Biomarkers and Surrogate Endpoints in Drug Development: A European Regulatory View

Wickström, Kerstin; Moseley, Jane

doi:10.1167/iovs.17-21778

Cited by 34 publications

(43 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…1,2 The emergence of precision medicine and high throughput precision technologies elevated aspirations for defining novel biomarkers that would accelerate improved treatment of diverse adverse health conditions by facilitating the identification of responders to promising novel or repurposed therapeutic strategies. 3,4 A cursory review of the medical literature [5][6][7] over the past 3 decades revealed the emergence of an increasing number of biomarker candidates. However, the exponential rate of initial discovery has now completely outpaced the ability of the biomedical community to successfully develop and validate the clinical utility of prospective biomarkers.…”

Section: Introductionmentioning

confidence: 99%

The acute respiratory distress syndrome biomarker pipeline: crippling gaps between discovery and clinical utility

Bime

Camp

Casanova

et al. 2020

Translational Research

View full text Add to dashboard Cite

Recent innovations in translational research have ushered an exponential increase in the discovery of novel biomarkers, thereby elevating the hope for deeper insights into "personalized" medicine approaches to disease phenotyping and care. However, a critical gap exists between the fast pace of biomarker discovery and the successful translation to clinical use. This gap underscores the fundamental biomarker conundrum across various acute and chronic disorders: how does a biomarker address a specific unmet need? Additionally, the gap highlights the need to shift the paradigm from a focus on biomarker discovery to greater translational impact and the need for a more streamlined drug approval process. The unmet need for biomarkers in acute respiratory distress syndrome (ARDS) is for reliable and validated biomarkers that minimize heterogeneity and allow for stratification of subject selection for enrollment in clinical trials of tailored therapies. This unmet need is particularly highlighted by the ongoing SARS-CoV-2/COVID-19 pandemic. The unprecedented numbers of COVID-19-induced ARDS cases has strained health care systems across the world and exposed the need for biomarkers that would accelerate drug development and the successful phenotyping of COVID-19infected patients at risk for development of ARDS and ARDS mortality. Accordingly, this review discusses the current state of ARDS biomarkers in the context of the drug development pipeline and highlight gaps between biomarker discovery and clinical implementation while proposing potential paths forward. We discuss potential ARDS biomarkers by category and by context of use, highlighting progress in the development continuum. We conclude by discussing challenges to successful translation of biomarker candidates to clinical impact and proposing possible novel strategies.

show abstract

Section: Introductionmentioning

confidence: 99%

The acute respiratory distress syndrome biomarker pipeline: crippling gaps between discovery and clinical utility

Bime

Camp

Casanova

et al. 2020

Translational Research

View full text Add to dashboard Cite

show abstract

“…Following the 2016 Endpoints Workshop on Age-Related Macular Degeneration and Inherited Retinal Diseases organized by the National Eye Institute (NEI, National Institute of Health, USA) and the United States Food and Drug Administration (FDA), it was noted that structural candidate endpoints require evident strong correlations with functional biomarkers in order to become accepted trial endpoints [12]. Patient-reported outcomes were proposed as meaningful secondary outcome measures for future trials in AMD by regulatory agencies and therefore should also be considered [45]. Despite available data on functional, structural, and patient-reported outcome measures in iAMD, surrogate biomarkers require careful validation to avoid potential harm as seen by regulators [46,47].…”

Section: Discussionmentioning

confidence: 99%

“…Surrogate endpoints are required in many slowly progressive diseases or when events of interest occur rarely or are difficult to capture [ 43 ]. They have been widely used in clinical trials in ophthalmology with, for instance, intraocular pressure being one of the most common endpoints in glaucoma trials [ 44 , 45 ]. In AMD clinical trials, only high-luminance, high-contrast best-corrected visual acuity has been available as a regulator accepted primary efficacy outcome measure for the approval of new therapeutics in ophthalmology [ 45 ], an outcome measure inapplicable to earlier AMD stages.…”

Section: Discussionmentioning

confidence: 99%

Clinical study protocol for a low-interventional study in intermediate age-related macular degeneration developing novel clinical endpoints for interventional clinical trials with a regulatory and patient access intention—MACUSTAR

Terheyden

Holz

Schmitz-Valckenberg

et al. 2020

Trials

View full text Add to dashboard Cite

Background: There is an unmet need for treatment options in intermediate age-related macular degeneration (iAMD). However, for any new interventions to be tested in clinical trials, novel currently unavailable clinical endpoints need to be developed. Thus, the MACUSTAR study aims to develop and evaluate functional, structural, and patient-reported candidate endpoints for use in future iAMD trials. Methods: The protocol describes a low-interventional clinical multicenter study employing a novel two-part design. The cross-sectional part (total duration, 1 month) and the longitudinal part (total duration, 36 months) include participants with iAMD and control groups with early/late/no AMD. The cross-sectional part's primary objective is a technical evaluation of functional, structural, and patient-reported candidate outcomes. The longitudinal part's primary objective is to assess the prognostic power of changes in functional, structural, and patient-reported outcomes for progression from iAMD to late AMD. All data will be used to support a biomarker qualification procedure by regulatory authorities. Discussion: The MACUSTAR study characterizes and evaluates much needed novel functional, structural, and patient-reported endpoints for future clinical trials in iAMD and will improve our understanding of the natural history and prognostic markers of this condition.

show abstract

“…Currently, we are unable to pinpoint individual local markers to a particular anterior-posterior location in OCT A-scans, and therefore the exact origin of the feature activation is yet unknown. Nevertheless, biomarker discovery such as reported here may become an important aspect in medical image interpretation as conventional markers are regarded to have substantial weaknesses in reflecting patient-centered outcomes (such as visual acuity), and pivotal drug developments fail presumably also due to the lack of reliable endpoints 15,16 .…”

Section: Discussionmentioning

confidence: 99%

Unbiased identification of novel subclinical imaging biomarkers using unsupervised deep learning

Waldstein

Seeböck

Donner

et al. 2020

Sci Rep

View full text Add to dashboard Cite

Artificial intelligence has recently made a disruptive impact in medical imaging by successfully automatizing expert-level diagnostic tasks. However, replicating human-made decisions may inherently be biased by the fallible and dogmatic nature of human experts, in addition to requiring prohibitive amounts of training data. In this paper, we introduce an unsupervised deep learning architecture particularly designed for OCT representations for unbiased, purely data-driven biomarker discovery. We developed artificial intelligence technology that provides biomarker candidates without any restricting input or domain knowledge beyond raw images. Analyzing 54,900 retinal optical coherence tomography (OCT) volume scans of 1094 patients with age-related macular degeneration, we generated a vocabulary of 20 local and global markers capturing characteristic retinal patterns. The resulting markers were validated by linking them with clinical outcomes (visual acuity, lesion activity and retinal morphology) using correlation and machine learning regression. The newly identified features correlated well with specific biomarkers traditionally used in clinical practice (r up to 0.73), and outperformed them in correlating with visual acuity (R 2 = 0.46 compared to R 2 = 0.29 for conventional markers), despite representing an enormous compression of OCT imaging data (67 million voxels to 20 features). In addition, our method also discovered hitherto unknown, clinically relevant biomarker candidates. The presented deep learning approach identified known as well as novel medical imaging biomarkers without any prior domain knowledge. Similar approaches may be worthwhile across other medical imaging fields. Medical imaging for precision medicine relies on biomarkers that capture patient and disease characteristics accurately, efficiently, reproducibly and interpretably. By tradition, the process to establish a biomarker starts with hypothesis generation based on professional experience or theoretical motivation, and concludes with hypothesis testing in specifically designed experiments, for instance by demonstrating the linkage between a marker and clinical outcomes. However, human experts are limited in discovering novel biomarkers because current dogmas may hinder unbiased hypothesis generation, or simply because they may not comprehend the phenotypes of patients and diseases in their full complexity. Recently, artificial intelligence (AI) has made a powerful entry into medical imaging by automatically replicating specific human tasks of biomarker identification and quantification with superhuman accuracy. For instance, artificial neural networks could autonomously diagnose skin cancer 1 , triage referable retinal diseases 2 and provide automated diagnoses of chest x-rays 3 or retinal images 4. When deep learning was targeted to clinical endpoints, it even enabled prediction of systemic cardiovascular parameters from photographs of the back of the eye 5. However, these so-called supervised deep learning approaches have critical disad...

show abstract

Biomarkers and Surrogate Endpoints in Drug Development: A European Regulatory View

Cited by 34 publications

References 22 publications

The acute respiratory distress syndrome biomarker pipeline: crippling gaps between discovery and clinical utility

The acute respiratory distress syndrome biomarker pipeline: crippling gaps between discovery and clinical utility

Clinical study protocol for a low-interventional study in intermediate age-related macular degeneration developing novel clinical endpoints for interventional clinical trials with a regulatory and patient access intention—MACUSTAR

Unbiased identification of novel subclinical imaging biomarkers using unsupervised deep learning

Contact Info

Product

Resources

About