Defining biomarker performance and clinical validity

J Med Biochem 2011; 30 (3) DOI: 10.2478/v10011-011-0028-0 UDK 577.1 : 61 ISSN 1452-8258 J Med Biochem 30: 193–200, 2011 Review article Pregledni ~lanak DEFINING BIOMARKER PERFORMANCE AND CLINICAL VALIDITY DEFINISANJE PERFORMANSI I KLINI^KE VALIDNOSTI BIOMARKERA Patrick M.M. Bossuyt Department of Clinical Epidemiology & Biostatistics, Academic Medical Center, University of Amsterdam, the Netherlands Summary: In the evaluation of biomarkers three questions can be answered: what is the analytical validity of the marker, what is its clinical validity, and does the marker have clinical utility? In most cases, clinical validity will be expressed in terms of the marker’s accuracy: the degree to which it can be used to correctly identify diseased patients or, more generally, patients with the target condition. Diagnostic accuracy is evaluated in studies in which the biomarker values are compared to the outcome of the clinical reference standard in the same patients. There are several ways in which the results of diagnostic accuracy studies can be summarized, reported, and interpreted. In this paper we summarize and present the available measures. We classify these as errorbased measures, information-based measures, and measures of the strength of the association. Clinical validity is linked to clinical utility. If the target condition is well defined and associated with unequivocal downstream management decisions, clinical validity, when defined in comparative terms, may sometimes act as a surrogate outcome measure for clinical utility. Kratak sadr`aj: Evaluacija biomarkera mo`e podrazumevati odgovore na tri pitanja: kakva je analiti~ka validnost markera, kakva je njegova klini~ka validnost i da li je marker klini~ki koristan. U ve}ini slu~ajeva, klini~ka validnost bi}e izra`ena kao ta~nost markera: do koje se mere on mo`e koristiti za ta~no identifikovanje obolelih pacijenata ili, uop{teno, pacijenata sa ciljanim stanjem. Dijagnosti~ka ta~nost procenjuje se u studijama u kojima se vrednosti biomarkera porede sa ishodom klini~kog referentnog standarda kod istih pacijenata. Postoji nekoliko na~ina da se sumiraju, predstave i tuma~e rezultati studija dijagnosti~ke ta~nosti. U ovom radu sumira}emo i predstaviti dostupne mere. Podelili smo ih na mere zasnovane na gre{kama, mere zasnovane na informacijama i mere ja~ine asocijacije. Klini~ka validnost povezana je sa klini~kom korisno{}u. Ukoliko je ciljano stanje ta~no definisano i povezano sa ispravnim odlukama o terapiji, klini~ka validnost, kada se defini{e u komparativnim terminima, ponekad mo`e biti surogat mera ishoda za klini~ku korisnost. Klju~ne re~i: biomarker, ta~nost, klini~ka validnost, klini~ka korisnost Keywords: biomarker, accuracy, clinical validity, clinical utility Introduction In recent decades the number of medical tests and biomarkers has been rising at a rapid pace. New markers are proposed at an increasing rate and the Address for correspondence: Patrick M.M. Bossuyt, Ph. D. Dept. Clinical Epidemiology & Biostatistics Academic Medical Center University of Amsterdam Room J1b-214 PO Box 22700; 1100 DE Amsterdam; the Netherlands p.m.bossuytªamc.uva.nl technology of existing tests is continuously being improved. Like any other medical technology, new medical tests and biomarkers should be thoroughly evaluated prior to their introduction into daily practice. A rigorous evaluation process of diagnostic tests before introduction into clinical practice will not only improve patients’ health but contribute to an efficient use of health care resources by preventing unnecessary testing. Unfortunately, the evaluation of medical tests is less advanced than that of treatments. The methodology has been less well developed. There is uncertainty about what methods to use, and what the sources of bias are in biomarker studies. 194 Bossuyt: Biomarker performance and validity The lack of progress in the methodology for biomarker evaluation can in part be attributed to the lower standards for the regulation of biomarkers. Unlike the evaluation of drugs, for which the threshold to marketing is relatively steep, entry to the market for developers of biomarkers has been less difficult for most products. In this paper we present a triad of questions that can be asked about any new biomarker. The three questions deal with the analytical validity of the marker, its clinical validity and the clinical utility. We discuss in more detail how the clinical validity of diagnostic markers is expressed, in terms of diagnostic accuracy. We close by describing when and how the clinical validity of a biomarker can be regarded as a proxy measure for the clinical utility. Three questions about biomarkers The evaluation of medical technology can be a time-consuming and costly process. An efficient use of resources calls for a well-planned evaluation strategy, in which more elaborate and therefore more expensive forms of evaluation are only performed if satisfactory results have been obtained in the previous steps of the evaluation process. Such a phased approach, moving gradually from small to larger studies, may also protect the rights and integrity of human volunteers and patients. Several comparable hierarchical models have been proposed for the evaluation of tests and biomarkers. Analogous to the 4-phase model for the evaluation of new drugs, these models require that in each phase certain conditions be fulfilled before the evaluation can continue with the subsequent phase. In a systematic review, we identified 19 phased evaluation schemes (1). One of the best known schemes consists of the levels of efficacy proposed by Fryback and Thornbury (2). Their scheme, originally developed for imaging, has also been used for other forms of testing. In genetics, the ACCE and EGAPP frameworks have become more widely used (3). The identified schemes showed substantial similarity. Table I presents a simplified summary of these schemes, translated as a set of three questions. The first question is »Can I trust the results of this marker?«. This is generally referred to as the ana- Table I Three questions in the Evaluation of Biomarkers. Question Feature Is it true? Analytical Validity Is it meaningful? Clinical Validity Is it useful? Clinical Utility lytical validity of the marker: the marker’s ability to measure what it is supposed to measure. The analytical validity of a test refers to its ability to accurately and reliably measure the entity or analyte of interest. There is no single statistic to express the level of analytical validity. Measures used include analytical sensitivity or limits of detection, precision, analytical specificity (cross-reactivity, interference), assay linearity, reliability and repeatability of test results, and assay robustness. The terms are not always used in an unambiguous way and there is little standardization in methods for this initial technical evaluation. Analytical validity is usually evaluated (...truncated)