Clinical specificity of two assays for immunoglobulin kappa and lambda free light chains.

OBJECTIVES: Free light chain (FLC) assays and the ratio of κ/λ are recommended for diagnosis, prognosis and monitoring of plasma cell dyscrasias (PCD). Limited data exists on FLC clinical specificity in patients diagnosed with other conditions. METHODS: We assessed the κ, λ, and κ/λ FLC ratio using the FreeLite assay and the Sebia FLC ELISA assay in 176 patients with clinical presentations of fatigue, anemia, polyclonal hypergammaglobulinemia, joint disorders, kidney disease and non PCD-cancers with no monoclonal protein observed on serum protein electrophoresis or MASS-FIX immunoglobulin isotyping. Manufacturer defined reference intervals (RI) and glomerular filtration rate (GFR) specific RI (renal RI) were utilized. RESULTS: For the κ/λ ratio, 68.7 % (121/176) of specimens on the FreeLite and 87.5 % (154/176) of specimens on the Sebia assay were within RI. For κ, 68.2 % (120/176) and 72.2 % (127/176) of results were outside RI for FreeLite and Sebia respectively. For λ, 37.5 % (66/176) and 84.1 % (148/176) of FreeLite and Sebia results were outside RI. With FreeLite and Sebia, patients with kidney disease (n=25) had the highest κ/λ ratios. 44 patients (25.0 %) had GFR <60 mL/min/BSA. When renal RI were applied, 13.6 % had a FLCr outside the renal RI with FreeLite, and 4.5 % with Sebia. CONCLUSIONS: In a cohort of patients with signs and symptoms suggestive of PCDs, but ultimately diagnosed with other conditions, Sebia FLC had improved clinical specificity relative to FreeLite, if one was using an abnormal κ/λ ratio as a surrogate for monoclonality.

GPT-4 Underperforms Experts in Detecting IV Fluid Contamination.

BACKGROUND: Specimens contaminated with intravenous (IV) fluids are common in clinical laboratories. Current methods for detecting contamination rely on insensitive and workflow-disrupting delta checks or manual technologist review. Herein, we assessed the utility of large language models for detecting contamination by IV crystalloids and compared its performance to multiple, but variably trained healthcare personnel (HCP). METHODS: Contamination of basic metabolic panels was simulated using 0.9% normal saline (NS), with (n = 30) and without (n = 30) 5% dextrose (D5NS), at mixture ratios of 0.10 and 0.25. A multimodal language model (GPT-4) and a diverse panel of 8 HCP were asked to adjudicate between real and contaminated results. Classification performance, mixture quantification, and confidence was compared by Wilcoxon rank sum. RESULTS: The 95% CIs for accuracy were 0.57-0.71 vs 0.73-0.80 for GPT-4 and HCP, respectively, on the NS set and 0.57-0.57 vs 0.73-0.80 on the D5NS set. HCP overestimated severity of contamination in the 0.10 mixture group (95% CI of estimate error, 0.05-0.20) for both fluids, while GPT-4 markedly overestimated the D5NS mixture at both ratios (0.16-0.33 for NS, 0.11-0.35 for D5NS). There was no correlation between reported confidence and likelihood of a correct classification. CONCLUSIONS: GPT-4 is less accurate than trained HCP for detecting IV fluid contamination of basic metabolic panel results. However, trained individuals were imperfect at identifying contaminated specimens implying the need for novel, automated tools for its detection.