By Arnold DuBell, Ph.D., M.B.A., Senior Consultant, Clinical & Scientific Assessment, Kantar Health and Stephanie Hawthorne, Ph.D., Vice President, Clinical & Scientific Assessment, Kantar Health
Immunotherapies have justifiably generated a lot of buzz in 2016. For lung cancer, much has changed in the past year. U.S. and European agencies approved Keytruda® (pembrolizumab, MSD) for second-line therapy in all histologies of PD-L1+ patients as well as Opdivo® (nivolumab, Bristol-Myers Squibb/Ono Pharmaceuticals) for use in second-line patients with all histologies (staggered launches by histology; note that this approval does not require expression of the biomarker PD-L1). Japan’s Ministry of Health, Labour and Welfare (MHLW) took a broader approach for Opdivo and granted the agent approval for the treatment of unresectable, advanced and recurrent non-small cell lung cancer (NSCLC) patients.
At the 2016 European Society for Medical Oncology (ESMO) Congress, results of two trials in Sunday’s Presidential Symposium highlighted the continuing battle between these two PD-1 inhibitors as they move to the frontline setting. First up was what may be at least the temporary victor. The Phase III KEYNOTE-024 trial1 randomized 305 previously untreated Stage IV NSCLC patients with a PD-L1 TPS2 ≥ 50% to Keytruda (200 mg IV Q3W for two years) versus four to six cycles of platinum-doublet chemotherapy. A variety of platinum-doublet options were offered to physicians but were dependent upon tumor histology. If the patient had non-squamous histology and was offered Alimta® (pemetrexed, Eli Lilly) plus a platinum agent combination as a control, that patient could also be administered Alimta maintenance therapy. Patients who progressed after being treated with the control arm were allowed to cross over to receive Keytruda at the same dosing as described above.
KEYNOTE-024 met its primary progression-free survial (PFS) endpoint, improving PFS by 4.3 months (10.3 months versus 6.0 months, HR 0.50, p<0.001). Of real note, since these data were an interim analysis, overall survival (OS) was also significantly improved in the Keytruda arm (medians not reached on either arm, HR 0.60, p=0.005). That this trial had met both of these endpoints had been known by the audience, as MSD announced the trial achieved its primary endpoint in a June 2016 press release, but this presentation was the first time the oncology community saw the magnitude of benefit. It was on the basis of an OS benefit at the interim that KEYNOTE-024 was stopped early by the trial’s data monitoring committee in June 2016. The overall response rate was also significantly improved in the Keytruda arm (45% versus 28%, p=0.0011).
Patients were treated for a longer time with Keytruda than chemotherapy (7.0 months versus 3.5 months). In spite of this longer exposure, Keytruda was better tolerated than chemotherapy (all grades: 73% versus 90%; Grade 3-4 adverse events: 26% versus 51%). Toxicities of note were nausea (10% versus 43%), anemia (5% versus 45%), fatigue (10% versus 27%), loss of appetite (10% versus 25%), vomiting (2% versus 23%) and neutropenia (1% versus 25%), which were all significantly decreased in the Keytruda arm; only pyrexia (Grades 1/2: 10% versus 4%) and diarrhea (all Grades: 14% versus 13%) were observed more frequently in the Keytruda arm compared with the chemotherapy arm.
At the end of this data presentation, the audience gave the presenter, Dr. Reck, a long and sustained ovation. This could not be said for Dr. Socinski, who next presented the results of the Phase III CheckMate 026 trial3 and received the usual polite ovation at his presentation’s end. CheckMate 026 randomized 541 patients to Opdivo 3 mg/kg Q2W versus histology-dependent chemotherapy. Maintenance Alimta was also allowed, as was crossover to Opdivo following progression on the control arm. Unlike KEYNOTE-024, patients in CheckMate 026 were only required to have tumors with PD-L1 expression ≥1%.
Unfortunately, CheckMate 026 did not meet its primary PFS (based on immune-related response criteria, IRRC) endpoint in patients whose tumors expressed PD-L1 at ≥5% (4.2 months versus 5.9 months, HR 1.15, p=0.2511). Interestingly, the PFS curves separated early (with the chemotherapy arm performing better than the Opdivo arm) but then later converged on each other after about seven months post-randomization, perhaps speaking to late and/or sustained effects of immunotherapy. Moreover, the OS in patients whose tumors expressed PD-L1 at ≥5% was not improved (14.4 months versus 13.2 months, HR 1.02), and the overall response rate was lower in these select patients treated with Opdivo (26.1% versus 33.5%).
The toxicity profile was improved with Opdivo, as there were fewer adverse events of any grade (71.2% versus 92.4%) and fewer Grade 3-4 adverse events (17.6% versus 50.6%). Common toxicities of all grades included fatigue (21.0% versus 35.4%), diarrhea (13.9% versus 12.9%), decreased appetite (12.0% versus 27.8%) and nausea (11.6% versus 48.3%).
The discussant to CheckMate 026, Dr. Rizvi, didn’t really have an explanation for why KEYNOTE-024 succeeded while CheckMate 026 didn’t. Although it may be that the molecules are ultimately just different – which is hard to justify given the preponderance of data in relapsed lung cancer and other tumors suggesting that in these other settings the efficacy benefits are similar – the best explanation lies in the design of the trial for CheckMate 026. It’s not the PD-L1 threshold difference, as subgroup analysis from CheckMate 026 suggested that patients whose tumors expressed PD-L1 at levels ≥ 50% also did not show significantly improved PFS or OS benefits. However, Dr. Rizvi did suggest some confusing issues in the trial design. First, he thought it very unusual that about 38% of patients had received prior radiation before enrolling in the trial. Moreover, he also thought that the time from diagnosis to initial treatment (two months) was rather long. Minor differences between the two trials were in the number of patients with squamous histology (19% in KEYNOTE-024 and 24% in CheckMate 026), slightly different choice of platinum doublets in the control arms of each trial (65% received Alimta-based in KEYNOTE-024 and 76% received Alimta-based in CheckMate 026), and subtle differences in the definition of the primary PFS endpoint (RECIST criteria for KEYNOTE-024, IRRC for CheckMate 026).
Based on these data, it is assumed that Keytruda will earn approval in the first-line setting and Opdivo will not. MSD has already filed with appropriate regulatory agencies for approval based on KEYNOTE-024. One unanswered question that was not addressed by these trials was whether physicians might sequence the agents – that is, use first-line Keytruda then second-line Opdivo. Unfortunately, although both KEYNOTE-024 and CheckMate 026 allowed use of checkpoint inhibitors (either as crossover or as a true second-line therapy) for patients on the control arm, neither trial reported more than a handful of patients on the respective PD-1 inhibitor arm received checkpoint inhibitor therapy again in later lines. Therefore, all that can be surmised is that until new agents are approved, Keytruda will be used first-line in patients with PD-L1 expression ≥ 50%. The patients with expression less than that threshold will likely be offered a PD-1 inhibitor in the second-line setting; based on current trends, the preferred second-line PD-1 inhibitor is Opdivo,4 although it remains to be seen if the divergent results of these two first-line trials will lead to a shift in standards in second-line.
Other immunotherapies have recently entered the fray. The PD-L1 inhibitor Tecentriq® (atezolizumab, Genentech/Roche) presented positive data from the Phase III OAK trialat this conference,5 and this agent is expected to be approved soon for use in relapsed patients. The PD-L1 inhibitors durvalumab (MedImmune /AstraZeneca) and avelumab (Pfizer/Merck KGaA) as well as the CTLA-4 inhibitors Yervoy® (ipilimumab, Bristol-Myers Squibb/Ono Pharmaceuticals) and tremelimumab (MedImmune/AstraZeneca) are in Phase III trials for several settings in NSCLC.
However, this monotherapy competition was overshadowed somewhat with the presentation of a third abstract at Sunday’s Presidential symposium at the ESMO conference. Prior to the presentation of KEYNOTE-024, data from the Phase II KEYNOTE-021 trial’s cohort G were presented.6 This trial is a multicohort study of Keytruda-based combination therapies in advanced NSCLC. Cohort G evaluated four cycles of carboplatin plus Alimta with or without Keytruda (200 mg Q3W for two years) in 123 patients. As with the other trials, Alimta maintenance and crossover in the control arm were allowed.
KEYNOTE-021 met its primary overall response rate (ORR) endpoint, with a 26% improvement in the response rate (55% versus 29%, p=0.0016). There were no complete responses (CRs) in either arm. The time to response was numerically improved for patients treated with the Keytruda combination (1.5 months versus 2.7 months). Subgroup analyses were performed, stratifying the data based on PD-L1 expression. A difference in the response for Keytruda combinations was based on PD-L1 expression levels, with the highest rates observed in patients with TPS ≥50% (<1% PD-L1 expression, 57% ORR; ≥1% PD-L1 expression, 54% ORR; 1-49% PD-L1 expression, 26% ORR; ≥50% PD-L1 expression, 80% ORR). Interestingly, different responses were also seen in the chemotherapy-alone arm based on PD-L1 expression levels (<1% PD-L1 expression, 13% ORR; ≥1% PD-L1 expression, 38% ORR; 1-49% PD-L1 expression, 39% ORR; ≥50% PD-L1 expression, 35% ORR). PFS was a secondary endpoint but was improved by 4.1 months with the addition of Keytruda (13.0 months versus 8.9 months, HR 0.53, p=0.0102); this is similar to the level of benefit observed with Keytruda versus platinum doublet in KEYNOTE-026 but in a broader patient population according to PD-L1 status. OS data were very immature, and no OS benefit was reported (HR 0.90).
Treatment-related adverse events of all grades were roughly similar between the two arms (93% versus 90%); however, Grade 3-4 adverse events were increased with Keytruda (39% versus 26%). Unlike KEYNOTE-024, this increase may be due in part to the increase in the longer time of treatment for the Keytruda combination (median exposure time, 8.0 months versus 4.9 months) as chemotherapy was part of this regimen. Adverse events that increased in the Keytruda arm include fatigue (64% versus 40%), nausea (58% versus 44%), rash (27% versus 15%), vomiting (27% versus 18%) and diarrhea (20% versus 11%); anemia was increased in the chemotherapy-alone arm (32% versus 53%).
It is the data from KEYNOTE-021 that suggest that Keytruda may not be, as the discussant to this trial, Dr. Soria, noted, the “800-pound gorilla.” Although these data appear somewhat similar at first glance, Dr. Soria also noted that the Kaplan-Meier PFS curves for KEYNOTE-021 separated at about the one month time point while the KEYNOTE-026 curves separated after three months, suggesting gently that the future may lie in combinations of checkpoint inhibitors with chemotherapy or other immunotherapies. AstraZeneca has three such trials for combinations of durvalumab with tremelimumab (MYSTIC and NEPTUNE in the first-line setting; ARTIC for relapsed patients). BMS has a Phase III trial, CheckMate-227, currently enrolling that looks at the combination of Opdivo alone, in combination with Yervoy, or in combination with chemotherapy, all compared with platinum-doublet chemotherapy. With these combination approaches being actively studied and potentially coming to market in the near future, Keytruda may not enjoy that “long ovation” for much longer. KEYNOTE-024 and CheckMate 026 are not the “end of the story” for immunotherapies for patients with NSCLC.
By Greg Wolfe, Ph.D., Senior Consultant, Clinical & Scientific Assessment, Kantar Health and Stephanie Hawthorne, Ph.D., Vice President, Clinical & Scientific Assessment, Kantar Health
Approximately three-quarters of breast cancers are hormone receptor-positive (HR+) and thus express estrogen receptors, progesterone receptors or both. Endocrine therapy is the mainstay of therapy for patients with HR+ and human epidermal growth factor receptor 2 (HER2)-negative breast cancer. Aromatase inhibitors (AI) are first-line standard of care for postmenopausal patients; however, the majority of women with advanced HR+, HER2-negative breast cancer develop disease that is resistant to these agents, and thus there is a great need for therapeutic agents that lengthen or restore sensitivity to endocrine therapy.
Growth of HR+ breast cancer depends upon cyclin D1, and cyclin D1 transcription is directly regulated through the estrogen receptor. Cyclin D1 activates cyclin-dependent kinases, including CDK4/6, which play a key role in cellular proliferation. To date, one selective CDK4/6 inhibitor, Ibrance® (palbociclib, Pfizer) is approved for treatment of HR+/HER2-negative advanced or metastatic breast cancer in combination with letrozole as initial endocrine-based therapy in postmenopausal women or in combination with fulvestrant in women with disease progression following endocrine therapy. Novartis is pushing for ribociclib (LEE011), through the MONALEESA-2 study, to become the next CDK4/6 inhibitor to market.
In May 2016, Novartis announced that the Independent Data Monitoring Committee recommended a halt to MONALEESA-2 as the trial had achieved its primary endpoint at a preplanned interim analysis; treatment with ribociclib plus letrozole showed a clinically meaningful improvement in progression-free survival (PFS) in 668 postmenopausal women who had received no prior therapy for advanced breast cancer. Results of MONALEESA-2 were presented today as part of the Presidential Symposium at the 2016 European Society for Medical Oncology (ESMO) Congress in Copenhagen.1,2 The Novartis-sponsored MONALEESA-2 trial is an international, placebo-controlled, Phase III trial that was designed to evaluate efficacy and safety of first-line ribociclib (600 mg/day, three weeks on/one week off) plus letrozole versus placebo plus letrozole in postmenopausal women with HR+/HER2-negative advanced breast cancer. At the time of this presentation, the median PFS (investigator assessment) had not been reached for the ribociclib arm and was 14.7 months for the placebo arm (HR=0.556; p=0.00000329), the overall response rate was 41% versus 28% (p=0.000155), respectively, and survival data were immature. Ribociclib treatment benefit was consistent across all patient subgroups. The most common Grade 3/4 adverse events associated with ribociclib included neutropenia in 60% of patients (versus 0.9% in the placebo arm) and leucopenia in 21% (versus 0.6% in placebo arm); however, febrile neutropenia occurred in only 1.5% of ribociclib-treated patients. The majority of nonhematologic adverse events in the ribociclib arm were reversible by dose interruptions and reductions, and just 7.5% of patients required permanent discontinuation of both ribociclib and letrozole because of adverse events.
In August 2016, Novartis announced that, based upon MONALEESA-2 results, the U.S. Food and Drug Administration (FDA) had awarded ribociclib in combination with letrozole Breakthrough Therapy Designation in postmenopausal women who had received no prior therapy for their advanced disease. Regulatory filing plans have not been announced but are presumably under preparation.
As Ibrance is already approved in the United States and is expected to be approved in Europe shortly, ribociclib will likely be the second CDK4/6 inhibitor to market. In terms of clinical activity, these two CDK inhibitors appear to be very similar, both offering strong PFS benefits in comparison with letrozole alone and both with similar adverse event profiles.1,2,3 With such similar profiles, ribociclib may find difficulty displacing Ibrance, which has already been strongly adopted into practice in the United States after having been launched in February 2015. In Europe, Ibrance will have less of a time advantage over ribociclib, and the competition between these two CDK4/6 inhibitors may be a tighter race. To help differentiate ribociclib from Ibrance, Novartis has initiated two additional Phase III studies, MONALEESA-3, to evaluate the combination of ribociclib plus fulvestrant in first or second line (similar in design to the PALOMA-3 trial for Ibrance4), and MONALEESA-7, to evaluate ribociclib plus tamoxifen plus goserelin in pre/perimenopausal women with advanced breast cancer (a unique indication from any of the other CDK inhibitors in development).
Among the 10 ambitious priorities a blue ribbon panel forwarded to the Cancer Moonshot Task Force, headed by Vice President Biden, the call for accelerated research to identify approaches to monitor and manage patient-reported symptoms and other patient-reported outcomes (PRO) brings this topic further prominence. This echoes a central theme at this year’s ASCO Annual Meeting, and the focus of an Executive Roundtable held during the meeting on patient-strategies for oncology research and treatment.
It may be surprising that symptom management and the use of PRO in cancer care hasn’t received more attention over the years, given that cancer is a disease of symptoms. Both disease symptoms (pain, fatigue, shortness of breath and loss of weight, among others) and toxicities of treatment have a disproportionate impact on the experience of someone with cancer relative to most other medical conditions. To date, clinical care and drug research have focused far more on outcomes like tumor growth or shrinkage, biomarkers, and measures of overall or progression-free survival, when in fact we are learning that patients may be willing to trade some measure of traditional outcome for relief from treatment-related symptoms, like fatigue, that would improve their quality of life.
As the 2016 CancerCare Patient Access and Engagement Report found, there is an “impact gap” in communication between patients and their clinical care teams. The survey found that in discussions with care providers, about half of patients downplayed their side effects and symptoms. Nearly 40% did not report symptoms or side effects because they did not want to “bother” their doctors. Making symptoms management a higher priority in the clinical encounter would facilitate these conversations between patients and the care team, in both directions.
It is incumbent on clinical development researchers to consider how best to measure the impact of new therapy on outcomes that matter to patients in order to support the overall Value assessment of the drug under study. This will require expertise in selecting appropriate, valid measures of relevant symptoms. Furthermore, regulatory agencies are increasingly engaging with multiple stakeholders to support consideration of the voice of the patient in their decisions.
The spotlight that’s being shone on this issue – and the potential research and implementation funding that may ensue – couldn’t come at a more pivotal moment for cancer research and treatment.
By Joe O’Connell, M.D., Vice President and Global Therapeutic Lead, Medical Affairs, inVentiv Health
By Cory Lewis, Ph.D., Manager, Clinical & Scientific Assessment, Kantar Health
We are in a unique time within cancer therapy, where significant resources are being used to develop drugs that target the immune system to fight cancer. Now we embark on a new endeavor in which most of the work is not discovering novel agents but developing a better understanding how to improve the use of therapies we have available.
Most of us are well aware of immunotherapy and its rise among the ranks of more than 3,400 cancer therapies. Immunotherapies targeting the immune checkpoint in cancer have led to the approval of Keytruda® (pembrolizumab, Merck), Opdivo® (nivolumab, Bristol-Myers Squibb), Yervoy® (ipilimumab, Bristol-Myers Squibb) and Tecentriq® (atezolizumab, Roche).
These immune checkpoint therapies are designed to inhibit those molecules that prevent the immune system pathway from acting on cancer cells, while immune costimulators are designed to activate the immune system.
Immune costimulators work by providing the signal that promotes the expansion and proliferation of killer CD8 and helper CD4 T-cells. While immune checkpoint therapies are “releasing the brakes,” costimulators are “stepping on the gas.” Both mechanisms are required for a robust immune system response.
The hope is that by combining checkpoint inhibitors with immune costimulators the immune response will be more robust and response rates will drastically increase. Many combinations are being evaluated to fine tune the immune system response while minimizing side effects.
Kantar Health’s newly developed CancerLandscapeTM platform identifies 23 immune checkpoint-targeted therapies in 893 trials in development, including 12 immune costimulators in 41 trials. Targets for immune costimulators include CD40L, GITR, 4-1BB, CD27, TNFRSF25, TMIGD2, ICOS, CD28 and OX40. Of this group, OX40 is the target furthest in development with over 12 ongoing clinical trials.
With all the new therapies and potential targets along with endless combination options it will be hard to decide which combinations are best to develop.
Some of these details are already being investigated:
These details are currently being investigated in animal models and in the clinic and are important to the patients and their response to immunotherapy. The hope would be we can even unlock tumors not thought to be able to respond to immunotherapies by perfecting the system. The surface has just been touched when it comes to costimulators, and we all await the answers.