Creating robust, reliable, clinically relevant classifiers from spectroscopic data

Abstract

I describe in detail the intimately connected feature extraction and classifier development stages of the data-driven Statistical Classification Strategy (SCS) and compare them with current practice used in MR spectroscopy. We initially created the SCS for the analysis of MR and IR spectra of biofluids and tissues, and subsequently extended it to analyze biomedical data in general. I focus on explaining how to extract discriminatory spectral features and create robust classifiers that can reliably discriminate diseases and disease states. I discuss our approach to identifying features that retain spectral identity and provisionally relate these features, averaged subregions of the spectra, to specific chemical entities ("metabolites"). Particular emphasis is placed on describing the steps required to help create classifiers whose accuracy doesn't deteriorate significantly when presented with new, unknown samples. A simple but powerful extension of the discovered features to detect metabolite-metabolite (feature-feature) interactions is also sketched. I contrast the advantages and disadvantages of using either spectral signatures or explicit metabolite concentrations derived from the spectra as sets of discriminatory features. At present, no clearcut preference is obvious and more objective comparisons will be needed. Finally, I argue that clinical requirements and exigencies strongly suggest adopting a two-phase approach to diagnosis/prognosis. In the first phase the emphasis ought to be on providing as accurate a diagnosis as possible, without any attempt to identify "biomarkers." That should be the goal of the second, research phase, with a view of providing prognosis on disease progression.

Clinical Proton MR Spectroscopy in Central Nervous System Disorders

“…Recommended cohort size will depend somewhat on the nature of the data, but anything less than several hundred subjects, both healthy subjects and those with the condition of interest, would not yield a classifier that would be certified for use by a regulatory agency. A detailed discussion of these issues has appeared recently (176,177).In addition, validation of MR spectroscopy biomarkers for clinical use requires their incorporation in robust prospective multicenter clinical trials, where patient selection and treatment meets prespecified criteria and the statistical methodology is set before the trial commences. This requires careful MR spectroscopy protocol design that can be adhered to at all the participating centers.…”

Diagnostic Applications of Nuclear Magnetic Resonance–Based Urinary Metabolomics

“…Such an algorithm has been part of a statistical classification strategy used in several studies involving NMR data. [57][58][59][60][61][62] A schematic diagram of the metabolomics workflow is depicted in Figure 1.…”

Mrs-based Metabolomics in Cancer Research

“…For details, consult Nikulin et al (1998) and Somorjai (2001). 2,101 Using proper feature extraction, even the simplest classifiers frequently outperform more sophisticated (eg, nonlinear) variants. 102,103 Because of its low complexity, and easy implementation, we generally employ the simple LDA (linear discriminant analysis) with leave-one-out crossvalidation.…”