E592
International Journal of Radiation Oncology Biology Physics
Conclusion: The concordance of abdominal wall motion to tumor motion can vary significantly. This may depend on the portion of the abdominal wall that is tracked. As there exists patients who have discordant motion of the abdominal wall and tumor, they may benefit from real-time MRI guided tracking during radiation treatment. Author Disclosure: A.P. Wojcieszynski: None. S.A. Rosenberg: None. C.R. Hullett: None. M.W. Geurts: None. Z.E. Labby: None. P.M. Hill: None. R.B. Bayliss: None. B.R. Paliwal: None. J. Bayouth: None. M. Bassetti: None.
Purpose/Objective(s): Photodynamic therapy (PDT) is an effective treatment modality for some superficial tumors, which uses laser light to activate photosensitizers that have been selectively absorbed by tumor cells. However, the clinical application of PDT has been limited by the finite penetration depth (typically 5mm) of laser light. Cerenkov emission is induced when a charged particle moves faster than the speed of light in a medium, which can excite fluorophore, protoporphyrin IX (PpIX) embedded in biological phantoms. This study investigates the excitation efficiency of PpIX by Cerenkov emission from 45MV photon beams on a LA45 accelerator for cancer therapy, which is termed as radiodynamic therapy (RDT). Materials/Methods: There are two different source components for Cerenkov emission at 45MV: (1) a fast component from the Compton electrons and pair productions e/e+, and (2) a slow component from positrons and electrons emitted by the radioactive decay nuclei produced by photonuclear reactions. Cerenkov emission energy and spatial distributions in excited target areas and their efficiency for PpIX were simulated by Monte Carlo simulations. The inner excitation of fluorophore by Cerenkov emission was theoretically compared to exotic excitation from the external laser light in PDT. A patented catalyst coenzyme was added as a substrate to increase the excitation efficiency. A specific probe of DMA (Singlet O2 fluorescent probe-9, 10-dimethylanthracene) was used to detect singlet oxygen. When fluorescent DMA was attacked by singlet oxygen, it will become a non-fluorescent product. Finally we compared our results with previous experimental results reported in the literature. Results: Our Monte Carlo results showed that the Cerenkov emission intensity induced at 45MV was 10 times of that at 6MV. The Cerenkov emission spectrum was peaked at 400-450nm. The PpIX excitation efficiency for fluorophore at 400-450nm (Soret Band) in RDT is 20 times of that for laser light at 630nm in PDT. The homogenous inner excitation from RDT is 20 times (continuous spectrum excitation and innerscattering) of the exotic excitation from exponential attenuation laser light in the target in PDT. Furthermore, the patented catalyst coenzyme enhanced the excitation efficiency of PpIX by 3-6 times. Our preliminary clinical results have showed favorable outcome of RDT for early and late stage cancers including GBM, H&N, breast, lung, liver, colon, prostate, etc. and as a systemic treatment modality for metastatic cancers. Conclusion: Our results indicated that the excite efficiency for PpIX from 45MV photon radiation was similar to that in PDT. RDT may be developed into a potential treatment modality for cancer of various stages and other diseases. Further experimental studies and clinical trials are warranted to investigate the mechanisms of RDT and quantify its clinical efficacy for cancer therapy. Author Disclosure: Z. Quanshi: None. Q. Sun: None. G. Xiao: None. J. Zeng: None. L. Chen: None. C. Ma: None.
3483 Characterization and Correction of Bragg Peak Location Uncertainties Resulting From Inherent Uncertainty in Proton Stopping Powers J.G. Hardie and C. Beltran; Mayo Clinic, Rochester, MN Purpose/Objective(s): To characterize the effects of systematic stopping power (SP) errors on Bragg peak position/proton range, and develop a procedure for measuring these discrepancies and correcting for them. Materials/Methods: The known proton SP are combinations of theoretical derivation, data fits, and empirical corrections, and are reasonably well known for elements and simple compounds, much less well known for complex materials like biological tissues. Uncertainties in SPs lead to range errors of 1-3%, similar to most of the geometric uncertainties corrected in robust optimization. In this study we simulated Bragg peaks with systematic shifts in SP at different positions. The locations of the Bragg peaks were determined, the shift from baseline measured and the resulting data analyzed for smoothness, regularity, and size of the effects. Using this analysis of our optimization search space, we designed an optimization to extract SP correction factors from measured beam energies after passing through a known set of materials. We simulated multiple targets made up of at most five materials, applied arbitrarily located systematic shifts to the SPs of these materials, and computed the exit energy of protons passing through these materials. An optimization function varied the baseline SPs to achieve the closest match to the “measured” proton exit energies. The derived SP scale factors were compared to the (known) SP uncertainties and used to correct the computed Bragg peak locations. Results: The relative shift of the Bragg peak is linearly dependent on the size of the SP error. For most materials tested and constant size regions, the size of the peak shift does not depend on where along the path the SP error occurs. The peak shift is directly dependent on the length of the material with incorrect SP. This is determined by the very mild energy dependence of SPs relative to, e.g., water. For most geometries, the relative peak shift is 1-2% for each 10% of proton path with SP uncertainties. An algorithm was developed to determine the SP error and correct for it. For multiple runs, with stopping power uncertainties up to +/- 20%, the Bragg peak range errors were +/- 15% of nominal Bragg peak depth. After correction, the Bragg peak range errors were reduced to 0.4% of nominal depth, independent of initial proton kinetic energy. Conclusion: This study demonstrates that inherent SP uncertainties contribute to proton range errors at about the same level of significance as many other sources of range uncertainty. The range errors are smoothly varying with input parameters and an optimization procedure can effectively correct these errors, reducing the shift in computed Bragg peak location to less than 0.5%. Subsequent work will apply this optimization to physical targets in our proton facility. Author Disclosure: J.G. Hardie: None. C. Beltran: None.
3484 Fluorophore, Protoporphyrin IX (PpIX), Excited From Cerenkov Emission Produced by 45MV Radiation for Cancer Therapy Z. Quanshi,1 Q. Sun,1 G. Xiao,1 J. Zeng,2 L. Chen,3 and C.M.C. Ma3; 1 Wuxi Yiren Tumor Hospital, Beijing, China, 2Wuxi Yiren Tumor Hospital, Wuxi, China, 3Fox Chase Cancer Center, Philadelphia, PA
3485 Set-up Accuracy and Intrafractional Motion for Cranial Stereotactic Radiosurgery: Frameless Versus Rigid Immobilization Systems A. Liu,1 M.D. Hall,2 S.V. Dandapani,1 T.E. Schultheiss,2 and J.Y.C. Wong2; 1Department of Radiation Oncology, City of Hope National Medical Center, Duarte, CA, 2City of Hope National Medical Center, Duarte, CA Purpose/Objective(s): The objective of this IRB-approved study was to compare the interfractional setup accuracy and intrafractional motion of patients treated with cranial stereotactic radiosurgery with frameless non-invasive vacuum-suction immobilization (VSI) versus rigid screw immobilization (RSI). Materials/Methods: 20 consecutive patients treated by tomotherapy were selected for data collection. The dose and number of fractions received by each patient ranged from 18Gy in 1 fraction (SRS) to 25Gy in 5 fractions (SRT). 12 patients were immobilized using a frameless VSI system and 8 patients were immobilized using the RSI system. Customized head cushions were used in all patients. 6 out of 12 VSI
Volume 93 Number 3S Supplement 2015 patients received pre and post treatment cone beam CT (CBCT) to evaluate the intrafractional motion of the VSI system. The intrafractional motion with the RSI system was reported to be negligible and not repeated in this study. All remaining patients received pretreatment CBCT or megavoltage CT (MVCT) to assess interfractional setup accuracy. Shifts to the final treatment position were determined based on matching bony anatomy in the pre-treatment setup CT and the planning CT. Setup CT and planning CT were registered retrospectively based on bony anatomy using image registration software to quantify rotational and translational errors. Results: Mean and standard deviation of the intrafractional motions were -0.50.7mm (lateral), 0.10.9mm (vertical), -0.50.6mm (longitudinal), -0.040.18 (pitch), -0.10.23 (yaw), and -0.030.17 (roll). Interfractional rotation errors were -0.100.25 (pitch), -0.200.41 (yaw), and -0.080.16 (roll) for RSI versus 0.200.69 (pitch), 0.350.82 (yaw), -0.340.56 (roll) for frameless VSI. In a 3D vector space, a tumor located 5cm from the center of image fusion would require a 0.9 mm margin with the RSI system and a 2.1 mm margin with VSI. Conclusion: With image guided radiation therapy, translational setup errors can be corrected by image registration between pretreatment setup CT and planning CT. However, rotational errors cannot be accounted for without 6 degree freedom couch. Our study showed that the frameless VSI immobilization system provided negligible intrafractional motion. The interfractional rotation setup error using VSI was larger than rigid immobilization with the RSI system. For single lesion far from the center of image registration or for multiple lesions, additional margin may be needed to account for the uncorrectable rotational setup errors. Author Disclosure: A. Liu: None. M.D. Hall: None. S.V. Dandapani: None. T.E. Schultheiss: None. J.Y. Wong: None.
3486 Direct Impact Analysis of Multileaf Collimator Leaf Position Error on Dose Error in Off-Axis and Distributions With Flattening Filter Free Beam J. Murakami,1 N. Wakai,2 I. Toshitaka,1 T. Youji,1 Y. Tadashi,1 F. Satoshi,1 T. Tamamoto,2 and M. Hasegawa2; 1Central Department of Radiology, Nara Medical University Hospital, Kashihara Nara 634-8522, Japan, 2 Department of Radiation Oncology, Nara Medical University, Kashihara Nara 634-8522, Japan Purpose/Objective(s): In this study, we investigated the impact of systematic multileaf collimator (MLC) position error on dose error for both normal flattening filtered (FF) and flattening filter-free (FFF) beam energies in off-axis and the impact on the dose distribution due to leaf position error with FF and FFF beams in clinical intensity modulated radiation therapy (IMRT) cases. Materials/Methods: Dosimetric leaf gap (DLG) values for 6-MV FF, 6MV FFF, 10-MV FF, and 10-MV FFF X-ray beams were measured using dynamic delivery at CAX and off-axis. The dose error due to leaf position error was calculated from these measurements. The measurements were performed using a medical linear accelerator equipped with an HDMLC with 120 leaves. The measurements were performed at several locations in the beam with a Farmer-type ion chamber. For the impact on dose distribution, we generated 10 prostate (10-MV FF and 10-MV FFF) and 10 head and neck (6-MV FF and 6-MV FFF) plans using IMRT with 7 static beams (original plan). Modified plans were generated to introduce leaf position error (0.2, 0.6 and 1.0 mm) into the original plans. The original plans were measured with a 2D measurement array and compared with the modified plans calculated using the same MU as in original plans. Gamma analysis was implemented using criteria of 2% / 2 mm. Results: In the case of a 10-mm gap width with leaf position error of 1 mm, the dose errors at off-center 0, 3, 5, and 7 cm were 9.7, 9.6, 9.6, and 9.6%, respectively, for 10-MV FF, compared to 9.7, 9.6, 9.6, and 9.6%, respectively, for 10-MV FFF. Similar dose errors were seen at 6-
Poster Viewing Session E593 MV FF and 6-MV FFF. The dose error was large for a small gap width and large gap error. The pass rates of gamma analysis with leaf position errors of 0, 0.2, 0.6, and 1.0 mm were 99.5 0.8 (mean SD), 99.6 0.6, 98.4 2.1, and 91.8 5.8%, respectively, for 10-MV FF, compared with 99.0 2.5, 97.1 3.0, 80.3 8.2, and 63.8 8.4%, respectively, for 10-MV FFF. The pass rates were markedly decreased with FFF plans in comparison with FF plans under the influence of leaf position error. Conclusion: The difference in the dose error due to leaf position error between FF and FFF beams was negligible. Similarly, there was no difference between the measurement at CAX and off-axis. In clinical IMRT plans, using an FFF beam with a high dose rate led to a low robustness in the presence of leaf position error, and the pass rate decrease significantly as leaf position error increased compared to the plans with the FF beam. Therefore, it is necessary for IMRT plans using an FFF beam with a high dose rate to set a tolerance level differing from that using an FF beam due to leaf position error. Author Disclosure: J. Murakami: None. N. Wakai: None. I. Toshitaka: None. T. Youji: None. Y. Tadashi: None. F. Satoshi: None. T. Tamamoto: None. M. Hasegawa: None.
3487 Assessment of Radiation Therapy Response in Cervical Cancer Using FDG PET-CT S.H. Lee,1 K. Sung,1 S.H. Lee,1 S.H. Ahn,1 J. Choi,1 K.C. Lee,1 and S.G. Kim2; 1Gachon University Gil Medical Center, Radiation Oncology, Incheon, South Korea, 2Gachon University Gil Medical Center, Nuclear Medicine, Incheon, South Korea Purpose/Objective(s): We retrospectively assessed the relationship between primary tumor 18F-fluorodeoxyglucose (FDG) uptake expressed as the maximum standardized uptake value (SUVmax), primary tumor response, and survival in patients with cervical cancer on pretreatment and posttreatment positron emission tomography-computed tomography (PETCT). Materials/Methods: We conducted a retrospective review of 28 cervical cancer patients (stage I-II: nZ15; stage III-IV: nZ13) treated with radiation therapy with or without chemotherapy from August of 2009 until February of 2014. All patients received external beam radiation therapy and intracavitary brachytherapy. All patients (mean age, 57.9 years; range, 37-81 years) had FDG PET-CT scans performed before and after radiation therapy (pre- and post-RT). The SUVmax were recorded from PET-CT scans performed pre- and post-RT. The correlation between SUVmax and treatment response was evaluated and analyzed. Measurement of primary tumor volumes (pTV) for the treatment response evaluation was performed by the diagnostic radiologist using magnetic resonance imaging or abdominopelvic CT. The changes of pTV and SUVmax between pre-RT and post-RT were analyzed. The correlation between SUVmax and the changes of pTV were also evaluated. Results: The median duration of follow-up was 21 months (range 5-40 months). The 3-year overall survival and progression free survival (PFS) in all patients were 80.4% and 69.5%, respectively. Of the 28 patients, 15 (53.5%) showed complete response on follow-up studies. The average SUVmax of cervical cancer pre- and post-RT were 11.265.31 and 2.981.04, respectively. After treatment for cervical cancer, the reduction of SUVmax (8.285.13) was significant (p<0.001). There was a significant correlation between pre-RT SUVmax and pre-RT pTV (rZ0.525, pZ0.004). A significant correlation was observed between post-RT SUVmax and the pTV reduction rate after RT (rZ-0.411, pZ0.030). The association between pTV reduction rate after RT and PFS was significant (pZ0.006). Conclusion: The SUVmax was significantly reduced after RT for cervical cancer. This study showed that the SUVmax after RT in patients with cervical cancer was significantly correlated with the primary tumor response showing a significant association with PFS. The results of this