Purpose: To assess and report the in vivo dose for a patient with a pacemaker being treated in left breast intraoperative radiation therapy (IORT). The ZEISS Intrabeam 50 kVp X‐ray beam with a spherical applicator was used. Methods: The optically stimulated luminescent dosimeters (OSLDs) (Landauer nanoDots) were employed and calibrated under the conditions of the Intrabeam 50 kVp X‐rays. The nanoDots were placed on the patient at approximately 15 cm away from the lumpectomy cavity both under and above a shield of lead equivalence 0.25 mm (RayShield X‐Drape D‐110) covering the pacemaker area during IORT with a 5 cm spherical applicator. Results: The skin surface dose near the pacemaker during the IORT with a prescription of 20 Gy was measured as 4.0±0.8 cGy. The dose behind the shield was 0.06±0.01 Gy, demonstrating more than 98% dose reduction. The in vivo skin surface doses during a typical breast IORT at a 4.5 cm spherical applicator surface were further measured at 5, 10, 15, and 20 cm away to be 159±11 cGy, 15±1 cGy, 6.6±0.5 cGy, and 1.8±0.1 cGy, respectively. A power law fit to the dose versus the distance z from the applicator surface yields the dose fall off at the skin surface following z^‐2.5, which can be used to estimate skin doses in future cases. The comparison to an extrapolation of depth dose in water reveals an underestimate of far field dose using the manufactory provided data. Conclusion: The study suggests the appropriateness of OSLD as an in vivo skin dosimeter in IORT using the Intrabeam system in a wide dose range. The pacemaker dose measured during the left breast IORT was within a safe limit.

Purpose: We report the depth dose measurements in air, in solid water, and in bone materials for the Intrabeam 50 kV x‐rays with a needle applicator. Methods: The absolute dose was measured using a PTW TN34013W soft x‐ray ion chamber. Gammex tissue equivalent materials of solid water, inner bone, and cortical bone slabs (minimum thickness of 2 mm) were used. In addition, the PTW solid water slabs with a minimum thickness of 1 mm were used. The manufactory calibrated depth dose data in water were compared. The x‐ray source together with a needle applicator was secured on an Intrabeam stand. The slabs lay on a 6 degrees of freedom treatment couch with a digitally controlled minimum step size of 0.1 mm. The depth of the source to the ion chamber was accurately and reproducibly adjusted by moving the couch up and down. Results: The depth dose measurements for the Intrabeam 50 kV x‐rays with a needle applicator were conducted up to 20 mm in depth. The values for the PTW solid water were close to those for water. The Gammex solid water demonstrated lower values compared to water, consistent with the observation of its positive CT number. At a depth of 10 mm, the dose rates of the system are 29.6, 3.6, 1.2, and 0.24 Gy/min in air, in water, in inner bone, and in cortical bone, respectively. The 10 mm water equivalent depths in inner and cortical bone are about 6.4 and 4.1 mm. A function of power law combining exponential was used to fit and interpolate data well. Conclusion: Direct depth dose measurements in different materials provide a basis for treatment calculation and planning taking into account the heterogeneous effect. The results can be used for verification of analytical and/or Monte Carlo dose calculation methods as well.

Purpose: To statistically determine the optimal tolerance level in the verification of delivery dose compared to the planned dose in an in vivo dosimetry system in radiotherapy. Methods: The LANDAUER MicroSTARii dosimetry system with screened nanoDots (optically stimulated luminescence dosimeters) was used for in vivo dose measurements. Ideally, the measured dose should match with the planned dose and falls within a normal distribution. Any deviation from the normal distribution may be redeemed as a mismatch, therefore a potential sign of the dose misadministration. Randomly mis‐positioned nanoDots can yield a continuum background distribution. A percentage difference of the measured dose to its corresponding planned dose (ΔD) can be used to analyze combined data sets for different patients. A model of a Gaussian plus a flat function was used to fit the ΔD distribution. Results: Total 434 nanoDot measurements for breast cancer patients were collected across a period of three months. The fit yields a Gaussian mean of 2.9% and a standard deviation (SD) of 5.3%. The observed shift of the mean from zero is attributed to the machine output bias and calibration of the dosimetry system. A pass interval of −2SD to +2SD was applied and a mismatch background was estimated to be 4.8%. With such a tolerance level, one can expect that 99.99% of patients should pass the verification and at most 0.011% might have a potential dose misadministration that may not be detected after 3 times of repeated measurements. After implementation, a number of new start breast cancer patients were monitored and the measured pass rate is consistent with the model prediction. Conclusion: It is feasible to implement an optimal tolerance level in order to maintain a low limit of potential dose misadministration while still to keep a relatively high pass rate in radiotherapy delivery verification.

Purpose: The objective of this study is to verify and analyze the accuracy of a clinical deformable image registration (DIR) software. Methods: To test clinical DIR software qualitatively and quantitatively, we focused on lung radiotherapy and analyzed a single (Lung) patient CT scan. Artificial anatomical changes were applied to account for daily variations during the course of treatment including the planning target volume (PTV) and organs at risk (OAR). The primary CT (pCT) and the structure set (pST) was deformed with commercial tool (ImSimQA‐Oncology Systems Limited) and after artificial deformation (dCT and dST) sent to another commercial tool (VelocityAI‐Varian Medical Systems). In Velocity, the deformed CT and structures (dCT and dST) were inversely deformed back to original primary CT (dbpCT and dbpST). We compared the dbpST and pST structure sets using similarity metrics. Furthermore, a binary deformation field vector (BDF) was created and sent to ImSimQA software for comparison with known “ground truth” deformation vector fields (DVF). Results: An image similarity comparison was made by using “ground truth” DVF and “deformed output” BDF with an output of normalized “cross correlation (CC)” and “mutual information (MI)” in ImSimQA software. Results for the lung case were MI=0.66 and CC=0.99. The artificial structure deformation in both pST and dbpST was analyzed using DICE coefficient, mean distance to conformity (MDC) and deformation field error volume histogram (DFEVH) by comparing them before and after inverse deformation. We have noticed inadequate structure match for CTV, ITV and PTV due to close proximity of heart and overall affected by lung expansion. Conclusion: We have seen similarity between pCT and dbpCT but not so well between pST and dbpST, because of inadequate structure deformation in clinical DIR system. This system based quality assurance test will prepare us for adopting the guidelines of upcoming AAPM task group 132 protocol.

Purpose: The current standard in dose calculation for intraoperative radiotherapy (IORT) using the ZEISS Intrabeam 50 kV x‐ray system is based on depth dose measurements in water and no heterogeneous tissue effect has been taken into account. We propose an algorithm for pre‐treatment planning including inhomogeneity correction based on data of depth dose measurements in various tissue phantoms for kV x‐rays. Methods: Direct depth dose measurements were made in air, water, inner bone and cortical bone phantoms for the Intrabeam 50 kV x‐rays with a needle applicator. The data were modelled by a function of power law combining exponential with different parameters. Those phantom slabs used in the measurements were scanned to obtain CT numbers. The x‐ray beam initiated from the source isocenter is ray‐traced through tissues. The corresponding doses will be deposited/assigned at different depths. On the boundary of tissue/organ changes, the x‐ray beam will be re‐traced in new tissue/organ starting at an equivalent depth with the same dose. In principle, a volumetric dose distribution can be generated if enough directional beams are traced. In practice, a several typical rays traced may be adequate in providing estimates of maximum dose to the organ at risk and minimum dose in the target volume. Results: Depth dose measurements and modeling are shown in Figure 1. The dose versus CT number is shown in Figure 2. A computer program has been written for Kypho‐IORT planning using those data. A direct measurement through 2 mm solid water, 2 mm inner bone, and 1 mm solid water yields a dose rate of 7.7 Gy/min. Our calculation shows 8.1±0.4 Gy/min, consistent with the measurement within 5%. Conclusion: The proposed method can be used to more accurately calculate the dose by taking into account the heterogeneous effect. The further validation includes comparison with Monte Carlo simulation.