Inferential modeling and predictive feedback control in real-time motion compensation using the treatment couch during radiotherapy.

Tumor motion induced by respiration presents a challenge to the reliable delivery of conformal radiation treatments. Real-time motion compensation represents the technologically most challenging clinical solution but has the potential to overcome the limitations of existing methods. The performance of a real-time couch-based motion compensation system is mainly dependent on two aspects: the ability to infer the internal anatomical position and the performance of the feedback control system. In this paper, we propose two novel methods for the two aspects respectively, and then combine the proposed methods into one system. To accurately estimate the internal tumor position, we present partial-least squares (PLS) regression to predict the position of the diaphragm using skin-based motion surrogates. Four radio-opaque markers were placed on the abdomen of patients who underwent fluoroscopic imaging of the diaphragm. The coordinates of the markers served as input variables and the position of the diaphragm served as the output variable. PLS resulted in lower prediction errors compared with standard multiple linear regression (MLR). The performance of the feedback control system depends on the system dynamics and dead time (delay between the initiation and execution of the control action). While the dynamics of the system can be inverted in a feedback control system, the dead time cannot be inverted. To overcome the dead time of the system, we propose a predictive feedback control system by incorporating forward prediction using least-mean-square (LMS) and recursive least square (RLS) filtering into the couch-based control system. Motion data were obtained using a skin-based marker. The proposed predictive feedback control system was benchmarked against pure feedback control (no forward prediction) and resulted in a significant performance gain. Finally, we combined the PLS inference model and the predictive feedback control to evaluate the overall performance of the feedback control system. Our results show that, with the tumor motion unknown but inferred by skin-based markers through the PLS model, the predictive feedback control system was able to effectively compensate intra-fraction motion.

Tracking the herd: resynchronization analysis of cell-cycle gene expression data in Saccharomyces cerevisiae.

Identification of genes expressed in a cell-cycle-specific periodical manner is of great interest to understand cyclic systems which play a critical role in many biological processes. However, identification of cell-cycle regulated genes by microarray gene expression data is complicated by the factor of synchronization loss, thus remains a challenging problem. Decomposing the expression measurements will allow to better represent the single-cell behavior and improve the accuracy in identifying periodically expressed genes. In this paper, we propose a resynchronization-based algorithm for identifying cell-cycle-related genes, where we present a simple synchronization loss model by modeling the gene expression measurements as a superposition of different cell populations growing at different rates. The underlying expression profile is reconstructed through resynchronization and is further fitted to the measurements in order to identify periodically expressed genes. Results from both simulations and real microarray data show that the proposed scheme is promising for identifying cycling genes and revealing underlying gene expression profiles.

Model-Based receptor quantization analysis for PET parametric imaging.

Dynamic PET (positron emission tomography) imaging technique allows image-wide quantification of physiologic and biochemical parameters. Compartment modeling is the most popular approach for receptor binding studies. However, current compartment-model based methods often either require the accurate arterial blood measurements as the input function or assume the existence of a reference region. To obviate the need for the input function or a reference region, in this paper, we propose to estimate the input function and the kinetic parameters simultaneously. The initial estimate of the input functions is obtained by the analysis of space intersections. Then both the input function and the receptor parameters, thus the underlying distribution volume (DV) parametric image, are estimated iteratively. The performance of the proposed scheme is examined by both simulations and real brain PET data in obtaining the underlying parametric images.