Radiotherapy planning in a prostate cancer phantom model with intraprostatic dominant lesions using stereotactic body radiotherapy with volumetric modulated arcs and a simultaneous integrated boost.

Aim: In the treatment of prostate cancer with radiation therapy, the addition of a simultaneous integrated boost (SIB) to the dominant intraprostatic lesions (DIL) may improve local control. In this study, we aimed to determine the optimal radiation strategy in a phantom model of prostate cancer using volumetric modulated arc therapy for stereotactic body radiotherapy (SBRT-VMAT) with a SIB of 1-4 DILs. Methods: We designed and printed a three-dimensional anthropomorphic phantom pelvis to simulate individual patient structures, including the prostate gland. A total of 36.25 Gy (SBRT) was delivered to the whole prostate. The DILs were irradiated with four different doses (40, 45, 47.5, and 50 Gy) to assess the influence of different SIB doses on dose distribution. The doses were calculated, verified, and measured using both transit and non-transit dosimetry for patient-specific quality assurance using a phantom model. Results: The dose coverage met protocol requirements for all targets. However, the dose was close to violating risk constraints to the rectum when four DILs were treated simultaneously or when the DILs were located in the posterior segments of the prostate. All verification plans passed the assumed tolerance criteria. Conclusions: Moderate dose escalation up to 45 Gy seems appropriate in cases with DILs located in posterior prostate segments or if there are three or more DILs located in other segments.

Large Prostate Volume Does Not Negatively Impact Health-Related Quality of Life in Patients with Prostate Cancer Treated with Ultrahypofractionated Stereotactic Body Radiotherapy.

BACKGROUND: Survival outcomes after primary radiotherapy for localized prostate cancer (PCa) are excellent, regardless of the specific treatment modality. For this reason, health-related quality of life (HRQOL) has come to play an ever more important role in treatment selection. Stereotactic body radiation therapy (SBRT) is increasingly used to treat patients with PCa. However, the impact of prostate volume on HRQOL is not clear. In this study, we aimed to determine whether a large prostate volume negatively influences HRQOL outcomes in patients undergoing ultrahypofractionated SBRT. MATERIAL AND METHODS: We conducted a prospective study of 530 men with low- and intermediate-risk localized PCa. All patients were treated from 2013 to 2017 with SBRT (Cyberknife system). HRQOL data were collected at baseline (pre-treatment), immediately after treatment, and at 12 and 24 months. QOL variables were assessed with the European Organization for Research and Treatment of Cancer QLQ-C30 and PR-25 module. Differences in the QLQ-C30 scales were considered clinically relevant when the change was >10 points. For the analysis, patients were classified into two groups according to prostate volume (≤60 vs. >60 cm3). RESULTS: The prostate volume was ≤60 cm3 in 415 patients (78.3%) and >60 cm3 in 115 (21.7%). No between-group differences were observed at baseline for any of the following variables: clinical stage; hormonal therapy; marital status; educational level; or employment status. No clinically-significant deterioration (functional and symptom scales) was observed in either group between the baseline and 24-month assessment. There were no clinically-relevant differences between the groups on any of the HRQOL variables, regardless of the prostate volume. CONCLUSIONS: This study shows that a large prostate volume (>60 cm3) does not appear to negatively impact HRQOL outcomes at two years in patients with localized prostate cancer treated with ultrahypofractionated SBRT administered with the CyberKnife system.

Cellular Damage in the Target and Out-Of-Field Peripheral Organs during VMAT SBRT Prostate Radiotherapy: An In Vitro Phantom-Based Study.

Hypo-fractionated stereotactic body radiation therapy (SBRT) is an effective treatment for prostate cancer (PCa). Although many studies have investigated the effects of SBRT on the prostate and adjacent organs, little is known about the effects further out-of-field. The aim of this study was to investigate, both in vitro and in a quasi-humanoid phantom, the biological effects (using a dose-scaling approach) of radiation in the out-of-field peripheral organs delivered by 6 MV volumetric modulated arc therapy (VMAT) SBRT in a prostate cancer model. Healthy prostate cells were irradiated in a phantom at locations corresponding to the prostate, intestine, lung, thyroid, and brain. Seven 10 Gy fractions of VMAT SBRT were delivered to the target in a single session without intermission (scaled-up method). Radiochromic films were used to measure the doses. The radiobiological response was assessed by measuring DNA breaks, the cell survival fraction, and differences in gene expression profile. Our results showed a strong, multiparametric radiobiological response of the cells in the prostate. Outside of the radiation field, the highest doses were observed in the intestine and lung. A small increase (not statistically significant) in DNA damage and cell death was observed in the intestines. Several gene groups (cell cycle, DNA replication) were depleted in the lung and thyroid (DNA replication, endocytosis), but further analysis revealed no changes in the relevant biological processes. This study provides extensive evidence of the types and extent of radiobiological responses during VMAT SBRT in a prostate cancer model. Additional research is needed to determine whether the radiobiological effects observed in the peripheral organs are validated in a clinical context.

Nontarget and Out-of-Field Doses from Electron Beam Radiotherapy.

In clinical radiotherapy, the most important aspects are the dose distribution in the target volume and healthy organs, including out-of-field doses in the body. Compared to photon beam radiation, dose distribution in electron beam radiotherapy has received much less attention, mainly due to the limited range of electrons in tissues. However, given the growing use of electron intraoperative radiotherapy and FLASH, further study is needed. Therefore, in this study, we determined out-of-field doses from an electron beam in a phantom model using two dosimetric detectors (diode E and cylindrical Farmer-type ionizing chamber) for electron energies of 6 MeV, 9 MeV and 12 MeV. We found a clear decrease in out-of-field doses as the distance from the field edge and depth increased. The out-of-field doses measured with the diode E were lower than those measured with the Farmer-type ionization chamber at each depth and for each electron energy level. The out-of-field doses increased when higher energy megavoltage electron beams were used (except for 9 MeV). The out-of-field doses at shallow depths (1 or 2 cm) declined rapidly up to a distance of 3 cm from the field edge. This study provides valuable data on the deposition of radiation energy from electron beams outside the irradiation field.

Influence of Specific Treatment Parameters on Nontarget and Out-of-Field Doses in a Phantom Model of Prostate SBRT with CyberKnife and TrueBeam.

The aim of the study was to determine the influence of a key treatment plan and beam parameters on overall dose distribution and on doses in organs laying in further distance from the target during prostate SBRT. Multiple representative treatment plans (n = 12) for TrueBeam and CyberKnife were prepared and evaluated. Nontarget doses were measured with anionization chamber, in a quasi-humanoid phantom at four sites corresponding to the intestines, right lung, thyroid, and head. The following parameters were modified: radiotherapy technique, presence or not of a flattening filter, degree of modulation, and use or not of jaw tracking function for TrueBeam and beam orientation set-up, optimization techniques, and number of MUs for CyberKnife. After usual optimization doses in intestines (near the target) were 0.73% and 0.76%, in head (farthest from target) 0.05% and 0.19% for TrueBeam and CyberKnife, respectively. For TrueBeam the highest peripheral (head, thyroid, lung) doses occurred for the VMAT with the flattening filter while the lowest for 3DCRT. For CyberKnife the highest doses were for gantry with caudal direction beams blocked (gantry close to OARs) while the lowest was the low modulated VOLO optimization technique. The easiest method to reduce peripheral doses was to combine FFF with jaw tracking and reducing monitor units at TrueBeam and to avoid gantry position close to OARs together with reduction of monitor units at CyberKnife, respectively. The presented strategies allowed to significantly reduce out-of-field and nontarget doses during prostate radiotherapy delivered with TrueBeam and CyberKnife. A different approach was required to reduce peripheral doses because of the difference in dose delivery techniques: non-coplanar using CyberKnife and coplanar using TrueBeam, respectively.

Criteria for Verification and Replanning Based on the Adaptive Radiotherapy Protocol "Best for Adaptive Radiotherapy" in Head and Neck Cancer.

No clear criteria have yet been established to guide decision-making for patient selection and the optimal timing of adaptive radiotherapy (ART) based on image-guided radiotherapy (IGRT). We have developed a novel protocol­the Best for Adaptive Radiotherapy (B-ART) protocol­to guide patient selection for ART. The aim of the present study is to describe this protocol, to evaluate its validity in patients with head and neck (HN) cancer, and to identify the anatomical and clinical predictors of the need for replanning. We retrospectively evaluated 82 patients with HN cancer who underwent helical tomotherapy (HT) and subsequently required replanning due to soft tissue changes upon daily MVCT. Under the proposed criteria, patients with anatomical changes >3 mm on three to four consecutive scans are candidates for ART. We compared the volumes on the initial CT scan (iCT) and the replanning CT (rCT) scan for the clinical target volumes (CTV1, referring to primary tumor or tumor bed and CTV2, metastatic lymph nodes) and for the parotid glands (PG) and body contour (B-body). The patients were stratified by primary tumor localization, clinical stage, and treatment scheme. The main reasons for replanning were: (1) a planning target volume (PTV) outside the body contour (n = 70; 85.4%), (2) PG shrinkage (n = 69; 84.1%), (3) B-body deviations (n = 69; 84.1%), and (4) setup deviations (n = 40; 48.8%). The replanning decision was made, on average, during the fourth week of treatment (n = 47; 57.3%). The mean reductions in the size of the right and left PG volumes were 6.31 cc (20.9%) and 5.98 cc (20.5%), respectively (p < 0.001). The reduction in PG volume was ≥30% in 30 patients (36.6%). The volume reduction in all of the anatomical structures was statistically significant. Four variables­advanced stage disease (T3−T4), chemoradiation, increased weight loss, and oropharyngeal localization­were significantly associated with the need for ART. The B-ART protocol provides clear criteria to eliminate random errors, and to allow for an early response to relevant changes in target volumes.

Does CyberKnife improve dose distribution versus IMRT and VMAT on a linear accelerator in low-risk prostate cancer?

BACKGROUND: Hypofractionated stereotactic body radiation therapy (SBRT) for prostate cancer (PCa) can be delivered with the robot-assisted CyberKnife (CK) system or on a linear accelerator using dynamic intensity-modulated radiotherapy (IMRT) or volumetric arc radiotherapy (VMAT). This retrospective study was performed to determine whether CK offers better dose distribution than IMRT and/or VMAT. MATERIALS AND METHODS: Treatment plans for three techniques were prepared using the same treatment parameters (36.35 Gy, 7.25 Gy/fr). We evaluated target coverage, conformity index (CI), homogeneity index (HI), gamma index (GI), and organs at risk (OAR) constraints. RESULTS: The mean planning target volume (PTV) dose for CK (39.58 Gy) was significantly greater than VMAT or IMRT (both 36.25 Gy). However, CK resulted in a wider dose range (31.48 to 45.89 Gy) vs. VMAT and IMRT (34.6-38.76 Gy). The mean dose to the rectum (V36Gy, mm3) was significantly lower (p < 0.001) in the CK plans (219.78 vs. 519.59 and 422.62, respectively). The mean bladder dose (V37Gy, mm3) was significantly greater for CK (3256 vs. 1090.75 for VMAT and 4.5 for IMRT (p < 0.001). CK yielded significantly better CI (1.07 vs. 1.17 and 1.25 for VMAT and IMRT, respectively; p < 0.01) and HI values (1.27 vs. 1.07 and 1.04; p < 0.01). GI values for the δd = 3mm, δ% = 3% criteria were 99.86 (VMAT), 99.07 (IMRT) and 99.99 (CK). For δd = 2mm, δ% = 2%, the corresponding values were 98.3, 93.35, and 97.12, respectively. CONCLUSIONS: For most variables, CK was superior to both VMAT and IMRT. However, dynamic IMRT techniques, especially VMAT, do not differ significantly from CK plans and are therefore acceptable alternatives to CyberKnife.

Development of a quasi-humanoid phantom to perform dosimetric and radiobiological measurements for out-of-field doses from external beam radiation therapy.

Our understanding of low dose, out-of-field radiation and their radiobiological effects are limited, in part due to the rapid technological advances in external beam radiotherapy, especially for non-coplanar and dynamic techniques. Reliable comparisons of out-of-field doses produced by advanced radiotherapy techniques are difficult due to the limitations of commercially available phantoms. There is a clear need for a functional phantom to accurately measure the dosimetric and radiobiological characteristics of out-of-field doses, which would in turn allow clinicians and medical physicists to optimize treatment parameters. We designed, manufactured, and tested the performance of a quasi-humanoid (Q-H) adult phantom. To test the physics parameters, we used computed tomography (CT) scans of assembled Q-H phantom. Static open field and dynamic techniques were measured both in- and out-of-field with ionization chambers and radiochromic films for two configurations (full solid and with water-filled containers). In the areas simulating soft tissues, lung, and bones, median Hounsfield units and densities were, respectively: 129.8, -738.7, 920.8 HU and 1.110, 0.215, 1.669 g/cm3 . Comparison of the measured to treatment planning systems (TPS) in-field dose values for the sample volumetric arc therapy (VMAT) (6 MV flattening filter-free (FFF)) plan, 96.4% of analyzed points passed the gamma evaluation criteria (L2%/2 mm, threshold (TH) 10%) and less than 1.50% for point dose verification. In the two phantom configurations: full poly(methyl) methacrylate (PMMA) and with water container, the off-axis median doses for open field, relative to the central axis of the beam (CAX) were similar, respectively: 0.900% versus 0.907% (15 cm distance to CAX); 0.096% versus 0.120% (35 cm); 0.018% versus 0.018% (52 cm); 0.009% versus 0.008% (74 cm). For VMAT 6 MV FFF, doses relative the CAX were, respectively: 0.667% (15 cm), 0.062% (35 cm), 0.019% (52 cm), 0.016% (74 cm). The Q-H phantom meets the International Commission on Radiation Units and Measurements (ICRU) and American Association of Physicists in Medicine (AAPM) recommended phantom criteria, providing medical physicists with a reliable, comprehensive system to perform dose calculation and measurements and to assess the impact on radiobiological response and on the risk of secondary tumor induction.

Ultra-Hypofractionated Proton Therapy in Localized Prostate Cancer: Passive Scattering versus Intensity-Modulated Proton Therapy.

Few studies have directly compared passive scattering (PS) to intensity-modulated proton therapy (IMPT) in the delivery of ultra-hypofractionated proton beams to the localized prostate cancer (PCa). In this preliminary study involving five patients previously treated with CyberKnife, treatment plans were created for PS and IMPT (36.25 CGE in five fractions with two opposing fields) to compare the dosimetric parameters to the planning target volume (PTV) and organs-at-risk (OAR: rectum, bladder, femoral heads). Both plans met the acceptance criteria. Significant differences were observed in the minimum and maximum doses to the PTV. The mean dose to the PTV was lower for PS (35.62 ± 0.26 vs. 37.18 ± 0.14; p = 0.002). Target coverage (D98%) was better for IMPT (96.79% vs. 99.10%; p = 0.004). IMPT resulted in significantly lower mean doses to the rectum (16.75 CGE vs. 6.88 CGE; p = 0.004) and bladder (17.69 CGE vs. 5.98 CGE p = 0.002). High dose to the rectum (V36.25 CGE) were lower with PS, but not significantly opposite to high dose to the bladder. No significant differences were observed in mean conformity index values, with a non-significant trend towards higher mean homogeneity index values for PS. Non-significant differences in the gamma index for both fields were observed. These findings suggest that both PS and IMPT ultra-hypofractionated proton therapy for PCa are highly precise, offering good target coverage and sparing of normal tissues and OARs.

Ionizing radiation exposure of stem cell-derived chondrocytes affects their gene and microRNA expression profiles and cytokine production.

Human induced pluripotent stem cells (hiPSCs) can be differentiated into chondrocyte-like cells. However, implantation of these cells is not without risk given that those transplanted cells may one day undergo ionizing radiation (IR) in patients who develop cancer. We aimed to evaluate the effect of IR on chondrocyte-like cells differentiated from hiPSCs by determining their gene and microRNA expression profile and proteomic analysis. Chondrocyte-like cells differentiated from hiPSCs were placed in a purpose-designed phantom to model laryngeal cancer and irradiated with 1, 2, or 3 Gy. High-throughput analyses were performed to determine the gene and microRNA expression profile based on microarrays. The composition of the medium was also analyzed. The following essential biological processes were activated in these hiPSC-derived chondrocytes after IR: "apoptotic process", "cellular response to DNA damage stimulus", and "regulation of programmed cell death". These findings show the microRNAs that are primarily responsible for controlling the genes of the biological processes described above. We also detected changes in the secretion level of specific cytokines. This study demonstrates that IR activates DNA damage response mechanisms in differentiated cells and that the level of activation is a function of the radiation dose.

Feasibility of intensity-modulated radiotherapy to treat gastric cancer.

AIM: To present a proposed gastric cancer intensity-modulated radiotherapy (IMRT) treatment planning protocol for an institution that have not introduced volumetric modulated arc therapy in clinical practice. A secondary aim was to determine the impact of 2DkV set-up corrections on target coverage and organ at risk (OAR). METHODS AND MATERIALS: Twenty consecutive patients were treated with a specially-designed non-coplanar 7-field IMRT technique. The isocenter-shift method was used to estimate the impact of 2DkV-based set-up corrections on the original base plan (BP) coverage. An alternative plan was simulated (SP) by taking into account isocenter shifts. The SP and BP were compared using dose-volume histogram (DVH) plots calculated for the internal target volume (ITV) and OARs. RESULTS: Both plans delivered a similar mean dose to the ITV (100.32 vs. 100.40%), with no significant differences between the plans in internal target coverage (5.37 vs. 4.96%). Similarly, no significant differences were observed between the maximal dose to the spinal cord (67.70 and 67.09%, respectively) and volume received 50% of the prescribed dose of: the liver (62.11 vs. 59.84%), the right (17.62 vs. 18.58%) and left kidney (29.40 vs. 30.48%). Set-up margins (SM) were computed as 7.80 mm, 10.17 mm and 6.71 mm in the left-right, cranio-caudal and anterior-posterior directions, respectively. CONCLUSION: Presented IMRT protocol (OAR dose constraints with selected SM verified by 2DkV verification) for stomach treatment provided optimal dose distribution for the target and the critical organs. Comparison of DVH for the base and the modified plan (which considered set-up uncertainties) showed no significant differences.

Patient safety in external beam radiotherapy, results of the ACCIRAD project: Recommendations for radiotherapy institutions and national authorities on assessing risks and analysing adverse error-events and near misses.

The ACCIRAD project, commissioned by the European Commission (EC) to develop guidelines for risk analysis of accidental and unintended exposures in external beam radiotherapy (EBRT), was completed in the year 2014. In 2015, the "General guidelines on risk management in external beam radiotherapy" were published as EC report Radiation Protection (RP)-181. The present document is the third and final report of the findings from the ACCIRAD project. The main aim of this paper is to describe the key features of the risk management process and to provide general guidelines for radiotherapy departments and national authorities on risk assessment and analysis of adverse error-events and near misses. The recommendations provided here and in EC report RP-181 are aimed at promoting the harmonisation of risk management systems across Europe, improving patient safety, and enabling more reliable inter-country comparisons.

Patient safety in external beam radiotherapy, results of the ACCIRAD project: Current status of proactive risk assessment, reactive analysis of events, and reporting and learning systems in Europe.

PURPOSE: To describe the current status of implementation of European directives for risk management in radiotherapy and to assess variability in risk management in the following areas: 1) in-country regulatory framework; 2) proactive risk assessment; (3) reactive analysis of events; and (4) reporting and learning systems. MATERIAL AND METHODS: The original data were collected as part of the ACCIRAD project through two online surveys. RESULTS: Risk assessment criteria are closely associated with quality assurance programs. Only 9/32 responding countries (28%) with national regulations reported clear "requirements" for proactive risk assessment and/or reactive risk analysis, with wide variability in assessment methods. Reporting of adverse error events is mandatory in most (70%) but not all surveyed countries. CONCLUSIONS: Most European countries have taken steps to implement European directives designed to reduce the probability and magnitude of accidents in radiotherapy. Variability between countries is substantial in terms of legal frameworks, tools used to conduct proactive risk assessment and reactive analysis of events, and in the reporting and learning systems utilized. These findings underscore the need for greater harmonisation in common terminology, classification and reporting practices across Europe to improve patient safety and to enable more reliable inter-country comparisons.

Carcinogenesis Induced by Low-dose Radiation.

BACKGROUND: Although the effects of high dose radiation on human cells and tissues are relatively well defined, there is no consensus regarding the effects of low and very low radiation doses on the organism. Ionizing radiation has been shown to induce gene mutations and chromosome aberrations which are known to be involved in the process of carcinogenesis. The induction of secondary cancers is a challenging long-term side effect in oncologic patients treated with radiation. Medical sources of radiation like intensity modulated radiotherapy used in cancer treatment and computed tomography used in diagnostics, deliver very low doses of radiation to large volumes of healthy tissue, which might contribute to increased cancer rates in long surviving patients and in the general population. Research shows that because of the phenomena characteristic for low dose radiation the risk of cancer induction from exposure of healthy tissues to low dose radiation can be greater than the risk calculated from linear no-threshold model. Epidemiological data collected from radiation workers and atomic bomb survivors confirms that exposure to low dose radiation can contribute to increased cancer risk and also that the risk might correlate with the age at exposure. CONCLUSIONS: Understanding the molecular mechanisms of response to low dose radiation is crucial for the proper evaluation of risks and benefits that stem from these exposures and should be considered in the radiotherapy treatment planning and in determining the allowed occupational exposures.

Prospective study on dosimetric comparison of helical tomotherapy and 3DCRT for craniospinal irradiation - A single institution experience.

AIM: This prospective study aims to assess feasibility of helical tomotherapy (HT) for craniospinal irradiation (CSI) and perform dosimetric comparison of treatment plans for both HT and 3D conformal radiotherapy (3DCRT). BACKGROUND: CSI is a challenging procedure. Large PTV size requires field matching due to technical limitations of standard linear accelerators, which cannot irradiate such volumes as a single field. HT could help to avoid these limitations as irradiation of long fields is possible without field matching. MATERIALS AND METHODS: Three adults were enrolled from 2009 to 2010. All patients received radiochemotherapy. Treatment plans in prone position for 3DCRT and in supine position for HT were generated. The superior plan was used for patients' irradiation. Plans were compared with the application of DVH, Dx parameters - where x represents a percentage of the structure volume receiving a normalized dose and homogeneity index (HI). RESULTS: All patients received HT irradiation. The treatment was well tolerated. The HT plans resulted in a better dose coverage and uniformity in the PTV: HI were 5.4, 7.8, 6.8 for HT vs. 10.3, 6.6, 10.4 for 3DCRT. For most organs at risk (OARs), the D(V80) was higher for HT than for 3DCRT, whereas D(V5) was lower for HT. CONCLUSIONS: HT is feasible for CSI, and in comparison with 3DCRT it improves PTV coverage. HT reduces high dose volumes of OARs, but larger volumes of normal tissue receive low radiation dose. HT requires further study to establish correlations between dosimetrical findings and clinical outcomes, especially with regard to late sequelae of treatment.

Patient safety in external beam radiotherapy - guidelines on risk assessment and analysis of adverse error-events and near misses: introducing the ACCIRAD project.

In 2011 the European Commission launched a tender to develop guidelines for risk analysis of accidental and unintended exposures in external beam radiotherapy. This tender was awarded to a consortium of 6 institutions, including the ESTRO, in late 2011. The project, denominated "ACCIRAD", recently finished the data collection phase. Data were collected by surveys administered in 38 European countries. Results indicate non-uniform implementation of event registration and classification, as well as incomplete or zero implementation of risk assessment and events analysis. Based on the survey results and analysis thereof, project leaders are currently drafting proposed guidelines entitled "Guidelines for patient safety in external beam radiotherapy - Guidelines on risk assessment and analysis of adverse-error events and near misses". The present article describes the aims and current status of the project, including results of the surveys.

Beam orientation in stereotactic radiosurgery using an artificial neural network.

BACKGROUND AND PURPOSE: To investigate the feasibility of using an artificial neural network (ANN) to generate beam orientations in stereotactic radiosurgery (SRS). MATERIAL AND METHODS: A dataset of 669 intracranial lesions was used to build, train, and validate three ANNs. In ANN1, Cartesian coordinates described the localization of the PTV and OARs. In ANN2, a genetic algorithm was used to optimize the model. In ANN3, vectors were used to define the distance between the PTV and OARs. In all ANNs, inputs consisted of the treatment plan parameters plus the patient's particular geometric parameters; outputs were beam and table angles. The ANN- and human-generated plans were then compared using dose-volume histograms, root-mean-square (RMS) and Gamma index methods. RESULTS: The mean volume of PTV covered by the 95% isodose was 99.2% in the MP's plan vs. 99.3%, 98.5% and 99.2% for ANN1, ANN2, and ANN3, respectively. No significant differences were observed between the plans. ANN1 showed the best agreement (Gamma index) with the human planner. While RMS errors in the three ANN models were comparable, ANN1 showed the lowest (best) values. CONCLUSION: ANN models were able to determine beam orientation in SRS. ANN-generated treatment plans were comparable to human-designed plans.