Zeitschrift fur medizinische Physik最新文献

{"title":"Value of the Acuros reconstruction CBCT algorithm for conventional (non-HyperSight) Ethos systems.","authors":"Erik B van Dieren, Anieck E Ter Braak, Esther Bosman, Sven Schwieters, Jeroen Veltman, Judith J Dasselaar","doi":"10.1016/j.zemedi.2026.03.002","DOIUrl":"https://doi.org/10.1016/j.zemedi.2026.03.002","url":null,"abstract":"<p><strong>Background and purpose: </strong>Since the introduction of CBCT in the field of radiotherapy, there has been an increasing interest to adapt to anatomy changes during treatment. Since 2019, CBCT based online adaptive radiotherapy is feasible with Varian's Ethos, but CBCTs have low contrast, in particular in the pelvis. This study examines the added value of the Acuros algorithm reconstruction, for phantom and clinical pelvis CBCTs, for non-HyperSight Ethos systems (NHES).</p><p><strong>Methods and materials: </strong>Data from two NHES systems were used. Firstly, bi-weekly CBCT scans were made using an anthropomorphic pelvis phantom, and CT contours were propagated. Using these, CBCTs, SNR, CNR, uniformity, and resolution were calculated. Secondly, CBCTs were analyzed for five prostate patients treated adaptively, using three sessions. The 2nd CBCT, just before beam on in the adaptive sessions, was reconstructed using Acuros. Besides analysis similar to the phantom study, seven observers were asked to score on various items.</p><p><strong>Results: </strong>For the phantom study, the Acuros reconstruction is consistently superior to iCBCT in terms of contrast to noise, signal to noise and uniformity, but spatial resolution is similar. For clinical CBCTs, Acuros results in better contrast to noise, uniformity, and resolution. In accordance, observers usually score Acuros images as better or similar (85-95%) for the adaptive task than iCBCT.</p><p><strong>Conclusions: </strong>In a pelvis phantom, Acuros outperforms standard reconstruction for most parameters, or it is equal. For clinical data, Acuros is usually better, and is generally preferred by observers. Hence, use of Acuros for NHES is recommended.</p>","PeriodicalId":101315,"journal":{"name":"Zeitschrift fur medizinische Physik","volume":" ","pages":""},"PeriodicalIF":4.2,"publicationDate":"2026-03-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147481486","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Multi-angle beam range measurement framework for carbon-ion radiotherapy using a commercial multi-layer ionization chamber.","authors":"Soorim Han, YongCheol Kwon, Taegeon Oh, Eunho Lee, Eun Kyu Kim, Tae Ho Kim, Dohyeon Yoo, Sangmin Lee, Changhwan Kim, Min Cheol Han, Chae-Seon Hong, Jin Sung Kim","doi":"10.1016/j.zemedi.2026.03.007","DOIUrl":"https://doi.org/10.1016/j.zemedi.2026.03.007","url":null,"abstract":"<p><p>Accurate beam range verification is critical in carbon-ion radiotherapy, where the sharp Bragg peak and complex material-dependent nuclear interactions pose greater measurement challenges compared to proton therapy; however, conventional water phantom methods for frequent and comprehensive checks required for routine quality assurance (QA) are extremely time-consuming. This study proposes and validates a framework using a commercial multi-layer ionization chamber (MLIC) to enable efficient, accurate, and comprehensive range verification. The framework integrates three key innovations: (1) a beam-model-based fitting method that uses reference integral depth-dose curves to reduce range determination uncertainty for carbon ions, (2) a time-optimized, trigger-free acquisition process, and (3) a custom gantry-compatible mount for stable multi-angle measurements. The framework's measurement uncertainty was evaluated via range-shift experiments, and its clinical feasibility was tested through year-long stability and multi-angle consistency measurements. The proposed method demonstrated a low range-determination uncertainty, with an expanded uncertainty of 0.24 mm (k = 3), and showed excellent agreement with standard water phantom measurements (mean deviation of 0.03 ± 0.15 mm). The framework reduced the measurement time for all 600 energy levels from >50 h to approximately 15 min, representing >100-fold improvement. Clinical validation confirmed high stability, with long-term and multi-angle deviations of 0.17 mm and 0.05 mm (k = 1), respectively. The system also successfully identified machine-related beam range inconsistencies that were not detectable using conventional QA protocols. The proposed MLIC-based framework provides a reliable, accurate, and highly efficient solution for routine beam range QA in carbon-ion therapy. Its successful clinical application demonstrates its feasibility for frequent and comprehensive verification across all energy levels and gantry angles.</p>","PeriodicalId":101315,"journal":{"name":"Zeitschrift fur medizinische Physik","volume":" ","pages":""},"PeriodicalIF":4.2,"publicationDate":"2026-03-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147439045","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Correction factors and uncertainties in in-vivo point dose measurements.","authors":"Julian Niedling, Ndimofor Chofor, Frank Hensley, Otto A Sauer, Sonja Wegener","doi":"10.1016/j.zemedi.2026.03.001","DOIUrl":"https://doi.org/10.1016/j.zemedi.2026.03.001","url":null,"abstract":"<p><p>In-vivo dosimetry is associated with high measurement uncertainties, partly due to deviations in detector response under clinical compared to calibration conditions, especially for IMRT/VMAT, stereotactic radiosurgery, and total body irradiation. This study proposes a procedure for detector calibration and correction, evaluates its feasibility, quantifies uncertainties, and provides guidance for determining correction factors. Correction factors for in-vivo ionization chambers (ICs), Metal-Oxide-Semiconductor Field-Effect Transistors (MOSFETs), and Optically Stimulated Luminescence Detectors (OSLDs) were obtained. Correction factors for field size, dose rate, angular, temperature dependence and response variation were derived experimentally. Correction factors were applied to measurements of irradiation plans simulating clinical treatments, including static fields, stereotactic, and total body irradiation using an anthropomorphic phantom. Measured data with and without applied correction factors were compared to doses calculated by treatment planning systems (TPS). Application of correction factors improved agreement between calculated and measured doses by up to 7.5% for MOSFETs, 3.1% for ICs, and 0.9% for OSLDs. Deviations between detectors were reduced by up to 0.6%. Maximum relative dose errors without correction factors compared to the uncertainties of the correction factors highlighted the benefit of corrections, particularly for MOSFETs (11.5% error vs. 4.4% uncertainty) and ICs (8.6% vs. 0.7%). For OSLDs, maximum uncertainties (6.5%) were slightly higher than the observed dose errors (4.5%). The results indicate, despite the uncertainties, both the necessity and benefit of applying the described correction factors.</p>","PeriodicalId":101315,"journal":{"name":"Zeitschrift fur medizinische Physik","volume":" ","pages":""},"PeriodicalIF":4.2,"publicationDate":"2026-03-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147439025","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Local diagnostic reference levels for routine chest X-ray examinations: first pilot study in Kazakhstan.","authors":"Moldir Aumalikova, Botagoz Nurseitova, Yerlan Kashkinbayev, Meirat Bakhtin, Moldir Aisauyt, Danara Ibrayeva, Nursultan Kusherbayev, Kuralay Ilbekova","doi":"10.1016/j.zemedi.2026.02.003","DOIUrl":"https://doi.org/10.1016/j.zemedi.2026.02.003","url":null,"abstract":"<p><p>X-ray imaging is one of the most frequently used diagnostic tools in Kazakhstan and globally. Despite its clinical benefits, it involves exposure to ionizing radiation, requiring dose optimization. The International Atomic Energy Agency (IAEA) recommends establishing Diagnostic Reference Levels (DRLs) to enhance radiation protection in diagnostic radiology. This study aimed to propose local DRLs for posteroanterior (PA) chest radiography in adult patients in selected medical centers in Astana, Kazakhstan. A secondary objective was to assess the quality control of X-ray systems used in routine clinical practice. The study was performed across three diagnostic centers, each using different radiography systems: Siemens Axiom Aristos MX (2009), Siemens Opti 150/30/50HC-100 (2007), and Floormounted DR Series ANGELL TECHNOLOGY DFM818 (2024). Standard chest radiographs were taken from 60 adult patients (≥18 years, 70 ± 20 kg; 20 per system). Entrance Surface Air Kerma (ESAK) and incident air kerma (Ki) were measured using calibrated RTI X-ray multimeters. Exposure parameters were recorded, and descriptive statistics and comparative analysis (Stata 16) were used to analyze the data. ESAK and Ki values showed significant differences among systems (p < 0.05), with older units yielding higher doses. These findings support the proposal of preliminary local DRLs for adult chest radiography in Astana. This study provides the first local DRLs for Kazakhstan and highlights the need for nationwide DRL implementation, aligned with IAEA safety standards (GSR Part 3, SSG-46), to improve radiation protection and diagnostic quality.</p>","PeriodicalId":101315,"journal":{"name":"Zeitschrift fur medizinische Physik","volume":" ","pages":""},"PeriodicalIF":4.2,"publicationDate":"2026-03-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147373649","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Initial validation and commissioning experience of Phoenix Plan treatment planning system for scanned carbon ion beams delivered by Heavy Ion Medical Machine.","authors":"Yunzhou Xia, Youfang Lai, Kai Wang, Zilu Wang, Zhiqiang Liu, Cheng Chang, Jian Shi","doi":"10.1016/j.zemedi.2026.02.006","DOIUrl":"https://doi.org/10.1016/j.zemedi.2026.02.006","url":null,"abstract":"<p><strong>Purpose: </strong>We present the commissioning and initial validation of Phoenix Plan, a GPU-accelerated treatment planning system (TPS) developed for raster-scanned carbon ion therapy delivered by the Heavy Ion Medical Machine (HIMM).</p><p><strong>Method: </strong>Phoenix Plan employs a quintuple-Gaussian pencil beam model for physical dose calculation and incorporates a subspot-splitting algorithm to account for lateral heterogeneities. RBE-weighted dose is computed using a linear-quadratic model based on the mixed-beam approach. We implemented a physically-informed interpolation method for integrated depth dose (IDD) curves, allowing the dose engine to be commissioned from a sparse subset of measured energies. GPU acceleration is achieved through a precomputed water-equivalent depth texture-based framework, enabling rapid and scalable dose calculations. The system was validated against a range of physical and biological benchmarks.</p><p><strong>Results: </strong>Comparisons between calculated and measured doses for spread-out Bragg peak (SOBP) plans showed agreement within 2.4 % in the target region and within 1 mm at distal fall-off. RBE-weighted doses at the SOBP centers were within 2 % of values reported in the literature, using a clinical factor of 1.38. Ten retrospective patient plans covering diverse anatomical sites achieved gamma passing rates above 95 % under both 3%/3 mm and 2%/3 mm criterion. The interpolation strategy reduced IDD measurement time from 41 h to 4.6 h without compromising accuracy. The beam modeling process was completed under five minutes.</p><p><strong>Conclusion: </strong>These results demonstrate that Phoenix Plan achieves accurate and efficient physical and biological dose calculation, supports streamlined commissioning workflows, and is ready for clinical deployment in conjunction with HIMM.</p>","PeriodicalId":101315,"journal":{"name":"Zeitschrift fur medizinische Physik","volume":" ","pages":""},"PeriodicalIF":4.2,"publicationDate":"2026-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146777029","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Simulation of B₀ magnetic field conditions in the human heart for improved diagnostic MRI.","authors":"Yun Shang, Sebastian Theilenberg, Michelle Castillo, Boyu Peng, Maggie Fung, Patrick Quarterman, Benjamin Navot, Sachin R Jambawalikar, Andrew J Einstein, Christoph Juchem","doi":"10.1016/j.zemedi.2026.02.005","DOIUrl":"https://doi.org/10.1016/j.zemedi.2026.02.005","url":null,"abstract":"&lt;p&gt;&lt;strong&gt;Background: &lt;/strong&gt;Cardiac balanced steady-state free precession (bSSFP) MRI suffers from dark band artifacts due to severe B&lt;sub&gt;0&lt;/sub&gt; inhomogeneity induced by air-tissue interfaces in the human body. Those artifacts remain uncontrollable despite scanner-provided B&lt;sub&gt;0&lt;/sub&gt; shimming based on low-order spherical harmonic shapes. To mitigate those artifacts for higher image fidelity, the development of advanced B&lt;sub&gt;0&lt;/sub&gt; shimming techniques is demanding and typically requires knowledge of cardiac B&lt;sub&gt;0&lt;/sub&gt; conditions across patient populations. We recently proposed a new simulation approach to derive cardiac B&lt;sub&gt;0&lt;/sub&gt; distributions from readily available chest-abdomen-pelvis CT structural images based on first principles of classical electrodynamics. This approach is able to calculate the air-tissue interface-induced magnetic field variation across the body from diverse subjects using regular CT images. While the computational method has been step-by-step validated in previous work, a comprehensive demonstration in patients has been missing.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Purpose: &lt;/strong&gt;The aim of this study is to validate our proposed structural CT-derived cardiac B&lt;sub&gt;0&lt;/sub&gt; magnetic field computation approach using a side-by-side comparison between simulation and in vivo measurement obtained from the same cohort of clinical subjects.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Methods: &lt;/strong&gt;Six subjects (female: 3, male: 3, age: 54.5 ± 17.2 years) with aortic stenosis or cancer who underwent clinically-indicated CT scans of the chest, abdomen, and pelvis were consented to undergo additional in vivo cardiac B&lt;sub&gt;0&lt;/sub&gt; measurement on a 3T MRI scanner. B&lt;sub&gt;0&lt;/sub&gt; distributions in the heart were computed based on CT images using our recently proposed approach and in vivo B&lt;sub&gt;0&lt;/sub&gt; maps acquired experimentally. CT images and MRI-based anatomical images were co-registered before field calculation, followed by the side-by-side field comparison after computation using correlation analysis of inhomogeneous field values, as well as the comparison of first-order spherical harmonic coefficients after decomposing field maps and the residual B&lt;sub&gt;0&lt;/sub&gt; inhomogeneities after shim from first to fifth order.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Results: &lt;/strong&gt;Simulated B&lt;sub&gt;0&lt;/sub&gt; maps show excellent agreement with in vivo measured B&lt;sub&gt;0&lt;/sub&gt; maps and strong average correlation (r = 0.92). After first and second-order shim analyses, both field maps show highly similar residual local B&lt;sub&gt;0&lt;/sub&gt; field patterns, especially near cardiophrenic angles and in the heart's inferior region, demonstrating the authenticity of localized features beyond overall field congruency. Both groups also demonstrate the trend of decreasing residual B&lt;sub&gt;0&lt;/sub&gt; inhomogeneity after shimming from first to fifth order, and their absolute differences in the mean of all subjects range from 0.25 to 2.3 Hz across different SH orders.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Conclusions: &lt;/strong&gt;The consistency between sim","PeriodicalId":101315,"journal":{"name":"Zeitschrift fur medizinische Physik","volume":" ","pages":""},"PeriodicalIF":4.2,"publicationDate":"2026-02-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146260550","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"MRI simulation and reconstruction framework for arbitrarily-oriented encoding and transmit/receive magnetic vector fields.","authors":"Fabian Bschorr, Thomas Hüfken, Tobias Lobmeyer, Volker Rasche","doi":"10.1016/j.zemedi.2026.02.002","DOIUrl":"https://doi.org/10.1016/j.zemedi.2026.02.002","url":null,"abstract":"<p><p>Conventional MRI is relying on the assumption of the magnetic field B₀ being homogeneous in direction and amplitude. However, with the growing interest in portable, affordable point-of-care MRI systems, these assumptions do not necessarily hold anymore due to compromises necessary to achieve a reduction in e.g. footprint, weight and portability. Simulation software may help by interpreting artifacts from concomitant fields or evaluation of new encoding schemes which are optimized for non-ideal hardware but also with the design of new scanner designs. In this work, a MATLAB-based simulation framework was developed capable of dealing with the full magnetic field vectors for B₀, the spatial magnetic encoding fields (SEMs), and the transmit/receive fields enabling the evaluation of arbitrary magnetic field configurations for encoding in MRI. For that, conventional matrix-based Bloch simulation is generalized, evaluated and validated, achieving substantially shorter simulation time. Furthermore, it is used to predict image quality in 2D gradient echo experiments with deflected magnetic fields. The comparison of numerical Bloch and the adapted matrix-based simulation revealed close agreement of both reconstructed images. Further, it was shown that compensation of the associated artifacts can be achieved by incorporating knowledge about the used magnetic fields into the reconstruction process. The presented and validated software package enables full consideration of arbitrarily-oriented SEMs, B₀ and transmit/receive magnetic fields used in MRI for signal simulation but also reconstruction removing the related artifacts. As such, the software might become a valuable tool for new low-field system designs and the investigation of new reconstruction algorithms.</p>","PeriodicalId":101315,"journal":{"name":"Zeitschrift fur medizinische Physik","volume":" ","pages":""},"PeriodicalIF":4.2,"publicationDate":"2026-02-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146204570","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Dynamic 3D MRI for vocal fold oscillation measurements.","authors":"Johannes Fischer, Paula L Jordan, Fiona Stritt, Bernhard Richter, Louisa Traser, Michael Bock","doi":"10.1016/j.zemedi.2026.01.005","DOIUrl":"https://doi.org/10.1016/j.zemedi.2026.01.005","url":null,"abstract":"<p><p>The source of the human voice lies in the rapid oscillations of the vocal folds (VF) inside the larynx. Currently, VF diseases are diagnosed with an endoscope providing a superficial view of the VFs but no further 3D information of the underlying VF structures. With magnetic resonance imaging (MRI) the full 3D spatial information could be acquired, but MRI so far lacks the necessary temporal resolution. We propose a new MRI method based on zero echo time (ZTE)-MRI to encode the rapid periodic VF oscillation. Combined with projection navigators to compensate for laryngeal motion and simultaneous voice recordings we are able to reconstruct the VF oscillation with a temporal resolution of 819μs and a spatial resolution of 0.77 mm. Our technique allows the first dynamic and 3D visualization of the VF oscillations and enables quantification of contact area and vertical displacement throughout the oscillation cycle. From the dynamic data, laryngeal tissue geometries can be extracted, such as the vertical displacement (1.2 mm) or the change in contact area (25.9 mm²) over the oscillation cycle. This non-invasive technique offers a new diagnostic tool for clinicians and for voice production studies.</p>","PeriodicalId":101315,"journal":{"name":"Zeitschrift fur medizinische Physik","volume":" ","pages":""},"PeriodicalIF":4.2,"publicationDate":"2026-01-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146088694","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"End-to-end evaluation of CyberKnife tracking accuracy with RayStation treatment planning system.","authors":"Nicolas Perichon, Marie Fargier-Voiron, Mireille Conrad, Shelley Bulling, Oscar Matzinger, Maud Jaccard","doi":"10.1016/j.zemedi.2026.01.003","DOIUrl":"https://doi.org/10.1016/j.zemedi.2026.01.003","url":null,"abstract":"<p><strong>Purpose/objective: </strong>The CyberKnife system uses different collimators to deliver stereotactic treatments. For accurate delivery, the dosimetric plan and the tracking algorithm both need to be handed correctly in the TPS. The established TPS for CyberKnife treatment planning is Precision from Accuray. Since June 2022, it has also been possible to plan CyberKnife treatments using RayStation from RaySearch Laboratories. The purpose of this work is to assess the tracking accuracy of RayStation TPS plan for CyberKnife by evaluating end-to-end tests (E2E) comparing with Precision. In a quality assurance context, this study that has never been carried out until now is very important because it allows to verify the correct implementation of targeting when using the treatment chain including the RayStation TPS and the CyberKnife treatment machine.</p><p><strong>Material/methods: </strong>E2E were performed for plans prepared in RayStation v12A (RS12A) for different collimator and tracking mode combinations. The results were compared to those obtained using Accuray Precision v3.3.1.0. E2E consisted of a concentric dose distribution around a ball target. Treatment was delivered after positioning with the target locating system for the appropriate tracking mode. Tests were performed at the same time for the plans from both TPS. The deviation on each axis and the radial error between the center of the ball target and the dose distribution were reported.</p><p><strong>Results: </strong>For all evaluated combinations of collimators and tracking modes, the RS12A deviation results were under international recommendations threshold (< 0.95 mm). In addition, RS12A deviation results for static targets (used for machine geometrical adjustments) were in agreement with the E2E Precision gold standard results, considering the 0.30mm method uncertainty.</p><p><strong>Conclusion: </strong>This study shows that E2E plans for CyberKnife can be generated in RS12A. The validation of the E2E results by comparison with the genuine treatment chain validates the targeting of the RayStation/CyberKnife treatment chain and also allows to use this method for periodic targeting checks and numerical robot position fine tune adjustment.</p>","PeriodicalId":101315,"journal":{"name":"Zeitschrift fur medizinische Physik","volume":" ","pages":""},"PeriodicalIF":4.2,"publicationDate":"2026-01-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146095457","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Qualifying optical biometry data before cataract surgery using autoencoders.","authors":"Achim Langenbucher, Peter Hoffmann, Alan Cayless, Nóra Szentmáry, Kamran Riaz, Damien Gatinel, Oliver Findl, Siegfried Priglinger, Seth Pantanelli, Tun-Kuan Yeo, Giacomo Savini, Jascha Wendelstein","doi":"10.1016/j.zemedi.2026.01.004","DOIUrl":"https://doi.org/10.1016/j.zemedi.2026.01.004","url":null,"abstract":"<p><strong>Purpose: </strong>To develop, implement and demonstrate a data driven strategy to identify 'suspects' and 'outliers' in datasets of biometric measurements taken before cataract surgery.</p><p><strong>Methods: </strong>A shallow autoencoder with one hidden layer and 3 neurons was trained on a large multicentre dataset (N = 152,397, 9 clinical centres) from the IOLMaster 700 including axial length (AL), central corneal thickness (CCT), anterior chamber depth (ACD), lens thickness (LT) and corneal front surface radius (R). Measurements were identified as 'suspects' or 'outliers' depending on the mean squared prediction error. Training was performed on 7 of the 9 datasets, and the remaining 2 datasets were used to evaluate the performance of the autoencoder.</p><p><strong>Results: </strong>Crossvalidation proved that the autoencoder with 3 neurons showed no noticeable overfitting. Measurements were marked as 'suspects' or 'outliers' based on the 95% and 99% quantiles as derived from the mean squared error with the training data. After piecewise linear correction of the autoencoder prediction error to correct for trend errors probably due to internal calibrations of the biometer, the autoencoder successfully identified potentially faulty measurements for the 2 test datasets.</p><p><strong>Conclusions: </strong>Autoencoders are quite popular for different application fields in engineering, and could also help in ophthalmology to identify potentially faulty biometric measurements which could lead to refractive surprises if used for intraocular lens power calculation. Further studies are required to validate this concept with other parameters or datasets from other biometers.</p>","PeriodicalId":101315,"journal":{"name":"Zeitschrift fur medizinische Physik","volume":" ","pages":""},"PeriodicalIF":4.2,"publicationDate":"2026-01-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146088756","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}