Diagnostic accuracy of intrathecal kappa free light chains compared with OCBs in MS

Abstract

Objective: To determine what kappa free light chain (KFLC) metric has the highest capacity to separate healthy patients from patients with MS, we evaluated the sensitivity, specificity, and the overall diagnostic accuracy of 4 different KFLC metrics. To assess the usefulness of KFLC in the diagnostics of MS, we compared the different KFLC metrics with oligoclonal bands (OCBs), the current gold standard biochemical method to demonstrate intrathecal antibody production.

Methods: CSF and plasma were collected from patients with confirmed or suspected MS, other neurological diseases, as well as symptomatic and healthy controls between May 2017 and May 2018 (n = 335) at the Department of Neurology, Karolinska University Hospital, as part of routine diagnostic workup. KFLC analysis and isoelectric focusing for the detection of oligoclonal bands (OCB) were determined and correlated with diagnosis. Receiver operating characteristic (ROC) curve analysis was used to determine accuracy.

Results: OCBs yielded a sensitivity of 87% and a specificity of 100%. All KFLC metrics showed a high sensitivity (89%-95%) and specificity (95%-100%). Using the optimal cutoff according to the Youden Index resulted for the KFLC intrathecal fraction in a cutoff of -0.41 with a sensitivity of 95% and a specificity of 97% and for CSF KFLC/CSF albumin with a cutoff of 1.93 × 10^-3 with a sensitivity of 94% and specificity of 100%.

Conclusion: All evaluated KFLC metrics have excellent accuracy, and both KFLC intrathecal fraction and CSF KFLC/CSF albumin are at least as good as OCB in separating patients with MS from a control group.

Classification of evidence: This study provides Class III evidence that CSF KFLC accurately distinguishes patients with MS from healthy controls.

Figure 1. STARD flow diagram of index test kappa free light chain intrathecal fraction (KFLC IF) and reference standard OCB resulting in a control group (n = 60) compared with patients with MS (n = 62)

KFLC IF below −0.5 = negative index test, KFLC IF −0/5 to 0 = inconclusive index test, KFLC IF above 0 = positive index test. n = 122 including inconclusive index test. CIS = clinically isolated syndrome; OCB = oligoclonal band; PPMS = primary progressive MS; SPMS = secondary progressive MS.

Figure 2. Distribution of KFLC in 6 different patient subgroups in the whole cohort (n = 335)

(A) KFLC IF. Scatterplots show (A) KFLC IF with the gray zone (−0.5 to 0) marked with a red square and (B) CSF KFLC/CSF albumin for 6 different patient subgroups in the whole cohort (n = 335). Bars show medians and interquartile ranges. CINDC = central inflammatory neurologic disease control; CIS = clinically isolated syndrome; IF = intrathecal fraction; KFLC = kappa free light chain; NINDC = noninflammatory neurologic disease control; PINDC = peripheral inflammatory neurologic disease control; RIS = radiologically isolated syndrome.

Figure 3. ROC curves of 4 different KFLC metrics in the STARD cohort

ROC curves for KFLC IF, KFLC index, CSF KFLC, CSF KFLC/CSF albumin, and oligoclonal band (OCB) from samples in the prospective cohort with values for all metrics including OCB. ROC analysis performed for the control group (n = 60) vs patients with MS (n = 62). AUC reported as percentage with 95% CIs in parentheses. AUC = area under the curve; IF = intrathecal fraction; KFLC = kappa free light chain; ROC = receiver operating characteristic.