Analyses of the Rationale and Implementation of Research Biopsies in Oncology Clinical Trials at a Tertiary Cancer Center

Abstract

Background

Biomarkers in clinical trials have led to massive incorporation of research biopsies, with potentially risks and no direct benefit for patients. In 2018, the American Society of Clinical Oncology (ASCO) released an ethical framework to provide guidance on incorporating research biopsies in cancer clinical trials.

Materials and Methods

We collected biopsy requirements of cancer clinical trials conducted at Institut Jules Bordet (IJB) between 2015 and 2019 to examine adherence with the ASCO Ethical Framework. We used logistic regression models to test the association between the request for biopsy, the request for tissue, and the adherence to the ASCO framework as well as some trial characteristics.

Results

Between January 2015 and December 2019, 178 oncological studies were conducted at IJB. Of these, 138 (78%) were sponsored by industry, 132 (74%) were phase II and III studies, and 141 (79%) concerned metastatic disease. Tissue was required for inclusion for 119 (67%) studies, among which 59 required at least one new biopsy. Adherence to ASCO's Ethical Framework was 67% for studies requiring tissue and went down to 39% for studies requiring at least one new biopsy. In multivariate analysis, requests for tissue or new biopsies increased in early-phase studies (p < .001, p < .001, respectively) and in studies investigating innovative treatments (immunotherapy or targeted therapies; p < .01, p = .02). Compliance to the ASCO framework significantly decreased with time (p < .001) and in early-phase studies (p < .001).

Conclusion

Numerous studies required tissue or new biopsies for exploratory objectives of unknown clinical utility. Requests for tissue increased over the years, whereas compliance to ASCO's Ethical Framework decreased.

Implications for Practice

In 2019, the American Society of Clinical Oncology (ASCO) developed an ethical framework to provide guidance on incorporating research biopsies in clinical trials. This study underlines the growing request for tissue in clinical trials with potentially no impact on drug development and no benefit to actual or future patients. Adherence to ASCO's Ethical Framework decreases through time. These results highlight the importance of improving the ethics of research biopsies. ASCO's Ethical Framework offers an opportunity to improve quality of care in clinical research by maximizing scientific utility and allowing for clinically meaningful correlative science and safe access to innovative treatments for a maximum number of patients.

Introduction

With the development of precision medicine and the advent of clinical trials testing targeted therapies and immunotherapies, biomarkers and the need for nondiagnostic biopsies have increased steadily in the past few years [1]. Patient-driven correlative science as a consequence of research biopsies is used in order to improve characterization of the pharmacodynamics effects of treatments and identify prognostic and predictive biomarkers with a goal to improve patients’ outcomes. In December 2018, the Food and Drug Administration published a guidance framework for biomarker qualification and development [2] based on the current drug development landscape and medical need, biomarker category biomarker test characteristics, and statistical considerations in order to improve the reliability and the relevance of biomarkers. Despite the optimism surrounding biomarkers, only few studies reported a clinically relevant biomarker result. Freeman et al. [3] reported that more than 40% of published early-phase studies did not include or included incomplete pharmacodynamics data. Another study demonstrated that a biomarker highlighted the assumed mechanism of action only in 39% of cases [4]. The request for tissue is an important limitation to access to clinical research and subsequently innovative drugs. Noninclusion of potential patients because of nonbiopsiable lesions, for example, in bone-only metastatic diseases, is problematic when the biopsy is aimed toward exploratory objectives. Such situations raise the issue of clinically meaningful inclusion criteria [5, 6]. Moreover, the necessity for a mandatory biopsy also implies a delay in treatment because most biopsies are demanded during a screening period in potentially symptomatic patients [7].

Whereas the clinical utility of tested biomarkers is often missing, the risk and/or burden for the patient undergoing a biopsy is not negligible [1]. Whereas biopsies of easily accessible sites such as skin and lymph nodes are safe, biopsies of organs such as lung or liver carry a higher risk, with between 4% and 6% of major complications described for lung biopsies [8] and 6% of complications, including less than 2% of severe complications, for liver biopsies [9]. Furthermore, informed consent forms (ICFs) often fail to provide the patient with accurate information about the nature and lack of benefit of research biopsies [10]. To compensate this unmet medical need, the American Society of Clinical Oncology (ASCO) has released an ethical framework [11] to provide guidance on incorporating research biopsies in clinical trials. These guidelines assign utility to biopsies according to the biomarkers’ potential to increase scientific knowledge, participants’ risk, and mandatory or optional nature of the biopsy. The ultimate goal is to maximize scientific utility, minimize participant risk, and improve oversight.

In this study, we retrospectively describe the characteristics of oncology clinical trials conducted between 2015 and 2019 within the Clinical Trials Conduct Unit at Institut Jules Bordet (IJB), a comprehensive cancer center in Brussels, in order to assess the compliance with the ASCO Ethical Framework. The ultimate goal is to encourage using the ASCO Ethical Framework as a backbone to design future clinical trials in which research biopsies would have a stronger scientific rationale, a higher expected utility that would justify the risk undertaken by patients.

Materials and Methods

The study was approved by the ethics committee of IJB. Consent was not applicable as no patient data were needed.

Identification of Trials

We searched within the IJB clinical trial database all oncology clinical trials conducted from January 1, 2015, to December 31, 2019. All trials included had been approved by the ethics committee of IJB. We identified 178 oncological studies conducted at IJB between 2015 and 2019.

Study Characteristics

Data were extracted from the center's clinical trials database in which details concerning the protocol and ICFs of all studies are available. Data were extracted by two assessors in one file with predefined variables. Potential inconsistencies were put aside for discussion among involved clinicians. Studies were characterized by type (primary tumor) and state (localized or metastatic) of cancer, phase (early vs. other), experimental treatment type (immunotherapy and targeted therapy vs. others), year of approval by ethical committee, and sponsorship (academic or industry). Early-phase studies included phase I, phase I/II, and phase I/III studies. Non-early-phase studies included phase III, phase II, phase II/III, observational, medical need program, and translational research studies. Observational studies were considered as non-early-phase studies as their purpose was to assess therapies already licensed or to describe clinical characteristics in a specific population. Immunotherapy and targeted therapies have been designated as “innovative treatments” for the analysis.

Identification of the Need for Biopsies and Biopsy Characteristics

We distinguished the studies that required at least one new biopsy or tissue only. Studies that required at least one new biopsy correspond to studies that requested a new biopsy for inclusion independently of the availability of archival tissue. Studies requiring tissue refer to studies for which archival tissue was needed for inclusion. If a study requested a new biopsy only in case of unavailability of archival tissue, the study was classified as studies requiring tissue only.

Evaluation of Tested Biomarkers

The evaluation of tested biomarkers was performed for studies requiring tissue and/or new biopsies. Biomarker analysis included type and utility of biomarkers considering the ASCO Ethical Framework. Integral biomarkers included biomarkers needed to assess eligibility and biomarkers that were part of the primary objective. Nonintegral biomarkers included biomarkers necessary for secondary objective analyses and exploratory objectives. We then assessed biomarkers’ utility considering the ASCO Ethical Framework as follows: biomarkers of expected utility encompassed biomarkers supported by a strong hypothesis with adequate statistical power, biomarkers of potential utility referred to biomarkers test including clear scientific plans to generate knowledge, and finally, biomarkers of unknown utility were biomarkers with unknown potential for advancing scientific knowledge.

Biopsy Risk for Participants

Biopsies were classified in three risk levels: low risk if the risk of severe complication is under 0.5% (breast, skin or superficial tissue, bone marrow), moderate risk if it is between 0.5% and 1.5% (intra-abdominal, muscular biopsy), and high risk if it is over 1.5% (intrathoracic and liver biopsy mostly).

We searched the clinical trial protocols for information concerning the risk level accepted in the study. If no information was specified in the protocol, we considered that high-risk biopsies were not excluded.

Compliance to the ASCO Ethical Framework

We assessed the compliance with the ASCO Ethical Framework by including the utility of biomarkers, the risk for patients, and the mandatory or optional nature of the biopsy. Low-risk biopsies were acceptable in all cases. Moderate-risk biopsies were acceptable for biomarkers of potential and expected utility. High-risk biopsies were tolerated only for biomarkers of expected utility.

Informed Consent Form Analysis

We searched within the informed consent forms for the presence of sufficient information about biopsy risk, benefit, and scientific rationale delivered to patients.

Statistical Analysis

Statistical analyses were performed to highlight the association between trial characteristics and the request of tissue and new mandatory biopsies. When material was requested, the association with the compliance to ASCO recommendations was investigated. The selected covariates were the year of approval of the study, the type of sponsor, the trial phase, and the type of experimental treatment. To test for the presence of association, logistic regression models were used for modeling the request for biopsy. After univariate analysis, multivariate analysis was performed using a backward selection for selecting the final model. Odds ratios (ORs) and 95% confidence intervals (CIs) are reported. Two-tailed p values <.05 were considered as showing a significant association between covariates and the request for tissue or new biopsies.

Results

General Characteristics

One hundred seventy-eight oncological studies conducted at IJB between January 2015 and December 2019 were analyzed. Of those, 138 (78%) were sponsored by industry, 132 (74%) were non–early-phase studies, 141 (79%) concerned metastatic disease, and 114 (64%) investigated an immunotherapy or targeted therapy. Additional study characteristics can be found in Table 1.

Of 178 studies, 119 required tissue (67%; 95% CI, 59%–74%) and 59 required at least one mandatory biopsy (33%; 95% CI, 26%–41%), as summarized in Figure 1.

Evaluation of the Utility of Biomarkers and Adherence to the ASCO Ethical Framework

Among 119 studies requiring new or archival tissue, 52 studies (44%) required tissue for integral biomarkers and 67 (56%) for nonintegral biomarkers. Biopsies were of expected utility in 63 (53%) studies, potential in 12 (10%), and unknown in 44 (37%) studies.

Among 59 studies requiring at least one new biopsy, 34 (58%) required only one biopsy whereas 23 (39%) required two biopsies and 2 (3%) up to three mandatory biopsies. Such biopsies were for integral biomarkers in 19 studies (33%) and for nonintegral biomarkers in 40 studies (68%). Biopsies were of expected utility in 36 (61%) studies, potential in 8 (14%), and unknown in 15 studies (25%). Most mandatory biopsies were required at screening (73%) and acceptable participants’ risk was not clearly stated in protocols in more than 80% of studies.

Overall, adherence to the ASCO Ethical Framework was 67% (n = 80) for studies requiring new or archival and went down to 39% (n = 23) when considering studies requiring at least one mandatory biopsy. Of note, among studies requesting more than one mandatory biopsy (n = 25), 84% were not in accordance with ASCO's Ethical Framework. Further details can be found in Table 2.

Biopsy Risk for Patients and Informed Consent Analysis

Biopsy-related risks were frequently specified in the ICF, especially for studies requiring a new biopsy (70%), but only mentioned in 13% of matched protocols. The rationale was described but the statistical analyses concerning biomarkers were lacking in 29% of studies requiring at least one new biopsy. Further details can be found in Table 3.

Association Between Trial Characteristics and Request for Tissue (Archival or New Biopsy)

Of 178 trials, 119 requested material, leading to a rate of 67% (95% CI, 59%–74%). In univariate analysis, the request for new or archival tissue was associated with the phase of the study (early vs. late: OR, 5.68; 95% CI, 2.11–15.30; p < .001) and the type of experimental treatment (immunotherapy or targeted therapy vs. others: OR, 2.89; 95% CI, 1.51–5.53; p = .001). There was no statistically significant difference between academic and industry-sponsored studies (OR, 1.94; 95% CI, 0.94–3.99; p = .07) and no statistically significant difference for year of trial initiation (OR, 1.17; 95% CI, 0.95–1.44; p = .14 for an increment of 1 year), meaning that the request for new or archival tissue did not significantly increase over time. In multivariate analysis, the request for new or archival tissue increased in early-phase studies (OR, 5.51; 95% CI, 2.00–15.14; p < .001) and in studies investigating innovative treatments (OR, 2.72; 95% CI, 1.38–5.39; p < .01). These results can be found in Table 4.

Association Between Trial Characteristics and Request for at Least One New Biopsy

Of 178 trials, 59 requested a new biopsy, leading to a rate of 33% (95% CI, 26%–41%). In univariate analysis, the request for a new biopsy was associated with the year of trial initiation (OR, 1.30; 95% CI, 1.05–1.61; p = .02 for an increment of 1 year), reflecting a growing request for new biopsies in most recent studies, the phase of the study (early vs. late: OR, 5.80; 95% CI, 2.81–11.96; p < .001), the type of treatment (immunotherapy or targeted therapy vs. others: OR, 2.65; 95% CI, 1.30–5.42; p < .01), and the randomized status (yes vs. no: OR, 3.37; 95% CI, 1.73–6.55; p < .001). There was no statistically significant difference between academic and industry-sponsored studies (OR, 1.96; 95% CI, 0.86–4.44; p = .11). In multivariate analysis, the request for new biopsies increased in early-phase studies (OR, 6.37; 95% CI, 2.93–13.83; p < .001) and in studies investigating innovative treatments (OR, 2.51; 95% CI, 1.15–5.45; p = .02) and increased over time (OR, 1.41; 95% CI, 1.11–1.80; p < .01 for an increment of 1 year). These results can be found in Table 5.

Association Between Trial Characteristics and Compliance to ASCO's Ethical Framework

Of 119 trials requiring new or archival tissue, 80 were in compliance with the ASCO Ethical Framework, with a compliance rate of 67% (95% CI, 58%–76%). In univariate analysis, compliance was associated with the year (OR, 0.63 for an increment of 1 year; 95% CI, 0.47–0.84; p = .002 for an increment of 1 year), the phase (early vs. late: OR, 0.14; 95% CI, 0.06–0.33; p < .001), and randomized status (yes vs. no: OR, 0.38; 95% CI, 0.17–0.86; p = .02). There was no statistically significant difference between academic and industry-sponsored studies (OR, 0.55; 95% CI, 0.19–1.61; p = .27). In multivariate analysis, compliance decreased through time (OR, 0.54; 95% CI, 0.37–0.77; p < .001 for an increment of 1 year) and in early-phase studies (OR, 0.10; 95% CI, 0.04–0.27; p < .001).

Discussion

Analysis of the studies conducted at IJB reveal that numerous studies require tissue for exploratory objectives with potentially no impact on future drug development and no benefit to current or future patients. Requests for tissue increase through the years, whereas compliance to ASCO's Ethical Framework decreases. Overall, adherence to the ASCO Ethical Framework was 67% (n = 80) for studies requiring tissue and 39% (n = 23) when considering studies requiring at least one mandatory biopsy. Of note, all studies included were designed prior to the diffusion of the ASCO Ethical Framework. It thus seems unreasonable to expect previously issued studies to abide by unpublished standards.

Our study underlines the growing request for tissue in clinical trials, especially in trials investigating innovative treatments or early-phase trials. In a similar study analyzing 57 clinical trials between 2005 and 2010 [12], 67% included at least one mandatory biopsy and 68% of the objectives of research biopsies were for correlative science. In comparison with our results, more studies required a new biopsy for an exploratory objective. However, 72% of the identified studies were early-phase studies when our sample included mostly non–early-phase studies (74%) which could explain the discrepancy. Of note, in our analysis we considered phase II studies non–early-phase studies as they intend to look for evidence of efficacy besides the analysis of safety data and are not from our perspective properly designed as early-phase studies. However, phase I/II and phase I/III studies were considered early-phase studies.

Data on risks of research biopsies are underreported [13] and highly depended on the site of biopsy [14]. Serious complications are estimated around 1%–2% [15–17], but data are highly heterogeneous and higher complication rates have been reported [18, 19]. Rates of adverse events seen with biopsies performed in clinical trials were similar to those in everyday clinical practice.

A recent analysis of procedure risks associated with 877 tumor biopsies from the MOSCATO-01 trial showed that complications occurred in 89 cases (10.1%) [19]. Most complications were grade I in 59 cases (66.3%); grade II in 16 (18%), and grade III in 14 (15.7%), with no grade IV complications and no procedure-related death. Complications were most common for lung biopsies. The objective of the MOSCATO-01 trial [20] was to identify actionable molecular alterations via high-throughput genomic analysis of fresh-frozen tumor biopsies for patients to have access to innovative targeted therapies. The rationale was strong and biopsies were therefore justified; however, the rate of 10% of complications is not negligible and invites us to reconsider how risk is portrayed to patients.

Reflecting on risks implies considering how a biopsy is perceived from a patient's point of view. Apart from the risk undertaken, would patients be willing to undergo such procedures if they were fully aware that they would have no direct impact on their health [21]? In one survey [10], 36% of patients stated that mandatory research-related biopsies would deter them from enrolling in a clinical trial. In the same study, 42% of patients believed that a nondiagnostic biopsy would influence their health and care, even when they were clearly informed that it was for research purposes only. In our study, biopsy-related risks were specified in 70% of ICFs when a new biopsy was required and in 55% of cases when tissue was required. Given the potential risks involved, it would be expected to be present in 100% of ICFs. This emphasizes the need for appropriate and precise ICFs as well as clear and transparent information delivered by investigators.

It is also important to explore the motivations for patients to consent to research biopsies [22–24]. Patients appear to be overall willing to undergo research biopsies [25, 26] with an acceptance of the associated risk. In a study aimed to assess the perceptions associated with mandatory biopsies [10], it appears that 20% of patients would accept a 5%–10% risk of a major complication whereas <1% of medical oncologists would accept the risk before approving such a study.

A study from MD Anderson Cancer Center [15] reviewing medical records of 155 patients who had one or more research biopsies in phase I trials demonstrated that the procedure was done in 86.6% of participants when the biopsy was mandatory but only in 4.4% of participants when optional. This tremendous difference questions not only the willingness of patients to undergo a biopsy but also the clinical benefit expected by the clinician, which influences the way the necessity of the biopsy is portrayed to the patient. Willingness of patients to undergo biopsies may differ in early-phase trials as patients have few other therapeutic options and are inclined to consent when a potential treatment is at stake.

Another issue is the use by clinicians of findings deriving from biopsies, in both the clinical and research setting. The term genomic confidence has been used to describe the attitude of an individual physician in their capacity to apprehend genomic concepts and make treatment recommendations based on genomic information. Gray et al. [27] surveyed 160 clinicians and reported that 22% of interviewees reported low confidence in their genomic knowledge, ability to explain genomic concepts to patients (14%), and ability to make treatment recommendations based on genomic data (26%). De Moor [28] found that oncologists who reported high levels of genomic confidence also tended to use these results more frequently to inform patient care and that confidence varied greatly by the type of tests used in their practice. Cohen [29] assessed the willingness to rebiopsy patients with relapsed disease for the purpose of tumor genomic profiling among 195 pediatric oncologists and found that it was highly conditioned by the probability of finding an actionable mutation. A better understanding of oncologists’ attitude toward genomic testing as well as its implications in terms of patient care and communication are essential to engage dialogue on the meaningful use of findings derived from biopsies.

Another point is the necessity to maximize scientific utility of research biopsies. In our study, a scientific rationale was detailed in the majority of protocols requiring biopsies. However, there was no clear statistical plan in around 30% of studies, suggesting that the objectives were not clearly defined. When considering mandatory biopsies for exploratory objectives, a strong scientific rationale ought to be documented along with a clear and precise statistical plan.

Moreover, results of research biopsies are largely underreported. In a study aimed at evaluating the publication and reporting of research biopsies [13] across 46 clinical trials, it appears that the majority of trials (61%) did not report research biopsy results. A larger study by the same team [30] identified 301 clinical trials, among which only 50.8% of all trials reported on biopsy-related results. Reporting of trials was associated with later-stage trials. Consequently, results being underreported has a minimal impact on the subsequent development of tested drugs. In a study analyzing whether biomarkers explored in 72 phase I studies had a subsequent impact on drug development [1], it appeared that only 5 studies resulted in a statistically significant biomarker result that was cited in publications and there was overall no impact on subsequent drug development of biomarker results.

The numerous issues revolving around the value of research biopsies questions the ethics of research in the personalized medicine era [31, 32]. Should there be a distinction between ethics in everyday practice compared with clinical research? Clinical care focuses on providing an individual patient with the most appropriate treatment at that time. The aim of clinical research is to increase generalizable knowledge with a final goal to improve care of the majority. Clinical research is based on the principle of equipoise [33, 34], implying that the outcome of a given intervention is uncertain to the investigator. Clinical equipoise is an essential foundation to the development of clinical research in order to ensure that subjects involved are not disadvantaged by their participation. Despite this theoretical separation in terms of ethics, it is important to underline that meaningful clinical research requires the context of clinical practice just as clinical practice is enriched by clinical research [35]. Patient care and clinical research are not to be separated, and ethics guiding everyday practice should conform to the same ethics principles [32]. Indeed, everyday clinical practice and research should rather be seen as a whole. Just as effective clinical practice cannot exist without clinical research, meaningful clinical research requires the context of clinical practice [35].

Simultaneously, every effort should be made to improve techniques such as liquid biopsies [36] to avoid risks associated with invasive procedures. However, the motives for collecting plasma or any other specimen, even if easily accessible, should be documented, justified, and detailed to the patient.

Conclusion

Future studies should follow ASCO's Ethical Framework, whose aim is to improve the ethics of research biopsies. It is a unique opportunity to maximize scientific utility and allow for clinically meaningful correlative science and safe access to innovative treatments for a maximum number of patients. A strong scientific rationale along with a precise statistical plan should be the basis for justifying the need for research biopsies as well as ensuring such procedures are done with minimal risks and clear and transparent information to patients.

Acknowledgments

Presented in part at the European Society for Medical Oncology Virtual Congress 2020 (poster # 588P).

N.O. is currently affiliated with Department of Medical Oncology, Centre Henri Becquerel, 1 Rue d'Amiens, 76038 Rouen, Haute-Normandie, France.

Author Contributions

Conception/design: Nathalie Olympios, Laetitia Collet, Marianne Paesmans, Philippe Aftimos

Provision of study material or patients: Nathalie Olympios, Laetitia Collet, Philippe Aftimos

Collection and/or assembly of data: Nathalie Olympios, Laetitia Collet

Data analysis and interpretation: Nathalie Olympios, Laetitia Collet, Marianne Paesmans, Christiane Jungels, Nuria Kotecki, Ahmad Awada, Philippe Aftimos

Manuscript writing: Nathalie Olympios, Laetitia Collet, Philippe Aftimos

Final approval of manuscript: Nathalie Olympios, Laetitia Collet, Marianne Paesmans, Christiane Jungels, Nuria Kotecki, Ahmad Awada, Philippe Aftimos

Disclosures

Ahmad Awada: Roche, Lilly, Amgen, EISAI, Bristol-Myers Squibb, Pfizer, Novartis, Merck Sharp & Dohme, Genomic Health, Ipsen, AstraZeneca, Bayer, Leo Pharma, Merck, Daiichi (C/A, RF [to institution], H); Philippe Aftimos: Boehringer Ingelheim, MacroGenics, amcure, Synthon, Radius, Servier, G1 Therapeutics, Roche, Novartis, Amgen (H); Merck Sharp & Dohme, Roche, Pfizer, Amgen (nonfinancial support). The other authors indicated no financial relationships.

(C/A) Consulting/advisory relationship; (RF) Research funding; (E) Employment; (ET) Expert testimony; (H) Honoraria received; (OI) Ownership interests; (IP) Intellectual property rights/inventor/patent holder; (SAB) Scientific advisory board

References

Author notes