Methodological and clinical challenges associated with biomarkers for psychiatric disease: A scoping review

Highlights

• For biomarker research to be successful, multiple techniques must be combined.

• There are intricacies unique to psychiatry that make successful biomarker identification difficult.

• Improved research methodologies that are easily combined are needed.

• Multidisciplinary teams should be involved at all stages of biomarker research.

Abstract

Over the past decade, psychiatric research has been on an important hunt for biomarkers of psychiatric disease. In psychiatry, the term “biomarker” is a broad umbrella term used to identify any biological variable that can be objectively measured and applied to a diagnosis; this includes genetic and epigenetic assessments, hormone levels, measures of neuro-anatomy and many other scientific modalities. However, despite hundreds of studies on the topic being published yearly and other medical specialties having success in discovering biomarkers, clinical psychiatric practice has not had the same success. This paper aims to consolidate the many opinions on the search for psychiatric biomarkers to suggest key methodological and clinical challenges that psychiatric biomarker research faces. Psychiatry as a specialty has many fundamental differences compared to other medical specialties in methods of diagnosing, underlying etiology and disease pathologies that may be limiting the success of biomarker research in itself and puts strict requirements on the research being conducted. The academic and clinical environment in which the research is being conducted also heavily influences the translation of the findings. Finally, once biomarkers are identified, more often than not they are inapplicable to clinical settings, unable to integrate into clinical practice and fail to outperform current diagnostic practices and guidelines. We also make six recommendations for more promising future research in psychiatric biomarkers.

Introduction

There is an urgent need for biomarkers of disease in various areas of medicine including pharmacology, oncology and, especially, psychiatry. Although the first reference to “biomarker psychiatry” occurred in 1966, the search for a biomarker in psychiatric research has skyrocketed in the last decade (Fig. 1). The vast number of studies seeking biomarkers alone suggests that the term “biomarker” is broad; on its own even the topic of what constitutes a biomarker has prompted numerous publications and working groups (e.g., Atkinson et al., 2001; Strimbu and Tavel, 2010). What does “biomarker” actually mean? By dictionary definition, it is “a distinct biochemical, genetic or molecular characteristic or substance that is an indicator of a particular biological condition or process” (Biomarker Definition, n.d.). By definition, a biomarker is an objective measurement of a specific psychiatric diagnosis. This use of the term biomarker may be in contrast to the historically subjective nature of psychiatry. Biomarker candidates currently being investigated include (but are not limited to) genotypes, epigenetic modifications, brain connectome characteristics, hormone levels and saccadic eye movement patterns. While there is no shortage of research aimed at identifying biomarkers, their clinical efficacy is, however, currently extremely limited (e.g., Carvalho et al., 2020; Schwarz et al., 2010; Lozupone et al., 2019).

Once identified, not all biomarkers have the same degree of impact on their respective clinical field. An ideal biomarker can be illustrated in the context of Huntington's Disease, which has a clear genetic cause. The presence of an elongated triple nucleotide repeat (36 or greater) in the Huntingtin gene indicates with certainty that the individual will develop Huntington's Disease (MacDonald et al., 1993). This genetic biomarker is ideal as there is no subjectivity associated with its interpretation. In contrast, when looking at cancer, the presence of BRCA1 and BRCA2 gene mutations confer an increased risk of developing breast cancer by approximately 65% and ovarian cancer by approximately 39% (Antoniou et al., 2003). The presence of either mutation alone does not definitively indicate the development of cancer (Antoniou et al., 2003).

Psychiatric conditions are inherently complex and often difficult to diagnose. For example, when an individual presents with symptoms of psychosis, it could be due to a number of disorders such as schizophrenia, bipolar disorder, major depressive disorder (MDD) with psychotic features or a reaction to illicit drugs. Furthermore, for most diagnostic categories, symptom presentation is highly heterogenous. For instance, to receive a diagnosis of MDD, an individual must meet five of nine listed symptoms with one symptom being either a depressed mood or anhedonia (American Psychological Association, 2013). The diagnostic challenges are further enhanced by a lack of objective, laboratory-based tests for psychiatric diagnoses. A clinician must therefore use clinical interviews to distinguish between these different diagnoses. All of these examples not only reinforce the immense drive but also highlight the need to search for psychiatric biomarkers.

Further support for the development of biomarkers is found with the American National Institute of Mental Health's (NIMH's) establishment of a strategic plan to develop evidence-based mental health care through the Research Domain Criteria (RDoC). The RDoC is a framework that aims to generate research in human function and behaviour from the micro (e.g., genetic) to the macro (e.g., self-reported symptoms) level and provides an ideal opportunity to discover biomarkers (Cuthbert, 2014). Biomarkers appear to be quite versatile; they seek to identify individuals at high-risk of diagnosis development, distinguish those with or without a diagnosis and predict patient treatment response.

Despite many research findings concerning biomarkers, results of individual empirical studies often remain unreplicated, and findings that were replicated have yet to impact any stage of clinical engagement. For example, in 2014, a blood test was scheduled for release on the market with the promise of being able to objectively aid in the diagnosis of MDD. MDDScore (Bilello et al., 2015) is an assay of nine MDD-associated biomarkers and controls for gender and body mass index. However, when considering the success of MDDScore, it is crucial to note that from 2015 to September 2020, no further studies have been published, there are no known psychiatric clinics utilizing MDDScore and it is not available for purchase.

The lack of utilization and acceptance of MDDScore illustrates the difficulty for psychiatric biomarker research to make a meaningful difference clinically. The goal of this paper is to highlight some of the key methodological and clinical challenges that impact biomarker research in terms of identification and successful implementation of biomarkers within the medical field. To the authors’ knowledge, this paper is the first to provide an extensive review of methodological and practical challenges facing biomarker research in psychiatry at various points in the research translation process. Outlining the challenges at both the methodological and practical levels could help the design, interpretation and implementation of psychiatric biomarker research.