Impact of Organ Function–Based Clinical Trial Eligibility Criteria in Patients With Diffuse Large B-Cell Lymphoma: Who Gets Left Behind?

Abstract

PURPOSE

Exclusion of patients needing urgent treatment or requiring novel biomarkers before enrollment has impacted the ability to enroll real-world patients in frontline trials of diffuse large B-cell lymphoma (DLBCL). The impact of baseline organ function–based eligibility criteria on this effect and clinical trial exclusion is less well-understood.

METHODS

Consecutive patients with newly diagnosed lymphoma were enrolled from 2002 to 2015 into the Molecular Epidemiology Resource (MER) of the University of Iowa and Mayo Clinic Lymphoma Specialized Program of Research Excellence. The current analysis includes 1,265 patients with DLBCL receiving standard immunochemotherapy. Organ function parameters were identified from criteria for hemoglobin, absolute neutrophil count, platelet count, creatinine clearance, and bilirubin, as reported in frontline DLBCL trials. Abstracted laboratory values from MER were used to determine the percent (%) of patients excluded. Outcomes and cause-of-death analyses comparing ineligible and eligible groups in MER were conducted. An interactive online tool was developed to estimate exclusions based on organ function for future trial design.

RESULTS

Between 9% and 24% of MER patients with DLBCL receiving standard immunochemotherapy were excluded on the basis of baseline organ function alone. Ineligible patients based on organ function had significantly inferior event-free survival (hazard ratios, 1.67-2.16), overall survival (hazard ratios, 1.87-2.56), and event-free survival at 24 months (odds ratio, 1.71-2.16). Ineligible patients were more likely to die from lymphoma progression than increased therapy-related complications.

CONCLUSION

Current national and international trials exclude up to 24% of patients from participation on the basis of organ function alone. A significant difference in the outcomes, notably lymphoma-related death, suggests issues with generalization and potential exclusion of high-risk patients. These data will help future clinical trial development and meet US Food and Drug Administration and ASCO recommendations to increase trial accrual.

INTRODUCTION

There is a critical need to enroll patients in clinical trials to improve patient outcomes and accelerate the drug development process. However, only approximately 3%-5% of adult patients with cancer in the United States participate in clinical trials.^1-3 Trial ineligibility was identified as a reason for nonparticipation in 21.5% (10.9%-34.6%) of patients in a recent meta-analysis.⁴ About two-thirds of clinical trial exclusion criteria pertain to patient's comorbidities, performance status, and overall health.^5,6 Some degree of exclusion in clinical trials is necessary to prevent drug-related toxicities and maintain patient safety. However, patients with organ dysfunction can be excluded from clinical trials, regardless of the specific metabolism of study drugs, and then treated off-study with similar standard regimens. The stakeholders from ASCO, Friends of Cancer Research, and the US Food and Drug Administration (FDA) recommended modernizing criteria related to baseline organ function and comorbidities, which are routinely used to exclude patients from cancer clinical trials.⁷ The goal of these recommendations was to make clinical trials more representative of the real-world population.

CONTEXT

• Key Objective

• The impact of baseline organ function–based eligibility criteria on clinical trial exclusion in diffuse large B-cell lymphoma (DLBCL) is unknown. We studied its effect on patient exclusion and its implications in a real-world DLBCL patient cohort.

• Knowledge Generated

• Up to 24% of patients were excluded from trial participation on the basis of organ function alone. Ineligible patients based on organ function had significantly inferior event-free survival and overall survival and were more likely to die from lymphoma progression than therapy-related complications.

• Relevance

• This study provides a better understanding of patient exclusion on frontline trials in DLBCL based on organ function eligibility criteria and calls for improvements in study design to improve generalization and avoid high-risk patients being left behind by current eligibility criteria.

Newly diagnosed diffuse large B-cell lymphoma (DLBCL) has been an active area of research with numerous randomized phase III clinical trials over the last 15 years. These studies have largely failed to improve outcomes in this setting.^8-11 Recently, trials have often required central pathology review and/or molecular phenotyping, which can delay trial enrollment.¹² Additionally, a perceived need for urgent treatment of some patients who are quite ill may have resulted in the nonenrollment of patients with more aggressive disease characteristics.^13,14 Exclusion of patients with DLBCL on the basis of organ function and comorbidities and its implications on the generalizability to the real-world population is less well- studied. The goal of this study was to evaluate the impact of laboratory-based organ function eligibility criteria on trial exclusion and its impact on outcomes from a well-established, prospective observational cohort of newly diagnosed patients with DLBCL treated with standard immunochemotherapy (IC).

METHODS

Patients were identified from the Molecular Epidemiology Resource (MER) of the University of Iowa/Mayo Clinic Lymphoma Specialized Program of Research Excellence (SPORE).¹³ This SPORE, from 2002 onward, has offered consecutive patients with newly diagnosed DLBCL (within 9 months of diagnosis) enrollment into the MER. Patients provided informed consent for participation in the SPORE and the use of their data for future research. These patients were managed per treating physician and contacted prospectively and systematically every 6 months for the first 3 years and then annually thereafter. Baseline clinical, laboratory, and pathological data were abstracted using a standard protocol; all outcome events (including patient-reported) were validated by review of medical records.

The current study includes patients with a new diagnosis of DLBCL diagnosed between March 2002 and June 2015 who received frontline R-CHOP or R-CHOP–like IC. Baseline organ function parameters were identified from recent clinical trials (Data Supplement, online only) in DLBCL. Recently reported phase II and III trials in DLBCL were selected as more stringent organ function criteria are required in the phase I dose-defining studies. The parameters included hemoglobin, absolute neutrophil count (ANC), platelet count, creatinine clearance (CrCl), and bilirubin. The cutoff values for different parameters reported in the clinical trial eligibility criteria were identified. Hemoglobin cutoff was not mentioned in four of the seven trials, so it was assumed not to affect exclusion. Each clinical trial that added a new drug (X) to the R-CHOP chemotherapy backbone was diligently reviewed to identify if any criteria were inappropriate based on the pharmacokinetic properties of the drug under evaluation or if any specific adverse events (AEs) were observed in the trials.

Organ function values were obtained via standard MER abstraction from the medical record and clinical laboratory value databases. Patients missing three or more of the five organ values were excluded from this analysis (flow diagram, Fig 1). The percentage of patients excluded on the basis of clinical trial criteria was determined for each organ value individually as well as across all organ values. Event-free survival (EFS) was defined as the time from diagnosis to relapse, progression, retreatment (second-line therapy), or death because of any cause. Overall survival (OS) was defined as the time from diagnosis until death because of any cause. EFS was reported at 24 months (EFS24), as previously described.¹⁵ Kaplan-Meier curves and Cox models were used to compare EFS and OS outcomes between eligible and ineligible patients. Logistic regression was used to compare EFS24 between eligible and ineligible patients. An interactive tool was developed in R-Shiny to allow users to estimate the percentage of patients who would be excluded by changing organ function cutoffs. All analyses were performed using R version 3.6.2, R-Shiny, and SAS version 9.4M5.¹⁶

FIG 1.

Flow diagram showing the patient selection process in the study.

RESULTS

A total of 1,265 newly diagnosed patients with DLBCL treated with R-CHOP or R-CHOP–like IC between 2002 and 2015 in the Mayo Clinic and University of Iowa Lymphoma SPORE database comprised the analysis data set. The baseline characteristics of the patients in the cohort are shown in Table 1. The median age of patients was 63 years (IQR, 52 to 71 years). A total of 123 patients (9.7%) were treated on various frontline clinical trials available at the time of presentation. At a median follow-up of 84.3 months (IQR, 58.9 to 120.7), 552 patients (43.6%) had an event and 440 patients (34.8%) died. EFS24 by Kaplan-Meier test was 71% (95% CI, 68.4 to 73.5).

TABLE 1.

Baseline Characteristics of Patients From MER Cohort Included in the Analysis

Impact of Organ Function–Based Criteria on Trial Exclusion in MER Cohort

Organ function–based criteria from recent DLBCL trials were identified (Table 2) and applied to the MER cohort. Between 9% and 24% of the MER patients were ineligible based on organ function (Table 3). Lenalidomide added to R-CHOP was evaluated in two trials, a randomized phase II ECOG 1412 and the phase III ROBUST.^11,17,18 Patients percent exclusion based on the criteria from the ROBUST trial was 10.0%. The addition of bortezomib to the R-CHOP backbone was evaluated in the phase III REMoDL-B trial.^10,19 Percent exclusion from the MER cohort was the lowest at 9.2%, with platelet criteria alone causing 4.7% of the exclusion. The GOYA trial compared R-CHOP with obinutuzumab (G) plus CHOP in patients with DLBCL.²⁰ Its criteria led to the exclusion of 15.9% of patients in the MER cohort. The hemoglobin of > 9 g/dL excluded 6.3% patients. Enzastaurin, an oral inhibitor of the PKC-β/PI3K/Akt pathway, is currently under evaluation for DLBCL frontline treatment in combination with R-CHOP in the phase III ENGINE study.²¹ Percent exclusion was highest at 24.1% in the MER cohort, mostly because of higher hemoglobin > 10 g/dL and CrCl > 50 ml/min criteria. The hemoglobin criteria alone caused the exclusion of 12.7% of patients.

TABLE 2.

Organ Function–Based Exclusion Criteria in Various Frontline DLBCL Trials

TABLE 3.

Percent Estimation of Patient Exclusion From MER Cohort Based on the Trial-specific Organ Function–Based Exclusion Criteria

Outcomes Analysis

The exclusion of patients based on organ function can also have potential implications on estimating the outcome of control patients, a key component of the trial biostatistical calculations. To further appreciate this impact of organ function–based exclusion, we compared outcomes and cause of death in the MER cohort on the basis of the eligibility criteria from different trials. Patients in the MER who were ineligible by trial criteria had significantly inferior EFS (hazard ratios [HRs], 1.67-2.16, Data Supplement), OS (HRs, 1.87-2.56, Data Supplement), and EFS24 (odds ratios [ORs], 1.71-2.16) (Table 4, Data Supplement). These results attenuated after adjusting for IPI (EFS [HRs, 1.19-1.61], OS [HRs, 1.30-1.89], and EFS24 [ORs, 1.13-1.52]) but remained clinically relevant, accentuating the impact of omission of these patients on the trial outcomes (Table 4, Data Supplement). Notably, patients who were trial-ineligible had a significantly increased risk of dying from progressive lymphoma (Data Supplement), whereas therapy-related deaths were similar between organ function–eligible and organ function–ineligible patients.^22,23 The ENGINE study had the most stringent criteria and maximum percent exclusion and is thus chosen here as an example. The EFS24 was 75% (95% CI, 72 to 78) in the eligible versus 60% (95% CI, 55 to 66) in the ineligible group (Fig 2A). Five-year survival was 68% (95% CI, 65 to 71) versus 48% (95% CI, 42 to 54) (Fig 2B) and the risk of death from progressive disease at 5 years was 15% (95% CI, 13 to 17) versus 25% (95% CI, 21 to 31) (Fig 2C).

TABLE 4.

HRs (95% CI) for EFS and OS in MER Patients Ineligible Versus Eligible (Reference) Based on Trial-Specific Eligibility Criteria

FIG 2.

ENGINE trial (A) OS, (B) EFS, and (C) cause of death between eligible (solid line) and ineligible (dotted line) MER patients. EFS, event-free survival; MER, Molecular Epidemiology Resource; OS, Overall survival.

Sensitivity Analyses

We performed a sensitivity analysis in the subset of MER patients who did not enroll in frontline trials (N = 1,142). Patients treated off-trial were slightly older (median age 63 v 61 years), more frequently had ECOG PS of 2-4 (18% v 6%), and had a slightly higher percentage of patients who failed to achieve EFS 24 status (29.6% v 23.6%; Data Supplement). Results were similar in this subset with the trial ineligibility based on organ function ranging from 10% to 25% (Data Supplement); similar differences in outcome were observed compared with the full cohort (data not shown).

Online tool.

An interactive tool was developed in R-Shiny to estimate the percentage of patients who would be excluded from trials on the basis of various organ function cutoffs (Data Supplement). The tool can be found at ref. 24.

DISCUSSION

Our study highlights the impact of baseline organ function in patients with DLBCL on EFS and OS and the contribution of each parameter in the clinical trial exclusion. This study also shows that patients with baseline organ dysfunction are at a higher risk of dying from lymphoma and, as such, comprise a group of patients who may benefit from different treatment approaches or special treatment arms. To our knowledge, these data have not been previously described and will be useful in future trial design. We also found that most trials in DLBCL have standard eligibility criteria for baseline organ function regardless of the drugs under evaluation.

Renal function in trials such as the ENGINE study was set at a cutoff value of > 50 mL/min when the drug enzastaurin predominantly undergoes hepatic metabolism via CYP3A and does not require renal elimination.²⁵ Obinutuzumab, a glycoengineered, type II, anti-CD20 monoclonal antibody, does not require any dose adjustments for renal or hepatic dysfunction. In the GOYA trial, the renal function alone excluded approximately 5.2% of MER patients. Bortezomib undergoes hepatic metabolism for clearance via CYP2C19 and CYP3A4 and does not require any dose modifications for renal impairment.²⁶ In the REMoDL-B trial, the renal function–based exclusion was limited to 2.0%, and there was no significant increase in grade 3 or higher neutropenia, febrile neutropenia, thrombocytopenia, or anemia. Regardless of higher treatment discontinuation rates in the experimental arm (13.1% v 9.4%), overall, a higher proportion of patients completed all six cycles of the treatment in both arms (91.3% in RCHOP v 87.1% in bortezomib-RCHOP).¹⁰ Lenalidomide is cleared primarily via excretion in the urine and has dose modifications for renal impairment, including dialysis patients, where the half-life of the drug is increased 4.5-fold.²⁷ Renal function criteria could have been more permissive in the trials with provisions for appropriate dose adjustments. The impact of renal function on trial exclusion is significant. Renal function excluded approximately 20.3%-45.9% of patients with cancer from participating in clinical trials if the CrCl was < 60 ml/min in a study evaluating more than 12,000 patients conducted by ASCO and FDA.⁷ Our study used the CrCl criteria and estimated exclusion of 2.0%-10.5% of MER patients with DLBCL from frontline clinical trials based on CrCl alone.

Hemoglobin when used as exclusion criteria with cutoff values of 9 or 10 g/dL significantly limited trial participation in some trials. No specific information regarding increased transfusion requirements or significant differences in anemia-related AEs compared with the R-CHOP arm was mentioned. In real-world practice, dose modifications are usually based on ANC or thrombocytopenia rather than hemoglobin alone, and patients are generally supported with blood transfusions or erythroid-stimulating agents. The most common grade 3-4 AEs (≥ 10%) were neutropenia (60% v 48%) and febrile neutropenia (14% v 9%) in the ROBUST study. This was also the most common reason for treatment discontinuation.¹¹ Similarly, in the ECOG 1412 trial, higher rates of grade ≥ 3 AEs were febrile neutropenia (25% v 12%) and thrombocytopenia (36% v 12%) in the R2CHOP arm versus R-CHOP arm, respectively. Despite this, 86% and 85% of the patients completed six cycles of treatment in the R2CHOP and RCHOP arm, respectively.¹⁸ Higher grade 3-4 thrombocytopenia was also reported in ECOG 1412, but it did not result in higher bleeding events. The hemoglobin cutoff of > 9 g/dL alone in the GOYA study excluded 6.3% of the patients in the MER cohort. Anemia ≥ grade 3 AE, however, was similar in both arms at 7.2 % versus 7.5% in G-CHOP and R-CHOP arms, respectively.²⁰ In the ENGINE study, the hemoglobin criteria alone caused the exclusion of approximately 12.7% of patients. Enzastaurin has no significant myelosuppression and minimally overlaps with the nonhematologic toxicities of R-CHOP.^28,29 In phase II study, similar rates of neutropenia and thrombocytopenia were noted, with slightly higher rates of anemia (11% in enzastaurin + RCHOP v 5% in R-CHOP) and febrile neutropenia (enzastaurin + RCHOP 16% v R-CHOP 5%). The toxicity profile was found to be extremely favorable, perhaps influenced by the selection criteria.^28,29

Outcomes analysis between eligible and ineligible groups from MER showed a statistically significant and clinically meaningful difference in OS and EFS across all the trials (Data Supplement). The odds of failure to achieve EFS24 status was up to 1.5 times higher in the ineligible group when adjusted for IPI. The cause-of-death analysis highlighted an increased risk of death because of progressive lymphoma in the organ function–based ineligible group in the MER.^22,23,30 This suggests that patients excluded from trials because of organ function comprise a high-risk group with poor disease-related outcomes with the standard-of-care therapy. These patients have an urgent need of novel therapies than a typical patient with DLBCL. Another consequence of their exclusion is that these approved regimens are then used in the real world on the basis of incomplete evidence of expected side effects and their management.

Recruitment to clinical trials remains an issue for all adult cancer types, including NHL. Loh et al¹⁴ evaluated the reasons for nonenrollment of patients with DLBCL in clinical trials and found an enrollment rate of only 18%. The reasons were that 37.5% did not meet the eligibility criteria and 17% required urgent treatment and could not wait to be registered.¹⁴ Additionally, the days from diagnosis to treatment initiation (DTI) is an important prognostic factor. Patients with DTI of < 14 days were discerned to have more aggressive disease features and worse EFS at 24 months.¹³ This suggests that higher-risk patients are treated off-study because they cannot wait until registration. Our study extends this observation, showing that patients excluded on the basis of organ function criteria could have higher-risk disease characteristics as they had a significantly higher risk of dying from progressive lymphoma than therapy-related deaths. Whether the organ function derangements that led to the patient exclusion in this study are because of high-risk disease or frailty needs further clarification. Therefore, modernizing eligibility criteria remains complex; however, several attempts are being made in the direction.

One approach would be to allow temporizing measures such as high-dose steroids with or without rituximab to reverse disease-related organ function derangements before trial enrollment. Second, designing trials that could potentially allow for one cycle of standard R-CHOP before initiating cell of origin, mutation, or other biomarker-based treatment. This was successfully accomplished in REMoDL-B and AMC-075 trial.^10,31,32 Third, organ function–based criteria should be carefully reviewed and tailored to the drug being tested. Fourth, the recruitment of patients with organ dysfunction onto separate arms of clinical trials with special provisions for monitoring and supportive care should be allowed. Nonmodifiable factors such as patient age, disease stage, and lack of trial availability at a treatment location cause considerable impediment to clinical trial enrollment. Focusing on modifiable factors such as eligibility criteria and innovative trial design will be essential for improving trial enrollment.

We estimated the exclusion from DLBCL trials of MER patients treated off-study with standard IC based solely on organ function eligibility criteria to be 9%-24%. These rates are consistent with a recently conducted meta-analysis.⁴ Based on this study's observations, we have developed an interactive application that estimates patient exclusion based on different parameters set in the clinical trial design. We hope that this will help design trials with eligibility criteria pertaining to the drug under evaluation and facilitate the inclusion of high-risk populations. The strengths of this study include a large, well-studied prospective patient cohort enrolled at academic medical centers, which are representative of patients considered for clinical trials. We evaluated seven different frontline DLBCL trials and observed consistent results on EFS and OS. Limitations include the clinical parameters in the observational MER cohort that were captured as performed under routine care. The laboratory parameters used for estimation of exclusion were abstracted from MER before initiation of treatment. There may have been changes in these clinical variables between diagnosis and treatment initiation. The assessment of the effect of this variation is beyond the scope of the current study. Disease assessment was not per a defined protocol as on a clinical trial, and the EFS end point in MER may not strictly align with the PFS definitions on the clinical trials we examined. Additionally, our study does not provide details on the frequency of dose modifications or delays between the two groups. Future studies to evaluate the differences in chemotherapy dosing between eligible and ineligible groups are warranted. Finally, the MER cohort is limited in racial and ethnic diversity (Caucasians 93.8%), and the trial exclusion based on laboratory parameters may not be directly applicable to other groups.²² Validation of these results in the larger and more diverse LEO Cohort Study (ClinicalTrials.gov identifier: NCT02736357) is planned.

In conclusion, eligibility criteria have a considerable influence on clinical trial participation and generalizability. Up to a quarter of patients in the population may be excluded because of organ function alone before applying other trial criteria yet still receive standard IC off-study. The eligibility criteria have a significant impact on trial exclusion and generalizability in patients with DLBCL. Innovative trial designs are required to capture a broader patient population and tailor the trials to both the patient and drug under evaluation. This study provides a better understanding and estimation of patient exclusion on frontline trials of DLBCL based on organ function eligibility criteria and calls for improvements in study design to improve generalization and avoid high-risk patients being left behind by current eligibility criteria.

AUTHOR CONTRIBUTIONS

Conception and design: Arushi Khurana, Raphael Mwangi, Grzegorz S. Nowakowski, Thomas M. Habermann, James R. Cerhan, Matthew J. Maurer, Thomas E. Witzig

Administrative support: Brian K. Link, James R. Cerhan

Financial support: James R. Cerhan

Provision of study materials or patients: Grzegorz S. Nowakowski, Thomas M. Habermann, Brian K. Link, James R. Cerhan

Collection and assembly of data: Arushi Khurana, Raphael Mwangi, Grzegorz S. Nowakowski, Thomas M. Habermann, Brian K. Link, James R. Cerhan, Matthew J. Maurer, Thomas E. Witzig

Data analysis and interpretation: Arushi Khurana, Raphael Mwangi, Grzegorz S. Nowakowski, Thomas M. Habermann, Stephen M. Ansell, Betsy R. LaPlant, Brian K. Link, James R. Cerhan, Matthew J. Maurer, Thomas E. Witzig

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

Impact of Organ Function–Based Clinical Trial Eligibility Criteria in Patients With Diffuse Large B-Cell Lymphoma: Who Gets Left Behind?

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/op/authors/author-center.

Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).