S Asbell

Low toxicity for lung tumors near the mediastinum treated with stereotactic body radiation therapy.

PURPOSEnTo report the local control, survival, and low toxicity observed at the Cooper University Hospital CyberKnife Center post stereotactic body radiation therapy (SBRT) in the treatment of lung tumors near the mediastinum.nnnMETHODS AND MATERIALSnTwenty-four medically inoperable lung cancer patients with tumors near the mediastinum were treated using the Accuray CyberKnife system (Accuray, Sunnyvale, CA) with Monte Carlo dose calculations and heterogeneity corrections from July 2008 to May 2010. The prescription dose ranged from 28.5 Gy to 60 Gy in 3-5 fractions. For conventional fractionation schemes, Emami et al(1) organized the dose tolerance limits into a unified format for clinical utility and partitioned them into 2 risk levels (5% and 50%) with preset volumes for most critical structures throughout the body. In contrast, statistical SBRT dose tolerance limits for mediastinal structures have not been established yet. We have sufficient experience at least to begin organizing a unified format with low-risk and high-risk partitions and preset volumes for 1-5 fractions exposing mediastinal structures. With the help of the (dose-volume histogram) DVH Evaluator, a software tool developed by our senior author, each treatment plan was assessed for safety and feasibility prior to treatment. The DVH Evaluator was also used to analyze the follow-up data and to create graphs of risk, called DVH Risk Maps, superimposing clinical data onto the unified SBRT dose tolerance limits.nnnRESULTSnIt was not feasible to prescribe the doses of peripheral lung lesions for all tumors near the mediastinum because of known toxicity. The crude local tumor control rate achieved in our series was 92%. Median survival was 26.8 months for the primary lung cases and 9.6 months for the metastatic cases. No patients experienced grade 3 or higher toxicities.nnnCONCLUSIONSnWe affirm that SBRT is feasible in the treatment of centrally located lung cancers when the dose tolerance limits of critical structures are diligently respected. The low adverse event rates that we have experienced, combined with a good local tumor control rate, are encouraging.

Small Bowel Dose Tolerance for Stereotactic Body Radiation Therapy

Inconsistencies permeate the literature regarding small bowel dose tolerance limits for stereotactic body radiation therapy (SBRT) treatments. In this review, we organized these diverse published limits with MD Anderson at Cooper data into a unified framework, constructing the dose-volume histogram (DVH) Risk Map, demonstrating low-risk and high-risk SBRT dose tolerance limits for small bowel. Statistical models of clinical data from 2 institutions were used to assess the safety spectrum of doses used in the exposure of the gastrointestinal tract in SBRT; 30% of the analyzed cases had vascular endothelial growth factor inhibitors (VEGFI) or other biological agents within 2 years before or after SBRT. For every dose tolerance limit in the DVH Risk Map, the probit dose-response model was used to estimate the risk level from our clinical data. Using the current literature, 21Gy to 5cc of small bowel in 3 fractions has low toxicity and is reasonably safe, with 6.5% estimated risk of grade 3 or higher complications, per Common Terminology Criteria for Adverse Events version 4.0. In the same fractionation for the same volume, if lower risk is required, 16.2Gy has an estimated risk of only 2.5%. Other volumes and fractionations are also reviewed; for all analyzed high-risk small bowel limits, the risk is 8.2% or less, and the low-risk limits have 4% or lower estimated risk. The results support current clinical practice, with some possibility for dose escalation.

Introduction and Clinical Overview of the DVH Risk Map

Radiation oncologists need reliable estimates of risk for various fractionation schemes for all critical anatomical structures throughout the body, in a clinically convenient format. Reliable estimation theory can become fairly complex, however, and estimates of risk continue to evolve as the literature matures. To navigate through this efficiently, a dose-volume histogram (DVH) Risk Map was created, which provides a comparison of radiation tolerance limits as a function of dose, fractionation, volume, and risk level. The graphical portion of the DVH Risk Map helps clinicians to easily visualize the trends, whereas the tabular portion provides quantitative precision for clinical implementation. The DVH Risk Map for rib tolerance from stereotactic ablative body radiotherapy (SABR) and stereotactic body radiation therapy (SBRT) is used as an example in this overview; the 5% and 50% risk levels for 1-5 fractions for 5 different volumes are given. Other articles throughout this issue of Seminars in Radiation Oncology present analysis of new clinical datasets including the DVH Risk Maps for other anatomical structures throughout the body.

Stereotactic Body Radiotherapy for Large (> 5 cm) Non-Small-Cell Lung Cancer.

Background Stereotactic body radiotherapy (SBRT) is a well‐established treatment option for early stage non–small‐cell lung cancer (NSCLC) tumors < 5 cm. There is limited information on tumors > 5 cm. Patients and Methods We performed retrospective data collection of patients enrolled onto a prospective SBRT registry study. Eligible patients for this study had node‐negative NSCLC measuring > 5 cm in any dimension. Data from 41 patients were analyzed. Median patient age was 75 years, and median tumor size was 5.6 cm (range, 5.0‐12.2 cm). Sixteen patients had squamous disease, 20 patients adenocarcinoma, and 1 mixed tumor; 4 patients had no biopsy. Median radiation dose per fraction was 50 Gy in 5 fractions. Radiation was prescribed to isodose line, median 66% (range, 50%‐84%). Results Before SBRT, 6 patients had previous chemotherapy and 7 patients had previous radiation. Median follow‐up for all patients was 15.2 months (range, 0.56‐48.1 months). At last follow‐up, 16 patients were still alive, with a median follow‐up of 16.1 months for surviving patients. The median survival was 17.5 months with 1‐ and 2‐year survivals of 65% and 34%. Two patients (4.8%) had local failure, and 13 patients (31%) had distant failure. Four patients (9.8%) had acute toxicity, and 7 patients (17.1%) had late toxicity, including 2 (4.8%) grade 3 late toxicities. Conclusion SBRT for tumors > 5 cm is effective, with good local control rates and acceptable toxicity. The main pattern of failure is distant, suggesting a possible role for systemic chemotherapy in these patients. Micro‐Abstract This study was undertaken to provide a better understanding of stereotactic body radiotherapy (SBRT) in nonoperable non–small‐cell lung cancer with a largest tumor dimension of > 5 cm. A retrospective analysis was conducted on a prospective SBRT registry, with analysis of 41 patients. SBRT results in good local control and acceptable rates of distant control and treatment‐induced toxicities in larger lung tumors.

SU‐E‐J‐147: Internal Brain Motion Between CT and MR Scanning

Purpose: To evaluate the internal brain motion between two imaging studies of CT and MRI. A study with 30 healthy volunteers with MRI scans in 4 different positions showed significant brain to skull motion up to 1 cm. Such motion among patients for radiotherapy to the brain is evaluated in this study. Methods: Twenty‐five patients underwent MRI and CT scans in the same day for radiotherapy planning were recruited. A whole brain fusion was first performed. Three to five pairs of control points were selected on both CT and MRI for starting an automated intensity based registration. The fusion was reviewed and fine‐tuned for the best skull‐to‐skull matching. To study potential internal brain motion, a subsequent fine‐tuning of the fusion was performed by matching the visible features, such as gyri, sulci and fissures, near the tumor site. The second fusion was reviewed and fine‐tuned by two physicists until the best visible feature matching could be agreed upon. The resulting rotation and translation between the whole brain and feature‐based fusions indicated potential internal brain motion between the two scans. Results: Between two fusions the mean internal shifts in x (LR), y (AP) and z (SI) were 0.34±0.95 mm, 0.21±1.18 mm and −0.34±0.8 mm, respectively. The mean overall shift was 1.4±1.1 mm, and the largest shift was 3.5 mm. The mean rotation angles were 0.22±1.32° (pitch), 0.14±0.4° (yaw) and 0.08±0.53° (roll), respectively. The pitch motion was predominant (head up and down) due to difference of couch tops of CT and MRI scanners. Conclusion: Our study showed small but measurable internal brain motion among radiotherapy patients with typical clinical setting for CT and MRI imaging. Therefore the CT‐MRI fusion should be carefully checked for internal structure matching. An additional treatment margin may be needed if an internal motion is observed.

SU‐E‐T‐883: Clinical Application of Monte Carlo for SBRT: Mediastinal Lung

Purpose: Treatment planning systems have provided Monte Carlo dose calculations for several years but many physicians are still hesitant to use them clinically due to lack of data. Therefore we recalculated 200 Ray Tracing treatment plans using Monte Carlo with heterogeneity corrections and compared to SBRT dose tolerance limits. Methods: From among these 200 CyberKnife cases, 25 mediastinal lung cases are presented in this study. An extensive literature review obtained 105 published SBRT dose tolerance limits for the mediastinal critical structures aorta, bronchi, esophagus, heart, and trachea. These limits were partitioned into high‐risk and low‐risk categories. The DVH Evaluator software tool was used to generate DVH Risk Maps for these critical structures, which superimpose a) published dose tolerance limits b) unified high‐risk and low‐risk trends and c) published adverse event doses, onto Monte Carlo patient data to assess risk of adverse events. Results: Recalculated treatment plan data is within the expected range of published SBRT dose tolerance limits, providing optimism for clinical use. None of the patients experienced any Grade 3 or higher adverse events. The low‐risk dose tolerance limits were exceeded 22 times in these cases with no severe adverse event, thus helping to validate their safety. Conclusions: The range of doses calculated by Monte Carlo for our historical patient data is compatible with published SBRT dose tolerance limits. SBRT dose tolerance limits should be fine‐tuned by Monte Carlo dose calculations in long‐term statistical followup studies. Disclosure: The first author has developed the DVH Evaluator software.

MO‐D‐BRB‐07: Phantom Validation and Clinical Application of Monte Carlo for Small Field SBRT

Purpose: Small field dosimetry is challenging in homogeneous medium and extremely difficult in an inhomogeneous medium. Monte Carlo dose calculation algorithms are considered as the most accurate for treatment planning. We present our validation of the Monte Carlo algorithm in the Accuray Multiplan system using measurements in a cork phantom. We also recalculated Ray Tracing treatment plans with the Monte Carlo algorithm and compared to SBRT dose tolerance limits. Methods: In our validation measurements with a cork phantom, an Exradin A16 ion chamber was used for collimators from 60mm to 20mm on a CyberKnife, and a PTW 60012 stereotactic diode for collimators from 60mm to 5mm. A literature review of more than 500 published SBRT dose tolerance limits was partitioned into high‐risk and low‐risk categories. Two hundred CyberKnife treatment plans were recalculated using Monte Carlo and compared to the dose limits. The DVH Evaluator software tool was used to generate DVH Risk Maps for 25 critical structures throughout the body, which superimpose a) published dose tolerance limits b) unified high‐risk and low‐risk trends and c) published adverse event doses, onto Monte Carlo patient doses to assess risk of adverse events. Results: The Monte Carlo calculations matched the Exradin A16 measurements to within 2.5% for field sizes down to 20mm, and matched the PTW 60012 measurements to within 2.5% for all field sizes down to 5mm. Recalculated treatment plan data is within the expected range of published SBRT dose tolerance limits, providing optimism for clinical use. Conclusions: The Accuray MultiPlan Monte Carlo algorithm is accurate even for small fields in heterogeneous media. The range of doses calculated by Monte Carlo for our patient data is compatible with published SBRT dose tolerance limits. SBRT dose tolerance limits should be fine‐tuned by Monte Carlo dose calculations in long‐term statistical followup studies. Disclosure: The first author has developed the DVH Evaluator software.

SU‐E‐T‐892: Calculation Uncertainty in CyberKnife Dosimetric Parameters of Brainstem

Purpose: It is important to understand the effect of dose calculation on the dosimetric parameters of a critical structure, which are often used to evaluate the risk of toxicity. We present the variation of some dosimetric parameters for brainstem as computed by two different dose algorithms. Methods: Both Ray Tracing and Monte Carlodose algorithms are commissioned for the CyberKnife planning system at Cooper. Study was performed for twenty seven intracranial patients treated with CyberKnife. Dose calculation was carried out in Multiplan workstation by Monte Carlo algorithm with heterogeneity corrections as well as Ray Tracing algorithm without heterogeneity corrections. Brainstem was delineated by the attending radiation oncologist during planning process. The same prescription and optimization as the actual treatment were employed in dose calculation for each patient. Results: Principle dosimetric parameters (D50, D10, D1cc, D0.1cc and maximum dose) of brainstem are derived for each patient. The difference in percentage between Ray Tracing and Monte Carlo calculation is plotted. Significant variation has been observed for each dosimetric parameter between two different dose algorithms. Also plotted is the variation of those dosimetric parameters against the volume of brainstem. Conclusions: The value of dosimetric parameters can be considerately different depending on dose algorithm used in calculation. The evaluation of dose tolerance for a critical structure using those dosimetric parameters needs to be cautious about dose computing methods.

SU‐C‐BRB‐02: A Phase Resolved Fiducial Setup Scheme for Stereotactic Body Radiation Therapy (SBRT)

Purpose: For CyberKnife based SBRT,CT scan of a single breathing phase is utilized for real‐time tumor tracking. However, x‐ray images for tracking are more often not in phase with the reference CT. Non‐rigid fiducial movement due to tumor deformation can induce significant uncertainties when matching images from different phases and this poses great difficulties for patient setup. A phase resolved fiducial setup scheme is developed that finds the best rigid transformation with tumor deformation corrected. Methods: Five cases (2 liver and 3 lung patients) were retrospectively analyzed in this study. For each patient, two sets of fiducials (at the ends of exhale (EOE) and inhale (EOI)) were first aligned by their centroids and linearly interpolated to generate fiducial sets for the phases in between. The fiducial set in 3D at each phase is iteratively registered to the fiducial set in x‐ray image until minimal residual error (RE) or mean distance between the projected fiducials and fiducials in the x‐ray image is reached. The phase with the smallest RE after registration of fiducial sets in 5 phases is determined as the phase of the x‐ray image. Results: For a liver case, 30 pairs of x‐ray images were registered to all 5 phases. The RE for registration with different phases was a smooth function with a distinct minimum. Fiducial matching for all the x‐ray images was also performed to the EOE phase. The REs resulted by our method were 4.2±0.58 mm, versus 5.3±1.48 mm with only EOE phase used for registration. The latter represented the RE currently achievable in current system. Similar results were also observed for the other 4 patients. Conclusions: In this study, a phase resolved fiducial setup scheme was developed and tested for 5 patients. It facilitates patient setup and tracking accuracy with reduced REs.

SU‐GG‐T‐538: Preliminary Estimates of Adverse Event Probability for Selected Critical Anatomical Structures in SBRT

Purpose: To make a preliminary estimate of the probability of adverse events for various critical anatomical structures for stereotactic body radiation therapy(SBRT). Method and Materials: An extensive literature search has uncovered a collection of more than 500 published dose tolerance limits for various anatomical critical structures for one to five fractions of SBRT. We have not found any TD5/5 or TD50/5 adverse event probabilities for SBRT yet, and an appreciable number of the authors do not even present minimum dosimetric information for the adverse events they report. We compiled a database from the available published data and performed statistical analysis to obtain a preliminary estimate of probability of adverse events. Results: Sixty five of the 500 dose tolerance limits correspond to a reported adverse event. For nineteen of these cases, the authors reported the number of patients exceeding the stated dose tolerance limit. For braintumors, details of adverse events were reported in optic chiasm and optic nerve, and for lungtumors details of adverse events were reported in bronchi, lungs, and ribs. For these critical structures we can estimate the adverse event probability, although in some cases the number of data points is quite limited. Conclusion: It would be quite helpful for the future if more authors would present dosimetric data corresponding to the adverse events they report. At a minimum, it would be helpful to report the maximum dose received by the involved critical structure, the doses received by a few selected volumes, and the number of patients that received these dose levels. In the meantime, this work provides preliminary estimates of adverse event probability for selected anatomical critical structures from the sparse data available.