Gene pathway identification is an open and active research area that has attracted the interest not only of biomedical scientists but also of a large number of researchers from disciplines related to knowledge discovery from biological data. In this paper, we used Structural Equation Modeling (SEM) in order to statistically investigate the Toll-Like Receptor (TLR) signaling pathway in Chronic Lymphocytic Leukemia (CLL). Specifically, we used Path Analysis, a special case of SEM which is a statistical technique for testing and confirming causal relations based on data and qualitative assumptions. The dataset consists of Real Time PCR data for 84 genes relevant to the TLR signaling pathway, that were obtained from 192 patients with CLL that have been classified based on the mutational status of their clonotypic antigen receptors as mutated CLL (M-CLL) or unmutated CLL (U-CLL). The causal effects among genes were estimated through regression weights. In each case, the initially assumed model was based on the KEGG pathway database which provides reference pathways. The initial models were tested with respect to the M-CLL and U-CLL datasets. Modifications were made according to the statistical results (statistically significant regression weights, modification indices), concluding to models with good fit. Models were consistent to the reference pathway mostly for M-CLL and much less for U-CLL. These results go along with the well-described differences in immune signaling between the two subgroups, and may indicate that signaling in U-CLL is more impaired and/or modulated by complex regulatory networks that remain to be elucidated.