The checkpoint immunotherapy revolution began in melanoma, with the approval of the anti-CTLA-4 ipilimumab in 2011, followed by the approvals of nivolumab and pembrolizumab, which both target PD-1, in 2014. Since then, these immunotherapies—and several others—have been approved for a variety of cancer types, but because melanoma was the first tumor type in which these therapies were widely tested, it still offers the best look at the long-term benefits of this approach.
On Day 4 of ASCO18, long-term follow up data from two early melanoma trials using PD-1 checkpoint immunotherapy were revealed. In the Keynote-001 trial, which was led by Antoni Ribas, M.D., Ph.D., a co-leader of the CRI-SU2C Dream Team and a member of the CRI Clinical Accelerator Leadership at UCLA, patients with advanced melanoma were treated with pembrolizumab. Five years after treatment, 34% of those patients are still alive, including 41% of those who weren’t previously treated with ipilimumab. Furthermore, 73% of all responding patients still maintain their response, including 82% of the ipilimumab-naïve responders.
The updated results of Keynote-001 were discussed by Harriet M. Kluger, M.D., of the Yale Cancer Center, who noted that 5% of the patients who responded on the trial remain on therapy, which led her to ask when is the right time to stop immunotherapy in patients who respond. She pointed to the Keynote-006 trial, in which patients with advanced melanoma were treated with pembrolizumab for up to two years, and suggested that with more follow-up time, that trial could help answer that question with respect to whether two years was enough time to support long-term, durable responses.
Speaking of Keynote-006, Georgina V. Long, M.D., Ph.D., a CRI Clinical Team Leader at the Melanoma Institute Australia, revealed the four-year follow up data from that trial during a talk Monday morning. In all, 556 patients began pembrolizumab treatment, and 42% of those patients responded, including 14% whose tumors disappeared completely. The overall 4-year survival rate for those 556 patients was 41.7% and their 3-year progression-free survival rate was 31.1%. Of the 556 patients who began the trial, 103 completed two years of the immunotherapy regimen and 86% were still progression-free twenty months after their last treatment. In some patients who progressed, re-treatment with pembrolizumab was able to offer them additional benefit.
Kluger discussed the results of three other posters as well—two in melanoma and one in squamous cell carcinoma that we’ll come back to later.
The first also involved Long, and focused on the use of PET (positron emission tomography) scans, which can measure the metabolic responses of tumors, as a means of predicting which patients with metastatic melanoma might become long-term responders after treatment with PD-1/PD-L1 immunotherapies, either alone or in combination. In the patients who survived at least one year after immunotherapy treatment, 68% were characterized as having complete metabolic responses via PET scans, while only 25% were characterized as having complete responses via CT (computed tomography) scans. In addition to helping doctors identify potential long-term responders to PD-1/PD-L1 immunotherapy, PET scans may also have the ability to identify non-responders, and thus could guide treatment discontinuation decisions; however, its value needs to be validated prospectively and at earlier intervals throughout treatment.
The other poster discussed by Kluger involved Long too, as well as fellow CRI Clinical Team Leader Richard A. Scolyer, M.D., of the Melanoma Institute Australia. This study found that at baseline, a type 1 immune response gene signature and high TBET expression could serve as a predictor of potential response to PD-1 monotherapy, whereas high EOMES gene expression and the presence of EOMES+ memory T cell populations could serve as predictors of response to combination PD-1 and CTLA-4 immunotherapy. Additionally, it was found that increased expression of interferon-γ and TBET could serve as predictors of response early during monotherapy, whereas greater numbers of cytotoxic memory T cells and increased expression of LAG3 and TIGIT could serve as predictors of response early during combination therapy.
PD-1 immunotherapy can also provide benefits in the adjuvant setting after resection of advanced melanoma, as revealed by Jeffrey S. Weber, M.D., of the Perlmutter Cancer Center at NYU Langone Health. Compared to patients who received ipilimumab after surgery, patients treated with nivolumab post-surgery experienced a 36% decrease in the risk of recurrence in the two years following treatment. In patients whose tumors had a PD-L1 expression level of 5% or higher, that reduction in risk was 46%, and in those with metastatic melanoma the median recurrence-free survival was 30.8 months for nivolumab-treated patients compared to 15.4 months for ipilimumab-treated patients. Nivolumab’s benefits also extended to patients with wild-type BRAF as well as those with mutated BRAF.
Due to the immunogenic nature of many melanomas, novel immunotherapy strategies—both in terms of new agents and introducing immunotherapy earlier in the course of disease—are often evaluated in this cancer type, too, as a way to determine their potential benefits.
As Joshua Mammen, M.D., Ph.D., of the University of Kansas, discussed, patients with advanced melanoma benefited from being treated with the oncolytic virus T-VEC prior to surgery, which led to a 14.7% clinical response rate and a 21.1% pathological complete response rate as well as a 45.6% disease control rate.
Mammen also spoke about a poster that highlighted the pre-surgical value of the combination of nivolumab and ipilimumab in high-risk resectable metastatic melanoma. Alone, nivolumab led to a 25% pathological complete response rate and a 25% clinical response rate, but when ipilimumab was added to the regimen, those rates jumped to 45% and 73%, respectively. Trends toward increased progression-free survival, distant metastasis-free survival, and overall survival were also evident; however, those improvements didn’t rise to the level of statistical significance.
Next, Alexander Menzies, M.B.B.S., Ph.D., of the Melanoma Institute Australia, discussed a poster from the ECHO-204 trial, which evaluated the combination of nivolumab with the IDO-1 inhibitor epacadaostat, in patients with advanced melanoma. Overall, the combination was associated with a 62% response rate that included an 18% complete response rate. Patients with PD-L1-positive tumors has higher response rates compared to those with PD-L1- tumors (75% versus 56%), and patients with IDO expression also fared better compared to those without (75% versus 50%). When looked at next to the data from CheckMate-067, which evaluated nivolumab alone in this setting, these results appear favorable; however, there was a higher proportion of patients with visceral metastases and elevated LDH levels in that monotherapy trial, so it might not be a fair comparison.
The second poster discussed by Menzies also tested a combination PD-1 and IDO-1 inhibition strategy, using pembrolizumab and indoximod, which led to a 53% response rate and 18% complete response rate. The PD-L1-positive group of patients did better with this combination as well. While these results also appear favorable next to pembrolizumab monotherapy data (from Keynote-006), the proportion of patients with visceral metastases and elevated LDH levels in this combination trial were not determined, thus preventing proper comparison.
Menzies covered one final study, in which UCLA’s Ribas was the first author and Long a contributor, that treated PD-1-naïve advanced melanoma patients with the combination of pembrolizumab and SD-101, a Toll-like receptor 9 (TLR9) agonist that stimulates a subset of dendritic cells and promotes increased antigen presentation. At the dose level of 2mg or less, 80% of patients responded and 76% were progression-free at six months.
After Menzies, Michael A. Postow, M.D., of Memorial Sloan Kettering discussed two posters that explored the use of immunotherapy combinations in patients who didn’t respond to prior PD-1 treatment.
The first, like Ribas and Long’s study above, also employed a TLR9 agonist (IMO-2125), this time in combination with ipilimumab, and led to a 38.1% response rate and 71.4% disease control rate. The second tested the combination of pembrolizumab and ipilimumab, which produced responses in 45% of treated patients and complete responses in 13.6%. Six of nine patients with elevated levels of LDH, as well as three of five with liver metastases, also responded.
Moving beyond melanoma, immunotherapy’s benefits in other skin cancers were also on display during Day 4 at ASCO18.
In Merkcel cell carcinoma (MCC), a rare and aggressive skin cancer commonly associated with the Merkel cell polyomavirus (MCPyV), Suzanne Topalian, M.D., Ph.D., a member of the CRI-SU2C Dream Team at the Johns Hopkins University School of Medicine, revealed the benefits of neoadjuvant (pre-surgical) nivolumab in patients with resectable disease. Of 25 evaluable patients, 40% had partial responses and additional 52% had their tumors stabilize, which translates to a 92% disease control rate. Only two patients experienced progressive disease. Benefits were experienced by patients whose tumors had minimal PD-L1 expression, too, as evidenced by one complete response, four partial responses, and six patients with stable disease. Only one patient with a PD-L1-negative tumor experienced disease progression. Importantly, immunotherapy did not delay surgery in any patients.
Following Topalian, Paul Nghiem, M.D., Ph.D., of the Fred Hutchinson Cancer Research Center and the University of Washington, gave two talks on immunotherapy in MCC.
First, he revealed that first-line pembrolizumab led to a 24% complete response rate, a 56% response rate, and a 66% disease control rate in patients with metastatic or recurrent/locally advanced MCC. Most responses had occurred within the first two to three weeks, are still ongoing, and have proven durable beyond two years past the start of treatment. The overall survival rate at the three-year mark was 64% (versus the historical 10% rate with first-line chemotherapy) and the median progression-free survival was 16.8 months (versus the historical 3.1-4.6 months seen with chemotherapy).
Second, he highlighted the benefits of avelumab, a PD-L1 checkpoint immunotherapy, in patients with metastatic, chemotherapy-resistant MCC. In addition to a 33% response rate, the two-year overall survival rate was 36%, which is relatively promising given that the historical overall survival rate for patients treated with chemotherapy drops to zero by nine months. Additionally, roughly two-thirds of the patients who responded remained progression-free for at least two years.
Following his second talk, Nghiem was asked by an audience member, “Are you curing Merkel cell carcinoma?”
After an initial hesitation, Nghiem volunteered that, “I don’t think I’m going too far out on a limb to say that it’s possible … but the follow up isn’t long enough [yet].”
Lastly, work in metastatic squamous cell carcinoma (mSCC) was discussed by Kluger (see above) and involved Regeneron’s Israel Lowy, M.D., Ph.D., a member of the steering committee of CRI’s Cancer Immunotherapy Consortium. In this study, cemiplimab, a PD-1 checkpoint immunotherapy, led to a 1-year progression-free survival rate of 52.5% and a 1-year overall survival rate of 80.6% in patients with mSCC.
While lung cancer immunotherapy didn’t dominate Day 4 quite the way it did Day 3, there were still plenty of trial readouts to go around.
Updates on IMpower131, which is exploring the use of chemotherapy with or without the PD-L1 checkpoint immunotherapy atezolizumab in first-line advanced squamous non-small cell lung carcinoma (NSCLC), were provided by Robert Jotte, M.D., Ph.D., of the Rocky Mountain Cancer Centers.
With a twelve-month progression-free survival rate of 24.7%, compared to 12% for chemotherapy alone, the combination demonstrated improved efficacy across all PD-L1 subgroups, with the benefits being even greater in the PD-L1-high patient population. While the response rates (49% versus 41%) weren’t drastically different between the combination and chemotherapy arms, respectively, there was a dramatic improvement seen in the PD-L1-high subgroup (60% versus 33%).
The same trend held for the median duration of response, which was 7.2 months versus 5.2 months in the combination and chemotherapy arms, respectively, and 18.7 months versus 5.3 months in patients with high PD-L1 expression. In terms of survival, the combination led to higher two-year rates compared to chemotherapy alone in the entire patient population (31.9% versus 24.1%), the PD-L1-high subgroup (47% versus 30%), and the PD-L1-negative subgroup (30% versus 16%), for some unknown reason the trend was reversed in the PD-L1-low subgroup (28% versus 37%). Other than perhaps an anomaly due to small patient sample size, there was no rationale or mechanism provided as a possible explanation.
The IMpower150 trial evaluated the above combination as well, only it added the VEGF-A-targeting antibody bevacizumab to the mix, and it evaluated the combination as a first-line treatment in non-squamous (as opposed to squamous) NSCLC. In his talk, Mark A. Socinski, M.D., of the Florida Hospital Cancer Institute, revealed that there was a significant overall survival benefit associated with adding atezolizumab to the chemotherapy-bevacizumab combination. At the eighteen-month mark in patients without EGFR mutations or ALK translocations, the progression-free survival rate of the triple combination was 27% versus 8% for the doublet, which corresponds to a 59% reduction in risk of progression, whereas the triplet had a median overall survival that was four and a half months longer than the doublet (19.2 months versus 14.7 months).
The overall survival rate and median overall survival were also greater in the triplet-treated population, across all PD-L1 subgroups, most significantly in the PD-L1-low subgroup, and regardless of EGFR mutational status or ALK translocation status. Of note, in patients with EGFR or ALK abnormalities, only 50% of the doublet-treated patients were still alive at 17.5 months, while at the 25 months more than 60% of the triplet-treated patients were still alive.
Hossein Borghaei, D.O., of the Fox Chase Cancer Center, also spoke about a first-line combination approach in advanced NSCLC, courtesy of the CheckMate-227 study, which evaluated nivolumab plus ipilimumab versus nivolumab plus chemotherapy versus chemotherapy alone in patients with PD-L1 tumor proportion scores of 1% or less. In this subgroup, the response rates for the three arms were 25.1% (nivolumab- ipilimumab), 36.7% (nivolumab-chemotherapy), and 23.1% (chemotherapy alone); however, the combinations were both much more effective in patients whose tumors had high mutational burdens, characterized as more than 10 mutations per DNA Megabase pair.
In the patients with high TMB and low PD-L1, the response rates were 36.8% (nivolumab-ipilimumab), 60.5% (nivolumab-chemotherapy), and 20.8% (chemotherapy alone). Interestingly, the nivolumab and ipilimumab combination (93%) far exceeded the nivolumab plus chemotherapy combination (33%) and chemotherapy only (less than 20%) in terms of the proportion of patient responses that lasted at least one year.
With respect to biomarkers that can predict progression-free and/or overall survival in advanced NSCLC, Kathryn Arbour, M.D., of Memorial Sloan Kettering, revealed that patients treated with baseline steroids were associated with poorer outcomes in patients treated with PD-1/PD-L1 immunotherapy. In two different patient cohorts—one at Memorial Sloan Kettering and one at Gustave Roussy in Paris, France—patients who received a daily dose of steroids had significantly lower response rates: 6% versus 19% at MSK, and 8% versus 18% at Gustave-Roussy. Fortunately, the use of steroids to manage immune-related adverse events during treatment didn’t appear to diminish the effectiveness of PD-1/PD-L1 immunotherapy in that context.
Thomas Gajewski, M.D., Ph.D., a former CRI CLIP Investigator and current member of our Scientific Advisory Council at the University of Chicago Medical Center, discussed a poster involving another biomarker that appeared to influence outcomes in PD-1/PD-L1-treated patients with NSCLC. In the study, it was found that an increased ratio of neutrophils to lymphocytes was associated with worse clinical outcomes in a range of cancer types, although no causal mechanism was identified so a purely correlative connection could not be ruled out.
Gajewski also discussed two other interesting concepts that were presented in posters.
The first involved using machine learning algorithms to identify patterns in complex data, such as that found within the tumor-immune microenvironment, and then learn to detect what the important factors are for given outcomes. In this case, the medical image computing program was able to confirm the importance of CD8+ tumor-infiltrating T cells in carrying out the anti-tumor activity enabled in response to PD-1/PD-L1 checkpoint immunotherapy, finding that, compared to patients with tumors characterized by a low radiomic-based CD8 T cell score, those with tumors characterized by a high CD8 T cell score had significantly superior overall survival (24.3 months versus 11.5 months), and were associated with a 42% reduction in risk of death.
The final study Gajewski discussed addressed the fact that not all disease progression appears to be equal. In fact, the authors broke progressive disease down into three categories—“no growth,” “mixed growth,” and “aggregate growth”—based on the growth of individual target and metastatic lesions. In melanoma patients from the Keynote-001 trial, the patients defined as having “no growth” progressive disease, in which no individual target lesion grew more than 20% from baseline, were found to have lower “on treatment” hazard ratios compared to the “aggregate growth” and “mixed growth” subgroups. In light of this, more precise classification of progressive disease could help make sense of the fact that many patients appear to benefit from continued PD-1 treatment even after disease progression, which would have important implications for clinical practice
Results from early phase clinical trials testing novel immunotherapy agents in other cancer types were also highlighted on Day 4.
Several such approaches were highlighted during a poster discussion session by Sherene Loi, M.D., Ph.D., of the Peter MacCallum Cancer Centre, including the ChAdOx1-MVA 5T4 vaccine (consisting of two engineered, replication-deficient viruses) in early stage prostate cancer, which was able to induce T cell responses against diverse 5T4 epitopes and was associated with increased CD8+ T cell infiltration into tumors. Loi also covered genetically modified T cells with engineered TCRs designed to target the HPV-16 E6 viral oncoprotein in advanced cancers. In addition to long-term persistence, these cells also produced two responses in the nine patients who received the highest dose level. In one of these responders, the treatment led to complete elimination of one lesion, and after the rest were resected, the patients have been disease-free for more than three years. Lastly, Loi discussed the combination of nivolumab and cabiralizumab, a CSF-1R inhibitor, in advanced tumors, which promoted responses across multiple tumor types and even four partial responses in patients with microsatellite stable (MSS) pancreatic cancer.
Charlotte Ariyan, M.D., Ph.D., of Memorial Sloan Kettering Cancer Center, discussed a poster presentation that involved the combination of nivolumab and IPI-549, which targets the PI3K pathway, in advanced solid tumors. Of twelve patients treated thus far, there have been two responses, one in gallbladder cancer and one in adrenocortical cancer, and both are ongoing. With the nivolumab dose kept constant, a PI3K-dose-related increase in the proliferation of exhausted CD8+ memory T cells was observed.
Finally, Stephen Shiao, M.D., Ph.D., of Cedars-Sinai Medical Center, discussed the impact of bacteria-related factors on the effectiveness of PD-1/PD-L1 checkpoint immunotherapy. In one study, it was shown that patients who responded to PD-1/PD-L1 immunotherapy had more diverse gut microbiomes and were particularly enriched with members of the Ruminococcaceae family of bacteria. In another, it was found that antibiotic treatment, especially more than two courses, was associated with worse survival after treatment with PD-1/PD-L1 immunotherapy. Shiao did mention a few caveats, particularly that antibiotics are not usually given to healthy people, so there could have been a slight selection bias in terms of patients predisposed toward worse outcomes.
Furthermore, other microbiome-related factors—such as diet, country of origin, and other medications—could be indirectly influencing immune activity and responses against cancer. Therefore, future studies will need to identify the precise mechanism for these effects. Then, these insights could allow for the development of antibiotic approaches that can favorably modulate the gut microbiome to mobilize the immune system and enhance immunotherapy’s effectiveness.
Check back with us tomorrow for our recap of the final day here at ASCO18!