Poster Viewing E651
Volume 99 Number 2S Supplement 2017
3553 Correlation of Parameterized ADC With Overall Survival for GBM Treatments Using Machine Learning A. Chu,1 T.K. Yanagihara,2 K.A. Cauley,3 and T.J.C. Wang2; 1Columbia University Medical Center, New York, NY, 2Department of Radiation Oncology, Columbia University Medical Center, New York, NY, 3Geisinger Medical Center, Danville, PA Purpose/Objective(s): To correlate parameters derived from apparent diffusion coefficient (ADC) volumes associated with the changes in specified abnormal white matter (WM) to overall survival (OS) following therapies in glioblastoma (GBM). The unsupervised machine learning was designed to reduce uncertainties on a relatively small size data for a prediction model. Materials/Methods: The study employed 41 GBM patients treated with radiotherapy (60 Gy) followed by Temozolomide. The ADC ranges for normal WM and gray matter (GM) were established from a pool of ADC voxels over 66 individual brains. The predictors were the averaged values of pre/post treated (TX) over the masked volumes, where (abnormal) preTX WMs were in normal GM ADC range, then applied the mask to the post-TX volume. The learning process has 2 main parts: (1) the probability of mapping from predictors to OS, (2) data resampling extracted the features of the mapping in parametric space with different complex levels of models. The number of coefficients in linear models can be from 3 to 7 (or more) for higher complexity. The probability was based on Bayesian with likelihood probability from institutional data and the prior information from published OS reports. The data were resampled with total number and location of datapoints being altered with bandwidth of kernel density estimation, cluster analysis in parametric space. Results: The characterized ADC volumes(1) provide more objective parameters for machine learning than manually contoured clinical tumor volume (CTV), (2) are within the sampled ADC volume based on FLAIR & T1-post highlighted regions, where the ADC histogram illustrates a unique pattern of pre-/post-TX dynamics; i.e. “WM looking like GM before treatment”. Without resampling in the learning process, the predicted OS had 0.57 of correlation coefficient (CC) with the data fitted with the simpler model. The model underestimated survival at late time points (>450 days) and overestimated survival in early time points. With resampling, the model CC improved (CC w 0.65) secondary to outlier suppression at later time points (survival: 200w450 days). The use of more sophisticated models did not improve the model fit. Conclusion: The subtle changes detected by ADC parameters could not be accurately quantified by T1, T2 or FLAIR suggesting that ADC offers additional information useful for prognostication. Despite high uncertainties presented in the small dataset, the correlation was non-trivial, and it can also be enhanced further with the learning process to feature the given data and a priori. Because the WM scaled as GM by ADC indicates WM abnormities, the relation of “quantitative abnormity” with OS could be pathologically meaningful. The algorithm appears to represent an optimum on the likely feature that data given as more complex models were limited by the small data size. Author Disclosure: A. Chu: None. T.K. Yanagihara: None. K.A. Cauley: None. T.J. Wang: Consultant; Doximity, Merck. Advisory Board; American Cancer Society North Jersey. Travel Expenses; Abbvie.
3554 Pulmonary Toxicity of Stereotactic Body Radiation Therapy Treatment of Multiple Lung Lesions Using a Frameless Robotic Radiosurgery System E. Chung,1 Y. Zhang,2 T.K. Podder,3 M. Yao,4 M. Machtay,1 and T. Biswas1; 1Case Western Reserve University School of Medicine, Cleveland, OH, 2Radiation Oncology, University Hospitals Seidman Cancer Center, Cleveland, OH, 3University Hospitals Seidman Cancer Center, Cleveland, OH, 4University Hospitals Cleveland Medical Center, Cleveland, OH
Purpose/Objective(s): Stereotactic body radiation therapy (SBRT) has become an acceptable treatment modality for medically inoperable stage I non-small cell lung cancer or oligometastatic cancer with pulmonary metastasis. We undertook this study to evaluate the dose-volume factors and their associated toxicity for SBRT treatment of multiple lung tumors. Materials/Methods: From our institutional data base between 2012-2016, we identified 15 patients who underwent SBRT for multiple lung tumors and compared these patients to 13 randomly selected patients with single lung tumors. All of these patients were treated in a single clinical facility using Cyberknife robotic SBRT treatment system. Mean age of the patients with multiple lesions was 76.2 years (range 58.9-89.9 years) while the mean age of the 13 patients with solitary lesions was 78.9 years (range 69.1-95.6 years). Mean prescription dose for multiple lesion patients was 50 Gy (range 30-60 Gy), delivered in 3-5 fractions every other day. All single lesion patients were delivered 50 Gy in 5 fractions. All SBRT plans were reviewed to obtain dosimetric parameters including mean lung dose (Gy), and V5, V10, V20 (%) of total lung (subtracting PTV). Relevant dosimetric parameters between the two groups were compared and correlated with pulmonary and other toxicities. Results: Mean follow up was 72 days (range 30-105 days) for multi-lesion patients vs. 76 days (range 41-125 days) for single lesion patients. Mean pretreatment combined PTV volume was 97.49 cc (range 18.5 cc e 277.6 cc) for patients with multiple lesions and 58.0 cc (range 19.1 cc e 159.2 cc) for patients with single lesion (p<0.001), while the total lung volumes for both groups were comparable. The mean V5 (49.8% vs. 23.9%, pZ.0005), V10 (27.3% vs. 12.9%, p<0.001), V20 (11.7% vs. 5.1%, p<0.001), mean lung dose (9.10 Gy vs. 4.38 Gy, p<0.001) were significantly greater in patients with multiple lesions compared to patients with single lesions, respectively. About 20% (3/15 patients) and 15.4% (2/13 patients) experienced fatigue in multi-lesion and single-lesion patients, respectively. One patient from both multi-lesion and single-lesion patients developed radiation pneumonitis (6.7% vs. 7.7%, ns). Conclusion: Although, one patient in each group developed radiation pneumonitis, no co-relation was found with radiation dose parameters. SBRT for multiple lung lesions appears to be feasible and safe without significant adverse effects. Author Disclosure: E. Chung: None. Y. Zhang: None. T.K. Podder: None. M. Yao: None. M. Machtay: Research Grant; Abbvie. Consultant; Abbvie, Stemnion Inc. Advisory Board; Stemnion Inc.; NRG Oncology Group, RTOG Foundation. T. Biswas: Employee; University Hospitals Siedman Cancer center.
3555 Online Advertising and Marketing Claims by Providers of Proton Beam Therapy: Are They Guideline Based? M.T. Corkum,1 W. Liu,1 D.A. Palma,1 R. Dinniwell,1 G.S. Bauman,1 A. Warner,1 M.V. Mishra,2 and A.V. Louie1; 1London Regional Cancer Program, London, ON, Canada, 2University of Maryland Medical Center, Baltimore, MD Purpose/Objective(s): The Internet has revolutionized how patients obtain information for new treatments such as proton beam therapy (PBT). Evidence of superiority for PBT exists only in certain disease sites, and often only for specific indications. As cancer patients frequently search the Internet for treatment options, and with hospital websites seen as reliable sources of knowledge, the purpose of this study was to evaluate direct-toconsumer advertising content and claims made by proton therapy centre (PTC) websites worldwide. Materials/Methods: Operational PTC websites in English were identified through the Particle Therapy Co-Operative Group website as of September 1, 2016. Supplementary searches did not reveal additional treatment centres. Data abstraction was performed independently by two authors using a standardized form, with discrepancies resolved through consensus. Adherence to American Medical Association Opinion on Advertising and Publicity standards was measured. Eight international guidelines were consulted to determine indications for PBT. Descriptive statistics were calculated using SAS.