One of the hot topics at this year’s ASCO Annual Meeting was broadening eligibility criteria to increase clinical trial access and enrollment.
Eligibility criteria are intended to ensure patient safety and define study population characteristics, but overly restrictive requirements limit the generalizability of results across broader “real-world” populations that ultimately receive the therapy.
Two studies published this spring highlight this conundrum in the context of therapy with immune checkpoint inhibitors (ICIs).
A study by Haslam and Prasad — published in May in JAMA Network Open — showed the percentage of patients with cancer eligible for immune checkpoint inhibitors (ICIs) increased from 1.54% in 2011 to nearly 50% in 2018. However, only 12.5% of patients treated with ICIs in 2018 responded to therapy — a substantially lower rate than that observed in registration trials for those agents.
Khozin and colleagues assessed real-world outcomes of patients with metastatic non-small cell lung cancer treated with PD-1 inhibitors.
The researchers analyzed electronic health records of 1,344 patients treated with nivolumab (Opdivo, Bristol-Myers Squibb) or pembrolizumab (Keytruda, Merck). Their results, published in May in The Oncologist, showed estimated OS in the real world (8 months) lagged what had been observed in registration trials (range, 9.2 months to 12.2 months).
Two years ago, ASCO and Friends of Cancer Research identified common trial exclusion criteria that — although intended to protect participants — likely do not affect safety. These included minimum age for enrollment, brain metastases, HIV infection, organ dysfunction, and prior or concurrent malignancies.
Many trials for ICIs excluded patients for similar reasons — specifically age, autoimmune disorders, HIV infection, prior transplantation or psychiatric conditions.
Although some exclusion criteria are based on legitimate safety concerns, others increase the potential that trial outcomes will not be reflective of clinical practice and leave providers confused about whether they can safely administer these agents to their patients.
Older adults remain underrepresented in cancer trials, including those involving ICIs.
The median age in most CheckMate and KEYNOTE trials of pembrolizumab and nivolumab ranged from 60 to 65 years, with few patients aged older than 75 years.
Immunosenescence — the change in immune system function due to age — is a significant barrier to enrolling older patients with cancer on ICI trials. Depletion in lymphocyte reserve, fewer naive CD4-positive and CD8-positive T cells, and changes in regulatory and memory T cells impair the ability to elicit an immune response.
Older adults may experience more immune-related adverse events than younger individuals, but this has not been proven.
In the CheckMate 017 and CheckMate 057 trials — which led to nivolumab’s approval for second-line or later therapy for advanced NSCLC — 7% to 10% of enrolled patients were aged 75 years or older, and a subset analysis suggested limited clinical benefit and shorter OS in this group.
An expansion trial designed to capture more older adults and those with poorer performance revealed no differences in treatment discontinuation, any-grade toxicity, grade 3 to grade 4 toxicity, or median OS.
A retrospective real-world analysis of patients with advanced NSCLC treated with ICIs, presented by de la Fuente and colleagues and European Lung Cancer Conference, suggested shorter PFS and OS among those aged 70 years or older, but toxicity was similar for older and younger patients.
At the same conference, Nosaki and colleagues presented a pooled analysis of three randomized trials that showed older adults with PD-L1-positive NSCLC achieved longer OS with pembrolizumab monotherapy than chemotherapy, with lower adverse event rates. Older adults and younger patients achieved similar 1-year OS rates with pembrolizumab; however, rates of immune-mediated adverse events and infusion reactions were higher among older adults.
Limited data from subset analyses suggest efficacy is similar for older and younger patients, although older patients — particularly those with poorer performance status — tend to have more adverse events.
A comprehensive geriatric assessment can improve quality of life and ability to receive cancer treatment. The concept of “comprehensive immune assessment” has been raised to understand how older adults might respond to treatment with ICIs, similar to “prehabilitation” before surgery.
Future studies may evaluate the role of immunosenescence on ICI response and toxicities. They also should examine how these data may guide development of future immunotherapies, or whether adjustments in administration may be more appropriate for older adults with cancer.
Prospective trials and more real-world data are required to further elucidate the safety and efficacy of ICIs for older adults, particularly as combination therapy approvals expand.
Patients with autoimmune diseases largely are excluded from immunotherapy clinical trials due to concerns about disease flares secondary to immune activation.
Clinicians who treat patients with autoimmune diseases have been reluctant to administer immunotherapies to this population. Further, will the need for immunosuppressive treatments to combat disease flares abrogate any potential benefit from ICIs?
An FDA-led retrospective analysis by Weinstock and colleagues examined 552 patients with a history of autoimmune disease enrolled on 22 clinical trials of ICIs. Autoimmune disease worsened among 6% to 16% of patients, depending on ICI therapy. Researchers reported two grade 4 cases of hyperglycemia among patients with diabetes, and grade 3 adverse events among patients with psoriasis, interstitial lung disease, hypothyroidism and ankylosing spondylitis.
Cortellini and colleagues examined real-world clinical outcomes of patients with advanced cancer and preexisting autoimmune diseases treated with ICIs. Of 751 patients, 85 had preexisting autoimmune diseases, most (82%) of which were inactive.
Incidence of any-grade immune-related adverse events was significantly higher among those with preexisting autoimmune diseases than without (66% vs. 40%), and both active and inactive cases were associated with any-grade immune-related adverse events. However, researchers observed no association between preexisting auto-immune disease and grade 3 or grade 4 toxicity, response rate, PFS or OS.
Khozin and colleagues used ASCO’s CancerLinQ database to conduct a retrospective observational study of 2,425 patients with stage III or stage IV NSCLC, 22% of whom had preexisting autoimmune disease.
Results showed no difference in median OS between those with or without autoimmune disease (11.5 months vs. 12.8 months). Researchers observed no difference in adverse events, although a subanalysis of those with active autoimmune disease revealed higher rates of select adverse events, including endocrine, gastrointestinal and blood disorders.
Overall, these data suggest clinicians should not withhold ICI treatment on the basis of preexisting autoimmune disease, even if it is active.
Due to a potential for increased flares, closer and more frequent monitoring for immune-related adverse events is required, with prompt intervention; however, it appears these patients experience clinical outcomes similar to those without autoimmune disease.
HIV and hepatitis
Significant improvements in HIV treatments have resulted in patients with HIV living with the disease for years. Cancer is the leading cause of death for HIV-infected individuals in the United States.
The two biggest drivers resulting in trial exclusion for this population are the potential for drug-drug interactions and altered immunity.
The drug interaction aspect should be based only on the pharmacological properties of the investigational therapy and concomitant drugs, not the underlying disease. Because ICIs activate the immune system, there is a concern that ICI administration may reactivate HIV.
A systematic review by Cook and Kim examined 70 HIV-positive individuals with cancer who received ICI therapy, six of whom experienced grade 3 adverse events. Among 34 with known paired pretreatment and posttreatment HIV loads, HIV remained suppressed among 26 of 28 (93%) with undetectable HIV load. Among 25 with CD4 counts, there was a mean increase of 12.3/µl. Response rates were similar to those observed among individuals with cancer who did not have HIV.
Ongoing prospective clinical trials will attempt to validate these findings but, until then, HIV-positive individuals should not be automatically excluded from ICI therapies. HIV loads and CD4 counts should be monitored at least every 3 months and in consultation with an infectious disease specialist.
Similar to HIV, those with preexisting hepatitis often are excluded from ICI clinical trials. However, data suggest PD-1 is upregulated in hepatitis C virus (HCV)-specific CD8-positive cells, suggesting that anti-PD-1 therapies may be a therapeutic target among patients with HCV.
A case report by Jang and colleagues of an HCV-positive patient with Merkel cell carcinoma who received anti-PD-1 therapy demonstrated rapid antitumor response, as well as rapid decline in HCV RNA without evidence of hepatocellular injury. A case series of seven hepatitis B virus- or HCV-positive patients with either metastatic melanoma or NSCLC — published in Melanoma Research — demonstrated similar response rates to anti-PD-1 therapies, without an increase in adverse events or virus reactivation.
Given the lack of prospective data in larger cohorts, these patients should be monitored closely for and in consultation with a hepatologist, but they should not be automatically excluded from receiving ICIs for cancer treatment.
Solid organ transplantation recipients also have been excluded from many ICI trials due to concerns of alloimmunity, organ rejection and immunosuppressive therapy. Notably, cancer is the second leading cause of death among solid organ transplantation recipients, as they are known to have increased risk for developing de novo cancer after transplant.
Abdel-Wahab and colleagues studied 39 patients who underwent allograft transplantation. Allograft rejection occurred among 41%, with median time to rejection of 21 days. Nearly one-third discontinued ICIs due to allograft rejection. Graft loss occurred among 81%, and 46% died.
Little data exist on the optimal use of ICIs for organ transplant recipients. Important factors to consider include the length of time between transplant and ICI administration, strength of immunosuppressants to prevent organ rejection, and immunogenicity of the grafted organ.
Real-world data on the use of ICIs by organ transplant recipients are needed to better understand who is at risk for poor clinical outcomes.
Administration of immunosuppressive corticosteroids before ICI initiation appeared associated with worse outcomes among patients with NSCLC.
In one study by Arbour and colleagues, 90 of 640 patients received corticosteroids (dosed at 10 mg prednisone equivalents daily) for various reasons, including dyspnea, fatigue and brain metastases. Baseline corticosteroid use was associated with decreased response rate, as well as shorter PFS and OS.
Corticosteroids often are prescribed for palliation of cancer-related symptoms independently associated with poor prognosis.
Ricciuti and colleagues investigated whether the indication for corticosteroids matters. Of 650 patients with NSCLC treated with ICIs, 93 received 10 mg prednisone equivalents or more daily.
These patients achieved shorter median PFS and OS than those who received less than 10 mg prednisone. Notably, this effect on survival was only observed among those who received 10 mg prednisone or more for palliative indications, not for those who received such a dose for reasons unrelated to cancer.
Thus, it appears the difference in survival outcomes may be driven by a poor-prognosis subgroup of patients who receive corticosteroids specifically for palliative indications.
To better understand how specific potentially vulnerable populations respond to new therapies, these subgroups should be included earlier in clinical trials rather than relying on real-world evidence and post-marketing data that become available years after approval.
Initiatives led by ASCO and others will help to broaden eligibility criteria in hopes of boosting enrollments rates and providing a clearer picture of how real-world patients will respond to novel agents.
The data highlighted in this column suggest many populations excluded from ICI registration trials likely will respond just as well as the included population.
However, there clearly are some patients — such as those with active autoimmune disease or those who recently underwent solid organ transplantation — who remain at high risk for complications, and careful consideration should be given to whether they are appropriate candidates for ICI treatments.
Prospective evaluation of high-risk patients will help answer these questions.