1 INTRODUCTION

Novel cancer therapies are an increasingly common, but important component of effective cancer treatment.¹ Over the last two decades there has been a dramatic rise in the number of novel cancer therapies, with >140 new approvals since 2000 alone.¹ Targeting specific molecular, or immune-based alterations, many of these drugs have been associated with dramatic improvements in survival.^2-5 However, emerging reports have suggested potential limitations in the wide applicability of trial findings supporting several therapies, particularly given differences in genetic, hormonal, and clinical utilization in broader populations.

Concurrently, national policies have encouraged increased focus on the representation of women in studies.^6-8 Historically, women were underrepresented in clinical trials due to misconceptions that trial observations among men were uniformly applicable to women, and fears of the impact of anticancer therapy on reproductive potential.⁹ However, sex-specific differences have increasingly been linked with pathophysiologic and clinical efficacy differences with many common therapies.^9-11 In addition, growing data support potential sex-based differences in clinical safety profiles of many novel anticancer therapies.¹²

In the US, novel anticancer therapies are reviewed by the Food and Drug Administration (FDA) for adequate representation and efficacy, with specific emphasis on supporting latter-phase clinical trials, prior to drug approval and use.¹² Incumbent on these clinical trials is the adequate representation and evaluation of target populations to allow informed assessments and decisions by providers and patients. Yet, whether women are well represented within these key anticancer clinical trials is unknown.

2 METHODS

3 RESULTS

Overall, there were 234 trials supporting 189 anticancer therapies, enrolling a total of 97,566 patients including 148 trials evaluating 60,216 patients (testing 99 discrete therapies) for non-sex specific disease(s); Supplementary Figure 1. Sex was reported in 187 (98.9%) clinical trials, of which 42,229 (43.4%) were women compared to 55,267 (55.6%) men; p < .01. Among non-sex-specific trials, 40.7% of participants were women, and the median age of study participants was 60.5 years (interquartile range: 55.0 to 63.0 years), with a median trial size of 300–400 participants. Biologic, targeted, or immune-based therapies were evaluated in the majority (79.1%) of trials. For analysis uniformity, sex was considered as the conventional binary gender assigned at enrollment. In total, women were under-enrolled in 66.9% (99 trials) of trials, including 47 trials (31.8%) with under 33% female participation, and two trials with <5% female participation. Women were underrepresented in 84% of immune checkpoint inhibitor trials. Women were consistently under-represented in gastrointestinal (gastric, liver, and pancreas; PPR = 0.76) and lung (PPR = 0.81) cancer trials; Figure 1. Outside of lymphomas, there was no over-representation of women in trial enrollment by disease type, trial phase, or cancer therapeutic class. In multivariable analysis, outside of cancer type, no specific trial characteristics were associated with the presence or absence of lower female enrollment (Table 1). There was no association between higher relative proportional cancer incidence in women and the percentage of women enrolled (odds ratio: 6.6, 95% CI 0.19 to 225.81).

3.3 Observed to expected female participation

Moreover, among the 148 trials where patients with sex-specific cancers were excluded, 2 (1.4%) did not report any female participation. There were 23,749 (39.4%) women enrolled from a total of 60,216 participants. Compared with the published incidence of 43.1%, this translated into an overall observed versus expected participation ratio of 0.91 (or 9.2% lower relative participation; p < .001; Figure 1). There was no difference in female representation by time of trial initiation; with a non-significant trend toward lower enrollment in the latter trials; Figure 2.

4 DISCUSSION

In this evaluation of pivotal clinical trials supporting the FDA approval of contemporary anticancer therapies, nearly 50% of trials did not report efficacy or safety by sex. In those trials reporting outcomes by sex, the noted rates of safety reporting were markedly low. This pattern remained, even after accounting for the presence or absence of potentially excess cancer subtype risk in women, and the use of purportedly targeted or immune-based therapies. Furthermore, among all trials, women were consistently underrepresented by nearly 10%, a pattern most amplified in gastrointestinal and lung cancer trials. There was no difference in drug efficacy or adverse event rates, by the presence or absence of higher female enrollment. This is concerning, particularly given the rapidly increasing number of perceived precision therapies and the lack of consistent tools to guide the interpretation of long-term efficacy and toxicity risk.

The observation of underreporting of sex-based efficacy in key cancer trials adds to an increasing body of evidence. In an analysis of hematologic malignancy trials, the enrollment rates of women consistently trailed males in acute leukemia and myeloma-based trials.¹⁹ Similarly, in a preliminary analysis of federally registered trials, women were underrepresented in several solid-tumor-based studies.²⁰ A similar relative absence of published sex-based efficacy analyses was observed in non-cancer populations.^{18, 21, 22} As noted by Murthy et al., the inclusion of women within pivotal cancer trials is still substandard.⁷ Although there have been at least some improvement in enrollment patterns over the last two decades, our data suggest continued efforts are needed. A comparable pattern amongst ophthalmic disease trials was also noted.²¹ Further, a similar trend was noted amongst cardiovascular disease trials wherein female participation was found to be substantially low or did not meet the targets.¹⁸ This suggests to us that even beyond cancer, other disease states may similarly see relative underrepresentation of women within pivotal clinical trials. Further, within the current analysis, we demonstrate for the first time that a similar sub-signal (relative absence of sex-based efficacy analyses) may exist in pivotal oncologic trials. Historically, clinical trials incorporate populations where disease prevalence and available subjects are prominent.²³ Our results provide significant insights into the nature of female inclusion and biologic investigation within pivotal cancer trials. Prior, more isolated, or disease-specific studies have suggested potential discrepancies in the reporting patterns of sex-based evaluation within cancer-focused trials. This is exemplified by a trial of pazopanib for renal cell carcinoma where the incidence in women is estimated to approach 37%, but where enrollment trailed expected by 22%.¹⁶ Similarly, initial reports suggested similar adverse event risk with nivolumab, an immune checkpoint inhibitor. However, although the exact risk is unclear, treatment appears to be associated with a nearly 40% relative increase for high-grade adverse event risk in women, depending on the population studied.²⁴ Despite our observation of some updates in the reporting of sex-based analyses among more recent studies, solutions for these discrepancies will require more systematic and prospective investigation. Yet, consideration for the implementation of a more diverse trial leadership structure, sex-specific patient navigators, planned assessment of sex-specific biologic effects, and the incorporation of sex-specific patient-reported outcomes may improve the capture of potentially relevant effects.

Clinicians and patients faced with difficult decisions on the initiation of potentially life-saving anticancer therapies rely on the rigor of reporting within supporting clinical trials. Often within this analysis, significant heterogeneity was observed in reporting sex-specific effectiveness. This was exemplified by the absence of sex-based safety reports within any “targeted” therapy trial, and by low rates of efficacy assessments in early immune checkpoint trials, treatment commonly marketed as precision therapies. Although some uptake was seen in the enrollment of women among more recent studies, well after the early to mid-2000s, a time when additional federal calls for gender-focused evaluations were widely circulated, many of the more recent trials evaluating therapies with apparent signals for sex differences did not report any investigation of potential efficacy or safety considerations.^{7, 25, 26} Similarly, drug companies may not be incentivized to publish potentially limiting or concerning sex-specific safety and/or efficacy data, particularly when signals are derived from underpowered (exploratory) studies.

Although the exact reasons for these observations are not clear, several plausible barriers to female recruitment and retention have been identified. Women have been underrepresented in clinical studies in part, anecdotally, due to underlying policies excluding women for reasons of reproductive health.²⁰ Whereas the intention was to protect women of childbearing age from potential adverse outcomes, the unintended exclusion of females, contributed to a gap in literature regarding drug effects in this population.²³ Even with adjustments in these policies, following recommendations by both NIH as well as FDA, data on whether outcomes are different between the sexes continues to be limited.^{25, 27} In a self-administered questionnaire study of non-cancer patients, underemployment and distrust in pharmaceutical institutions were more common among female participants.²⁶ Lymphoma, a disease wherein enhanced awareness and robust social support groups are increasingly available, saw higher female proportional enrollment, suggesting these strategies may be advantageous.²⁸ Childcare commitments and time away from work have been cited as barriers to participating in clinical trials in younger women.^{29, 30} Similarly, women of childbearing age may be excluded from participation due to concerns of potential harm. The fear of adverse effects on women's health was also cited as a reason for refusal to participate in trials.³¹ Moreover, in several countries, women may exhibit less self-autonomy in decisions to enroll in research, as frequently women are required to seek permission from a husband or parent.³² These barriers have been recognized to drive disproportionate enrollment of women in other populations. However, with cancer trials, elucidation of these barriers may require additional community-based and prospective studies.

While the reasons for the underreporting of sex-based efficacy and safety are not well understood, there are some potential theories to explain why. In a 2010 Society for Women's Health Research evaluation of 11 major journals, only 18.2% (two) had a requirement for sex-based data reporting. Similarly, editorial discomfort may have contributed to the relative deemphasis of sex-based differences, particularly where patterns were not clear.³³ Additionally, some investigators may have elected not to publish underpowered sex-based safety or efficacy data if the data were more difficult to interpret. Accordingly, it may be reasonable to suggest that trial outcomes be disaggregated by sex when larger cohort sizes are achieved (e.g., >300 participants). Further, it may be reasonable to mandate the publication of sex-based efficacy for latter-phase clinical trials by medical journals. Beyond measures that have already been initiated since the landmark publication by Murthy et al.,⁷ it is imperative to consider policy benchmarks wherein the publication of any larger, phase II, III, or IV clinical trial (e.g., >200 enrolled subjects), or trials linked to the FDA's decision on oncologic drug approval, should be mandated to publish or make available online their sex-based efficacy data, even if exploratory. Similar to NIH's policy for all studies to include consideration of sex as a biologic variable, we also suggest a relative mandate by the journal for the publishing (or online availability) of sex-based efficacy and safety outcomes data supported by NIH funding (to further support broader implementation of these goals and standards). Yet, given the growing appreciation of the consequences of differences in safety profile with many novel targeted and immune therapies by sex, added priority to stratified data presentation is warranted.³⁴ Despite efforts like the NIH Revitalization Act, which encourages adequate inclusion of women in clinical studies, challenges persist regarding the effectiveness of this reform.^{35, 36} Equitable enrollment of women is critical to elucidate potential biological differences, as has already been suggested in several fields of medicine. The lack of sex-based analysis and reporting in clinical studies undermines the ability to adequately deliver more precision-based care. This specifically hampers the development and implementation of targeted precision medicine approaches,³⁷ while also potentially masking important sex differences in treatment response, toxicity, and adverse effects. Considering the potential broad public health implications (e.g., potentially lower efficacy or safety in women, data transparency, incomplete knowledge base), improved strategies are needed.

5 CONCLUSION

In summary, among clinical trials supporting contemporary cancer therapies, women were frequently underrepresented, and sex-based analyses were commonly not performed. This pattern remained even after accounting for trial characteristics, including the time of trial initiation. Given the potentially serious impacts of inadequate clinical and safety information, increased focus on the consistency and systematic nature of female enrollment and sex-based reporting is needed.

AUTHOR CONTRIBUTIONS

All of the authors had full access to all of the data in the study, reviewed drafts of the manuscript, and approved the final version. Concept and design: MAB, JMB, and DA. Acquisition, analysis, or interpretation of data: SuK, SG, BeO, MAB, JMB, and DA. Drafting of the manuscript: all authors. Critical revision of the manuscript for important intellectual content: All authors. Statistical analysis: SuK, EM, MAB. Administrative, technical, or material support: DA. Supervision: DA. The work reported in the paper has been performed by the authors, unless clearly specified in the text.

FUNDING INFORMATION

This work was supported in part by National Cancer Institute (NCI) Grant (P30 CA016058); as well as by National Institutes of Health (NIH) (K01HL142848 [KB], R01HL159216 [KB], R56HL159216 [KB], and L30HL148881 [KB], KL2-TR002733 [JB], K12-CA133250 [DA], K23-HL155890 [DA], R01HL170038 [DA], R01HL168045 [DA]). Dr. Breathett also receives funding from the Health Resources and Services Administration (HRSA). Dr. Carter was supported in part by charitable contributions from the Mary H. and J. Churchill Hodges Clinical Prevention Program Fund at the Ohio State University Wexner Medical Center. Mr. Kola-Kehinde was supported by Ohio State University Comprehensive Cancer Center's Pelotonia grant funds. Dr. Paskett has received research funding from the Breast Cancer Research Foundation, and The American Cancer Society. Dr. Addison was also supported by a Robert Wood Johnson Foundation (Harold Amos)- American Heart Association grant. Support also received from a 2022 American Society of Clinical Oncology (ASCO) Merit award (Dr. Babcook). All other authors have reported no relationships relevant to the contents of this paper to disclose. The funder did not play a role in the design of the study; the collection, analysis, and interpretation of the data; the writing of the manuscript; and the decision to submit the manuscript for publication.

CONFLICT OF INTEREST STATEMENT

Dr. Paskett received research funding from NCI, Merck Foundation, Pfizer, Genentech, Guardant Health and Astra Zeneca and is an Advisory Board member for GSK and Merck. The other authors have no conflict to declare.