A novel and robust approach to defining population-based reference ranges for a serum free light chain assay in a clinical laboratory. (9th November 2022)

• Record Type: Journal Article
• Authors: Treger, Rebecca; Mathias, Patrick; Hutchinson, Kathleen; Bryan, Andrew; Chaudhary, Anu; Fink, Susan; Wener, Mark; Morishima, Chihiro
• Is Part Of: American journal of clinical pathology. Volume 158(2022)Supplement 1
• Page Start: S2
• Page End: S3
• Publication Date: 2022-11-09
• Subjects: Diagnosis, Laboratory -- Periodicals; Pathology -- Periodicals; 616.07
• DOI: 10.1093/ajcp/aqac126.003
• Languages: English
• ISSNs: 0002-9173

Abstract

Abstract: Serum protein electrophoresis and immunofixation are used to detect the presence of monoclonal heavy and light immunoglobulin chains for the diagnosis of plasma cell neoplasms (PCN). In conjunction, the Freelite® assay has been used to detect excess free kappa and lambda light chains produced by PCN, where the kappa:lambda free light chain ratio (KLR) serves as a surrogate marker of clonality and disease. The manufacturer-defined KLR reference range of 0.26-1.65 is used in internationally recognized criteria for PCN diagnosis and monitoring, and KLR results affect clinical decision making, disease categorization, and response assessment. Our laboratory has verified the manufacturer-defined Freelite® KLR reference range three times using healthy control sera since assay implementation in 2006. Despite this, high percentages of abnormal KLR have consistently been observed among immunofixation-defined monoclonal-negative specimens. We therefore retrospectively interrogated 48, 012 serum free light chain clinical results reported at our institution between 2010-2020 by extracting results from our laboratory information system and analyzing population distributions using the R programming language. For all specimens, the distribution of KLR values was shifted towards the towards the upper limit of the manufacturer-defined reference range, indicating that it poorly describes both our monoclonal-negative and monoclonal-positive patient populations. Furthermore, >80% of monoclonal-negative specimens exhibited KLR above the median of the manufacturer-defined reference range (0.26-1.65), while 24% (1, 226/5, 057) unexpectedly exhibited KLR >1.65. We found that this phenomenon has been present consistently since 2010 and is not affected by patient sex or age, impaired renal function, lot-to-lot variation, or assay drift. We therefore sought to establish an institution-appropriate KLR reference range by defining a reference population from within our existing dataset. We leveraged International Classification of Diseases codes, estimated glomerular function data, and serum/urine immunofixation results to exclude specimens from individuals with PCN- or renal-related diagnoses, impaired renal function, and prior, current, or subsequent monoclonal components. The resultant 1, 536 monoclonal-negative reference specimens were then classified according to the manufacturer-defined KLR reference range, where 13% remained abnormally elevated, consistent with the published false positive rate in healthy volunteers. Lastly, we determined a new KLR reference range of 0.66-2.21 using the central 95% of KLR values from these reference specimens. When applied to our dataset, this institution-specific KLR reference range improved the accurate classification of specimens containing monoclonal free light chains by immunofixation, while reducing the false-positive rate in monoclonal-negative specimens. Together, our findings support the use of historical KLR and immunofixation data from routine clinical testing to derive institution-appropriate KLR reference ranges, which can then be used for continuous quality improvement. Implementation of this more-accurate KLR reference range should improve subsequent test utilization by non-specialist providers, reduce concern for/over-diagnosis of kappa-involved disease, and improve the detection and monitoring of lambda-involved disease.