Error analysis of the in vw 0 xrf measurement of lead in bone by monte carlo simulation

Q. Ao,S. H. Lee,R. P. Gardner

1999-01-01

Abstract:The possible sources of error that may affect the in vivo X-ray fluorescence measurement of lead in bone are identified and estimated by Monte Carlo simulation in this paper. The specific pux:se Monte Carlo code CEARXRF has been used for error analysis for the benchmark Cd based KXRF and polarized X-ray beam based LXRF lead measurement systems. Monte Carlo simulation indicates that the thickness of soft tissue overlying the bone, the distance from the device to the bone sample, the distribution of lead in bone, and bone dimensions significantly impact the accuracy of these methods. Use of the ratio of the Keel X ray to the coherent scattered source photon peak for the KXRF technique minimizes most measurement errors so that the accuracy is improved significantly by this approach. Using the L, and Lp X rays to measure lead content and skin thickness simultaneously provides a method for minimizing the skin thickness effect on the LXRF technique. The LXRF technique is found to be more sensitive. Introduction Two X-ray fluorescence .techniques have been used for the in vivo measurement of lead in bone; K X-ray fluorescence (KXRF) and L X-ray fluorescence (LXRF). The KXRF technique has been more widely used and validated [l]. It measures lead approximately 37 mm into bone. Therefore, it provides data on the total amount of lead throughout the bone. The LXRF technique, which has been used only in some pediatric studies, measures lead only 2 to 3 mm into the bone [2]. Unlike the KXRF method, this technique is highly sensitive to slight variations in thickness of skin overlying the bone. The theoretical advantage of LXRF is that it may provide specific information on lead levels in subperiosteal bone that may be more easily mobilizable than lead in dense cortical bone and may therefore more rapidly reflect environmental and physiological changes. A number of interfering factors have impacted the relatively poor minimum detection limit (MDL) that is possible with both techniques. These include: (1) statistical counting rate fluctuations, (2) signal-to-noise ratio, (3) covering skin thickness, (4) distance from the device to the bone sample, (5) bone and skin densities, (6) bone dimension (shape and size), (7) lead distribution in bone, and (8) interfering elements in bone and skin. Improvements in the MDC for both KXRF and LXRF methods are highly desirable by minimizing the errors due to the listed interfering factors. However, the systematic analysis of errors for the in viva lead XRF measurement still remains little studied. Our present work attempts to identify and quantify the errors in the KXRF and LXRF measurement of lead in bone so that the optimization of system design and analytical procedure can be achieved. Copyright (C) JCPDS-International Centre for Diffraction Data 1999