Computational Algorithm Based On Serum Immunoassays for the Identification of Haematological Disease May Identify Patients with Other Acute Disorders

Abstract

AbstractAbstract 4001 IntroductionSerum free light chain (FLC) measurements are used routinely to identify monoclonal immunoglobulin (M-Ig) production in patients. In addition, summated kappa + lambda concentrations have recently been shown to have prognostic value in chronic lymphocytic leukaemia (CLL), Hodgkins lymphoma and HIV patients; presumably reflecting immune stimulation. A recently developed companion assay to FLC testing measures intact immunoglobulin heavy chain/light chain (HLC) pairs in serum calculating Ig’κ/Ig’λ ratios as a reflection of clonality. Here we report on a computational approach combining FLC and HLC measurements to identify clonal disease, hyper- and hypo-gammaglobulinemia, and as markers of immune dysfunction. Furthermore, we comment on the potential of such an algorithm to risk stratify haematological and non-haematological malignancies. Patients, Materials and Methods1468 patients referred to the Royal Wolverhampton Hospital for haematological evaluation were enrolled on to the study. Median patient age was 64 years, with a slight female preponderance (60%). Patient sera were analysed at presentation using routine electrophoresis tests (serum protein electrophoresis, SPEP and immunofixation, IFE). FLC and HLC (IgG, IgA and IgM) were measured by nephelometric immunoassay. Each immunoassay result was assessed with respect to individual normal ranges (FLCκ = 3.3–19.4mg/L, FLCλ = 5.71–26.30mg/L, IgGκ = 4.03–9.78g/L, IgGλ = 1.97–5.71g/L, IgAκ = 0.48–2.82g/L, IgAλ = 0.36–1.98, IgMκ = 0.29–1.82g/L, IgMλ = 0.17–0.94g/L) and ratios (FLCκ/FLCλ = 0.26–1.65, Gκ/Gλ = 0.98–2.75, Aκ/Aλ = 0.8–2.04, Mκ/Mλ = 0.96–2.3). The results were used to create an algorithm which assessed the degree of abnormality individually and with respect to the other results generated. Diagnosis was recorded ∼3 months after the analysis and patients were followed for up to 3 years. Results293/1468 (19%) samples had an abnormal SPEP of which 95/293 were confirmed by IFE including: 10 intact immunoglobulin Multiple Myeloma (MM, 6 IgGκ, 1 IgGλ, 2 IgAκ, 1 IgAλ), 6 light chain MM (LCMM, 4 FLCκ, 2 FLCλ), 2 Waldenstrom macroglobulinemia (2 IgMκ), 1 IgMκ cryoglobulinemia, 3 CLL, 1 mantle cell lymphoma, 3 other lymphoma, 1 IgGκ plasmacytoma, and 68 MGUS patients, 26/68 were confirmed MGUS (1 FLCκ, 13 IgGκ, 4 IgGλ, 4 IgAκ, 2 IgMκ, 2 IgMλ), 42 were laboratory findings requiring follow up (1FLCκ, 3 FLCλ, 19 IgGκ, 5 IgGλ, 4 IgAκ, 1 IgAλ, 8 IgMκ, 1 IgMλ).FLC/HLC algorithm identified 85/95 IFE positive patients; of the 10 patients reported not identified by the algorithm, 3 had oligoclonal banding indicative of infection and 7 had monoclonal bands secondary to other diagnoses (including 2 colon cancer patients and 2 rheumatoid arthritis patients). In addition the algorithm identified 15 IFE negative patients with M-Ig production including: 2 AL amyloid patients, 1 asymptomatic LCMM, 1 patient with ∼1g/L FLCκ (lost to follow up), 1 follicular lymphoma patient and 1 CLL; diagnosis was not available for 9/15 patients. Furthermore, the algorithm identified 205/1468 patients as having elevated polyclonal FLC (cFLC) >50mg/L without evidence of clonal production. In a subset analysis comparing matched numbers of patients with elevated or normal cFLC concentrations, cFLC >50mg/L predicted all cause mortality (logistic regression odds ratio 9.96, 95% CI 4.72–21.0, p<0.001), and was associated with poorer overall survival (Kaplan Meier p<0.001, Cox regression hazard ratio 8.73, 95% CI 4.47–17.02). DiscussionMM presents with disparate and vague symptoms pervasive in an elderly population. Current guidelines recommend SPEP and FLC analysis reflexing to IFE for confirmation in the event of abnormality. However, SPEP and IFE require interpretation making the assessment variable. Standardised immunoassays (FLC+HLC) may obviate the need for interpretation and in this study identified additional haematological malignancies. Furthermore, standard assessments can be used to identify M-Ig, but in the absence of this finding may offer little information pertaining to the symptoms that prompted referral. By contrast this algorithm identified patients with elevated cFLC which in agreement with other reports was associated with increased risk of mortality. Further work is required to assess the algorithm, and determine if cFLC measurements should be used as a prompt for additional diagnostic investigations. Disclosures:Faint:The Binding Site Group, Ltd.: Employment. Harding:The Binding Site Group Ltd.: Membership on an entity's Board of Directors or advisory committees. Mirbahai:The Binding Site Group, Ltd: Employment.