In more ways than one, the timing could not have been worse for Genentech Inc. to report failure of a confirmatory phase III trial for its PD-1/PD-L1 inhibitor, Tecentriq (atezolizumab), in the lead indication of locally advanced or metastatic urothelial cancer (mUC), the most common form of bladder cancer. A day earlier, officials at parent company Roche Holding AG, of Basel, Switzerland, had touted the drug's market clout, crediting the launch of Tecentriq for fueling much of the company's 3 percent gain in first-quarter pharma sales. The overall survival (OS) miss in Genentech's IMvigor211 study also came on the heels of accelerated approval in the same indication for Bavencio (avelumab), the PD-L1 inhibitor developed by EMD Serono, the biopharmaceutical unit of Merck KGaA, of Darmstadt, Germany, and partner Pfizer Inc., of New York.

The data release from IMvigor211 was thin. Genentech reported only "the initial observation" that the chemotherapy arm outperformed Tecentriq in previously treated individuals with mUC. The company said findings from the trial generally were consistent with those observed in a similar group of patients in the phase II IMvigor210 study. Safety also was deemed consistent with previously observed findings about the drug.

IMvigor211 enrolled 931 participants with previously treated mUC who progressed during or following a platinum-based regimen, randomized to Tecentriq or chemotherapy (vinflunine, paclitaxel or docetaxel), administered every three weeks. The study was designed to test OS, the primary efficacy endpoint, in successive fashion in study populations defined by PD-L1 expression – first in individuals with the highest levels of PD-L1 expression, next in those with any level of PD-L1 expression and last in the intent-to-treat (ITT) population. The protocol called for the achievement of statistically significant OS in the highest PD-L1 expressers to evaluate the entire PD-L1 expresser population for statistical significance and, in turn, for that population to achieve statistically significant OS to evaluate the ITT population.

Genentech said it will present full data from IMvigor211 at an undisclosed conference later this year.

On Wednesday, shares of Roche (SWISS:ROG) fell CHF5 (US$4.96), or about 1.8 percent, to close at CHF267.60 (US$265.28).

Tecentriq was granted accelerated approval by the FDA last year, based on tumor response rate and duration of response (DOR) in IMvigor210, an open-label, multicenter, two-cohort study that evaluated the drug's safety and efficacy in people with locally advanced or mUC, regardless of PD-L1 expression. Its approval marked the first for a checkpoint inhibitor in mUC along with the first companion diagnostic, the Ventana PD-L1 (SP142) assay, to detect PD-L1 protein expression levels on tumor-infiltrating immune cells by staining and scoring the cells within the tumor microenvironment. (See BioWorld Today, May 19, 2016.)

Tecentriq was predicted by some analysts to capture the highest proportion – up to 40 percent – of estimated peak market sales of about $2.3 billion among PD-1/PD-L1 inhibitors targeting bladder cancer. "However, this news may place this assumption at risk," observed Leerink Partners LLC analyst Seamus Fernandez, noting that the "surprise" phase III miss could threaten the drug's label in second-line bladder cancer.

Last month, Tecentriq gained accelerated approval to treat first-line bladder cancer patients who are ineligible for cisplatin-based chemotherapy. The confirmatory trial (IMvigor130) for that population is ongoing, with results expected in 2019.

"We have several other ongoing studies of Tecentriq in different types of bladder cancer, including the phase III IMvigor010 study of Tecentriq as an adjuvant therapy in people with muscle-invasive bladder cancer and a phase Ib study of Tecentriq in non-muscle invasive bladder cancer," Genentech spokeswoman Austine Graff told BioWorld Today. Data from IMvigor211 "will be further examined in an effort to better understand these results" in light of the ongoing trials.

'WE CONTINUE TO SEE ANTI-PD-L1S HAVING LIMITED UTILITY'

Fernandez expects Keytruda (pembrolizumab), the anti-PD-1 from Merck and Co. Inc., of Kenilworth, N.J., to capture 30 percent of peak bladder cancer sales, followed by the anti-PD-1 Opdivo (nivolumab) from New York-based Bristol-Myers Squibb Co. (BMS) at 20 percent.

The increasingly crowded mUC market also includes Imfinzi (durvalumab), the breakthrough-designated PD-L1 inhibitor from London-based Astrazeneca plc that won accelerated approval earlier this month. (See BioWorld Today, May 2, 2017.)

Fernandez suggested Imfinzi could capture 10 percent of peak bladder cancer sales, with latecomer Bavencio potentially gaining a modest share of sales, as well.

But the race remains too close to call, as currently marketed PD-1/PD-L1 inhibitors all have accelerated approval in second-line bladder cancer. Fernandez predicted that Keytruda will gain full approval on or before its June 14 PDUFA date, based on data from the confirmatory phase III KEYNOTE-045 trial.

Although Tecentriq is the only agent in the class with an accelerated indication for first-line mUC patients who are not eligible for cisplatin-containing therapy, Keytruda also may gain accelerated approval in the setting by its PDUFA date, according to Fernandez.

Like most early checkpoint inhibitors, Tecentriq is part of an extensive development program that includes 17 ongoing or planned phase III studies. In addition to bladder cancer, they include kidney, skin, breast, colorectal, prostate, ovarian, blood and several kinds of lung cancers, alone and in combination with other agents. In addition to bladder cancer, the therapy has accelerated approval in non-small-cell lung cancer (NSCLC).

Roche, which had reported Tecentriq sales of $158.6 million for full-year 2016, posted first-quarter Tecentriq sales of CHF109 million (US$108.1 million), up from CHF78 million (US$77.3 million) in the fourth quarter of 2016, with growth driven mainly by the uptake of Tecentriq in second-line bladder cancer and second- and third-line lung cancer.

On the company's earnings call, Daniel O'Day, chief commercial officer, credited bladder cancer for 70 percent of Tecentriq's quarterly growth, maintaining that the drug "has established itself as a standard of care" in the indication.

Analysts were less convinced. Given the failure of IMvigor211 to show an OS benefit, Evercore ISI analyst Umer Raffat agreed that Tecentriq's accelerated approval in mUC could be at risk. Noting in an email that BMS and Pfizer also received approval for their immuno-oncology (I-O) drugs as monotherapies in mUC on an accelerated basis, "Is there a broader risk to I-O monotherapy in this indication?" he asked. Although data on Opdivo showed a slightly higher overall response rate (ORR) in mUC than Pfizer and Roche, "we also acknowledge that we do NOT have full visibility into duration of response," Raffat wrote, "and it's still unclear how ORR/DOR correlates with survival."

And Raffat asked hypothetically whether Keytruda may, indeed, be the better drug, pointing out that "the big news" in I-O last year was that the monotherapy trial of Keytruda in first-line lung cancer succeeded while the CHECKMATE -026 experiment by BMS in Opdivo did not. At the time, analysts were stunned by the -026 failure, with some blaming an aggressive trial design. (See BioWorld Today, Aug. 8, 2016.)

But "we now have a second cancer indication where MRK's Keytruda worked but a competitor PD-1/PD-L1 didn't," Raffat wrote in an email, noting that the KEYNOTE -045 trial hit an OS benefit with a hazard ratio of 0.73. "This is one of those data points that makes you pause for a second," he added.

The trial designs appeared similar, according to Raffat, who said the difference in outcomes between the Merck and Roche trials couldn't be attributed to PD-L1 expression "because although we know [the] Merck trial outperformed in high PD-L1 expressers, Roche's primary analysis in this IMvigor211 was on highest PD-L1 expression – and it didn't work."

Genentech still needs to answer important questions about its matinee I-O drug, according to Raffat, who said he's "curious" to learn what percentage of chemo patients enrolled in IMvigor211 received the study drug upon crossover – the proportion was 12 percent in the Merck trial, he wrote. That data point is important, he added, "because Merck didn't show [progression-free survival] PFS benefit either . . . it was all about OS."

Additionally, Raffat wondered about the percentage of patients in IMvigor211 who were metastatic, noting the hazard ratio on survival was lower in the second-line metastatic subgroup in Merck's KEYNOTE study.

In an extensive look at the I-O space following disclosure of the Tecentriq OS findings, independent analyst Amit Roy of Foveal LLP made similar observations. He called the Tecentriq outcome in bladder cancer "a predictable failure," noting that "in contrast, anti-PD-1 Keytruda has shown a 27 percent overall survival benefit (p=0.002) vs. the same chemotherapy in its phase III study KEYNOTE-045 in the same line of bladder cancer."

The findings, he added, provide more evidence that anti-PD-L1s are consistently weaker than PD-1s.

"Bladder cancer represents the third line of cancer showing lower efficacy data for all anti-PD-L1s compared to anti-PD-1s, with a similar pattern in first- and second-line NSCLC," Roy wrote. While the cause of that difference remains unknown, "with weaker data and later entry and no differentiation on safety, we continue to see anti-PD-L1s having limited utility."

Tellingly, late Wednesday Merck received FDA accelerated approval for the combination of Keytruda and chemotherapy to treat first-line metastatic nonsquamous NSCLC, irrespective of PD-L1 expression. The company said the expanded indication was approved based on tumor response rate and PFS.