Co-authored by Elizabeth Clarke, Ph.D., M.P.H., Stephanie Hawthorne, Ph.D., Len Kusdra, Ph.D. and Gregory Wolfe, Ph.D.
Immunotherapy has dominated the conversation in oncology for the past several years, and the excitement still hasn’t waned at all — in fact, it continues to grow. New drugs have entered development in this space, new mechanisms of action have emerged, and new tumor types have become the focus of our attention. We have previously seen various levels of evidence to support the activity of checkpoint inhibitors in melanoma, non-small cell lung cancer, renal cell carcinoma, bladder cancer, breast cancer, gastric cancer, head and neck cancer and Hodgkin’s lymphoma. At this year’s annual meeting of the American Society of Clinical Oncology (ASCO), we observed first reports of the activity of several PD-1 and PD-L1 inhibitors in multiple new tumor indications, adding to the growing body of evidence that suggests immunotherapy is closely approaching panacea status.
Opdivo® (nivolumab, Bristol-Myers Squibb/Ono Pharmaceuticals) was studied in a global Phase I/II trial in patients with advanced hepatocellular carcinoma (HCC).1 This was a dose-escalation study, and it also stratified patients according to their hepatitis B/C viral (HBV, HCV) infection status (uninfected vs. HCV-infected vs. HBV-infected). The majority of patients (75%) had received prior systemic therapy for their disease, with most (68%) having received prior Nexavar® (sorafenib, Onyx/ Amgen/Bayer). At this interim analysis of 42 evaluable patients, single-agent Opdivo produced a 19% objective response rate (ORR; which was measured across all dose cohorts), including two patients (5%) with a complete response; an additional 48% of patients had stable disease. Responses were seen in all patient cohorts, but the response rate was higher in those with HCV infections (36%). It was postulated that perhaps patients with higher viral load/disease burden are more sensitive to PD-1 inhibition and speculated that HCV T-cells may naturally express higher levels of PD-L1, making them more sensitive to PD-1 inhibition; both of these observations are very preliminary in nature, however, and need more data to fully understand whether there is a correlation with activity and, if so, why. The duration of responses were very encouraging (out to one year in some patients) and the one-year overall survival (OS) rate was 62% — unheard of in metastatic, relapsed HCC! These are exciting data in a tumor of very high unmet need, and we’re looking forward to following this trial for the final results and monitor for future clinical development in this indication.
There has been little development of targeted therapies for esophageal cancer, making it a tumor of small incidence but large unmet need. Immunotherapy agents have already shown promise in gastric and gastroesophageal junction adenocarcinoma, so it was nice to see a presentation describing the first data for this class in patients with metastatic esophageal cancer. An expansion cohort of the Phase Ib KEYNOTE-028 trial enrolled 23 patients with primarily PD-L1-positive, heavily pretreated esophageal cancer (a few patients had gastroesophageal carcinoma).1 Treatment with single-agent Keytruda® (pembrolizumab, Merck & Co.) was associated with a 30.4% partial response rate (29.4% in the 17 patients with squamous cell carcinoma and 40.0% for the five patients with adenocarcinoma). Responses are very durable, with an average duration of response of 40 weeks. More than half of patients showed some degree of tumor shrinkage. These are excellent outcomes, especially considering that a 30% response rate isn’t much different than that usually seen in the first-line setting for gastroesophageal cancer. However, the discussant, Dr. Fuchs, emphasized slight caution due to the lack of central review in this trial. Similar to other trials for Keytruda, adverse events (AEs) were very limited, with only four patients (17.4%) experiencing Grade 3 toxicities. During the discussion, Dr. Fuchs hypothesized on why these patients seemed to respond so well to Keytruda. Although the patients with the Epstein-Barr virus (EBV)-positive and microsatellite instability (MSI)-positive2 adenocarcinoma subtypes might be responding to increased immune activity due to association of these subtypes with increased PD-L1 expression, the responses in patients with squamous cell carcinoma may be associated with the high degree of DNA focal copy number alterations. Further research is warranted to evaluate these hypotheses as well as confirm the encouraging activity of Keytruda in these patients.
Today results were presented from two separate Phase Ib studies that are evaluating safety and efficacy of PD-1 checkpoint inhibitors in patients with relapsed/refractory ovarian cancer. In the first of the two trials,3 75 heavily pretreated ovarian cancer patients enrolled in an expansion cohort were treated with the PD-L1 inhibitor avelumab (Merck KGaA/Pfizer). Patients were not selected based on PD-L1 status. There were no complete responses, and eight patients achieved partial responses for an ORR of 10.7%. Stable disease was the best response in 33 patients, for a clinical benefit rate of 54.7%. Responses were ongoing in five of eight patients, and the median duration of response was nine weeks. Grade 3/4 adverse events were reported in six patients (8%), and the most common treatment-related AEs of any grade included fatigue (16.0%), chills (12.0%), nausea (10.7%) and diarrhea (10.7%). The author noted that a Phase III trial of avelumab is currently being planned. The second presentation reported results from the ovarian cancer cohort (n=26) of the Phase 1b KEYNOTE-028 trial.4 Heavily pretreated, PD-L1-positive patients were enrolled in KEYNOTE-028 and received the anti-PD-1 antibody Keytruda, and they were treated for 24 months or until disease progression. The ORR was 11.5% and included one complete response and two partial responses. Another six patients achieved stable disease for a disease-control rate of 34.6%. Median duration of response was not reached. Grade 3/4 AEs were reported in 3.8% of the patients, and the most common treatment-related AEs included arthralgia (23.1%), diarrhea (11.5%) and nausea (11.5%). The authors are further analyzing their data to discern whether there is a relationship between PD-L1 expression and activity. These two studies demonstrate that both anti-PD-1 and anti-PD-L1 antibodies are well-tolerated and active in heavily pretreated ovarian cancer patients. Interestingly, the ORR was similar between the two trials; therefore, as with other indications, physicians may need to wait for follow-up data to determine the ultimate choice of immunotherapy option.
Another example where immunotherapies may make an impact is in recurrent glioblastoma multiforme (GBM), which has a poor prognosis when treated with current therapies and thus represents a considerable unmet need. PD-L1 expression has been associated with high-grade but not low-grade gliomas.5 Also, antitumor activity using a non-commercial anti-PD-1 monoclonal antibody was observed in mouse glioma models.6 Given this preliminary data, Bristol-Myers Squibb (BMS) initiated the Phase III CheckMate-143 trial, which will randomize patients to Opdivo or Avastin® (bevacizumab, Genentech/Roche/Chugai) after progression following surgical resection, radiation and temozolomide. The design of CheckMate-143 included a safety run-in phase evaluating Opdivo with or without Yervoy® (ipilimumab, BMS), which reportedly helped influence the Phase III trial design; some of this data from 20 patients were presented in a poster session at ASCO 2015.7 Note that the arms containing Yervoy were included based on activity observed for Yervoy in patients with melanoma and brain metastases.8 In the data just presented, treatment-related serious AEs occurred in two patients in the monotherapy arm and seven patients treated with the combination. Moreover, as might be expected, the combination regimen had more treatment-related discontinuations (40% vs. 0%). With this said, Opdivo monotherapy was very well tolerated. Opdivo monotherapy was associated with a 10% partial response (PR) rate and a 50% clinical benefit rate; six- and nine-month OS rates were 70% and 60%, respectively. In contrast, Opdivo plus Yervoy was associated with no PRs and a 40% stable disease rate; six- and nine-month OS rates were 80% and 60%, respectively. Given this data, Opdivo monotherapy was chosen to be evaluated in the Phase III trial initiated last year. With the limited options currently available for GBM patients, there should be a high level of enthusiasm to quickly see the outcome from this trial.
Small Cell Lung Cancer
Checkpoint inhibitors are active in other thoracic tumors. Opdivo is already approved for use in non-small cell lung cancer (NSCLC) patients with squamous histology and recently excited ASCO attendees with the CheckMate-057 data in patients with non-squamous histology. Keytruda also showed strong promise in both NSCLC in the KEYNOTE-001 trial and malignant pleural mesothelioma in an expansion cohort of the Phase Ib KEYNOTE-028 trial. Given the level of competition between the two molecules in these other thoracic tumors, it is not surprising to see data for both agents in small cell lung cancer (SCLC). A different cohort from Keynote-028 examined the safety and preliminary efficacy of Keytruda in 20 patients with PD-L1-positive relapsed SCLC.9 The toxicity profile was mild, with 10% of patients having Grade 3 or greater AEs; one fatality occurred due to Grade 5 colitis. There was promising evidence of efficacy, with seven (35%) evaluable patients achieving a partial response. Median time to response was 8.6 weeks, with six of seven responses ongoing at the time of data cutoff; one patient has exhibited a response of at least 32 weeks. Opdivo was evaluated in the CheckMate-032 trial, a Phase I/II trial randomizing 90 heavily pretreated SCLC patients to Opdivo with or without Yervoy.10 The combination was evaluated at two dose levels: Yervoy dosed at 1 mg/kg or 3 mg/kg. The most common Grade 3/4 side effects in patients receiving Opdivo plus Yervoy included diarrhea (8.5%), increase in lipase (6.4%), rash (4.3%) and vomiting (4.3%); the total incidence of Grade 3/4 toxicities was 15% for the monotherapy and 34% for the combinations. The initial ORR for Opdivo monotherapy was 18% (CR 0%) and for both combinations was 17% (CR 2%). The ORR for the combination increased with continued treatment, as seven patients with initial stable disease were upgraded to confirmed partial responses, resulting in an ORR of 32.6%. Subgroup analysis showed that this level of response was found in both platinum-sensitive and platinum-resistant patients. Median duration of response was 6.9 months for the combination and was not reported for the monotherapy. Median OS was 4.4 months for Opdivo monotherapy and 8.2 months for the combination. Neither trial showed evidence of a linkage between increased PD-L1 expression and activity. Although Keytruda showed a higher level of activity than Opdivo monotherapy (35% versus 18%), the numbers of patients and the inability at this stage to assess how similar the patient populations are will force physicians to wait for further data in this indication to help them make choices as to which immunotherapy to offer their SCLC patients.
In the original Phase I trial for Opdivo presented at ASCO 2012,11 no responses were observed among 19 patients with colorectal cancer (CRC). At ASCO 2015, however, data presented from a new Phase I trial suggests a subset of these patients may actually benefit from PD-1 inhibition. In a multicohort Phase II trial, patients with pretreated CRC were enrolled into two cohorts – those with tumors proficient for mismatch repair (MMR; n=25, n=25 evaluable) and those with MMR-deficient tumors (n=25; n=13 evaluable); a third cohort enrolled patients with MMR-deficient non-colorectal pretreated solid tumors (n=21; n=10 evaluable). Activity was strikingly different among these cohorts of patients following treatment with Keytruda. The ORR was 62% in MMR-deficient CRC patients and 60% ORR in MMR-deficient non-colorectal cancers, but the ORR was 0% in patients with MMR-proficient CRC. The difference can’t be blamed on the depth of response, because disease control including stable disease was equally different: 92% in MMR-deficient CRC, 70% in MMR-deficient non-colorectal cancers, and only 16% in MMR-proficient CRC. Progression-free survival and OS were also longer in patients with MMR-deficient CRC or other cancer compared to patients with CRC-proficient tumors. Although the sample sizes of MMR-deficient CRC and MMR-deficient non-CRC cohorts were small, and there was a slight imbalance in patient demographics between the cohorts (MMR-proficient CRC patients were older and slightly more pretreated), the degree of differential efficacy seems large enough to suggest the correlations may be real. Additional analysis shows that MMR-deficient tumors have greater density of invasive CD8+ T-cells, greater density of PD-L1-positive invasive CD8+ T-cells, and more somatic mutations per tumor than MMR-proficient tumors. Together, these data point to a possible mechanistic explanation for the differential activity of Keytruda, wherein higher tumor mutational burden leads to greater immune system recognition and tumor infiltration, leading to improved activity of PD-1 inhibitors.
This theory explains the exquisite activity of checkpoint inhibitors that has been observed in melanoma, NSCLC, bladder cancer, SCLC, gastric cancer, esophageal cancer and head and neck cancer: These tumors have the highest somatic mutational burden according to data from the Cancer Genome Project.12 Although mutational load is unlikely to be the only factor that influences likelihood of response to checkpoint blockade, new data from ASCO 2015 suggest it certainly plays a role and may need to be considered for future development plans and eventually patient selection across a number of solid tumor types.
1. El-Khouery AB, Melero I, Crocenzi TS, et al.; “Phase I/II safety and antitumor activity of nivolumab in patients with advanced hepatocellular carcinoma (HCC): CA209-040;” J Clin Oncol 33, 2015 (suppl; abstr LBA101).
2. MSI: microsatellite instability; EBV: Epstein-Barr Virus
3. Disis ML, Patel MR, Pant S, et al.; “Avelumab (MSB0010718C), an anti-PD-L1 antibody, in patients with previously treated, recurrent or refractory: a phase Ib open label expansion trial;” J Clin Oncol, 33 (suppl., abstr 5509), 2015.
4. Varga A, Piha-Paul SA, Ott PA, et al.; “Antitumor activity and safety of pembrolizumab in patients with PD-L1 positive ovarian cancer: interim results for a phase 1b study;” J Clin Oncol, 33 (suppl., abstr 5510), 2015
5. Yao Y, Tao R, Wang X, et al.; “B7-H1 is correlated with malignancy-grade gliomas but is not expressed exclusively on tumor stem-like cells;” Neuro Oncol, 11:757–766, 2009.
6. Zeng J, See AP, Phallen J, et al.; Anti-PD-1 blockade and stereotactic radiation produce long-term survival in mice with intracranial gliomas;” Int J Radiat Oncol Biol Phys, 86: 343-349, 2013.
7. Sampson JH, Vlahovic G, Sahebjam S, et al.; “Preliminary safety and activity of nivolumab and its combination with ipilimumab in recurrent glioblastoma (GBM): CheckMate-143; J Clin Oncol, 33 (15_supp), Abst 3010, 2015.
8. Margolin K, Ernstoff MS, Hamid O, et al.; “Ipilimumab in patients with melanoma and brain metastases: an open-label phase 2 trial;” Lancet Oncol, 13:459-465.
9. Ott PA, Fernandez MEE, Hiret S, et al.; “Pembrolizumab in patients with extensive-stage small cell lung cancer: Preliminary safety and efficacy results from KEYNOTE-028;” J Clin Oncol. 33 (suppl. Abstr 7502), 2015.
10. Antonio SJ, Bendell JC, Taylor MH, et al.; “Phase I/II study of nivolumab with or without ipilimumab for treatment of recurrent small cell lung cancer: CA209-032;” J Clin Oncol. 33 (suppl. Abstr 7503), 2015.
11. Le D, Uram J, Wang H, et al.; “PD-1 blockade in tumors with mismatch repair deficiency;” J Clin Oncol 33, 2015 (suppl; abstr LBA100).
12. Alexandrov LB, Nik-Zainal S, Wedge DC, et al.; “Signatures of mutational processes in human cancer;” Nature 2013, 500: 415-421.