Identifying immunoglobulin abnormalities in the presence of variable polyclonal background

Modern mass spectrometry-based methods for the isotyping and quantitation of monoclonal proteins (M-proteins) are now supplanting traditional gel-based methods such as serum protein electrophoresis (SPEP). The resultant improvement in resolution poses new signal processing challenges in the separation of M-protein from normal variations in polyclonal immunoglobulin background. We model normal variations in this background by means of a singular value decomposition and apply a coupled alternating-projection algorithm to separate normal and abnormal spectral components, both of which are variable. Ion simulation results guide our algorithm details and highlight limitations that can exist when testing M-proteins of high concentration. We show experimental quantitation results on clinical submitted sera with acceptable results. Finally, we outline enhancements to the fundamental method which have resulted in an enhanced analytical measuring range for clinical care of patients with plasma cell disorders.