Evaluation of multiple-vendor AI autocontouring solutions.

BACKGROUND: Multiple artificial intelligence (AI)-based autocontouring solutions have become available, each promising high accuracy and time savings compared with manual contouring. Before implementing AI-driven autocontouring into clinical practice, three commercially available CT-based solutions were evaluated. MATERIALS AND METHODS: The following solutions were evaluated in this work: MIM-ProtégéAI+ (MIM), Radformation-AutoContour (RAD), and Siemens-DirectORGANS (SIE). Sixteen organs were identified that could be contoured by all solutions. For each organ, ten patients that had manually generated contours approved by the treating physician (AP) were identified, totaling forty-seven different patients. CT scans in the supine position were acquired using a Siemens-SOMATOMgo 64-slice helical scanner and used to generate autocontours. Physician scoring of contour accuracy was performed by at least three physicians using a five-point Likert scale. Dice similarity coefficient (DSC), Hausdorff distance (HD) and mean distance to agreement (MDA) were calculated comparing AI contours to "ground truth" AP contours. RESULTS: The average physician score ranged from 1.00, indicating that all physicians reviewed the contour as clinically acceptable with no modifications necessary, to 3.70, indicating changes are required and that the time taken to modify the structures would likely take as long or longer than manually generating the contour. When averaged across all sixteen structures, the AP contours had a physician score of 2.02, MIM 2.07, RAD 1.96 and SIE 1.99. DSC ranged from 0.37 to 0.98, with 41/48 (85.4%) contours having an average DSC ≥ 0.7. Average HD ranged from 2.9 to 43.3 mm. Average MDA ranged from 0.6 to 26.1 mm. CONCLUSIONS: The results of our comparison demonstrate that each vendor's AI contouring solution exhibited capabilities similar to those of manual contouring. There were a small number of cases where unusual anatomy led to poor scores with one or more of the solutions. The consistency and comparable performance of all three vendors' solutions suggest that radiation oncology centers can confidently choose any of the evaluated solutions based on individual preferences, resource availability, and compatibility with their existing clinical workflows. Although AI-based contouring may result in high-quality contours for the majority of patients, a minority of patients require manual contouring and more in-depth physician review.

Design and implementation of automated notification systems and an electronic whiteboard for radiation therapy planning monitoring.

PURPOSE: Radiotherapy (RT) treatment and treatment planning is a complex process prepared and delivered by a multidisciplinary team of specialists. Efficient communication and notification systems among different team members are therefore essential to ensure the safe, timely delivery of treatments to patients. METHOD: To address this issue, we developed and implemented automated notification systems and an electronic whiteboard to track every CT simulation, contouring task, the new-start schedule, and physician's appointments and tasks, and notify team members of overdue and missing tasks and appointments. The electronic whiteboard was developed to have a straightforward view of current patients' planning workflow and to help different team members coordinate with each other. The systems were implemented and have been used at our center to monitor the progress of treatment-planning tasks for over 2 years. RESULTS: The last-minute plans were relatively reduced by about 40% in 2023 compared to 2021 and 2022 with a p-value < 0.05. The overdue contouring tasks of more than 1 day decreased from 46.8% in 2019 and 33.6% in 2020 to 20%-26.4% in 2021-2023 with a p-value < 0.05 after the implementation of the notification system. The rate of plans with 1-3 day planning time decreased by 20.31%, 39.32%, and 24.08% with a p-value < 0.05 and the rate of plans with 1-3 day planning time due to the contouring task overdue more than 1 day decreased by 49.49%, 56.89%, and 46.52% with a p-value < 0.05 after the implementation. The rate of outstanding appointments that are overdue by more than 7 days decreased by more than 5% with a p-value < 0.05 following the implementation of the system. CONCLUSIONS: Our experience shows that this system requires minimal human intervention, improves the treatment planning workflow and process by reducing errors and delays in the treatment planning process, positively impacts on-time treatment plan completion, and reduces the need for compressed or rushed treatment planning timelines.

Case Report: Bilateral targeted intraoperative radiotherapy: a safe and effective alternative for synchronous bilateral breast cancer.

Background: The incidence of bilateral breast cancer (BBC) ranges from 1.4% to 11.8%. BBC irradiation is a challenge in current clinical practice due to the large target volume that must be irradiated while minimizing the dose to critical organs. Supine or prone breast techniques can be used, with the latter providing better organ sparing; both, however, result in lengthy treatment times. The use of Intra-operative radiotherapy (IORT) in breast cancer patients who choose breast conservation has been highlighted in previous studies, but there is a scarcity of literature analyzing the utility and applicability of IORT in BBC. This case series aims to highlight the applicability of administering bilateral IORT in patients with BBC. Case reports: Five patients with bilateral early-stage breast cancer (or DCIS) were treated with breast-conserving surgery followed by bilateral IORT. Of the 10 breast cancers, 8 were diagnosed as either DCIS or IDC, while the other 2 were diagnosed as invasive lobular carcinoma and invasive carcinoma, respectively. During surgery, all patients received bilateral IORT. Furthermore, 1 patient received external beam radiation therapy after her final pathology revealed grade 3 DCIS. The IORT procedure was well tolerated by all five patients, and all patients received aromatase inhibitors as adjuvant therapy. Additionally, none of these patients showed evidence of disease after a 36-month median follow-up. Conclusion: Our findings demonstrate the successful use of IORT for BCS in patients with BBC. Furthermore, none of the patients in our study experienced any complications, suggesting the feasibility of the use of IORT in BBC. Considering the benefits of improved patient compliance and a reduced number of multiple visits, IORT may serve as an excellent patient-centered alternative for BBC. Future studies are recommended to reinforce the applicability of IORT in patients with BBC.

Organ-at-risk dose prediction using a machine learning algorithm: Clinical validation and treatment planning benefit for lung SBRT.

OBJECTIVE: To quantify the clinical performance of a machine learning (ML) algorithm for organ-at-risk (OAR) dose prediction for lung stereotactic body radiation therapy (SBRT) and estimate the treatment planning benefit from having upfront access to these dose predictions. METHODS: ML models were trained using multi-center data consisting of 209 patients previously treated with lung SBRT. Two prescription levels were investigated, 50 Gy in five fractions and 54 Gy in three fractions. Models were generated using a gradient-boosted regression tree algorithm using grid searching with fivefold cross-validation. Twenty patients not included in the training set were used to test OAR dose prediction performance, ten for each prescription. We also performed blinded re-planning based on OAR dose predictions but without access to clinically delivered plans. Differences between predicted and delivered doses were assessed by root-mean square deviation (RMSD), and statistical differences between predicted, delivered, and re-planned doses were evaluated with one-way analysis of variance (ANOVA) tests. RESULTS: ANOVA tests showed no significant differences between predicted, delivered, and replanned OAR doses (all p ≥ 0.36). The RMSD was 2.9, 3.9, 4.3, and 1.7Gy for max dose to the spinal cord, great vessels, heart, and trachea, respectively, for 50 Gy in five fractions. Average improvements of 1.0, 1.4, and 2.0 Gy were seen for spinal cord, esophagus, and trachea max doses in blinded replans compared to clinically delivered plans with 54 Gy in three fractions, and 1.8, 0.7, and 1.5 Gy, respectively, for the esophagus, heart and bronchus max doses with 50 Gy in five fractions. Target coverage was similar with an average PTV V100% of 94.7% for delivered plans compared to 97.3% for blinded re-plans for 50 Gy in five fractions, and respectively 98.4% versus 99.2% for 54 Gy in three fractions. CONCLUSION: This study validated ML-based OAR dose prediction for lung SBRT, showing potential for improved OAR dose sparing and more consistent plan quality using dose predictions for patient-specific planning guidance.

Comprehensive Patient-Specific Intensity-Modulated Radiation Therapy Quality Assurance Comparing Mobius3D/FX to Conventional Methods of Evaluation.

Purpose To determine the appropriateness of implementing Mobius3D/FX (Varian Medical Systems, Inc., Palo Alto, CA, USA) as not only a pretreatment secondary check but as an alternative to measurement-based patient-specific intensity-modulated radiation therapy (IMRT) quality assurance (QA). Methods Mobius3D/FX was commissioned and stock beam models were tweaked so that an independent recalculated 3D dose distribution can be obtained. Then, 50 patient-specific treatment plans for various indications were delivered across a 2D ion chamber array, radiochromic film setup, and electronic portal imager and analyzed with MobiusFX and gamma analysis. The concordance of plans scored as passing between MobiusFX and the conventional methods of QA was determined. Results All analyzed treatment plans passed with a gamma passing rate >90% across all conventional QA methods, most commonly using a 3%/3mm gamma criterion except for film measurements where a 5%/3mm criterion was applied. There was good agreement and concordance between MobiusFX and conventional methods when using a 3%/3mm criteria for MobiusFX, whereas a 2%/2mm criteria appeared too stringent as it failed treatment plans deemed clinically acceptable using conventional methods. Conclusions Using a 50-sample subset of clinically delivered treatment plans this non-inferiority-type comparison shows Mobius3D/FX based on log file analysis to be a suitable alternative to conventional QA methods when utilizing the 3%/3mm gamma criterion. Methods based on log file analysis can provide an opportunity for resource sparing, improving the efficiency, and workflow for evaluating IMRT treatment plans.

Validation of accuracy deformable image registration contour propagation using a benchmark virtual HN phantom dataset.

Virtual anatomic phantoms offer precise voxel mapping of the variation of anatomy with ground truth deformation vector fields (DVFs). Dice similarity coefficient (DSC) and mean distance to agreement (MDA) are the standard metrics for evaluating geometric contour congruence when testing deformable registration (DIR) algorithms. A HN virtual patient phantom data set was used for a kVCT-kVCT automatic propagation contour validation study employing the Accuray DIR algorithm. Furthermore, since TomoTherapy uses MVCT images of the relevant anatomy for adaptive monitoring, the kVCT image data set quality was transformed to an MVCT image data set quality to study intermodal kVCT-MVCT DIR accuracy. The results of the study indicate that the Accuray DIR algorithm can be expected to autopropagate HN contours adequately, on average, within tolerances recommended by TG-132 (DSC 0.8-0.9, MDA within voxel width). However, contours critical to dosimetric planning should always be visually proofed for accuracy. Using standard reconstruction MVCT image quality causes slightly less, but acceptable, agreement with ground truth contours.

A skin dose prediction model based on in vivo dosimetry and ultrasound skin bridge measurements during intraoperative breast radiation therapy.

PURPOSE: Using in vivo measurements from optically stimulated luminescence dosimeters (OSLDs) to develop and validate a prediction model for estimating the skin dose received by patients undergoing breast intraoperative radiation therapy (IORT). METHODS AND MATERIALS: IORT was performed using INTRABEAM-600 with spherical applicators placed in the lumpectomy cavity. Ultrasound skin bridge measurements were used to determine the applicator-to-skin distance, with OSLDs placed to measure the skin surface dose at the corresponding points. The OSLD response was calibrated for the 50 kVp INTRABEAM-600 output. Models were fit to describe the dose fall-off with increasing applicator-to-skin distance and the best fitting model was chosen for estimating skin dose. RESULTS: Twenty four patients with 25 lumpectomy cavities were included, and the average skin dose recorded was 1.18 Gy ± 0.88 Gy, ranging from 0.17 Gy to 4.77 Gy, with an average applicator-to-skin distance of 19.9 mm ± 5.1 mm. An exponential-plateau model was found to best describe the dose fall-off with a root-mean-square error of 0.73. This model was then validated prospectively using skin dose measurements from five consecutive patients. Validation measurements were well within the 95% prediction limits of the model, with a root-mean-square error of 0.52, showing that the prediction model accurately estimates skin dose using ultrasound skin bridge measurements. CONCLUSIONS: This prediction model constitutes a useful tool for estimating the skin dose received during breast lumpectomy IORT. The model and accompanying 95% confidence intervals can be used to establish a minimum allowable skin bridge distance, effectively limiting the maximum allowable skin dose.

Evaluating which plan quality metrics are appropriate for use in lung SBRT.

OBJECTIVE: Several dose metrics in the categories-homogeneity, coverage, conformity and gradient have been proposed in literature for evaluating treatment plan quality. In this study, we applied these metrics to characterize and identify the plan quality metrics that would merit plan quality assessment in lung stereotactic body radiation therapy (SBRT) dose distributions. METHODS: Treatment plans of 90 lung SBRT patients, comprising 91 targets, treated in our institution were retrospectively reviewed. Dose calculations were performed using anisotropic analytical algorithm (AAA) with heterogeneity correction. A literature review on published plan quality metrics in the categories-coverage, homogeneity, conformity and gradient was performed. For each patient, using dose-volume histogram data, plan quality metric values were quantified and analysed. RESULTS: For the study, the radiation therapy oncology group (RTOG) defined plan quality metrics were: coverage (0.90 ± 0.08); homogeneity (1.27 ± 0.07); conformity (1.03 ± 0.07) and gradient (4.40 ± 0.80). Geometric conformity strongly correlated with conformity index (p < 0.0001). Gradient measures strongly correlated with target volume (p < 0.0001). The RTOG lung SBRT protocol advocated conformity guidelines for prescribed dose in all categories were met in ≥94% of cases. The proportion of total lung volume receiving doses of 20 Gy and 5 Gy (V20 and V5) were mean 4.8% (±3.2) and 16.4% (±9.2), respectively. CONCLUSION: Based on our study analyses, we recommend the following metrics as appropriate surrogates for establishing SBRT lung plan quality guidelines-coverage % (ICRU 62), conformity (CN or CIPaddick) and gradient (R50%). Furthermore, we strongly recommend that RTOG lung SBRT protocols adopt either CN or CIPadddick in place of prescription isodose to target volume ratio for conformity index evaluation. Advances in knowledge: Our study metrics are valuable tools for establishing lung SBRT plan quality guidelines.

Interstitial Surgical Cavity Sizing Applicators for the Treatment of the Breast Lumpectomy Cavity with Intraoperative Radiation Therapy.

Surgical cavity sizing applicators were developed for utilization prior to intraoperative radiation therapy (IORT) of the breast lumpectomy cavity with the Zeiss INTRABEAM (Carl Zeiss Meditec AG, Jena, Germany) device. The use of these applicators minimizes the number of sterilizations of the treatment applicator, which is currently limited to 100 sterilizations per applicator. This maximizes the number of patients who can be treated with each applicator, resulting in cost savings for the treating institution.

Right Care for the Right Patient Each and Every Time.

PURPOSE: To implement a biometric patient identification system in the field of radiation oncology. MATERIALS AND METHODS: A biometric system using palm vein scanning technology has been implemented to ensure the delivery of treatment to the correct patient each and every time. By interfacing a palm vein biometrics system (PVBS) (PatientSecure®, Imprivata, Lexington, Massachusetts) with the radiation oncology patient management system (ROPMS) (ARIA®, Varian Medical Systems, Palo Alto, California) one can integrate patient check-in at the front desk and identify and open the correct treatment record of the patient at the point of care prior to the initiation of the radiation therapy treatment. RESULTS: The learning time for the use of the software and palm scanner was extremely short. The staff at the front desk and treatment machines learned the procedures to use, clean, and care for the device in one hour's time. The first key to the success of the system is to have a policy and procedure in place; such a procedure was created and put in place in the department from the first day. The second key to the success is the actual hand placement on the scanner. Learning the proper placement and gently reminding patients from time to time was found to be efficient and to work well. CONCLUSION: The use of a biometric patient identification system employing palm vein technology allows one to ensure that the right care is delivered to the right patient each and every time. Documentation through the PVBS database now exists to show that this has taken place.

An efficient Volumetric Arc Therapy treatment planning approach for hippocampal-avoidance whole-brain radiation therapy (HA-WBRT).

An efficient and simple class solution is proposed for hippocampal-avoidance whole-brain radiation therapy (HA-WBRT) planning using the Volumetric Arc Therapy (VMAT) delivery technique following the NRG Oncology protocol NRG-CC001 treatment planning guidelines. The whole-brain planning target volume (PTV) was subdivided into subplanning volumes that lie in plane and out of plane with the hippocampal-avoidance volume. To further improve VMAT treatment plans, a partial-field dual-arc technique was developed. Both the arcs were allowed to overlap on the in-plane subtarget volume, and in addition, one arc covered the superior out-of-plane sub-PTV, while the other covered the inferior out-of-plane subtarget volume. For all plans (n = 20), the NRG-CC001 protocol dose-volume criteria were met. Mean values of volumes for the hippocampus and the hippocampal-avoidance volume were 4.1 cm(3) ± 1.0 cm(3) and 28.52 cm(3) ± 3.22 cm(3), respectively. For the PTV, the average values of D(2%) and D(98%) were 36.1 Gy ± 0.8 Gy and 26.2 Gy ± 0.6 Gy, respectively. The hippocampus D(100%) mean value was 8.5 Gy ± 0.2 Gy and the maximum dose was 15.7 Gy ± 0.3 Gy. The corresponding plan quality indices were 0.30 ± 0.01 (homogeneity index), 0.94 ± 0.01 (target conformality), and 0.75 ± 0.02 (confirmation number). The median total monitor unit (MU) per fraction was 806 MU (interquartile range [IQR]: 792 to 818 MU) and the average beam total delivery time was 121.2 seconds (IQR: 120.6 to 121.35 seconds). All plans passed the gamma evaluation using the 5-mm, 4% criteria, with γ > 1 of not more than 9.1% data points for all fields. An efficient and simple planning class solution for HA-WBRT using VMAT has been developed that allows all protocol constraints of NRG-CC001 to be met.

Spine stereotactic body radiation therapy plans: Achieving dose coverage, conformity, and dose falloff.

We report our experience of establishing planning objectives to achieve dose coverage, conformity, and dose falloff for spine stereotactic body radiation therapy (SBRT) plans. Patients with spine lesions were treated using SBRT in our institution since September 2009. Since September 2011, we established the following planning objectives for our SBRT spine plans in addition to the cord dose constraints: (1) dose coverage­prescription dose (PD) to cover at least 95% planning target volume (PTV) and 90% PD to cover at least 99% PTV; (2) conformity index (CI)­ratio of prescription isodose volume (PIV) to the PTV < 1.2; (3) dose falloff­ratio of 50% PIV to the PTV (R(50%)); (4) and maximum dose in percentage of PD at 2 cm from PTV in any direction (D(2cm)) to follow Radiation Therapy Oncology Group (RTOG) 0915. We have retrospectively reviewed 66 separate spine lesions treated between September 2009 and December 2012 (31 treated before September 2011 [group 1] and 35 treated after [group 2]). The χ(2) test was used to examine the difference in parameters between groups. The PTV V(100% PD) ≥ 95% objective was met in 29.0% of group 1 vs 91.4% of group 2 (p < 0.01) plans. The PTV V(90% PD) ≥ 99% objective was met in 38.7% of group 1 vs 88.6% of group 2 (p < 0.01) plans. Overall, 4 plans in group 1 had CI > 1.2 vs none in group 2 (p = 0.04). For D(2cm), 48.3% plans yielded a minor violation of the objectives and 16.1% a major violation for group 1, whereas 17.1% exhibited a minor violation and 2.9% a major violation for group 2 (p < 0.01). Spine SBRT plans can be improved on dose coverage, conformity, and dose falloff employing a combination of RTOG spine and lung SBRT protocol planning objectives.

Single institution experience treating 104 vestibular schwannomas with fractionated stereotactic radiation therapy or stereotactic radiosurgery.

UNLABELLED: The pupose of this study is to assess the long-term outcome and toxicity of fractionated stereotactic radiation therapy (FSRT) and stereotactic radiosurgery (SRS) for 100 vestibular schwannomas treated at a single institution. From 1993 to 2007, 104 patients underwent were treated with radiation therapy for vestibular schwannoma. Forty-eight patients received SRS, with a median prescription dose of 12.5 Gy for SRS (range 9.7-16 Gy). For FSRT, two different fraction schedules were employed: a conventional schedule (ConFSRT) of 1.8 Gy per fraction (Gy/F) for 25 or 28 fractions to a total dose of 45 or 50.4 Gy (n = 19); and a once weekly hypofractionated course (HypoFSRT) consisting of 4 Gy/F for 5 fractions to a total dose of 20 Gy (n = 37). Patients treated with FSRT had better baseline hearing, facial, and trigeminal nerve function, and were more likely to have a diagnosis of NF2. The 5-year progression free rate (PFR) was 97.0 after SRS, 90.5% after HypoFSRT, and 100.0% after ConFSRT (p = NS). Univariate analysis demonstrated that NF2 and larger tumor size (greater than the median) correlated with poorer local control, but prior surgical resection did not. Serviceable hearing was preserved in 60.0% of SRS patients, 63.2% of HypoFSRT patients, and 44.4% of ConFSRT patients (p = 0.6). Similarly, there were no significant differences in 5-year rates of trigeminal toxicity facial nerve toxicity, vestibular dysfunction, or tinnitus. CONCLUSIONS: Equivalent 5-year PFR and toxicity rates are shown for patients with vestibular schwanoma selected for SRS, HypoFSRT, and ConFSRT after multidisciplinary evaluation. Factors correlating with tumor progression included NF2 and larger tumor size.

High dose per fraction dosimetry of small fields with gafchromic EBT2 film.

PURPOSE: Small field dosimetry is prone to uncertainties due to the lack of electronic equilibrium and the use of the correct detector size relative to the field size measured. It also exhibits higher sensitivity to setup errors as well as large variation in output with field size and shape. Radiochromic film is an attractive method for reference dosimetry in small fields due to its ability to provide 2D dose measurements while having minimal impact on the dose distribution. Gafchromic EBT2 has a dose range of up to 40 Gy; therefore, it could potentially be useful for high dose reference dosimetry with high spatial resolution. This is a requirement in stereotactic radiosurgery deliveries, which deliver high doses per fraction to small targets. METHODS: Targets of 4 mm and 12 mm diameters were treated to a minimum peripheral dose of 21 Gy prescribed to 80% of the maximum dose in one fraction. Target doses were measured with EBT2 film (both targets) and an ion chamber (12 mm target only). Measured doses were compared with planned dose distributions using profiles through the target and minimum peripheral dose coverage. RESULTS: The measured target doses and isodose coverage agreed with the planned dose within +/-1 standard deviation of three measurements, which were 2.13% and 2.5% for the 4 mm and 12 mm targets, respectively. CONCLUSIONS: EBT2 film is a feasible dosimeter for high dose per fraction reference 2D dosimetry.

A technique for stereotactic radiosurgery treatment planning with helical tomotherapy.

The purpose of this study was to develop an efficient and effective planning technique for stereotactic radiosurgery using helical tomotherapy. Planning CTs and contours of 20 patients, previously treated in our clinic for brain metastases with linac-based radiosurgery using circular collimators, were used to develop a robust TomoTherapy planning technique. Plan calculation times as well as delivery times were recorded for all patients to allow for an efficiency evaluation. In addition, conformation and homogeneity indices were calculated as metrics to compare plan quality with that which is achieved with conventional radiosurgery delivery systems. A robust and efficient planning technique was identified to produce plans of radiosurgical quality using the TomoTherapy treatment planning system. Dose calculation did not exceed a few hours and resulting delivery times were less than 1 hour, which allows the process to fit into a single day radiosurgery workflow. Plan conformity compared favorably with published results for gamma knife radiosurgery. In addition, plan homogeneity was similar to linac-based approaches. The TomoTherapy planning software can be used to create plans of acceptable quality for stereotactic radiosurgery in a time that is appropriate for a radiosurgery workflow that requires that planning and delivery occur within 1 treatment day.