Real-World Evidence Studies in Oncology Therapeutics: Hope or Hype?

Abstract

Randomized controlled trial (RCT) remains a gold standard in evidence-based medicine for assessing the efficacy and safety of cancer therapies. However, due to some inherent methodological limitations of RCT, such as stringent inclusion criteria, highly specific treatment, ethical and scientific compromise in rare cancer, and inability to adequately assess safety, real-world evidence (RWE) has been adjudged as a suitable option to complement data obtained from RCT. Moreover, in the context of cancer therapeutics, few notable merits pertain to developing a novel product for rare cancer subtypes, establishing new indications for already approved drugs, optimization of treatment regimen and sequence, a better description of long-term safety, and supporting the reimbursement-related decision. However, the implementation of RWE for the aforementioned purposes will be limited by various challenges, especially in the context of developing economies such as India. Special attention should be given to the availability of data, maintaining the quality standard, and establishing stringent regulations for privacy and security along with active regulatory engagement with relevant stakeholders. Such activities will be key to facilitating the use of RWE in cancer therapeutics.

Introduction

The recent advancements in research and development in modern medicine including oncology, prompted by science and technology, have been noteworthy. The traditional randomized controlled trial (RCT) has been adjudged as the gold standard to generate efficacy and safety of a new drug attributed to its inherent design and high internal validity [1]. However, to improve the efficiency and also hasten the novel drug approval process in the context of growing technological prowess, FDA has promulgated the use of real-world evidence (RWE) studies [2]. RWE studies can be used to complement RCTs by depicting the usage, benefit, and risk of a new therapy derived from various data-linked patient populations in clinical practice. Such diverse data sources, described as real-world data, can be electronic medical records (EMR), patient registries, claim databases, and patient surveys [3]. Therefore, apart from being an endeavor based on lower cost and resources with less time, it is more generalizable and context specific in defining effectiveness, i.e., in the case of subgroups being underrepresented or excluded in RCT. Moreover, in a situation where conduction of RCT is challenging, i.e., recruitment difficulties in a rare cancer population, RWE emerges as a viable alternative [4, 5]. Therefore, despite being classified as level III and IV evidence (moderate to low strength), the ASCO guidance framework encourages the adoption of RWE as a valuable tool in oncological research where RCT is considered not feasible or ineffective [6]. The differences in between RCT and RWE are depicted in Table 1.

Table 1.

Key differences in between randomized controlled trial (RCT) and real-world evidence (RWE) study

Impact as study attribute	Randomized controlled trial (RCT)	Real-world evidence (RWE) study
Acceptance as an evidence base	Wide acceptance	Relatively less than RCT
Representativeness among patient population	Homogenous patient population in a controlled setting	Diverse patient population represented in a clinical care setting
Study validity	High internal validity	Low due to potential bias and confounding
Application	Specifically used to address efficacy and safety	Diverse application pertinent to health care related decision making

The recent progress in innovative study design and analytical methodologies render RWE better poised to assess effectiveness. Therefore, RWE has been largely endorsed by the EU and FDA related to drug approval and label expansion of various oncological drugs [7, 8]. Concomitantly, it is gaining traction in developing regions, i.e., in the Asia Pacific (APAC) have been at the forefront of leveraging RWE studies [9]. However, despite the recent spurt in these studies, the full potential remains largely unexplored due to a few limitations that need to be addressed and solved accordingly.

Expanding Role of RWE in Oncology

As previously mentioned, the applicability of RWE studies could be of immense potential in oncology. There are instances when RWE studies are done to complement previous RCT results in oncology for better treatment-related decision-making.

For example, neo adjuvant chemoradiotherapy (CRT) using carboplatin and paclitaxel followed by resection in primary resectable non-metastatic esophageal cancer (EC) have been implemented in the Netherlands following by CROSS-trial. A retrospective study was done using medical records of 145 such patients encountered in daily practice being on the same treatment protocol guided by the CROSS trial. This resulted in a 3-year overall survival (OS) of 49.6% and PFS of 45.6%, compared to that of 58% and 51% in the CROSS trial. Such slightly poorer survival was attributed to the less favorable patient and tumor characteristics (low WHO performance score, increase age, larger tumor length, more involved lymph node), which is common in daily practice [10].

Similarly, FDA approved the use of first-line pembrolizumab to standard chemotherapy of pemetrexed and a platinum-based drug for metastatic non-squamous non-small-cell lung cancer (NSCLC) without epidermal growth factor (EGFR) or anaplastic lymphoma kinase (ALK) mutations followed by KEYNOTE- 189 trial. To explore the effect of such therapy, a RWE study was undertaken using anonymized, electronic health record databases of 283 eligible patients in the USA. The median OS reported in this study (16.5 months) was somewhat shorter than the KEYNOTE-189 trial (22.0 months). Again reasons for such shorter OS can be attributed to higher median age, presence of comorbidities, and tendency of stopping pemetrexed earlier. The finding of higher OS, real-world progression-free survival (rw PFS), and real-world tumor response rate (rw TRR), concomitant with higher programmed death ligand 1 (PD-L1) expression, was similar to that of KEYNOTE-189 trial [11].

Another study done in the Netherlands tried to compare outcomes for immune checkpoint inhibitors in patients of stage IV NSCLC patients in 2015–2018 in six large Dutch teaching hospitals. For patients treated by first-line pembrolizumab and second-line nivolumab, PFS times were comparable between real-world and trials [HR 1.08 (95% CI 0.75–1.55) and HR 0.91 (95% CI 0.74–1.14), respectively]. OS was significantly shorter in the study for patients on 1st-line pembrolizumab therapy in comparison to the trial (HR 1.55; 95% CI 1.07–2.25). Therefore, efficacy effectiveness (EE) gap was not evident in PFS, contrary to that of OS for 1st-line pembrolizumab [12].

Increased Study Generalizability

RCTs generally include patients who can significantly differ from the heterogenous population that any physician encounters. This is largely due to stringent eligibility criteria favoring a more healthy and younger population having good performance status or without any comorbidities, not representative of the actual population in a clinical setting [13, 14]. For example, according to a report, less than 5% of adult cancer patients participate in a trial [15]. Also, for common cancer, the proportion of patients taking part in a clinical trial is low. For example, it has been reported that in breast cancer, only 3% of the patients participate in a clinical trial [16].

The analysis of the investigational new drug application submitted to the FDA in 2015 for cancer and oncological products showed that 60% of the trial necessitated Eastern Cooperative Oncology Group performance status 0 to 1 (excluding significant symptomatic and unfit patients) devoid of any metastasis or significant comorbidities, that in most of the times patients tend to harbor in the practice setting [17]. Therefore, the clinical effect observed in RCT tends to get attenuated in clinical practice. For example, patients with advanced hepatocellular carcinoma showed significant improvement in overall survival treated with sorafenib in an RCT [18]. However, subsequent analysis of data in a clinical setting indicated shorter survival, raising doubts over the reproducibility of the erstwhile RCT [19].

RWE studies can be used to complement the study findings of RCT in a more inclusive and broader patient population, including those who otherwise would have been excluded in an RCT [13, 14]. A real-world study compared the effect of eligibility criteria of 10 RCTs to that of a more inclusive set of criteria on overall survival (OS) in more than 60,000 patients of non-small-cell lung cancer [20]. With the relaxation of the eligibility criteria having a significant impact on overall survival, investigators found that the average study population doubled over each trial (from an average of 1553 patients to 3209 patients). These findings indicate that patients, who otherwise got excluded from the RCT, could benefit from the intervention in the clinical setting [20]. Moreover, researchers identified particular eligibility criteria that will not have any substantial impact on OS [20]. Therefore, RWE could assist clinicians to extrapolate the study findings to a broader patient population who might benefit from the treatment in question.

Role in Rare Cancer with High Unmet Need

The application of RWE in oncology is deemed significant in the context of approval of novel drugs for a rare or orphan indication. For instance, from 1983 to 2019, oncology represented 37% of the total orphan designation granted by the FDA among all therapeutic areas [21]. There has been a paradigm shift toward the advent of precision medicine which is based upon underlying genomic alteration rather than the organ of origin. As a consequence, the common cancers are classified as rare subtypes. Recruitment constraint along with difficulties to power a traditional phase III RCT over an acceptable time limit in the context of very few numbers of patients suffering makes RWE a pertinent alternative. Concomitantly, there are instances of the FDA regularly accepting single arm studies for approval of drugs for such conditions [22]. In this context, the use of external controls or synthetic controls, derived from an external data source from a patient cohort, can be compared against an interventional therapy, hence supporting the approval of any new drug indication. If initial single arm studies indicate a substantial treatment effect of any drugs, assessment of treatment effect in similar groups of patients utilizing RWD can yield a reliable assessment of the comparative effectiveness and safety [5]. For example, metastatic Merkel cell carcinoma (mMCC) is an uncommon, yet aggressive skin cancer for which there was a dearth in the development of evidence based standard treatment. A study performed by the US Oncology Network, using the iKnowMed database (an oncology-specific EHR), compared a substantial number of patients diagnosed and treated with chemotherapy for mMCC to that of a single-arm, phase 2 trial experimenting avelumab, a checkpoint inhibitor. The improved ORR in the patient treated with avelumab than that of external control paved the way for the first FDA approval of first-line therapy in this patient population [23]. Similarly, single arm trial of blinatumomab for patients suffering from Philadelphia chromosome negative relapsed or refractory B cell precursor ALL, a rare variant, was supported by a historical control group derived from a chart review of patients from the USA and Europe, followed by which it was approved for aforementioned indication in 2014 by FDA as an orphan drug [24].

Role in Label Expansion

Oncology therapies approved for a specific indication are commonly used for related malignancies. RWE can be used to leverage broader indications, which can accelerate access to patients, saving the costs incurred by RCTs. For example, blinatumomab was approved in 2016 for all indication in the pediatric population, supported by data from a single arm, open-label, expanded access protocol including 41 children under the age of 18 years [25]. Similarly, palbociclib, which is indicated for female breast cancer, got supplementary approval by FDA for the use in male breast cancer. A retrospective outcome analysis was done using EHR data from the Flatiron Health Analytic Database and IQVIA pharmacy and medical claim database following which the FDA approved the supplemental indication of palbociclib in 2019 [26].

Role in the Depiction of Treatment Safety

The evaluation of safety signals can be well depicted in a real-world post marketing safety surveillance. A limited number of patients in a clinical trial without a longer follow-up period would only allow common adverse events to be detected; hence, rare toxicities would be elusive. Furthermore, the stringent inclusion criteria in RCT would limit the likelihood of participants experiencing any adverse event [27].

RWE could evaluate medication over a longer period in a broader and more diverse patient population. Hence, apart from concomitant ascertainment of the frequency and pattern of the common adverse event, identification and quantification of rare adverse event would also be characterized [28]. For example, an increased risk of veno occlusive disease within the first year of approval of gemtuzumab for the treatment of relapsed acute myelogenous leukemia prompted FDA to withdraw the drug from the market [29]. The cardiac toxicity of immune checkpoint inhibitors is another rare event to be appreciated in a routine trial setting but has been better characterized by postmarketing data [30]. FDA mandates the use of RWE for persistent collection of effectiveness as well as safety data to support any novel oncological product once they are marketed [3].

Apart from the aforementioned role leveraged by RWE studies, RWE has been used for various treatment-related decision-making. Characterizing patient population and optimization of treatment choice is an important feature that RWE can impart. For example, Hemato Oncology Latin America (HOLA) characterized multiple myeloma patients in Latin America and their differing responses toward bortezomib-based therapy defined on age [ages > 65 years were less likely to respond to bortezomib than chemotherapy] [31].

Optimizing proper treatment regimen and treatment sequencing in a clinical setting, which cannot be delineated by RCTs due to high resource requirements, could be well informed by RWE studies. RCTs often compare a new drug with the standard of care, which is subjected to change in a given disease or a subpopulation or any geographical location [32, 33]. For example, the treatment option in BRAF mutated metastatic melanoma is rapidly evolving given seven new therapies were approved between 2014 and 2018 [34]. A retrospective real-world observational study was designed derived from the analysis of 600 patients from a US-based Cardinal Health Oncology Provider network (a community involving over 7000 oncologists), which showed that the sequence of targeted therapy followed by immunotherapy was associated with a better response rate than the reverse sequence for BRAF-mutated metastatic melanoma [35].

Other Role

RWE can also help health care professionals to strategize their dosing in a defined population of a particular ethnicity for better optimization of efficacy as well as safety. For example, an Asian RWE study examined lowered dose of sunitinib therapy than the conventional dose in the context of prior research showing high toxicities in the Asian population with metastatic renal cell carcinoma. This study corroborated the fact that the attenuated dose regimen had comparable rw-PFS and real-world overall survival (rw-OS) with a better safety profile as compared to the conventional dose [36].

Since comparative values and cost-effectiveness are measured by RWE studies, it has been extensively used in cancer drug submission by National Institute for Health and Care Excellence (NICE) in the UK from 2011 to 2018 to facilitate reimbursement decision followed by the formation of practice guideline and shaping related health policies [37]. Real-world pharmacoeconomic studies are gaining much importance specially in developing countries with limited resources. For example, a recent systematic review of cost-effective analysis (CEA) of breast cancer medications from 2009 to 2019 in low- and middle-income countries (LMIC) showed most of the drugs were not cost-effective. The review also reiterated the requirement of high-quality and setting-specific pharmacoeconomic evaluation for recent drugs in different disease stages of breast cancer in developing countries [38].

The role of RWD can also be leveraged in a post marketing studies (phase IV) to address specific information related to quality, efficacy, and safety concerns that can be specific to a certain subpopulation through large-scale use for an extended period. Such studies, usually mandated by regulatory bodies, are known as postmarketing surveillance (PMS) activities [39]. The PMS setting comprises many activities including clinical trials, spontaneous reporting system, nonrandomized database studies, or registry studies [40]. The later study is done utilizing health insurance claims data which contain structured diagnosis and capture patient experience across the specific treatment. Such studies could be used to perform confirmatory studies as well as in safety monitoring and surveillance of products [41]. However, regulatory bodies such as FDA endorse such studies in safety monitoring and surveillance in relation to PMS activities [40].

Limitations and Challenges to Overcome for Conducting RWE Studies

Despite the diverse promising role promulgated by RWE, several limitations can hinder the full utilization of RWE for clinical decision-making in oncology. Such limitations could rise from the way RWD is generated, analyzed, and subsequently utilized by different stakeholders, especially in developing countries.

The availability of RWD is limited due to the dearth of an available secondary database in emerging economies which can be a barrier to the adoption of RWE. The implementation of the Health Information Technology for Economic and Clinical Health (HITECH) act in 2009 in the USA led to the access of EMR data as an RWD source [9]. However, in countries such as India, any policy or legislation is either absent or at a nascent stage and the use of EMR has been largely limited to only a few private or corporate healthcare providers, restricting its use only in special situations. Moreover, medical claims in India scarcely contribute to the cause of medical research as very less people are insured with no coverage extended to daily outpatient visits or day consultations [42]. However, few patient registries such as National Cancer Registry are a pertinent choice for leveraging real-world data. Moreover, the adoption of digital initiatives such as the National Digital Health Mission and National eHealth Authority can facilitate data integration and operability [42].

Subsequently, there might be issues regarding data quality and integrity. Limited understanding and inadequate skill development related to data collection or entry methods, due to lack of proper training, will lead to the incompleteness of data, errors, or inconsistencies, which could not be managed by robust statistical approaches. Data collection might be fragmented or contextual based on data provided by health-related apps or personal devices. This ultimately will reduce data credibility or reliability leading to mistrust for RWE [43]. Additionally, the prospect of collecting data from different sources would prompt concerns about data privacy or sharing in emerging economies [44, 45]. Such apprehension is justified in the lack of established acts or regulations in contrast to that of the European Union or the USA, where data protection is governed by the General Data Protection Regulation (GDPR) and Health Insurance Portability and Accountability Act of 1996 (HIPAA) [46]. However, clear policies mandating the specific purpose for data collection with patient consent are missing. India is no exception with the lack of robust method and structured database; a high doctor-to-patient ratio exposes the physician over burdened with a huge patient load. Therefore, physicians are unable to spend much time completing information for medical reports without technical assistance or support staff.

To leverage the full potential of real world data in relevant research including oncology, few practical issues need to be considered. Similar to the conduct of RCT, herein framing a meaningful research question is pertinent which entails the utilization of real-world data, concomitant with appropriate study design which takes into account the potential data limitation and measures to contain bias [47, 48]. Pre-registration of protocols with prespecified analysis plan with subsequent adoption of guideline and research framework for reporting results can uplift the scientific integrity of RWE studies. For example, ISPOR-ISPE recently launched a real-world evidence registry to facilitate pre-registration of real-world research protocols [49]. It also makes recommendations on good procedural practice for real-world data studies. For reporting results, templates such as STaRT-RWE (structured template for planning and reporting real-world evidence studies) and RECORD (reporting of studies conducted using observational routinely collected data) have been established [50, 51].

Collection of high-quality, relevant, and fit-for-purpose data compatible with internationally recognized data standards like Systematized Nomenclature of Medicine Clinical Term (SNOMED-CT) or International Classification of Diseases (ICD) is imperative [52]. There should be clear documentation of data completeness and provenance, for example, synoptic reporting [53, 54]. A clear provision should be laid regarding data ownership and privacy, aligned with the regulation and act specified before, to preserve trust in healthcare systems [47]. Regulatory assistance is imperative to provide clear guidance for regulatory submission to utilize RWE to its full potential. Therefore, they need to get engaged for framework or policy shaping adapted to the regional consideration. Such policy guidelines should be precise over issues such as guidance over data ownership, protection, transparency, quality assurance, monitoring, and approval of data collection based on the intended use and patient consent [9].

As discussed earlier, RWE studies tend to be more inclusive and representative of the diverse population to fill the knowledge gap encountered in conventional trials. However, in the developed world like the USA, there are reports of racially biased datasets [55]. Patients from lower socioeconomic groups may seek care in smaller community hospitals which are deprived of RWD curation, which occurs in major academic networks [56]. This disparity magnifies in lower and middle income countries in context of availing healthcare as well as any new innovation. For example, RWD derived from medical wearables increases such gap between those with or without interconnected devices [57]. For policymakers, it is to ensure that RWD studies should not be restricted to specific privileged groups, instead, it needs to be more inclusive. Likewise, coding needs to be done for demographic data in HER to ensure representativeness [52]. Proper incentivization for RWD infrastructure can bring more value to the downstream users of such data like patients, clinicians, and providers; ultimately establishing a positive feedback loop for RWD development (52).

Therefore, it is for all different stakeholders including research authorities, policymakers, oncologists, and patient advocacy groups to collaborate and adopt such practices to ascertain the prospect of better understanding and utilization of RWE in cancer therapeutics.

Comments

Despite being a relatively new concept in clinical research, the application of RWE is poised to increase in therapeutics, especially in oncology. Though it is commonly used to complement the safety and efficacy data generated by RCT, it can be imperative in various situations where RCT is not feasible enough to generate evidence. However, there are few limitations and challenges that need to get addressed, especially in developing economies such as India. Various efforts should be made by relevant stakeholders to leverage its full potential in the therapeutic landscape.

Declarations

Conflict of Interest

The author declares no competing interests.