Feasibility of automated pancreas segmentation based on dynamic MRI

OBJECTIVE

MRI-guided radiotherapy is particularly attractive for abdominal targets with low CT contrast. To fully utilize this modality for pancreas tracking, automated segmentation tools are needed. A hybrid gradient, region growth and shape constraint (hGReS) method to segment two-dimensional (2D) upper abdominal dynamic MRI (dMRI) is developed for this purpose.

METHODS

2D coronal dynamic MR images of two healthy volunteers were acquired with a frame rate of 5 frames per second. The regions of interest (ROIs) included the liver, pancreas and stomach. The first frame was used as the source where the centres of the ROIs were manually annotated. These centre locations were propagated to the next dMRI frame. Four-neighborhood region transfer growth was performed from these initial seeds before refinement using shape constraints. RESULTS from hGReS and two other automated segmentation methods using integrated edge detection and region growth (IER) and level set, respectively, were compared with manual contours using Dice's index (DI).

RESULTS

For the first patient, the hGReS resulted in the organ segmentation accuracy as a measure by the DI (0.77) for the pancreas, superior to the level set method (0.72) and IER (0.71). The hGReS was shown to be reproducible on the second subject, achieving a DI of 0.82, 0.92 and 0.93 for the pancreas, stomach and liver, respectively. Motion trajectories derived from the hGReS were highly correlated to respiratory motion.

CONCLUSION

We have shown the feasibility of automated segmentation of the pancreas anatomy on dMRI.

ADVANCES IN KNOWLEDGE

Using the hybrid method improves segmentation robustness of low-contrast images.

Valosin-containing protein regulates the proteasome-mediated degradation of DNA-PKcs in glioma cells

DNA-dependent protein kinase (DNA-PK) has an important role in the repair of DNA damage and regulates the radiation sensitivity of glioblastoma cells. The VCP (valosine-containing protein), a chaperone protein that regulates ubiquitin-dependent protein degradation, is phosphorylated by DNA-PK and recruited to DNA double-strand break sites to regulate DNA damage repair. However, it is not clear whether VCP is involved in DNA-PKcs (DNA-PK catalytic subunit) degradation or whether it regulates the radiosensitivity of glioblastoma. Our data demonstrated that DNA-PKcs was ubiquitinated and bound to VCP. VCP knockdown resulted in the accumulation of the DNA-PKcs protein in glioblastoma cells, and the proteasome inhibitor MG132 synergised this increase. As expected, this increase promoted the efficiency of DNA repair in several glioblastoma cell lines; in turn, this enhanced activity decreased the radiation sensitivity and prolonged the survival fraction of glioblastoma cells in vitro. Moreover, the VCP knockdown in glioblastoma cells reduced the survival time of the xenografted mice with radiation treatment relative to the control xenografted glioblastoma mice. In addition, the VCP protein was also downregulated in ∼25% of GBM tissues from patients (WHO, grade IV astrocytoma), and the VCP protein level was correlated with patient survival (R²=0.5222, P<0.05). These findings demonstrated that VCP regulates DNA-PKcs degradation and increases the sensitivity of GBM cells to radiation.

MO-A-213AB-05: Synthesis of High Quantum Yield Nano-Scintillators for Simultaneous Photodynamic Therapy in Radiotherapy

PURPOSE

Radioresistant tumors provide one of the biggest challenges for improving radiation therapy efficacy. We have demonstrated that semiconductor quantum dots can be used as 'impedance matching' devices converting high energy X-rays to visible photon range for a simultaneous and mechanistically independent treatment; photodynamic therapy (PDT). Application of quantum dots was limited by it toxicity and relatively low quantum yield with X-ray excitation. To further improve the combined therapy strategy, we synthesize and screen for safer and more efficient nanoparticles.

METHODS

Colloidal GdSe and Gd2O3 nanoparticles were synthesized in octadecene using octylamine and oleic acid as surfactants. The average diameter of the nanoparticles was 10 nm. The samples in organic solution was secured in a light tight box and irradiated by 300 kVp X-rays with a dose rate of 300 cGy/min. The visible photon yield is collected by a lens and transferred to the photomultiplier tube via optic fibers. The average photon counts for 100 seconds were compared between GdSe, Gd2O3 and CdSe quantum dots.

RESULTS

Different than quantum dots with fixed band gap and emission peaks, the fluorescent peaks of GdSe and Gd2O3 were excitation light dependent. Shorter emission wavelengths were resulted from higher incident photon energies but the spectra overlap well with the Soret band of porphyrin photosensitizers. While their fluorescence is substantially weaker compared with QDs under UV excitation, fluorescence from both GdSe and Gd2O3 is one order of magnitude stronger than that of QD when excited by 300 kV X-rays.

CONCLUSIONS

Metal oxide nanoparticle scintillators are more efficient X-ray scintillators than Quantum dots. Moreover, Metal oxide nanoparticle can utilize the Soret band of photosensitizers for more effective energy transfer and excitation. These improvements will reduce the required radiation dose and drug concentration for simultaneous photodynamic therapy.

SU-D-BRB-05: Small Animal Lung Compliance Imaging: Assessment System for Tissue Sensitivity to Radiation Induced Lung Injury

PURPOSE

Recent clinical trials and animal studies have indicated that the tissue sensitivity to radiation induced lung injury (RILI) may be region- specific. In this study, we propose a new 4D cone beam CT (CBCT) basedcompliance imaging method to measure regional pulmonary function change in precisely irradiated small animal under CBCT guidance on small animal radiation research platform (SARRP) to facilitate our understanding of region-specific tissue sensitivity to RILI.

METHODS

Four Sprague-Dawley rats underwent prospective pressure gated 4D CBCT on SARRP. Three animals were selected as control group which underwent a second 4D CBCT scan. The fourth animal was irradiated in the central lung (24 Gy) using 3 × 3 mm collimating cone 2 months prior to the scan. The specific compliance (Csp) was calculated via the real time pressure measurement from the ventilator and displacement field from 3D B-spline image registration between the end of inhale and end of exhale phases from the 4D CBCT scan. The 3D Csp maps from the control animal group were mapped to the irradiated animal as a Csp functional atlas for statistical analysis. We alsoevaluated the repeatability of the Csp measurement on a voxel-by-voxel basis.

RESULTS

No significant Csp difference is found after two month of radiation between the irradiated rat (0.22±0.05) and the functional atlas (0.21±0.07). The observation is consistent with previous publications. The averaged linear correlation coefficient between the voxel-by-voxel Csp measurements from initial and repeat scans in control group is 0.98.

CONCLUSIONS

We proposed a method that uses 4D CBCT based compliance imaging to measure region-specific tissue sensitivity of RILI. We compared the irradiated animal two months after radiation with the control group. Our study shows an excellent robustness of the proposed method for regional lung tissue specific compliance measurement. This work was supported in part by UVa George Amorino Pilot Grant.

TU-G-BRB-01: Continuous Path Optimization for Non-Coplanar Variant SAD IMRT Delivery Using C-Arm Machines

PURPOSE

To develop and investigate a continuous path optimization methodology to traverse prescribed non-coplanar IMRT beams with variant SADs, by orchestrating the couch and gantry movement with zero-collision, minimal patient motion consequence and machine travel time.

METHODS

We convert the given collision zone definition and the prescribed beam location/angles to a tumor-centric coordinate, and represent the traversing path as a continuous open curve. We proceed to optimize a composite objective function consisting of (1) a strong attraction energy to ensure all prescribed beams are en-route, (2) a penalty for patient-motion inducing couch motion, and (3) a penalty for travel-time inducing overall path-length. Feasibility manifold is defined as complement to collision zone and the optimization is performed with a level set representation evolved with variational flows. The proposed method has been implemented and tested on clinically derived data. In the absence of any existing solutions for the same problem, we validate by: (1) visual inspecting the generated path rendered in the 3D tumor-centric coordinates, and (2) comparing with a traveling-salesman (TSP) solution obtained from relaxing the variant SADs and continuous collision-avoidance requirement.

RESULTS

The proposed method has generated delivery paths that are smooth and intuitively appealing. Under relaxed settings, our results outperform the generic TSP solutions and agree with specially tuned versions.

CONCLUSIONS

We have proposed a novel systematic approach that automatically determines the continuous path to cover non-coplanar, varying SAD IMRT beams. The proposed approach accommodates patient-specific collision zone definition and ensures its avoidance continuously. The differential penalty to couch and gantry motions allows customizable tradeoff between patient geometry stability and delivery efficiency. This development paves the path to achieve safe, accurate and efficient non-coplanar IMRT delivery with the advanced robotic controls in new-generation C-arm systems, enabling practical harvesting of the dose benefit offered by non-coplanar, variant SAD IMRT treatment.

SU-E-J-192: Static Breath-Hold MRI Based Measurement of Change in Pulmonary Function Following a Course of Radiation Therapy

PURPOSE

Radiation Therapy (RT) induced pulmonary function change may depend on the location, underlying function of that lung prior to radiations, radiation dose/fractionation and other factors. We propose to evaluate the radiation induced pulmonary function change using static breath-hold MRI scans with vascular information and 3D deformable image registration which can provide pulmonary function relative to RT dose on a regional basis.

METHODS

A MRI scan pair near the end of inhale and near the end of exhale with breath hold were acquired for one lung cancer patient before RT and 6 months after RT. The patient was treated with SBRT with 55 Gy to PTVs in the right and the left lung respectively. B-spline based vesselness preserving image registration algorithm was applied to register the MRI pair for the calculation of local lung expansion as a measurement of regional pulmonary function (PF). The PF maps before RT and after RT were then mapped to the planning CT using the same algorithm tuned for MRI-CT registration. The pulmonary function change was calculated via the PF ratio between two MRI pairs.

RESULTS

Strong spatial correlation was found between the irradiated lung region and the region with greatly decreased PF. Based on dose and PFC distribution, no strong determinant factor was found for PF lost in the left lung while the right lung shows that all the lung tissue receiving dose larger than 28 Gy will have a decreased PF.

CONCLUSIONS

We demonstrated a method that uses static breath-hold MRI based lung imaging to evaluate radiation induced pulmonary function change which can be applied to study the dose and the pulmonary function change in a regional basis. This work is supported by NIH grant support 1R21CA144063.

Serine/threonine protein phosphatase 6 modulates the radiation sensitivity of glioblastoma

Increasing the sensitivity of glioblastoma cells to radiation is a promising approach to improve survival in patients with glioblastoma multiforme (GBM). This study aims to determine if serine/threonine phosphatase (protein phosphatase 6 (PP6)) is a molecular target for GBM radiosensitization treatment. The GBM orthotopic xenograft mice model was used in this study. Our data demonstrated that the protein level of PP6 catalytic subunit (PP6c) was upregulated in the GBM tissue from about 50% patients compared with the surrounding tissue or control tissue. Both the in vitro survival fraction of GBM cells and the patient survival time were highly correlated or inversely correlated with PP6c expression (R²=0.755 and −0.707, respectively). We also found that siRNA knockdown of PP6c reduced DNA-dependent protein kinase (DNA-PK) activity in three different GBM cell lines, increasing their sensitivity to radiation. In the orthotopic mice model, the overexpression of PP6c in GBM U87 cells attenuated the effect of radiation treatment, and reduced the survival time of mice compared with the control mice, while the PP6c knocking-down improved the effect of radiation treatment, and increased the survival time of mice. These findings demonstrate that PP6 regulates the sensitivity of GBM cells to radiation, and suggest small molecules disrupting or inhibiting PP6 association with DNA-PK is a potential radiosensitizer for GBM.