By: Len Kusdra, Analyst, Clinical & Scientific Assessment, Kantar Health, Tatiana Spicakova, Consultant, Clinical & Scientific Assessment, Kantar Health, Greg Wolfe, Senior Consultant, Clinical & Scientific Assessment, Kantar Health and Stephanie Hawthorne, Senior Director, Clinical & Scientific Assessment, Kantar Health
The role of Programmed Death-1 (PD-1) pathway in suppression of antitumor immunity has become one of the hottest topics in oncology over the past couple of years. PD-1 is a key immune checkpoint receptor expressed on activated T-cells, and binding of PD-1 to its ligand (PD-L1) results in the suppression of the immune response. While the PD-1 pathway normally plays a protective role by attenuating immune-mediated destruction of healthy tissue, the pathway can be exploited by cancer cells to protect themselves from the attack by tumor-specific T-cells. A number of immune checkpoint inhibitors are in late-stage clinical development; these agents work by blocking the interaction between PD-1 and PD-L1, thereby activating the immune system against cancer cells.
PD-1 Pathway Inhibitors in Phase III Development
|Drug||Manufacturer||Antibody||Target||Development Stage||Ongoing pivotal trials|
|Keytruda||Merck||Humanized IgG4||PD-1||Approved in Yervoy-treated melanoma (U.S.); Phase III||Melanoma, NSCLC, Head and Neck|
|Opdivo||BMS||Human IgG4||PD-1||Filed in Yervoy-treated melanoma (US/EU)
Filed in squamous 3rd line NSCLC (US/EU); Phase III
|Melanoma, NSCLC, RCC, Head and Neck|
|MPDL3280A||Roche||Human IgG1||PD-L1||Phase III||Bladder, NSCLC|
|MEDI-4736||AstraZeneca||Human IgG1||PD-L1||Phase III||NSCLC|
For these novel immune checkpoint inhibitors, the fight to capture the lead in high-profile tumor types such as melanoma and non-small cell lung cancer (NSCLC) is becoming increasingly fierce. But their clinical development does not end there and, in fact, continues to expand into other indications, such as renal cell carcinoma (RCC), urothelial cancer, head and neck cancer, gastrointestinal cancer and hematological malignancies. Data continued to report promising activity across different tumor types at the annual 2014 European Society for Medical Oncology (ESMO), as summarized below.
Opdivo improves response rate versus chemotherapy in Yervoy-treated melanoma in a Phase III randomized trial – is its efficacy superior to the already approved Keytruda?
(Weber, Abstract LBA3)
Past oncology meetings have highlighted the promise of Opdivo™ (nivolumab, Bristol-Myers Squibb) in melanoma. Opdivo demonstrated impressive early clinical data that suggested improved overall response rate (ORR) with fewer toxicities when compared with historical data for Yervoy® (ipilimumab, Bristol-Myers Squibb), a CTLA-4 immune checkpoint inhibitor. At ASCO 2014, another anti-PD-1 monoclonal antibody, Keytruda® (pembrolizumab, Merck), took center stage with impressive Phase II data that ultimately led to its approval in the U.S. in September 2014 as therapy for Yervoy-treated melanoma patients.
Recognizing the urgency of getting to market as soon as possible, BMS submitted U.S./EU regulatory applications for Opdivo for Yervoy-treated melanoma, with expected U.S. approval in March 2015. Once Opdivo becomes approved, physicians will look to efficacy and safety data to drive their decision when choosing between the two agents. At ESMO, highly anticipated data from a randomized Phase III study (Opdivo versus chemotherapy of choice) were presented for 405 previously treated (including Yervoy) unresectable Stage III/IV melanoma patients.
The ORR was 32% for Opdivo versus 11% for chemotherapy, with benefit observed across all subgroups, including PD-L1 negative patients. Grade 3/4 toxicities were reported in 9% of patients in the Opdivo arm versus 31% in the chemotherapy arm, with no new safety signals reported. Unfortunately, the progression-free survival (PFS) and overall survival (OS) data were not yet available at the time of analysis, but duration of response among responding patients suggests prolonged benefit (3.6 months in the chemotherapy arm versus median not reached in the Opdivo arm but with 95% of responders still benefiting at six months). So how does Opdivo compare to Keytruda in this patient population? As shown below, the activity appears very similar between the drugs, and no dramatic differences exist in their safety profiles.
Efficacy in Previously-Treated Metastatic Melanoma
|Keytruda (n=197)1||28%||2%||5.6 mos|
1Ribas, Abstract LBA9000, ASCO 2014
While response rate is a valid and informative endpoint about the drug’s efficacy, PFS and OS data ultimately might help decide which of the two will become the winner. This is especially important for Opdivo’s trial, in which OS was the co-primary endpoint. Will the improved response rates translate into a statistically significant OS benefit that supports the regulatory applications? The expected answer is yes, although due to immature data we now continue to wait with bated breath for this data. While Keytruda already demonstrated that PD-1 inhibitor works in Yervoy-treated melanoma patients, the results from the Opdivo trial were significant in that it presented, for the first time, data from a randomized Phase III study. Commercially, Keytruda will have a six-month time-to-market advantage over Opdivo and will likely amass a strong foothold in the Yervoy-pretreated setting by the time Opdivo launches. Will having overall survival data from a randomized study (assuming it is available by the time of launch) sway physicians to forgo Keytruda in favor of Opdivo, or will physicians interpret a positive survival benefit for Opdivo to be a surrogate for similar expectations with Keytruda?
Is Keytruda leading the PD-1 pack in NSCLC?
(Garon, Abstract LBA43)
Safety and efficacy data for Keytruda were reported for previously treated and treatment-naive NSCLC patients enrolled in Phase I study expansion cohorts. In 262 patients, treatment-related adverse events included (any grade; Grade 3/4) fatigue (20%; <1%), pruritus (9%; 0%) and pneumonitis (4%, 2%). In 236 evaluable patients, the ORR was 21% (26% in treatment-naïve and 20% in previously treated patients). Curiously, current/former smokers appeared to achieve a better response compared to never smokers (27% vs 9%), although the underlying reasons for this observation remain unknown. Median PFS and OS also appeared better for treatment-naïve versus previously treated patients (mPFS: 27 weeks vs. 10 weeks; mOS: not reached vs. 8.2 months).
The study evaluated whether the expression of PD-L1 biomarker correlates with outcomes. Strong PD-L1 expression was defined as 50% or higher membrane staining in tumor cells and weak expression as 1-49%. The ORR correlated with the level of PD-L1 expression (37% for strong positive, 17% for weak positive and 10% for negative). PFS was also longer for patients with strong versus weak PD-L1 expression (HR 0.52), as was OS (HR 0.59). While the efficacy certainly appears better for PD-L1 positive patients, it is not a black-and-white scenario, and many unresolved issues remain. Namely, differences exist in how the PD-L1 biomarker is measured in the different assays that the various manufactures are developing (what antibody is used), how tissue is collected (old, frozen tissue versus fresh biopsy), where it is measured (tumor cells versus tumor-infiltrating immune cells) and which cutoff threshold is used (1% versus 5% versus others). More importantly, a subset of PD-L1 negative patients still derive a benefit from these inhibitors, yet many of the pivotal trials in NSCLC are conducted in PD-L1 positive tumors. How will that affect the real-world patients, and will the biomarker-negative patients be denied treatment if the drugs receive PD-L1 positive labels (especially outside of the U.S., where payers tend to impose stricter rules)? Keytruda is exclusively targeting PD-L1 positive patients in its pivotal NSCLC trials, while BMS chose to target both PD-L1 positive patients as well as all-comers (including squamous and non-squamous histologies). If those trials are positive, perhaps the inclusion of these patients will enable BMS to have a competitive marketing edge over Keytruda.
While the data presented for Keytruda was certainly very encouraging, how does it compare with the other key competitors that are also vying to win the NSCLC space? At a first glance, it appears that the agents have similar levels of activity in terms of response rates and survival. But the data are still premature, and the efficacy and safety could start differentiating as more patients are enrolled and longer follow-up becomes available. For now, there is no clear winner in NSCLC, and it has yet to be determined whether targeting the receptor versus ligand may prove to be a more efficacious or safer strategy.
|NSCLC ORR (pretreated)||mPFS||mOS|
|MPDL3280A1||23% (n=53)||ND||23% (n=53)||ND||ND|
|MEDI-47362||16% (n=58)||ND||16% (n=58)||ND||ND|
|Opdivo3||17% (n=129)||30% (n=20)||17% (n=129)||2.3 mos||9.9 mos|
|Keytruda4||21% (n=236)||26% (n=42)||20% (n=194)||2.5 mos||8.2 mos|
1Soria, 1322P, ESMO 2014
2Brahmer, Abstract 8021, ASCO 2014
3Gettinger, Abstract 8024, ASCO 2014
4Garon, LBA43, ESMO 2014
In terms of approval timelines in NSCLC, BMS has an advantage over the other agents as it already submitted a regulatory application in the EU and a rolling submission in the U.S. for Opdivo as a third-line therapy in squamous NSCLC. The submissions were based on data from a Phase II study, and the drug could be approved in both regions next year.
Keytruda shows promising activity in gastric and urothelial cancers
Like BMS, Merck is also pursuing an aggressive development strategy for Keytruda that spans multiple tumor types. Promising preliminary results were reported at ESMO for the gastric cancer (Muro, Abstract LBA15) and urothelial cancer (Plimack, Abstract LBA23) cohorts from the KEYNOTE-012 Phase Ib study. In the gastric cancer presentation, 162 patients with recurrent or metastatic adenocarcinoma of the stomach or gastroesophageal junction (GEJ) were screened to identify 65 (40%) patients with PD-L1 positive tumors, who were then treated with 10 mg/kg Keytruda every two weeks. Keytruda showed an acceptable safety profile in 39 evaluable patients, although there was one incidence each of Grade 4 pneumonitis and Grade 5 (fatal) hypoxia. Keytruda achieved 30.8% ORR (no complete responses) and reduction in tumor size in 41% of patients. Efficacy was similar in Asian and non-Asian patients, and responses were durable. Based on these results, initiation of a Phase II trial in advanced gastric cancer is expected in in the first quarter of 2015.
In the urothelial cancer cohort, 95 patients with recurrent or metastatic cancer of the renal pelvis, ureter, bladder or urethra were screened to identify 61 (64.2%) patients with PD-L1 positive tumors. Thirty-three patients with PD-L1 positive tumors were treated with 10 mg/kg Keytruda every two weeks. Grade 3 or higher adverse events were observed in four patients and included AST increase, dehydration, neuromyopathy, macropapular rash, pruritic rash, rhabdomyolysis, thrombocytopenia, and toxic encephalopathy (n=1 for each, some patients had multiple Grade ≥3 adverse events). The ORR was 24.1% (including three complete responses), and a reduction in tumor size was observed in 64% of patients. Six of the seven responses were ongoing, and median duration of response was not reached at a median follow-up of 11 months. These promising results served as the impetus for the Phase III KEYNOTE-045 trial scheduled to begin by the end of 2014.
MPDL3280A shows promising activity in monotherapy activity in bladder, and activity in RCC in combination with Avastin, but will it succeed in CRC?
Roche/Genentech was awarded breakthrough therapy status for its MPDL3280A in urothelial bladder cancer (UBC) in May 2014. Unlike Opdivo and Keytruda, this antibody targets the ligand (PD-L1) rather than receptor, but it is too early to tell whether this strategy will translate into better efficacy and/or a better safety profile. Perhaps recognizing the increasing crowdedness and being too late to the melanoma market, Roche’s clinical development strategy has focused on other tumor types, including NSCLC and RCC as well as the less commercially attractive (in the eyes of big pharma) urothelial/bladder space. Its first potential approved indication is likely to be bladder, in which they could file for an accelerated approval based on data from the ongoing Phase II study, thus potentially beating Keytruda to the market in the urothelial space. It would certainly give physicians the chance to become familiar with the agent in the real-world setting so that by the time the agent is approved (if successful) in larger indications, it may ease the adoption process by clinicians; being approved on the market could also aid in off-label utilization following positive Phase III data in other tumor types, bypassing a delay for supplemental regulatory approval. In terms of its activity in metastatic RCC as monotherapy, data were presented from a Phase I expansion cohort that included 69 patients (McDermott, Abstract 809O). MPDL3280A was very well-tolerated in these patients with Grade 3 or higher fatigue reported in two patients; all other adverse events were Grade 1/2. In 62 evaluable patients, the ORR was 20% in patients with PD-L1 expression of IHC1+/2+/3+ (including one complete and six partial responses) and 10% in patients who were PD-L1 negative (IHC0).
While the inhibitors have been primarily tested as monotherapy, the next wave of trials will examine their efficacy in combination with other regimens, such as chemotherapy, targeted therapy and other immune modulators. Not surprisingly, Roche has already begun the clinical development of MPDL3280A in combination with Avastin® (bevacizumab) and reported preliminary data from a Phase Ib study (Lieu, Abstract 1049O). The study had two arms: Arm A (n=35) evaluated MPDL3280A IV every three weeks plus Avastin 15mg/kg IV every three weeks; Arm B (n=36) evaluated MPDL3280A IV every two weeks plus Avastin 10mg/kg every two weeks and FOLFOX. Dose expansion cohorts in Arm A included patients with colorectal cancer (CRC) and other solid tumors, including breast, melanoma, NSCLC and RCC; Arm B included oxaliplatin-naïve CRC (with or without liver lesions) as well as other solid tumors, including RCC and breast cancer. In Arm A, the combination with Avastin achieved 40% ORR in the first-line RCC cohort (n=10) and 8% ORR in CRC patients. In Arm B, the combination with Avastin plus chemotherapy achieved 36% ORR in CRC.
Avastin plus chemotherapy has been a blockbuster regimen in the treatment of CRC, so it is not surprising that Roche is pushing the combination with its PD-L1 inhibitor. Notably, initial data from Opdivo’s Phase I trial did not show activity in CRC (or prostate cancer), so CRC has not been on the forefront of pivotal studies for other PD-1 inhibitors. While the 36% ORR is certainly encouraging, one must question how much of it was contributed by Avastin plus chemotherapy alone. The response rate appears similar to that observed in pivotal trials for Avastin plus chemotherapy, so the early readout seemed perhaps a bit underwhelming.
But the response rate may not be the best endpoint to evaluate the efficacy of the combination regimen as immunotherapy does not always follow the standard response kinetics and can be associated with delayed responses. So it is still possible that longer follow-up might show improved survival, but will it produce truly stellar results as we have seen so far for the combination of two checkpoint inhibitors in melanoma (Sznol, Abstract LBA9003, ASCO 2014)? In the absence of randomized data, it is difficult to say whether this combination will pan out in CRC. The data in RCC, on the other hand, appear much more promising for the Avastin combination, although the number of patients was small. Avastin has shown efficacy as monotherapy in frontline RCC by achieving approximately 13% ORR, so the reported 40% ORR when combined with PD-L1 inhibitor provides encouraging evidence that the improved efficacy is driven by the addition of MPDL3280A. The activity of the combination regimen certainly appears higher compared to MPDL3280A alone, as discussed above, and might be the strategy going forward, although no such plans have been announced yet. It certainly would help Avastin garner greater utilization in RCC, where it is approved but not a key competitor. However, MPDL3280A plus Avastin will find itself competing with Opdivo, which is currently being studied in combination with Yervoy as first-line therapy for RCC in a head-to-head trial versus Sutent® (sunitinib, Pfizer).
The amount of data presented at ESMO on the role of immune checkpoint inhibitors was impressive and quite overwhelming. It has become very clear that PD-1 targeted drugs are considered a major breakthrough in oncology with the promise to make a meaningful impact in the lives of patients with various malignancies. To date, data has focused on their activity in solid tumors, but we expect to see the first sets of data for these agents in hemtatologic malignancies at the upcoming American Society of Hematology (ASH) conference. It is the beginning of exciting times in oncology and we can all sit back and enjoy the fight for the PD-1 space spanning multiple tumors.