-cell lymphoma (DLBCL), spotlighting numbers that will likely allow the pharma giant to plunge head-to-head in the market against its main initial rival, Kite Pharma. But Novartis also found itself defending its manufacturing operations Wednesday morning, as analysts zeroed in on a number of patients who couldn’t be treated in the study.
The figures — drawn from its eagerly anticipated Phase II JULIET study — that will attract everyone’s attention first: At three months the overall response rate hit 45% among 23 of 51 patients evaluated with an impressive 37% achieving a complete response and 8% achieving a partial response. And those CRs continued beyond the three months to the data cutoff in prepping for the announcement, offering a glimpse of durability that will help make Novartis’ case.
“For us as a medical company and from my organization’s standpoint, we want to have a dramatic impact,” Novartis $NVS chief medical officer Vas Narasimhan tells me. “This is, of course, transformative… It gives us the conviction to move ahead.”
Narasimhan in particular highlighted early evidence of persistence for this therapy, which will be crucial to differentiating their drug from all the CAR-T therapies to come, from any angle.
The best reason to suggest why CTL019 could be a standout, he says, points to their drug’s use of the 4-1BB costimulatory domain, distinguished from the CD28 domain used by Kite’s rival drug and Juno’s initial lead therapy, now scrapped in favor of a drug that uses 4-1BB. That 4-1BB domain may explain why the initial ramp up of T cell expression is slightly slower, adding to longterm persistence. And it may also explain Novartis’ relatively clean safety profile, with no major issues (so far) related to neurotoxicity, which will likely be given careful scrutiny at the FDA.
All that, though, has yet to be determined.
Purists hate cross-trial comparisons, noting hard-to-score comparisons in the patients studied. But Wall Street’s first response will be to line up the data from the pharma giant against the numbers for Kite’s KTE-C19 (axicabtagene ciloleucel) at 3 months. Their drug hit an ORR of 39% among 51 patients, with 33% achieving a complete response — comparable, though, marginally less significant results that will leave these two companies battling it out.
Narasimhan declined to touch my question about comparing numbers in the pivotal studies, which was to be expected. But Twitter was abuzz this morning.
— Brad Loncar (@bradloncar) June 7, 2017
As expected, cytokine release syndrome was common, with 57% experiencing CRS — although no one died from it. No deaths were attributed to the drug.
Kite’s numbers have matured beyond that to the 6-month mark, slipping slightly but remaining strongly competitive, as Kite CEO Arie Belldegrun predicted it would. And after the JULIET data were initially released in an abstract early Wednesday, circulating on Twitter, Kite’s shares jumped 6% after investors got a chance to evaluate how the two drugs stacked up.
Novartis will get the first crack at an approval for children with acute lymphoblastic leukemia, with an FDA expert panel slated for July 12. And with a priority review, the agency is committed to a quick response with a decision due later this year. Kite is about two months behind with a PDUFA date of November 29.
Another area that Novartis, Kite and any future rivals will compete on is manufacturing time.
Says Narasimhan: “Vein-to-vein time (the stretch from extracting cells to getting them back in as a therapy) we expect to have in the range of 20 days. We were slower in the clinical trials,” with the manufacturing arm bringing that down from 29 days.
Cowen got the discussion about manufacturing started Wednesday morning by noting the abstract’s figures on patients in the study who never made it to the infusion point. Novartis responded with a statement, noting that they had factored in a high dropout rate due to the very advanced stage of an aggressive disease and the rapid deterioration you would expect in their status.
Analysts also noted that a number of patients never were treated because of manufacturing problems — a potentially serious snafu. But Novartis says it is getting a better handle on that and downplayed the issue, adding:
Only 6% (9 of 141) of enrolled patients were discontinued due to inability to manufacture an adequate dose of CART cells. Over the course of JULIET, with continuous process improvements, manufacturing success rate improved to 97% for the last 30 patients.
In JULIET, patients could receive bridging chemotherapy, reflecting the aggressive nature of disease in patients with r/r DLBCL, so even patients in poor condition could be enrolled and infused. The cryopreserved leukapheresis used in JULIET gives physicians the flexibility to schedule apheresis at a time that is in the best interest of their patients, including times far in advance of manufacturing. The trial design and population infused reflects the real world challenges in treating patients with an aggressive cancer such as r/r DLBCL.
As for manufacturing, we are confident we will be able to meet the required manufacturing demands moving forward.
The actual cell processing time is 10-12 days. We anticipate that the time from manufacture start to product release (including Quality assessments) will be 22 days at the time of commercial launch. We have a rigorous quality assessment process to ensure a GMP compliant product is released to patients.
Significantly, Kite CMO David Chang told me at ASCO that they have vein-to-vein down to about 17 days, and would still like to shave a few days off that.
Notably, Novartis’ manufacturing chief for CAR-T recently resigned to take a similar job at Seattle Genetics, raising concerns within the pharma giant’s R&D operations that the continuing exodus of execs out of Novartis — particularly following last summer’s reorganization — could hamper its performance.
Novartis execs, though, have insisted that the multinational company is completely committed to making this all work commercially, noting their lead role in the field.
Analysts have been debating for months now how the competition will shake out. The lead CAR-T at Juno imploded last year after it killed five patients due to cerebral edema. Then Kite rattled the market recently with the news that one of its patients also died from cerebral edema. Some have theorized that the added safety threat could be due to the different costimulatory domains used in the drug, with Kite and Juno’s first lead using CD28 and Novartis focused on 4-1BB. It’s never been established, though, as researchers continue to gain new insights on the best way to design these CAR-Ts.
Novartis has made CTL019 one of its top oncology programs, pushing hard to gain first-market advantage for a new kind of cancer treatment that reengineers patients cells into a therapy. Its best early promise lies in liquid tumors with investigators working to tackle a broad slate of these cancers. Next stop: finding better ways to go after solid tumors with this technology. Behind that, Cellectis and others will try to develop off-the-shelf therapies that can be directly administered to patients.
Narasimhan also mentioned that the pharma giant has been working on freezing cells for later use, perhaps preparing therapies that can be used at a later stage, if a patient needs them to battle drug resistance. How that sort of operation could be squared with payers, though, he freely admits will be a complex challenge.
This first round of data marks the cutting edge of the competition. There’s much, much more to come.