BACKGROUND:Abdominal compression (AC) using a belt-based system is widely employed in radiation therapy to limit respiration-induced target motion for abdominal and hepatic malignancies. While AC has demonstrated efficacy in reducing intrafraction motion amplitude, the interfractional anatomical consequences of daily compression belt (CB) placement variability remain poorly characterized. A clearer understanding of this variability is necessary to determine whether online adaptive radiotherapy (oART) is required to maintain dosimetric robustness throughout the treatment course. PURPOSE:This study aims to quantify interfractional anatomical variability associated with CB use in a clinical MR-guided radiotherapy cohort and to compare these findings against a matched cohort of non-compressed abdominal patients to determine the clinical significance of CB-induced setup variability. METHODS:Daily treatment MR images were retrospectively collected from 17 CB patients and 17 matched non-CB controls treated on an MR-Linac. External contours were delineated at the center of the planning target volume (PTV) and at the region of maximum compression for CB patients. Interfractional variability was assessed using cross-sectional area (CSA), Dice Similarity Coefficient (DSC), Mean Distance to Agreement (MDA), and Hausdorff Distance (HD). Interfractional organ displacement of the liver and kidneys was quantified via rigid registration. Linear mixed-effects models (LMEMs) evaluated the association between patient demographics-including age, BMI, sex, subcutaneous adipose tissue (SAT), visceral adipose tissue (VAT)-and interfractional variability metrics. RESULTS:Interfractional DSC over the PTV region was significantly lower in the CB cohort compared with non-CB controls (p = 0.01 and p = 0.03 for comparisons to fraction one and sequential fraction comparisons, respectively), and MDA was significantly elevated in the CB cohort (p < 0.001). No significant difference in CSA variability was observed between cohorts. In the maximum compression region, median HD was 1.6 cm relative to fraction one, with no significant correlation between compression-region and PTV-region CSA variability (p = 0.99). Older age, male sex, and higher BMI were associated with improved belt-contact shape reproducibility (all p < 0.05), though these demographic effects did not translate to PTV-region stability. LMEM analysis revealed that fixed demographic effects explained only a small fraction of PTV variability (R2 m = 0.06-0.29), with the remainder attributable to patient-specific random effects. Interfractional organ displacement ranged from 5.91 to 9.15 mm across organs and cohorts, with no statistically significant difference between groups. CONCLUSIONS:CB use introduces measurable but modest increases in interfractional anatomical variability at the PTV level compared with standard abdominal setups. Compression-site variability does not propagate directly to the treatment target, and patient demographics alone are insufficient to predict CB-induced setup variability. Daily MR-based image guidance and selective adaptive replanning remain the appropriate management strategy for patients treated with abdominal compression.
BACKGROUND:Signal-to-noise ratio (SNR) measurements are important for quality assurance, especially when quantitative MRI markers or new hardware are being developed or evaluated. Regions of low signal or assumed uniform signal are used for SNR estimation in a variety of MR image types, but this is unreliable due to spatial heterogeneity of noise arising from the organization of multiple-detector arrays and sensitivity encoding (SENSE) reconstruction. PURPOSE:We present a universally applicable method for empirical noise mapping by propagating simulated noise through the image-generating reconstruction pipeline, to identify optimal regions of interest (ROIs) for determining noise levels in prostate diffusion-weighted imaging (DWI). METHODS:Thirty-six men were imaged at 3T with DWI. K-space data were propagated through an off-line reconstruction pipeline including sensitivity encoding (SENSE), to generate DW images. Simulated complex Gaussian noise generated from prescan per-channel measurements was similarly processed to produce pure noise maps. Per-voxel standard deviations (SD) of spatial distributions of noise, averaged over prostate, internal obturator muscle (IOM), low signal (anterior and posterior), rectum, and bladder ROIs were calculated and compared to evaluate suitability for SNR estimation in prostate DWI. RESULTS:IOM ROI noise levels, averaged bilaterally, were not significantly different from those in the prostate. Bladder ROI noise levels were comparable. Noise levels measured in low signal areas were highly dependent on SENSE algorithm regularization strength. DISCUSSION:The proposed method can be used-on any scanner-to accurately calculate noise levels in MR images, without repeated acquisition or interference from SENSE use. CONCLUSION:Bladder and averaged IOM ROIs are appropriate for estimating DWI noise levels in the prostate. Low signal ROIs should not be used, due to the strong dependence on the SENSE regularization strength, and by extension, on the SENSE reconstruction algorithm. The presented method can be used instead to obtain accurate spatial maps of noise levels.
BACKGROUND:Osteonecrosis of the femoral head (ONFH) frequently progresses to femoral head collapse, resulting in irreversible joint dysfunction and the need for total hip arthroplasty. Reliable assessment of magnetic resonance imaging (MRI)-defined collapse status remains limited with conventional assessment methods. PURPOSE:To develop and validate a multi-modal nomogram integrating MRI-based radiomics signatures and clinical risk factors for the assessment of MRI-defined collapse status in patients with osteonecrosis of the femoral head (ONFH). METHODS:This retrospective study analyzed 149 ONFH patients (99 non-collapsed, 50 collapsed), partitioned into training (n = 104) and testing (n = 45) sets. Radiomics features were extracted from MRI-based necrotic lesion segmentations. A Radscore, derived via least absolute shrinkage and selection operator (LASSO) regression, was integrated with selected clinical variables to construct a combined nomogram. RESULTS:The combined clinical-radiomics model demonstrated favorable discriminative performance, achieving an AUC of 0.940 (95% CI: 0.858-0.994) in the testing set, which outperformed the clinical-only model (AUC = 0.589) and the radiomics-only model (AUC = 0.818). Using the cutoff determined by the Youden index in the training cohort, the combined model exhibited a sensitivity of 80.0%, specificity of 90.0%, and overall accuracy of 86.7% in the testing cohort. Calibration curves showed acceptable agreement between model-predicted probabilities and observed MRI-defined collapse status. CONCLUSION:The integrated clinical-radiomics nomogram showed promising performance for non-invasive assessment of femoral head collapse status in patients with ONFH. This quantitative approach may assist individualized collapse-status stratification, but external validation and prospective clinical workflow studies are required before routine clinical implementation.
BACKGROUND:A common method to estimate the effective dose in CT is to multiply the dose length product in a CT scan by a conversion factor (k). The k-factors are determined by dividing the effective dose (calculated using a weighted summation of organ doses) by the dose length product (DLP) for specific digital phantoms. The k-factors determined in this way may vary substantially depending on the effective dose calculation methods and phantom models used. PURPOSE:The purpose of this study was to estimate variation associated with different acquisition and patient characteristics in the quantification of k-factors. METHODS:K-factors were calculated using VirtualDoseCT for a diverse set of computational phantoms representing different patient models, scanner models, and seven CT exams. The tube potential varied from 80 to 120 kV. The scan length was varied by ± 5 cm relative to the default setting. The CTDIvol was set to 10 mGy when tube current modulation was not used; otherwise, the software default was used. A multivariable regression method was employed to derive an empirical formula for k-factors, along with model-uncertainty analysis and Sobol's sensitivity analysis to identify the influence of each variable. RESULTS:K-factors for all exams increased with kV. Negative dependence of k-factors on BMI was observed, with lower values in males than in females. Moderate to high variation was observed in the quantification of k-factors without TCM, which remained unchanged with TCM. Based on Sobol's indices (0.8-1), k-factors were influenced most by scan length and patient size for adults, and scan length and age for pediatric phantoms. DISCUSSION:The study showed that a large spread in k-factors was predominantly influenced by scan length, patient size and age. The heterogeneity in k-factors may lead to large heterogeneity in the effective dose estimate if the DLP×k method is used. These results indicate that age and size specific k-factors to estimate effective dose from DLP may be warranted. CONCLUSIONS:A set of k-factors should be established to account for these factors and improve effective dose estimations.
BACKGROUND:Cone-beam computed tomography (CBCT) is increasingly used in adaptive radiotherapy (ART), yet the dosimetric accuracy and image quality of advanced CBCT imaging platforms-such as Halcyon HyperSight 4.0 (HH), which combines an upgraded detector panel with advanced iterative reconstruction, scatter-correction, and metal-artifact-reduction algorithms-remain insufficiently characterized clinically. PURPOSE:To evaluate HH CBCT reconstruction algorithms for ART by assessing image quality, dose accuracy, and in vivo verification using optically stimulated luminescence dosimeters (OSLDs). METHODS:Task-based image quality was evaluated with ImQuest per AAPM TG-233 using a Gammex 464 ACR phantom on three systems: Canon Aquilion Large Bore CT (LB, reference), TrueBeam (TB), and HH. Seven algorithms were compared (LB, TB-FDK, TB-iCBCT, HH-FDK, HH-iCBCT, HH-Acuros, HH-MAR), with d' (the detectability index, a task-based image-quality metric per AAPM TG-233) derived for four insert materials (polyethylene, acrylic, bone, air). A volumetric modulated arc therapy (VMAT) plan was independently re-optimized (using identical optimization objectives/constraints as the LB reference plan) on HH-iCBCT, HH-Acuros, and HH-MAR images and verified by portal dosimetry (1%/1 mm gamma), delivered to an Alderson RANDO phantom with dose measured at 15 points (2 clinical target volume [CTV], 7 organ-at-risk [OAR], and 6 lung points) using BeO OSLDs and compared to LB (± 5% criterion). Mean d' per algorithm was compared to LB via one-sample t-tests (p < 0.05). RESULTS:HH-iCBCT (d' = 4.70, p = 0.016), HH-Acuros (d' = 4.61, p = 0.005), and HH-MAR (d' = 4.71, p = 0.024, numerically highest) showed significant detectability improvement over LB (d' = 4.42), with HH-iCBCT most consistent across materials. Mean absolute OSLD dose errors were 1.71%, 2.09%, and 2.90%; all 15 points met the ± 5% criterion except Lt_lung2 in HH-MAR (-5.32%). CONCLUSION:HH-iCBCT demonstrated clinically acceptable image quality and dose accuracy for CBCT-based ART, supported by in vivo BeO OSLD verification in an anthropomorphic phantom.
BACKGROUND:Total skin electron therapy (TSET) is a standard treatment for cutaneous T-cell lymphoma. Due to the complex patient positioning and irregular body contours, robust in-vivo dosimetry (IVD) is essential to verify dose uniformity. While thermoluminescent dosimeters (TLDs) are traditional standard, their utility is hindered by labor-intensive, manual processing. PURPOSES:This study evaluates the clinical implementation of a Gafchromic™ EBT4 film-based IVD system integrated with a bespoke, open-source analysis platform for automated batch-processing. To address the inherent orientation dependence and loss of film orientation frequently encountered when preparing small-format IVD films with a manual paper cutter, we developed an orthogonal dual-scan protocol. By averaging pixel values from two perpendicular scans for both calibration and clinical measurement, this protocol effectively mitigates orientation-dependent uncertainties and ensures dosimetric robustness. METHODS:The film-based IVD system was clinically assessed in five TSET treatments and compared to six treatments utilizing TLD-based IVD. Dosimetric accuracy was evaluated by comparing normalized fractional doses to the prescription. Statistical analysis was performed using the Mann-Whitney U test with the Benjamini-Hochberg procedure to control the false discovery rate across 20 anatomical sites. Workflow efficiency was quantified by the total time required for data readout and analysis per patient. RESULTS:The EBT4 film-based IVD demonstrated dosimetric accuracy comparable to the TLDs, with mean fractional doses of 102.7% ± 9.6% and 100.4% ± 9.7%, respectively (p = 0.25). Clinical implementation of the film-based approach significantly enhanced efficiency, reducing the total processing time from approximately 50 min to 15 min per treatment. CONCLUSIONS:The proposed EBT4 film-based IVD system, supported by open-source software and an orthogonal dual-scan protocol, offers a robust, cost-effective, and time-efficient alternative to traditional TLDs. This approach streamlines the clinical workflow without compromising dosimetric accuracy, making it a viable solution for TSET and broader radiotherapy IVD applications.
BACKGROUND:TG-43 formalism dose calculation for brachytherapy with the homogeneous water-based assumption is widely used, which can lead to dose inaccuracy in treatment planning. PURPOSE:Extremity malignancies are also treated with high-dose-rate (HDR) 192Ir interstitial brachytherapy, a treatment site that has not been specifically investigated in prior comparative dosimetric studies, and the proximity of source dwell positions to bone and air introduces significant heterogeneity. BrachyPlanCheck, an in-house brachytherapy second dose verification program for Monte Carlo (MC) simulations, incorporates these heterogeneities and is used for a comparative dosimetric study on extremity malignancies. METHODS:BrachyPlanCheck was commissioned and validated in both a homogeneous water phantom and a heterogeneous breast phantom. The clinical evaluation included 8 patient-specific lower limb brachytherapy plans, comparing TG-43 and MC calculation using BrachyPlanCheck. Dosimetric differences were evaluated using dose comparisons, gamma index analysis, and dose-volume histogram (DVH) metrics for clinical target volume (CTV) and organ-at-risk (OAR), including bone and skin. RESULTS:In commissioning Level 1, MC simulations using BrachyPlanCheck demonstrated point dose agreement within ± 0.9%. Gamma passing rates exceeded 99.9% with criteria 1 mm/1% in commissioning Level 2. In clinical evaluation, TG-43 calculation overestimated CTV V150 by up to 10.50% and CTV D90 by up to 7.25%. OAR doses were generally higher with TG-43, with mean differences of 1.09% for bone D1cc (p = 0.109), 1.38% for bone D0.1cc (p = 0.078), 7.30% for skin D2cc (p = 0.008), and 5.68% for skin D0.1cc (p = 0.008). CONCLUSION:Validated by phantom and patient study, BrachyPlanCheck demonstrated robust performance for independent secondary dose verification. TG-43 calculation overestimated dose to CTV and OAR in extremity malignancy cases compared with BrachyPlanCheck.
BACKGROUND:Accurate delineation of target volumes and organs-at-risk (OARs) is a critical yet labor-intensive component of rectal cancer radiotherapy. While deep learning (DL)-based automatic contouring systems are increasingly used to address inter-observer variability and improve efficiency, artificial intelligence models require rigorous quality assurance and updates to reflect current technology. However, high-level evidence regarding the longitudinal real-world impact of implementing and iteratively updating these systems in clinical workflows is currently lacking. PURPOSE:This study aimed to evaluate the real-world clinical impact of implementing and updating a DL-based automatic contouring system in rectal cancer radiotherapy to generate high-quality evidence of iterative updates. METHODS:This longitudinal retrospective analysis included 150 patients divided into three cohorts: pre-implementation (n1 = 50), post-implementation (n2 = 50), and post-update (n3 = 50). Geometric similarities between unedited-automatic and final treatment contours were compared across cohorts. Failure rates were systematically analyzed. Six oncologists contoured 21 additional cases through manual, first-generation (Auto1), and second-generation (Auto2) system-assisted methods to evaluate contouring time, inter-observer consistency, and accuracy. Additionally, a 5-point Likert scale was used by two blinded senior oncologists to assess the clinical acceptability of the generated contours. RESULTS:The mean Dice similarity coefficient (DSC) values of clinical target volume (CTV) before and after implementing the automatic contouring system were 0.87 ± 0.04 and 0.88 ± 0.04 (P = 0.067), while those of OARs were 0.80 ± 0.06 and 0.88 ± 0.05 (P < 0.001), respectively. Following the system update, they improved from 0.88 ± 0.04 to 0.93 ± 0.04 for CTV (P < 0.001) and from 0.88 ± 0.05 to 0.95 ± 0.02 for OARs (P < 0.001). The system update achieved an approximately 80.6% reduction in the mean failure rate. Auto2-assisted method decreased the total time by approximately 58.8% compared with the manual method, and 21.9% compared with the Auto1-assisted method. This method also demonstrated optimal inter-observer consistency (0.95 ± 0.03) and accuracy (0.94 ± 0.03) for CTV. In the blinded clinical evaluation, 99.2% (125/126) of the oncologist-revised final contours received a Likert score of ≥ 4, and Auto2-generated unedited contours showed significantly higher clinical acceptability than Auto1 (4.02 ± 0.25 vs. 3.26 ± 0.49, P < 0.001) CONCLUSIONS: Implementing an automatic contouring system provided crucial guidance for clinical practice. Its iterative update significantly reduced workload and inter-observer variation while enhancing contouring efficiency and quality.
BACKGROUND:Pulmonary ventilation imaging has become an increasingly important component in thoracic radiotherapy, as it is applied to functional avoidance radiotherapy, where radiation doses are strategically minimized to preserve functional lung regions. PURPOSE:We developed a framework that generates computed tomography ventilation images (CTVI) from planning CT (PCT) images and demonstrated its estimation performance. METHODS:The subjects were a patient cohort consisting of 102 locally-advanced non-small-cell lung cancer (NSCLC) patients who received radiotherapy from 2014 to 2023. The PCT images were acquired while ensuring the absence of baseline drift, frequency variation, amplitude changes, and additive random observation noise using a real-time position management (RPM) system to minimize the introduction of inaccuracies into PCT images. On the day of PCT scan, CTVI based on four-dimensional computed tomography images ( CTVI 4 DCT DIR ) images was also generated using both deformable image registration and the computing methods employed in the VAMPIRE study. RESULTS:The CTVI based on PCT ( CTVI PCT DNN ) estimated by a model trained via hyperparameter optimization of a U-net deep neural network with 5-fold cross-validation was compared with CTVI 4 DCT DIR . In 5-fold cross-validation, the average voxel-wise Spearman's correlation coefficient (rs) ± one standard deviation between CTVI PCT DNN and CTVI 4 DCT DIR was 0.77 ± 0.08. The Dice similarity coefficient (DSC) was computed for three functional regions (high, moderate, and low), each delineated by approximately equal volumes, obtaining DSChigh, DSCmoderate, and DSClow values of 0.68 ± 0.06, 0.51 ± 0.08, and 0.75 ± 0.04, respectively. CONCLUSIONS:We successfully developed a framework that estimates CTVI PCT DNN , and demonstrated its estimation performance in terms of Spearman's correlation coefficient and DSC on locally-advanced NSCLC patients.
BACKGROUND:Frameless thermoplastic masks are standard for intracranial stereotactic radiosurgery (SRS), with the Encompass (CQ Medical) widely used on the Varian HyperArc platform. A candidate mask must minimize beam attenuation, permit accurate dose calculation, and enable reproducible setup, yet these are rarely assessed together. Comprehensive comparative data for the proposed Environ (MacroMedics) mask are limited. PURPOSE:To compare the Environ and Encompass masks across three domains: beam attenuation, dose-calculation accuracy under different algorithm and mask-contouring conditions, and interfraction setup reproducibility. METHODS:Attenuation was measured with an ionization chamber in a head phantom (TrueBeam, 6 FFF) at seven static gantry angles and a 0°-180° arc, with an estimated measurement uncertainty budget. Planning accuracy was assessed using HyperArc plans on phantom CT for 1-15 mm PTV margins (single-target) and 3/5/7 mm (multi-metastasis), calculated in Eclipse with AAA and Acuros XB (AXB), each with and without the mask in the body contour. Setup reproducibility was analyzed retrospectively from CBCT six-degree-of-freedom corrections (Encompass: 67 fractions/21 patients; Environ: 72 fractions/24 patients). Because repeated fractions are not independent, positioning data were analyzed at the patient level (Mann-Whitney U, Levene's test), by within-patient reproducibility, and with a confirmatory mixed-effects model. RESULTS:The Encompass attenuated 1.13% ± 0.59% more than the Environ for static beams (p = 0.002); the arc difference (0.06%) was within uncertainty (∼0.4%, k = 1). The Encompass required mask contouring with AXB for acceptable agreement (2.38% without), whereas the Environ achieved sub-1% AXB agreement without contouring (0.24%). The Environ showed a smaller mean vertical correction (mixed-effects p = 0.016) and reduced within-patient variability in vertical (0.31 vs. 1.25 mm, p < 0.001) and pitch (0.50 vs. 0.99°, p = 0.009); other axes were comparable. CONCLUSION:The Environ performed comparably or favorably across all three domains, supporting it as a clinically acceptable SRS alternative. It achieved acceptable AXB agreement without mask contouring under the conditions tested, though this depends on institution-specific settings and should be verified locally. The setup-reproducibility advantages, most robust vertically, are hypothesis-generating given the retrospective design and absence of intrafraction data.
BACKGROUND:Ultra-high-resolution (UHR) photon-counting CT (PCCT) with small-pixel detectors improves resolution and noise, but its dependence on diverse imaging conditions remains insufficiently characterized. PURPOSE:To systematically evaluate the small-pixel effect in UHR PCCT by assessing noise properties and high-contrast spatial resolution under varying acquisition and reconstruction conditions. METHODS:Images of an ACR CT accreditation phantom with a body ring simulating a medium-sized patient were acquired on a clinical PCCT system (NAEOTOM Alpha.Peak, Siemens Healthineers, Germany) using standard-resolution (STD) and UHR modes. Scans were performed at dose levels ranging from 3 to 24 mGy. Reconstructions included slice thicknesses from 0.2/0.4 to 3.0 mm, convolution kernels from Br36 to Br72, quantum iterative reconstruction both off and at strength levels 1-4, matrix sizes of 512 to 1024, and field-of-view (FOV) settings from 210 to 350 mm. Both low-energy threshold (T3D) and virtual monoenergetic images were evaluated. Image noise was quantified using noise magnitude, noise power spectrum (NPS), and average spatial frequency. High-contrast spatial resolution was evaluated using the modulation transfer function and visual assessment of resolution patterns. RESULTS:Compared with the STD mode, the UHR mode consistently reduced image noise, with the greatest reductions observed at thinner slice thicknesses and with sharper kernels. In contrast, dose level, QIR strength, matrix size, and spectral image type had minimal influence on the relative noise differences between the two modes. NPS analysis further demonstrated lower noise magnitude and, depending on the imaging conditions, a shift of the noise spectrum toward lower spatial frequencies, indicating relative suppression of high-frequency noise in the UHR mode. High-contrast spatial resolution was more strongly influenced by imaging sampling conditions, particularly matrix size and FOV, and remained overall comparable between the STD and UHR modes. CONCLUSIONS:The UHR mode of PCCT consistently achieves lower image noise than the STD mode without compromising high-contrast spatial resolution. Its greatest benefit is observed in high-spatial-resolution imaging protocols employing thin slice thicknesses and sharp reconstruction kernels, whereas the incremental benefit is limited for routine CT imaging. These findings support a task-based implementation of the UHR mode to optimize protocol selection according to clinical imaging requirements.
BACKGROUND:In computed tomography (CT), radiation dose is commonly evaluated using system-displayed metrics such as CTDIvol; however, variations in X-ray beam quality may influence organ absorbed dose even under identical CTDIvol conditions. PURPOSE:To investigate the impact of X-ray beam quality, including tube voltage and beam-shaping filters, on organ absorbed dose in chest computed tomography (CT) under identical CTDIvol conditions, with particular emphasis on spectral characteristics and energy-dependent dose conversion. METHODS:Half-value layer (HVL) measurements were performed using the aluminum attenuation method to determine effective energy under different tube voltage conditions (80, 100, 120, and 135 kV, and 120 kV with a silver [Ag] filter). X-ray spectra were measured using a CdTe-based spectrometer with a 90°Compton scattering configuration, and mean photon energy was calculated from the measured spectra. Absorbed doses in the breast and lung were evaluated using optically stimulated luminescence (OSL) dosimeters placed in an anthropomorphic phantom. Scan parameters were adjusted to maintain a constant CTDIvol (6.2 mGy) across all conditions. Absorbed dose was calculated from air kerma using f-factors derived from both effective energy and mean photon energy. RESULTS:Effective energy increased with tube voltage and was markedly elevated by the Ag filter. Spectral measurements demonstrated substantial reduction of low-energy photons and beam hardening under the Ag condition. Despite identical CTDIvol, both air kerma and absorbed dose varied with beam quality. Absorbed dose generally increased with effective energy in both tissues; however, under the 120 kV + Ag condition, the two tissues diverged: the lung absorbed dose decreased despite the higher effective energy, whereas the breast absorbed dose continued to increase. When the mean photon energy was used instead, this discrepancy was resolved differently: the absorbed dose decreased under the 120 kV + Ag condition in both the breast and the lung. This tissue-dependent sensitivity to the choice of energy metric was attributable to the energy dependence of the f-factor, which approached or exceeded unity in the breast under the Ag condition. CONCLUSIONS:Organ absorbed dose in chest CT varies with beam quality even under identical CTDIvol conditions, and this variation is tissue-dependent. Effective energy and mean photon energy can lead to different, and in the breast even opposite, conclusions regarding the direction of dose change under the same irradiation condition. Specifically, under standard (non-filtered) conditions, mean-photon-energy-based absorbed doses in the breast were approximately 4%-5% higher than effective-energy-based estimates, whereas under the 120 kV + Ag condition this relationship reversed, with mean-photon-energy-based doses approximately 1.6% lower. Although the magnitude of this reversal was relatively small, it was consistent and reproducible across repeated measurements. These findings highlight that the choice of energy metric used to convert air kerma to absorbed dose can meaningfully affect organ dose assessment in CT dosimetry, particularly when the f-factor approaches or exceeds unity.
BACKGROUND:In Stereotactic Central Ablative Radiotherapy (SCART), the volume of tumor receiving the full prescribed ablative dose decreases rapidly as the prescription dose increases, posing a critical challenge. To address this issue, this study investigates the dosimetric impact of combining various photon energies and flattening filter modes available on modern linear accelerators (linacs), aiming to provide more options for clinical application. METHODS:Fifty-one cases of abdominal large-volume tumors treated with conventional radiotherapy at the Second People's Hospital of Changzhou from January 2021 to August 2023 were retrospectively selected. Plans were created using the Monaco 5.11 treatment planning system. All plans employed volumetric modulated arc therapy (VMAT) with two coplanar 360° arcs (clockwise + counterclockwise). The two arcs used different combinations of photon energies/flattening filter modes: 6 MV & 6 MV, 6 MV & 10 MV, 6 MV & 6 MV flattening-filter-free (FFF), 10 MV & 10 MV, 6 MV-FFF & 6 MV-FFF, and 6 MV-FFF & 10 MV, while keeping all other planning parameters consistent. RESULTS:Compared to the standard 6 MV & 6 MV combination, the 6 MV & 10 MV combination reduced the dose to normal tissues without a statistically significant change in ablation volume. The 6 MV & 6 MV-FFF combination yielded significantly larger ablation volumes, albeit with a statistically significant increase in normal tissue dose, which remained within clinically acceptable limits. The 6 MV-FFF & 6 MV-FFF combination also yielded a significantly larger ablation volume, but the increase in normal tissue dose was not statistically significant. The 10 MV & 10 MV and 6 MV-FFF & 10 MV combinations reduced normal tissue dose while increasing ablation volume, though the increase was less than that achieved with the 6 MV & 6 MV-FFF and 6 MV-FFF & 6 MV-FFF combinations. CONCLUSION:The choice of energy/mode combination in SCART can be tailored to clinical priorities. For paramount organ-at-risk sparing, 6 MV & 10 MV is recommended. For maximizing ablation volume, 6 MV-FFF & 6 MV-FFF is recommended, followed by 6 MV & 6 MV-FFF. For a balanced approach, the 6 MV-FFF & 6 MV-FFF combination demonstrated the best overall performance, followed by the 6 MV-FFF & 10 MV combination.
BACKGROUND:Low-dose-rate prostate seed brachytherapy (BT) is an established guideline-conformal treatment option for low-risk prostate cancer. Precise seed placement is essential for achieving adequate dose coverage while minimizing exposure of surrounding healthy tissue. MRI guidance offers superior soft-tissue contrast for target visualization, but technical challenges remain for accurate transrectal seed implantation under in-bore conditions. PURPOSE:The Remote Control Manipulator (RCM) is an MRI-compatible robotic guidance system originally designed for in-bore transrectal targeted prostate biopsies. This work investigates its feasibility for MRI-guided transrectal targeted focal prostate seed BT. We demonstrate the intervention workflow and quantify seed placement accuracy in a preclinical phantom experiment. METHODS:The proposed intervention workflow defines the setup steps, interaction of the treatment planning system (TPS) with the robot's navigation software, MR-imaging as well as seed- and needle-guide registration. We performed and analyzed the total workflow in an anthropomorphic pelvis phantom. Seed positions corresponding to the initial treatment plan, the expected positions after needle-guide alignment, and the final released seed positions were assessed in detail. Their geometric displacement was decomposed into contributions from robotic alignment, seed release, and registration uncertainty. Dosimetric consequences were evaluated using the target's V100% dose coverage. RESULTS:Robotic alignment produced a mean expected seed displacement of 0.6 ± 0.2 mm. Total deviation from planned to released seed positions was 2.6 ± 0.9 mm, with 0.7 ± 0.4 mm along-trajectory displacement. MRI-only seed registration differed from CT-verified reconstruction by 2.8 ± 1.4 mm. After two released needles, V100% declined from 100% to 92.9% using MRI-only data and to 85.7% using CT-verified seed positions. CONCLUSIONS:The workflow was successfully executed and demonstrated sub-millimeter robotic targeting precision. Seed displacement during release and limited MRI-only seed localization remain dominant sources of error impacting dose coverage. Active needle guidance during imaging and improved MRI-only seed reconstruction could enable monitoring the live dose distribution and adaptive trajectory correction.
AIM:Current radiotherapy (RT) guidelines for Cardiac Implantable Electronic Devices (CIEDs) are largely based on clinical management strategies and retrospective reports. This systematic review adopts a medical physics perspective to provide a critical, quantitative analysis of the technical evidence required for developing robust safety action plans. MATERIALS AND METHODS:Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, a systematic search was conducted (January 2014-December 2024) across five major databases. Sixteen physicists from the Associazione Italiana di Fisica Medica e Sanitaria (AIFM) Working Group (WG) performed a double-review and standardized data extraction, focusing on physics-specific variables for the irradiation of Left Ventricular Assist Devices (LVADs), Pacemakers (PMs), and Implantable Cardioverter-Defibrillator (ICD) devices. Data were grouped into five topics: photon/electron RT, hadron therapy, modeling RT effects (Treatment Planning System [TPS]/Monte Carlo [MC]), imaging, and specific considerations regarding Magnetic Resonance Imaging Linac (MRI-Linac) systems. The Malfunction Rate (MR)-the percentage of devices with a clinically relevant malfunction-was calculated, and a Z-test was carried out to highlight safer irradiation conditions. Additionally, a narrative synthesis was performed. RESULTS:Of the 60 selected articles, 47 were co-authored by physicists. The studies considered a total of 1769 real devices, while 13 studies reported on virtual devices (i.e., simulations, electronics, or measurements only). Reported irradiation beams included photons (encompassing Intra-Operative RadioTherapy [IORT], Total Body Irradiation [TBI], and Flattening Filter Free [FFF] beams), electrons, protons, and carbon ions. Thirty-seven manuscripts investigated CIED malfunctions, yielding an overall MR of 12.6%. The MR was significantly lower for devices irradiated under scatter conditions and when non-neutron-producing beams were used. Information on the entire RT process (imaging, dose calculation, planning, and irradiation) was analyzed, and a flowchart was developed to guide medical physicists in their daily clinical practice. CONCLUSION:The review concludes that the medical physicist's role is non-negotiable in developing and implementing advanced planning and dosimetry strategies to prevent CIED malfunction and ensure patient safety.
A patient with castration-resistant prostate cancer (CRPC) was referred to our clinic seeking palliative management of a painful metastasis in the left pelvis. At the time of referral, the patient was receiving [177Lu]Lu-PSMA 617 (Pluvicto®) therapy at an outside institution, with the next scheduled administration approaching within days. As the patient's next Pluvicto® cycle could coincide with the planned SBRT, the radiation oncologist consulted with the medical health physics team to evaluate potential radiation safety concerns and to determine an appropriate treatment timeline. This case report details the radiation safety considerations, dose estimations, and the measured dose rate readings on the patient to balance the optimal patient care and to minimize staff exposure.
BACKGROUND:Low-cost Geiger-Müller detectors may support clinical radiation monitoring, but their response varies with radiation type and measurement conditions. PURPOSE:To evaluate an Arduino-based Geiger-Müller monitoring system in representative clinical radiation environments and determine whether condition-specific calibration reduces disagreement with a reference survey meter. METHODS:The system was compared with a reference energy-compensated Geiger-Müller survey meter (Ludlum Model 3005) under computed tomography scatter, Ir-192 brachytherapy, and 6 MV linear accelerator maze conditions. Calibration coefficients, expressed as counts per minute (CPM) per µSv/h, were estimated using zero-intercept regression, and method agreement was assessed using bias analysis and Bland-Altman plots. Portable deployment and user acceptability were also evaluated. The effective dead time of the counting chain was estimated using the first-order two-source approximation, and the intrinsic photon energy response was characterized using discrete-energy gamma-emitting sources (Co-57, Cs-137, and Co-60). Calibration-coefficient stability was also assessed at a fixed maze position across nominal dose rates of 100-600 MU/min to isolate count-rate effects. RESULTS:Calibration coefficients ranged from 124 to 133 for computed tomography scatter, 99 to 179 for Ir-192, and 145 to 233 for the 6 MV linear accelerator maze, demonstrating marked condition dependence. Applying condition-specific calibration reduced systematic bias within the evaluated environments. The effective dead time was approximately 258 µs, and the detector showed greater response per unit dose at lower photon energies, consistent with the known energy dependence of uncompensated Geiger-Müller tubes and providing a plausible explanation for the condition-dependent calibration coefficients. At a fixed maze position, the calibration coefficient varied little across nominal dose rates of 100-600 MU/min (coefficient of variation, 3.04%), with no evidence that count-rate dependence dominated within this range. In the linear accelerator maze, mean bias approached zero after distance-specific calibration, whereas the limits of agreement widened with increasing distance. Among 27 respondents, intention for continued use was more strongly associated with perceived stability and lower perceived risk than with numerical agreement alone. CONCLUSIONS:With condition-specific calibration, the low-cost monitoring system showed low mean bias relative to the reference instrument in the evaluated low-to-moderate dose-rate clinical environments. These findings support further evaluation of the system as a supplementary tool for routine radiation monitoring and safety checks under comparable clinical conditions.
BACKGROUND:Precise target definition in prostate radiotherapy requires accurate computed tomography (CT)-magnetic resonance imaging (MRI) image registration using gold fiducial markers (GFMs). While multi-echo gradient recalled echo (ME GRE) sequences facilitate GFM identification, the quantitative impact of acquisition parameters on physical measurement accuracy remains insufficiently understood. PURPOSE:This study evaluates how MRI slice thickness and marker orientation angle jointly affect GFM visualization and geometric accuracy to characterize the resulting localization uncertainties. METHODS:A kiwifruit-based phantom (T1: 1603.3 ms; T2: 72.3 ms) that mimics the relaxation times of the human prostate was imaged with 3.0T MRI. Imaging was performed using a 3D ME GRE sequence with three slice thicknesses (1.0, 2.0, and 3.0 mm) and 12 orientation angles ranging from horizontal to 90°. Five multidisciplinary observers independently identified marker coordinates while blinded to the acquisition parameters. Subjective confidence scores were also recorded. RESULTS:Two-way analysis of variance (ANOVA) revealed significant main effects for slice thickness (F(2, 144) = 32.03, p < 0.001) and orientation angle (F(11, 144) = 4.06, p < 0.001), with significant interaction between them (F (22, 144) = 2.72, p < 0.001). Mean measurement errors with 95% confidence intervals (CI) were 0.96 mm (95% CI: 0.70-1.22) for 1.0 mm, 1.49 mm (95% CI: 1.23-1.75) for 2.0 mm, and 2.00 mm (95% CI: 1.74-2.26) for 3.0 mm slices. The peak error reached 3.10 mm under the combination of a 3.0 mm slice and 15° orientation. Subjective confidence scores showed a significant negative correlation with physical measurement errors (r = -0.296, p < 0.001). CONCLUSIONS:MRI slice thickness and marker orientation significantly impact GFM visualization accuracy. The results suggest that a 1.0 mm slice thickness offers a technical advantage for maintaining sub-millimeter localization accuracy across various marker orientations and highly desirable for high-precision radiotherapy planning.
BACKGROUND:Medical physics residency programs require reliable systems for organizing rotation files, documenting educational activities, recording competency sign-offs, and monitoring resident progression. Current residency standards from the Commission on Accreditation of Medical Physics Educational Programs (CAMPEP) require clearly defined training schedules, rotation objectives, resident progress evaluation, and regular documentation of trainee development. Many programs continue to rely on static documents, spreadsheets, or multiple disconnected platforms, which may create version-control problems, increase administrative burden, and reduce transparency. Prior reports have described dedicated educational-management software (Typhon Group and MedHub) for this purpose, but these require separate licensing and onboarding. RadMachine (Radformation, Inc., New York, NY) is widely used in radiation oncology departments for machine quality assurance (QA) management; repurposing such an existing QA platform for residency management has not been previously described. PURPOSE:To describe the repurposing of RadMachine as a platform for medical physics residency rotation organization, activity tracking, competency sign-offs, and compliance with residency documentation needs. METHODS:A residency management framework was developed within the RadMachine platform. Each resident was configured as a virtual device. Test lists were created and assigned to each resident to represent clinical rotations, an operation sign-off list, a proficiency checklist, and a comprehensive examination. Individual tests represented, but were not limited to, learning objectives, practical tasks, reading topics, or sign-off requirements. Completion status was recorded in various ways, including binary completion states, supervisor sign-offs, detailed comments, or report upload. The prior Microsoft Word-based workflows were qualitatively compared with the proposed RadMachine-based framework. RESULTS:The RadMachine-based framework transformed static training documents into dynamic, trackable test lists. Compared with the prior workflow, the new framework improved version control, reduced administrative effort, simplified updates, allowed transparency, and ensured access to the most current curriculum. Residents could view completed and pending requirements in real time, while faculty could rapidly review trainee progress and outstanding competencies. CONCLUSIONS:Implementing the machine QA software RadMachine for residency education provides a practical, scalable, and low-overhead solution for rotation management and competency tracking. Because it reuses cloud-based QA software already licensed and deployed in the department for routine clinical QA-that is, existing departmental infrastructure-the approach requires no additional procurement, licensing, or information-technology onboarding. This approach is particularly useful for residency programs seeking efficient alternatives to document-based systems and offers strong potential for future expansion.
BACKGROUND:Clinical shift scheduling for medical physicists is challenging given multidisciplinary roles and competing clinical, research, and service demands. Manual workflows often lack the flexibility and transparency needed for this complex field. PURPOSE:To develop, deploy, and evaluate a preference-driven, optimization-based scheduling platform for an academic radiation oncology medical physics division. METHODS:Using a graphical user interface, each physicist encoded half-day shift slots as unavailable (0), available but not preferred (0.5), available (1), or preferred (2) for each two-month scheduling block over two years of clinical use. Shift requirements were defined across ten service categories with weighted effort units, and individual targets were adjusted for non-clinical effort (e.g., research, teaching, administrative service) and planned time off. Scheduling was formulated as a constrained binary integer optimization problem, with coverage, sequencing, and availability constraints, solved using a genetic algorithm to minimize a preference-based fitness function. Final schedules were exported to clinical calendars. Retrospective evaluation computed descriptive statistics per scheduling period and overall, including the distribution of assigned shifts by preference category and concordance between assignments and stated availability/preferences. Qualitative feedback from users and scheduling leads was collected to further assess operational impact. RESULTS:From January 2024 through February 2026, 4980 shifts were assigned across fifteen scheduling blocks for 23 physicists. Across 25,796 preference entries, 95% of assignments matched stated preference and availability, with non-preferred assignments accounting for only 5% of total assignments. All clinical coverage requirements were met, and no staff were scheduled for shifts marked as unavailable. Manual interventions were rare, and user feedback indicated improved transparency, satisfaction, and reduced administrative burden. CONCLUSIONS:Preference-integrated optimization reliably produced feasible, clinically robust schedules for medical physics shift scheduling. This approach is readily implementable and supports efficient, equitable scheduling in multidisciplinary settings.
