摘要
We previously demonstrated immune activation in the maternal peripheral circulation associated with preterm labor (PTL). There was an elevation in WBC mRNA of anti-inflammatory complement decay-accelerating factor (CD55) and the innate-immune response activating toll-like receptor 4 (TLR4). These findings suggested that collectively, these two molecules might serve as useful biomolecules to aid in the diagnosis of PTL. In this study, we used a combined marker approach to determine whether a dual marker model utilizing both CD55 and TLR4 mRNA levels to classify PTL would increase diagnostic accuracy compared to either molecule alone. Two methods were evaluated;a linear discriminant (LD) method and a distribution free (DF) method, in order to find the optimal linear combination of TLR4 and CD55 data to diagnose PTL accurately. Our results indicated that a combined CD55-TLR4 dual marker model could provide statistically significant improve- ments compared to CD55 or TLR4 single marker models for PTL classification performance.
We previously demonstrated immune activation in the maternal peripheral circulation associated with preterm labor (PTL). There was an elevation in WBC mRNA of anti-inflammatory complement decay-accelerating factor (CD55) and the innate-immune response activating toll-like receptor 4 (TLR4). These findings suggested that collectively, these two molecules might serve as useful biomolecules to aid in the diagnosis of PTL. In this study, we used a combined marker approach to determine whether a dual marker model utilizing both CD55 and TLR4 mRNA levels to classify PTL would increase diagnostic accuracy compared to either molecule alone. Two methods were evaluated;a linear discriminant (LD) method and a distribution free (DF) method, in order to find the optimal linear combination of TLR4 and CD55 data to diagnose PTL accurately. Our results indicated that a combined CD55-TLR4 dual marker model could provide statistically significant improve- ments compared to CD55 or TLR4 single marker models for PTL classification performance.
