Archive | 2019
Expression and diagnostic utility of single and combined CD200, CD148 and CD160 markers in mature B‑cell neoplasms as revealed by ROC and SVM analyses
Abstract
Timely and accurate diagnosis of mature B-cell neoplasm (MBN) subtypes is crucial for initiating a proper management plan. The aim of this study was to evaluate the diagnostic utility of CD200, CD148 and CD160 in different MBN subtypes using two discriminatory statistical techniques, namely receiver operating characteristic (ROC) curve analysis and support vector machine (SVM). This study included 86 patients with MBNs, 61 males and 25 females, whose data of medical histories and clinical examinations were collected. Fresh blood/bone marrow specimens were also collected and subjected to detailed morphological, cytogenetic and flow cytometric immunophenotypic examination. Diagnosis was established according to the 2017 WHO guidelines. The discriminatory performance of single and combined markers was assessed using ROC analysis. The discriminatory accuracy of the best markers was analyzed using SVM and compared to ROC analysis. The results revealed that CD148/CD200 demonstrated the best diagnostic utility in discriminating mantle cell lymphoma (MCL) from chronic lymphocytic leukemia (CLL), and hairy cell leukemia (HCL) from lymphoplasmacytic lymphoma (LPL) plus splenic marginal zone lymphoma (SMZL), with a sensitivity and specificity of 100% at signal ratios of >2.6 and <0.15, respectively. CD200 demonstrated equivalent diagnostic performance only in discriminating HCL from LPL plus SMZL at a signal cut‐off value of >279. Discrimination between SMZL and LPL was unfeasible with any of the used markers. The discriminatory accuracy of SVM using CD148 and CD200 was comparable to that of ROC analysis. On the whole, the findings of this study indicate that CD148/CD200 demonstrated the best diagnostic utility in discriminating MCL from CLL, and HCL from LPL plus SMZL. Combined markers offer a diagnostic value, particularly in difficult cases with anomalous marker expression. SVM could be used for the efficient analysis of flow cytometric data.