Digital Holographic Microscopy for Nanoscale Dose Calculation and Assessing Gold Nanoparticle Uptake in Live Cells

Volume 96  Number 2S  Supplement 2016 with results of the weekly CBCT and portal images from EPID. Motion curves of surface Intersection point and surface principal components were also compared with breathing volume curves from the ABC system. Results: Daily target setup errors were reduced from 0.9 cm to <0.4 cm by using the 3D/4D video surface-guidance that was always consistent with results of portal images and CBCT if there were no big body motion during CBCT scanning. Higher accuracy and precision of 0.2 cm and longer stability without hardware change for clinical usage in four years were achieved for the video camera operated at the imaging speed and tracking rate of 15 frames per second. The novel intersection point and surface motion trackings were also more sensitive and accurate than that of commercial RPM and ABC systems. UP to 2 cm body sag on the slant breast board during longer IMRT treatments and surface deformation for five large breasts and some patients undertaken deep breath hold irradiation were detected. Conclusion: Accurate and precise optical image-guided setup of breast and chest wall targets was clinically validated with thousands of portal images and CBCT scans on 140 patients. More importantly, one can reliably track irregular target motion during daily dose delivery and detect gross target deformation during the entire course of treatment. The video guidance provides us all-time IGRT of breast and chest wall cancer without large volume low dose irradiation in taking daily CBCT and/or portal images. Author Disclosure: S. Li: None. C. Miyamoto: None. K. Reilly: None. A. Padmanabhan: None. B. Micaily: None.

1012 Differences in Dose Coverage and Uniformity in Fractionated HighDose-Rate Interstitial Breast Brachytherapy Based on EMT Measurements M. Kellermeier,1 B. Hofmann,2 V. Strnad,1,2 and C. Bert1,2; 1Radiation Oncology, University Clinic Erlangen, Erlangen, Germany, 2FriedrichAlexander-University Erlangen-Nu¨rnberg, Erlangen, Germany Purpose/Objective(s): To analyze if electromagnetic tracking (EMT) can be used to measure the implant geometry in fractioned HDR interstitial brachytherapy (HDR-iBT) in the treatment of breast cancer. Furthermore, based on EMT-measured catheter traces, we tested if a dose distribution can be reconstructed for each treatment fraction, i.e. up to 10 times over 5 days. Materials/Methods: Between 01/2015 and 01/2016, 24 patients were accrued within an IRB approved study. The geometry of interstitial singleleader catheters (mean: 17 pcs) inside of the breast were measured first on the CT table just after acquiring the treatment planning scan (EMTCT) and second on the HDR treatment table directly after each treatment fraction (EMTFx). Each measurement was performed with 5 degrees of freedom (DoF) sensor manually and consecutively inserted in the catheters. The data were corrected for breathing motion using the information of three 6 DoF sensors on the patient’s chest. Dwell positions (DPs) were determined from the measured catheter traces. For each patient, the determined DPs from EMTCT and EMTFx were registered to the CT-derived DPs from treatment planning. EMT-determined DPs were used for dose calculation (TG-43 formalism). The planning target volume (PTV) and dwell time data were taken from treatment planning. Data were analyzed by determining the coverage index (CI Z VPTV(Dref) / VPTV) for the PTV and the dose non-uniformity ratio (DNR Z VD150% / VD100%). For each patient, the quality indexes from EMTFx were normalized to that from EMTCT indicated by index n: CIn and DNRn. Results: EMT of the whole catheter implant per patient took per fraction on average 6.8 minutes plus a few minutes for setup. For all patient measurements, the CIn ranged from 0.85 to 1.1 and the median CIn was slightly reduced to 0.99 (25th/75th percentile: 0.94 / 1.01). In reference to the treatment fractions, the median CIn showed a maximum deviation on the second day (CIn Z 0.96). The DNRn ranged from 0.72 to 1.38 over all dose calculations. The median DNRn showed a reduction to 0.99 (25th/ 75th percentile: 0.89 / 1.11). The maximum deviation for median DNRn to

ePoster Sessions S169 0.91 was also found on the second day. In addition to dose volume histograms, the 3D dose distributions were analyzed for local dose deviations. Conclusion: EMT measurements in 24 HDR-iBT breast patients were feasible within the clinical workflow and well-tolerated in this group of patients. Maximal dose deviations on the second day might impact boost treatments with two fractions only. Based on EMT determined dose calculations adaptive treatment protocols and tests for possible treatment delivery errors should be implemented. Further work is required on the point-based registration of EMT and CT coordinate systems. Author Disclosure: M. Kellermeier: Research Grant; Elekta. B. Hofmann: None. V. Strnad: Consultant; Elekta. C. Bert: Research Grant; Elekta.

1013 Digital Holographic Microscopy for Nanoscale Dose Calculation and Assessing Gold Nanoparticle Uptake in Live Cells C. Grassberger,1 P. Dinkelborg,2 A. McNamara,2 J.P. Schuemann,2 S.J. McMahon,2 H. Willers,3 H. Paganetti,1 and M. Wang3; 1Massachusetts General Hospital, Boston, MA, 2Massachusetts General Hospital, Harvard Medical School, Boston, MA, 3Department of Radiation Oncology, Massachusetts General Hospital, Boston, MA Purpose/Objective(s): To assess whether the three-dimensional density distribution and localization of gold nanoparticles (GNP) within live cancer cells can be obtained with Digital Holographic Microscopy (DHM). The exact location of GNP is crucial for their radiosensitizing properties due to the short range of emitted secondary electrons. Materials/Methods: DHM is a novel commercially available technique allowing non-invasive, quantitative cell tomography, measuring the refractive index throughout a cell in three dimensions with a lateral resolution of 100 nm and a depth resolution of w300 nm. The device allows real-time imaging of live cells at unparalleled resolution. Illumination is provided by a 520 nm low-power laser rotating around the sample at a 45 deg angle, a reference beam is split off to enable measurement of the phase-shift. Reconstruction of the acquired holograms using a complex deconvolution algorithm enables imaging past the optical diffraction limit with high sample contrast. We imaged a panel of cell lines including human fibroblasts and NCI-H1703, A549, Calu6, and MCF7 cancer cell lines in their cell-specific media. Results: All cell lines showed excellent sample contrast and we performed long-term imaging of the same cells >48 hours. Cytoplasm, endoplasmic reticulum, organelles (mitochondria, lysosomes, and ribosomes), nuclear membrane, and nucleoli could all be easily distinguished, demonstrating that the technique is able to quantitatively measure small changes in density of sub-cellular structures over time and compare them in absolute terms among cells. The optical density differences between mitochondria and cytoplasm were largest, 1.19  0.04 vs. 1.72  0.04 (mean  SD) and similar among cell lines. Such large density differences lead to heterogeneity in local energy deposition within cells during radiation therapy. Within the nucleus, large differences were observed between chromatin phases, with variations up to 0.32  0.04 with an average nuclear density of 1.4. The densest organelles observed have optical densities w2.0 and are suspected to be ribosomes. When H1703 cells are grown as tumor spheroids we detected significantly larger areas of condensed chromatin in the nucleus compared to traditional 2D-grown cells. 1.9 nm GNP added to H1703 and MCF7 cell lines in concentrations of 10-250 mg/mL could be localized within cells due to their high absorption of light around 510 nm. The images indicate that GNP preferentially accumulates around organelles close to the nuclear envelope, but the exact distribution varies among cells. Conclusion: DHM is a valuable tool to provide structural input for nanometer-scale dose calculations and to measure density differences within cells in vitro. The strong absorbing properties of GNP at the laser’s wavelength could enable localization of very small amounts of nanometersize GNP in live samples.

S170

International Journal of Radiation Oncology  Biology  Physics

Author Disclosure: C. Grassberger: None. P. Dinkelborg: None. A. McNamara: None. J.P. Schuemann: None. S.J. McMahon: None. H. Willers: None. H. Paganetti: None. M. Wang: None.

(BOV). This 2D image was calculated from the daily CBCT and was registered to the corresponding 2D range image calculated from the planning CT or the CBCT acquired on the first day of treatment. The range images for positioning were calculated from (1) anterior-posterior BOV to extract the lateral and longitudinal translations and yaw angle (couch angle), (2) lateral BOV to extract the anterior-posterior and longitudinal translations and pitch angle, and (3) a series of BOVs for oblique beams around the longitudinal axis to extract the roll angle. The energy of each beam was chosen so that the ranges were in the target region. In this initial investigation, in order to avoid the CBCT scatter effect on the range accuracy, the range based registration was examined between those from CBCTs only. Twenty sets of CBCTs from two cases, one medulloblastoma and one pelvic osteosarcoma, were investigated. Results: The registration results from the (volumetric) CBCT number based registration and the range image based registration were within 0.5  1.5 mm and 0.5  0.7 degree for the pelvis case, and within 0.3  1.1 mm and 0.1  0.4 degrees for the brain case. The extracted longitudinal translations from the range images (1) and (2) were consistent. The rootmean-square-difference of the corresponding ranges was 1.4  0.3 mm and 1.2  0.2 mm for the pelvis and brain cases, respectively. The anatomic structures matched better on the side where the range was calculated than the other side. In addition, the difference of two range images indicated the tissues between these two ranges. The image contrast of the range images was better than that of the radiography because the range contains the patient’s anatomic information along the beamlet before the beamlet reaches its range, whereas the X-ray radiography is the sum of the CT numbers for the whole of the patient’s body along the X-ray path. Conclusion: Range image based registration introduces a new, important dimension to improve patient positioning and range verification before beam is delivered. CBCT scatter must be corrected for a reliable range image and the registration. Author Disclosure: W. Yao: None. J.B. Farr: None. C. Hua: None. V. Moskvin: None. L. Zhao: None. M. Krasin: None. J.T. Lucas: None. C.L. Tinkle: None. T.E. Merchant: None.

1014 Monte Carlo Simulation of Microscopic Dose Enhancement of Glucose Conjugated Gold Nanoparticles for the I-125 Radioactive Seeds Brachytherapy R. Yang,1 Y. Chen,2 H. Wang,1 X. Zhang,1 J. Wang,1 and J. Li2; 1Peking University Third Hospital, Beijing, China, 2Tsinghua University, Beijing, China Purpose/Objective(s): To investigate the microscopic energy deposition effect of glucose conjugated gold nanoparticles (Glu-GNPs) within a tumor irradiated by I-125 radioactive seeds on the nanometer/cell scale, and to quantify the corresponding microscopic dose enhancement factor (mDEF) around Glu-GNPs based on the real intra-cellular localization and spatial distribution of Glu-GNPs. Materials/Methods: The spectra of secondary electrons from atoms of gold and molecules of water under I-125 seeds irradiation of colorectal CL-187 cells cultured with culture medium added with Glu-GNPs was simulated with the Monte Carlo code Geant4. The simulation was based on the real Glu-GNPs intra-cellular localization and distribution observed with a transmission electron Microscope (TEM). The gold/water electron dose point kernels and corresponding mDEF was computed based on the simulated spectra of secondary electrons. Results: The fluence of secondary electron was increased by a factor up to 100 over radial distances of 10 um, when Glu-GNPs was added. For the tested Glu-GNPs size and concentration, the microscopic energy deposition increased by a factor up to 500 and 100 in the area immediately surrounding the Glu-GNPs and 5 um far from the Glu-GNPs. The mDEF around the Glu-GNPs ranged from 10 to 100 within 5-30 um far from GluGNPs, and decreased to less than 5% beyond a radial distance of 50 um. The formation of intracellular nanoparticle clusters increased the maximum mDEF by more than 100% within 10 um from the clusters. Conclusion: Glu-GNPs can significantly enhance the microscopic energy deposition for I-125 radioactive seeds brachytherapy of tumors. The significant microscopic dose enhancement effect is limited within 30 um from the Glu-GNPs. The active tumor targeting strategy using Glu-GNPs is beneficial to maximize the radiobiological benefit from brachytherapy using I-125 radioactive seeds, especially the formation of Glu-GNPs clusters within the tumor cells. Author Disclosure: R. Yang: Research Grant; National Natural Science Foundation of China. Y. Chen: None. H. Wang: Research Grant; National Natural Science Foundation of China. X. Zhang: None. J. Wang: None. J. Li: None.

1015 A Feasibility Study on Proton RangeeBased Registration for Patient Positioning in Proton Therapy W. Yao,1 J.B. Farr,1 C.H. Hua,1 V. Moskvin,1 L. Zhao,1 M. Krasin,1 J.T. Lucas, Jr,2 C.L. Tinkle,1 and T.E. Merchant1; 1St. Jude Children’s Research Hospital, Memphis, TN, 2St Jude Children’s Research Hospital, Memphis, TN Purpose/Objective(s): Registration of image data between the planning computed tomography (CT) and cone beam CT (CBCT) of the patient on the couch by using the CT numbers is widely used in patient positioning. The registration goal in photon therapy is to match the tumor targets and surrounding tissues, but this goal is insufficient for proton therapy, where the target coverage depends on the proton range along the proton beam path, not just the matching of tumor targets and surrounding tissues. Therefore, we hypothesize that proton range-based registration, in addition to CT numberebased registration, can improve patient positioning in proton therapy. Materials/Methods: The range of a mono-energetic proton beam in the patient body forms a two-dimensional (2D) image on the beam of view

1016 Autonomous Quality Assurance for Spot-Scanning Proton Therapy and SRS/SBRT Using Radioluminescent Phantoms, Optical Imaging, and Machine Vision C.H. Jenkins,1,2 Y. Yang,1 S.J. Yu,1 H. Yu,3 Y. Matsuzaki,4 T. Yoshimura,5 Y. Fujii,5 K. Umegaki,6,7 H. Shirato,5,7 and L. Xing1,7; 1Department of Radiation Oncology, Stanford University, Stanford, CA, 2Department of Mechanical Engineering, Stanford University, Stanford, CA, 3Department of Electrical Engineering, Stanford University, Stanford, CA, 4Proton Beam Therapy Center, Hokkaido University Hospital, Sapporo, Japan, 5 Department of Radiation Medicine, Hokkaido University Graduate School of Medicine, Sapporo, Japan, 6Faculty of Engineering, Hokkaido University, Sapporo, Japan, 7Global Station for Quantum Medical Science and Engineering, Global Institution for Collaborative Research and Education (GI-CoRE), Hokkaido University, Sapporo, Japan Purpose/Objective(s): Routine quality assurance (QA) is critical for ensuring proper system operation and accurate treatment delivery. The recent development of phosphor based visualization of radiation beams offers a unique opportunity for autonomous evaluation of radiation delivery. The purpose of this study was to investigate the feasibility of a fully autonomous, self-calibrated system for performing QA measurements for spot scanning proton therapy systems and linear accelerators (LINAC) used for stereotactic radiosurgery (SRS) and stereotactic body radiosurgery (SBRT) applications. Materials/Methods: An autonomous QA system, consisting of a custom radioluminescent phantom, an optical imaging system, and image processing software, was developed to perform the mechanical alignment tests specified in TG-142. The phantom, fabricated on a 3D printer and coated with a mixture of Gd2O2S:Tb and PDMS, was placed on the treatment couch near isocenter (+/- 1 cm). A CMOS digital camera was used to image the geometric shape of the irradiating proton or X-ray beam as well