AACR In Review: IDO Pushes Ahead, CTLA-4 Combo Lags Behind

Though momentum is building for IDO/PD-1 combos, it's still early days for new immunotherapy programs. Clinicians and investors weigh in on the latest data from the American Association for Cancer Research meeting and the state of the immuno-oncology field.

Merck & Co. Inc. arguably got a leg up in its bid to lead immuno-oncology coming out of the American Association for Cancer Research meeting, between the rise of IDO as a target and new data showing disappointing efficacy for Bristol-Myers Squibb Co.'s CTLA-4/PD-1 combination in melanoma compared to PD-1 monotherapy. But development activity is intense and it's unclear where it will all end.

At this year's AACR annual meeting, held April 1-6 in Washington D.C., Bristol reported that its Opdivo/Yervoy combination was superior to the CTLA-4 inhibitor Yervoy (ipilimumab) alone in terms of overall survival (OS) in the Phase III CheckMate 067 study in metastatic melanoma. The data should support converting the accelerated approval to full FDA approval, a win for Bristol. But the data also raised questions about whether performance of Opdivo/Yervoy compared to the PD-1 inhibitor Opdivo (nivolumab) as a monotherapy are good enough to justify the high toxicities associated with the combination.

In the trial, median OS at the two-year mark was 64% for the combination versus 59% for Opdivo monotherapy and 45% for Yervoy alone. The Grade 3/4 adverse event rate for the combination regimen was 58% vs. 21% for Opdivo monotherapy and 28% for Yervoy.

"That becomes the issue when staring at the survival outcome. Is the small observed difference in overall survival worth the magnified, multiplied toxicity experience with that regimen?" Keith Flaherty, director of developmental therapeutics at Massachusetts General Hospital, said in an interview at the meeting.

Turning To Testing

Exploratory analyses suggest that the combination performed better in those with lower levels of PD-L1 expression – the combination was similar to monotherapy with a cutoff of 1% PD-L1 – and who were BRAF-mutant.

That suggests it may be possible to select patients for treatment based on subsets. PD-L1 hasn't been an important biomarker in melanoma in the past, but the '067 data have raised its profile.

Prescribers are now more likely to have PD-L1 testing done in a patient where the decision about whether to get the combination or monotherapy is borderline, Michael Atkins, deputy director of the Georgetown-Lombardi Comprehensive Cancer Center, said after the meeting. They also are more likely to consider studies with a PD-1 inhibitor and another agent, particularly for PD-L1-positive patients, said Atkins, who is also co-chair of the scientific advisory committee for the Melanoma Research Foundation.

Having a regimen that works in the notable subset of PD-L1-low patients would be valuable, but with about 315 patients per arm in the three-arm trial, the CheckMate 067 study wasn't actually powered to show a difference in overall survival in the subsets. Royal Marsden Hospital's James Larkin, the lead investigator of the trial, was reluctant to draw conclusions for patient management based on the subset figures, despite repeated questions at the meeting.

Flaherty said that he would like to see more data on how the combination works in PD-L1-low patients.

In the meantime, in the frontline setting, patients may choose PD-1 monotherapy or a clinical trial of a combination including PD-1 antibodies with novel agents instead, he suggested.

"PD-1 antibodies plus novel treatments is the most popular clinical trial category, not just for melanoma but lung cancer and many other cancer types," he commented.

Trials are in high demand and patients are very enthusiastic about pursuing them, Flaherty added.

According to the Trialtrove database, some 2,927 oncology immunotherapy trials are ongoing, of which 1,472 are combination studies. Bristol has the largest number of ongoing overall IO studies (377) followed by Merck (266), but in terms of agents, Merck's Keytruda is involved in the most studies compared to Opdivo 255 vs. 203 for Opdivo, according to this database.

Interest In IDO

The combination of inhibitors of indoleamine 2,3-dioxygenase 1 (IDO1) and PD-1 in particular had a high profile at this year's AACR meeting. NewLink Genetics Corp. released Phase II safety and efficacy data for its IDO inhibitor indoximod with Merck & Co. Inc.'s anti-PD1 Keytruda (pembrolizumab) and Bristol aired Phase I combination safety data for its IDO inhibitor BMY-986205 with Opdivo.

Both datasets suggested the IDO/PD-1 combination is well tolerated, complementing data already released for Keytruda with Incyte Corp.'s epacadostat, the most advanced IDO inhibitor in development, and whetting the appetite of investors. (Also see "IDO Emerges As Clean Combo Partner, Rising Star At AACR" - Scrip, 4 Apr, 2017.) Separate to indoximod, NewLink has another IDO inhibitor called GDC-0919 that is partnered with Roche, in Phase I.

In melanoma, Atkins notes that if nivolumab monotherapy as a frontline treatment has a 60+% survival rate in PD-L1+ patients, it will be hard for another combination to beat. Furthermore, the regimen may change to improve tolerability. In the trial, dosing for the combination included 1 mg/kg of Opdivo with 3 mg/kg Yervoy, followed by Opdivo 3 mg every two weeks.

Atkins said that the toxicity with Opdivo/Yervoy is manageable and may help prevent the need for a salvage treatment after PD-1 monotherapy for many patients.

Lessons For Lung Cancer?

Roy Herbst, a lung cancer expert and chief of medical oncology at the Yale Cancer Center, noted that there were many fewer events than expected in the CheckMate 067 study, which means it is too early to make a call about overall survival, but speaks to how well the therapies worked, both monotherapy and combination regimens.

The patient populations and dosing regimens for the combination in the crucial indication of first-line non-small cell lung cancer vs. melanoma are different so read-through from one indication to the next is limited.

Bristol's CheckMate 227 first-line NSCLC study tests 1 mg/kg Yervoy with 3 mg/kg Opdivo, the reverse of '067, and tolerability is better. PD-L1 status has also been a more significant biomarker in lung cancer.

However, speaking generally about nivolumab/ipilimumab, Howard (Jack) West, medical director of thoracic oncology at the Swedish Cancer Institute in Seattle, commented that there is reason to be dubious about the combination based on the challenging toxicity profile of the combination overall, particularly in the community setting.

"I am concerned that any tolerability data obtained in clinical trials won't necessarily be generalizable to broader populations of typically older, less fit patients who may be less amenable to greater to greater side effects than the selected patients who choose to pursue clinical trials, and these patients are still going to be treated by physicians with far less support and experience than those who participated in the clinical trials," he told Scrip.

Herbst said that as co-leader for the Lung Master Protocol (Lung MAP) trial, which tests multiple drugs, he has been impressed that the combination can be managed well in the community setting, with education of oncologists.

The '067 study in melanoma "doesn't in any way affect my thoughts as we await lung cancer data," Herbst told Scrip.

Results from AstraZeneca PLC's MYSTIC study of its CTLA-4 inhibitor tremelimumab with its PD-L1 inhibitor durvalumab in first-line lung cancer are due mid-year and Bristol's CheckMate 227 study is expected early in 2018.

West also noted recent changes to plans for the CTLA-4/PD-1 combinations. Bristol announced Jan. 19 that it was not going to seek accelerated approval for the Yervoy/Opdivo combination after all, based on “a review of data available at this time." This followed news that AstraZeneca was changing the design of its MYSTIC study to include overall survival as a co-primary endpoint along with progression-free survival and give more weight to durvalumab as a monotherapy. (Also see "AstraZeneca Updates MYSTIC Plan In Bid To Challenge Merck In NSCLC" - Scrip, 17 Jan, 2017.)

"We will need to see what the actual results in Phase III lung cancer trials demonstrated, but I don't think that we should presume that such a combination is likely to fundamentally change the treatment landscape … Even if the trial results are positive, I strongly suspect that the combination would be used selectively, likely quite sparingly, in community practice, given how many patients have a relatively compromised functional status," West said.

Meanwhile, Merck awaits a decision on a filing for accelerated approval of its PD-1 inhibitor Keytruda for use in combination with chemotherapy in first-line lung cancer. (Also see "Game Changer: Merck's Stealth Keytruda-Chemo Filing Stirs NSCLC Market Dynamics" - Scrip, 11 Jan, 2017.)

AACR 2017: 'The Year Of IDO'

Brad Loncar, CEO of Loncar Investments, commented that Yervoy has been falling out of favor – the toxicity of the combination "is the biggest bugaboo," though people have always known it was toxic. What really attracted attention was Bristol's release about the change in plans for the '227 study, which to investors implied that the company had seen some data that looked disappointing, Loncar told Scrip. AstraZeneca's decision to postpone the release of MYSTIC data from January to the middle of this year and change the trial design to have a larger monotherapy component was also concerning.

In addition to NewLink's positive Phase II data for the IDO inhibitor indoximod at AACR, both Bristol and Merck said they will be expanding collaborations to develop their PD-1 inhibitors with Incyte's epacadostat, in announcements coinciding with the meeting. Merck already had a Phase III study up and running of epacadostat in melanoma and will be running registrational studies in six additional tumor types, including two in NSCLC. Bristol will be launching two new registrational studies of Opdivo/epacadostat, one in first-line NSCLC and one in first-line head and neck cancer. This aggressive push by two leaders in immuno-oncology leads investors to believe the IDO/PD-1 combination has real legs.

At the minimum Bristol's new development plans with Incyte makes it look like they are "spreading their bets," and at the maximum it looks like PD-1/IDO is the more appealing option and the company doesn't have the luxury of time to wait around for its own IDO inhibitor to go through the whole development process, Loncar observed.

This AACR "was the year of IDO," with PD-L1/CTLA-4 falling a bit more out of favor and PD-L1/IDO emerging as the most hopeful replacement for it, Loncar suggested.

BMO Capital Markets analysts concluded that the IDO/PD-1 combination looks better tolerated than the combination of PD-1 with chemotherapy or Yervoy and that Merck gained an edge coming out of the meeting.

BMO analysts are expecting more promising data for the PD-1/IDO combination at the American Society of Clinical Oncology annual meeting in June, "which will probably provide the rationale for Merck and Bristol’s decisions to advance it to pivotal trials," Arfaei said.

Mass General's Flaherty said that the data presented at AACR reinforce the promise of the IDO/PD-1 combination, especially in terms of safety, but also acknowledged the usual hazards of drawing conclusions based on small uncontrolled trials, in that the results could be due to statistical chance. The epacadostat/Keytruda combination was moved fast to Phase III in melanoma based on data from a single-arm study of 19 patients. NewLink had single-arm data for 60 patients at AACR. It would be easier to get more excited about the efficacy results from small, uncontrolled combination trials if they focused on patients with features that make them less responsive to PD-1 inhibitors, Flaherty said.

Yale's Herbst commented that a lot of novel classes are being studied right now in combination trials – such as IDO, LAG-33, TIM-3 and CD127 – but they are all still in early stages of development. PD-1/CTLA-4 is by far at the top and there is not enough data to suggest that any of the new combinations could compete, the clinician said. There's a lot of work ahead to figure out which to move forward to big Phase III studies, he said. "The entire community is trying to figure that out right now," Herbst said.

BMO analysts advised that the IO market is rapidly evolving and will become increasingly fragmented, as evidenced by the broad range of possible IO combinations for multiple tumors.