Automatic online and real-time tumour motion monitoring during stereotactic liver treatments on a conventional linac by combined optical and sparse monoscopic imaging with kilovoltage x-rays (COSMIK).

The purpose of this study was to develop, validate and clinically demonstrate fully automatic tumour motion monitoring on a conventional linear accelerator by combined optical and sparse monoscopic imaging with kilovoltage x-rays (COSMIK). COSMIK combines auto-segmentation of implanted fiducial markers in cone-beam computed tomography (CBCT) projections and intra-treatment kV images with simultaneous streaming of an external motion signal. A pre-treatment CBCT is acquired with simultaneous recording of the motion of an external marker block on the abdomen. The 3-dimensional (3D) marker motion during the CBCT is estimated from the auto-segmented positions in the projections and used to optimize an external correlation model (ECM) of internal motion as a function of external motion. During treatment, the ECM estimates the internal motion from the external motion at 20 Hz. KV images are acquired every 3 s, auto-segmented, and used to update the ECM for baseline shifts between internal and external motion. The COSMIK method was validated using Calypso-recorded internal tumour motion with simultaneous camera-recorded external motion for 15 liver stereotactic body radiotherapy (SBRT) patients. The validation included phantom experiments and simulations hereof for 12 fractions and further simulations for 42 fractions. The simulations compared the accuracy of COSMIK with ECM-based monitoring without model updates and with model updates based on stereoscopic imaging as well as continuous kilovoltage intrafraction monitoring (KIM) at 10 Hz without an external signal. Clinical real-time tumour motion monitoring with COSMIK was performed offline for 14 liver SBRT patients (41 fractions) and online for one patient (two fractions). The mean 3D root-mean-square error for the four monitoring methods was 1.61 mm (COSMIK), 2.31 mm (ECM without updates), 1.49 mm (ECM with stereoscopic updates) and 0.75 mm (KIM). COSMIK is the first combined kV/optical real-time motion monitoring method used clinically online on a conventional accelerator. COSMIK gives less imaging dose than KIM and is in addition applicable when the kV imager cannot be deployed such as during non-coplanar fields.

Cone-Beam Computed Tomography Internal Motion Tracking Should Be Used to Validate 4-Dimensional Computed Tomography for Abdominal Radiation Therapy Patients.

PURPOSE: To demonstrate that fiducial tracking during pretreatment Cone-Beam CT (CBCT) can accurately measure tumor motion and that this method should be used to validate 4-dimensional CT (4DCT) margins before each treatment fraction. METHODS AND MATERIALS: For 31 patients with abdominal tumors and implanted fiducial markers, tumor motion was measured daily with CBCT and fluoroscopy for 202 treatment fractions. Fiducial tracking and maximum-likelihood algorithms extracted 3-dimensional fiducial trajectories from CBCT projections. The daily internal margin (IM) (ie, range of fiducial motion) was calculated for CBCT and fluoroscopy as the 5th-95th percentiles of displacement in each cardinal direction. The planning IM from simulation 4DCT (IM4DCT) was considered adequate when within ±1.2 mm (anterior-posterior, left-right) and ±3 mm (superior-inferior) of the daily measured IM. We validated CBCT fiducial tracking as an accurate predictive measure of intrafraction motion by comparing the daily measured IMCBCT with the daily IM measured by pretreatment fluoroscopy (IMpre-fluoro); these were compared with pre- and posttreatment fluoroscopy (IMfluoro) to identify those patients who could benefit from imaging during treatment. RESULTS: Four-dimensional CT could not accurately predict intrafractional tumor motion for ≥80% of fractions in 94% (IMCBCT), 97% (IMpre-fluoro), and 100% (IMfluoro) of patients. The IMCBCT was significantly closer to IMpre-fluoro than IM4DCT (P<.01). For patients with median treatment time t < 7.5 minutes, IMCBCT was in agreement with IMfluoro for 93% of fractions (superior-inferior), compared with 63% for the t > 7.5 minutes group, demonstrating the need for patient-specific intratreatment imaging. CONCLUSIONS: Tumor motion determined from 4DCT simulation does not accurately predict the daily motion observed on CBCT or fluoroscopy. Cone-beam CT could replace fluoroscopy for pretreatment verification of simulation IM4DCT, reducing patient setup time and imaging dose. Patients with treatment time t > 7.5 minutes could benefit from the addition of intratreatment imaging.

Automated patient setup and gating using cone beam computed tomography projections.

In radiation therapy, fiducial markers are often implanted near tumors and used for patient positioning and respiratory gating purposes. These markers are then used to manually align the patients by matching the markers in the cone beam computed tomography (CBCT) reconstruction to those in the planning CT. This step is time-intensive and user-dependent, and often results in a suboptimal patient setup. We propose a fully automated, robust method based on dynamic programming (DP) for segmenting radiopaque fiducial markers in CBCT projection images, which are then used to automatically optimize the treatment couch position and/or gating window bounds. The mean of the absolute 2D segmentation error of our DP algorithm is 1.3 ± 1.0 mm for 87 markers on 39 patients. Intrafraction images were acquired every 3 s during treatment at two different institutions. For gated patients from Institution A (8 patients, 40 fractions), the DP algorithm increased the delivery accuracy (96 ± 6% versus 91 ± 11%, p < 0.01) compared to the manual setup using kV fluoroscopy. For non-gated patients from Institution B (6 patients, 16 fractions), the DP algorithm performed similarly (1.5 ± 0.8 mm versus 1.6 ± 0.9 mm, p = 0.48) compared to the manual setup matching the fiducial markers in the CBCT to the mean position. Our proposed automated patient setup algorithm only takes 1-2 s to run, requires no user intervention, and performs as well as or better than the current clinical setup.

Using dynamic programming to improve fiducial marker localization.

Fiducial markers are used in a wide range of medical imaging applications. In radiation therapy, they are often implanted near tumors and used as motion surrogates that are tracked with fluoroscopy. We propose a novel and robust method based on dynamic programming (DP) for retrospectively localizing radiopaque fiducial markers in fluoroscopic images. Our method was compared to template matching (TM) algorithms on 407 data sets from 24 patients. We found that the performance of TM varied dramatically depending on the template used (ranging from 47% to 92% of data sets with a mean error <1 mm). DP by itself requires no template and performed as well as the best TM method, localizing the markers in 91% of the data sets with a mean error <1 mm. Finally, by combining DP and TM, we were able to localize the markers in 99% of the data sets with a mean error <1 mm, regardless of the template used. Our results show that DP can be a powerful tool for analyzing tumor motion, capable of accurately locating fiducial markers in fluoroscopic images regardless of marker type, shape, and size.

High-resolution photoacoustic tomography of resting-state functional connectivity in the mouse brain.

The increasing use of mouse models for human brain disease studies presents an emerging need for a new functional imaging modality. Using optical excitation and acoustic detection, we developed a functional connectivity photoacoustic tomography system, which allows noninvasive imaging of resting-state functional connectivity in the mouse brain, with a large field of view and a high spatial resolution. Bilateral correlations were observed in eight functional regions, including the olfactory bulb, limbic, parietal, somatosensory, retrosplenial, visual, motor, and temporal regions, as well as in several subregions. The borders and locations of these regions agreed well with the Paxinos mouse brain atlas. By subjecting the mouse to alternating hyperoxic and hypoxic conditions, strong and weak functional connectivities were observed, respectively. In addition to connectivity images, vascular images were simultaneously acquired. These studies show that functional connectivity photoacoustic tomography is a promising, noninvasive technique for functional imaging of the mouse brain.

Molecular determinants of pentamidine-induced hERG trafficking inhibition.

Pentamidine is an antiprotozoal compound that clinically causes acquired long QT syndrome (acLQTS), which is associated with prolonged QT intervals, tachycardias, and sudden cardiac arrest. Pentamidine delays terminal repolarization in human heart by acutely blocking cardiac inward rectifier currents. At the same time, pentamidine reduces surface expression of the cardiac potassium channel I(Kr)/human ether à-go-go-related gene (hERG). This is unusual in that acLQTS is caused most often by direct block of the cardiac potassium current I(Kr)/hERG. The present study was designed to provide a more complete picture of how hERG surface expression is disrupted by pentamidine at the cellular and molecular levels. Using biochemical and electrophysiological methods, we found that pentamidine exclusively inhibits hERG export from the endoplasmic reticulum to the cell surface in a heterologous expression system as well as in cardiomyocytes. hERG trafficking inhibition could be rescued in the presence of the pharmacological chaperone astemizole. We used rescue experiments in combination with an extensive mutational analysis to locate an interaction site for pentamidine at phenylalanine 656, a crucial residue in the canonical drug binding site of terminally folded hERG. Our data suggest that pentamidine binding to a folding intermediate of hERG arrests channel maturation in a conformational state that cannot be exported from the endoplasmic reticulum. We propose that pentamidine is the founding member of a novel pharmacological entity whose members act as small molecule antichaperones.

Resolution of crossing fibers with constrained compressed sensing using diffusion tensor MRI.

Diffusion tensor imaging (DTI) is widely used to characterize tissue micro-architecture and brain connectivity. In regions of crossing fibers, however, the tensor model fails because it cannot represent multiple, independent intra-voxel orientations. Most of the methods that have been proposed to resolve this problem require diffusion magnetic resonance imaging (MRI) data that comprise large numbers of angles and high b-values, making them problematic for routine clinical imaging and many scientific studies. We present a technique based on compressed sensing that can resolve crossing fibers using diffusion MRI data that can be rapidly and routinely acquired in the clinic (30 directions, b-value equal to 700 s/mm2). The method assumes that the observed data can be well fit using a sparse linear combination of tensors taken from a fixed collection of possible tensors each having a different orientation. A fast algorithm for computing the best orientations based on a hierarchical compressed sensing algorithm and a novel metric for comparing estimated orientations are also proposed. The performance of this approach is demonstrated using both simulations and in vivo images. The method is observed to resolve crossing fibers using conventional data as well as a standard q-ball approach using much richer data that requires considerably more image acquisition time.

Resolution of Crossing Fibers with Constrained Compressed Sensing using Traditional Diffusion Tensor MRI.

Diffusion tensor imaging (DTI) is widely used to characterize tissue micro-architecture and brain connectivity. Yet, DTI suffers serious limitations in regions of crossing fibers because traditional tensor techniques cannot represent multiple, independent intra-voxel orientations. Compressed sensing has been proposed to resolve crossing fibers using a tensor mixture model (e.g., Crossing Fiber Angular Resolution of Intra-voxel structure, CFARI). Although similar in spirit to deconvolution approaches, CFARI uses sparsity to stabilize estimation with limited data rather than spatial consistency or limited model order. Here, we extend the CFARI approach to resolve crossing fibers through a strictly positive, parsimonious mixture model. Together with an optimized preconditioned conjugate gradient solver, estimation error and computational burden are greatly reduced over the initial presentation. Reliable estimates of intra-voxel orientations are demonstrated in simulation and in vivo using data representative of typical, low b-value (30 directions, 700 s/mm(2)) clinical DTI protocols. These sequences are achievable in 5 minutes at 3 T, and the whole brain CFARI analysis is tractable for routine analysis. With these improvements, CFARI provides a robust framework for identifying intra-voxel structure with traditional DTI and shows great promise in helping to resolve the crossing fiber problem in current clinical imaging studies.