Journal of Applied Clinical Medical Physics最新文献

{"title":"Characterization, commissioning, and clinical evaluation of a commercial BeO optically stimulated luminescence (OSL) system","authors":"Joseph P. Kowalski,&nbsp;Brett G. Erickson,&nbsp;Qiuwen Wu,&nbsp;Xinyi Li,&nbsp;Sua Yoo","doi":"10.1002/acm2.70057","DOIUrl":"10.1002/acm2.70057","url":null,"abstract":"<p>This article investigates the performance of a commercial BeO optically stimulated luminescent (OSL) dosimetry system (myOSLchip, RadPro GmbH International, Remscheid, Germany) through the application of the commissioning framework for luminescent dosimeters as described in the American Association of Physicists in Medicine Task Group 191 (AAPM TG191) report. Initial clinical experiences and dosimetric results are also presented. The following properties of the system were characterized: linearity correction factors ranged from -0.5% to +3% for dose levels spanning 0.1 to 20 Gy. Beam quality correction factors (relative to 6 MV) ranged from -4.5% (2.5FFF) to +4.5% (15MV) for photon beams and +1.9% (6 MeV) to +4.3% (20 MeV) for electron beams. An average (µ) signal loss per reading of -2.13% ± 0.20% was measured, however greater signal loss was observed in the first reading (µ = -2.6% ± 0.46%). An initial decline in individual element sensitivity relative to baseline was observed from 0–15 Gy cumulative dose (µ = -1.98% ± 0.55%), with negligible further deterioration from 15–32 Gy (µ = -2.38% ± 0.85%). Post-irradiation, there was a transient OSL signal which faded with a half-life of 1.8 min; this signal enhancement was +5% at 5 min post-irradiation and +1% at 15 min relative to 24 h. Dosimeter response was not dependent on average dose rate in the range of 100–2500 MU/min. With respect to clinical testing, equal or superior performance compared with aluminum oxide OSLs (nanoDots) is shown for a range of clinical techniques and modalities including TSET, TBI, en-face electrons, and pacemaker/out-of-field measurements. The feasibility of myOSLchip to serve as a primary clinical in vivo dosimetry system and direct replacement for Landauer's microStar system is demonstrated.</p>","PeriodicalId":14989,"journal":{"name":"Journal of Applied Clinical Medical Physics","volume":"26 4","pages":""},"PeriodicalIF":2.0,"publicationDate":"2025-02-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/acm2.70057","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143476612","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Quantitative beam optimization for radiotherapy of peripheral lung lesions: A pilot study in stereotactic body radiotherapy","authors":"Hamed Hooshangnejad,&nbsp;Jina Lee,&nbsp;Leslie Bell,&nbsp;Russell K. Hales,&nbsp;Khinh Ranh Voong,&nbsp;Sarah Han-Oh,&nbsp;Kai Ding,&nbsp;Reza Farjam","doi":"10.1002/acm2.70029","DOIUrl":"10.1002/acm2.70029","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Background</h3>\u0000 \u0000 <p>To quantify beam optimization for stereotactic body radiotherapy (SBRT) of peripheral lung lesions.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Method</h3>\u0000 \u0000 <p>The new beam optimization approach was based on maximizing the therapeutic gain (TG) of the beam set by minimizing the average physical depth of the lesion with respect to the beam's eye view (BEV). The new approach was evaluated by replanning the 25 SBRT lesions retrospectively to assess if a better plan is achievable in all aspects. Difference in 25 Gy isodose line volume (IDLV₂₅ _Gy), IDLV₂₀ _Gy, IDLV₁₅ _Gy, IDLV₁₀ _Gy, and IDLV₅ _Gy between the two plan cohorts were calculated as a measure of plan size and fitted in a linear regression model against the changes in the lesion depth with respect to the BEV to assess the relationship between the changes in the treatment depth and that of the plan size.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>Beam optimization achieved a better plan in all cases by lowering the depth of treatment with an average of % 20.03 ± 12.30 (3.66%–45.78%). As the depth of treatment decreases, the size of the plan also decreases. We observed a reduction of % 4.64 ± 4.55 (0.02%–21.58%, <i>p</i> &lt; 3.8 × 10⁻⁵), %5.16 ± 5.54 (0.03%-24.68%, <i>p</i> &lt; 0.005), %6.46 ± 6.95 (−1.35%-29.05%, <i>p</i> &lt; 0.009), %12.83 ± 9.06 (0.89%–37.65%, <i>p</i> &lt; 0.0001), and %14.01 ± 9.87 (1.43%–41.84%, <i>p</i> &lt; 4.5 × 10⁻⁶) in IDLV₂₅ _Gy, IDLV₂₀ _Gy, IDLV₁₅ _Gy, IDLV₁₀ _Gy, and IDLV₅ _Gy, respectively.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusion</h3>\u0000 \u0000 <p>Physical depth of the lesion with respect to the BEV is inversely proportional to the TG of a beam-set and can be used as a robust and standard metric to select an appropriate beam-set for SBRT of the peripheral lung lesions. Further evaluation warrants the utility of such concept in routine clinical use.</p>\u0000 </section>\u0000 </div>","PeriodicalId":14989,"journal":{"name":"Journal of Applied Clinical Medical Physics","volume":"26 4","pages":""},"PeriodicalIF":2.0,"publicationDate":"2025-02-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/acm2.70029","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143476630","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A knowledge-based planning model to identify fraction-reduction opportunities in brain stereotactic radiotherapy","authors":"Shane McCarthy,&nbsp;William St. Clair,&nbsp;Damodar Pokhrel","doi":"10.1002/acm2.70055","DOIUrl":"10.1002/acm2.70055","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Objective</h3>\u0000 \u0000 <p>To develop and validate a HyperArc-based RapidPlan (HARP) model for three-fraction brain stereotactic radiotherapy (SRT) plans to then use to replan previously treated five-fraction SRT plans. Demonstrating the possibility of reducing the number of fractions while achieving acceptable organs-at-risk (OAR) doses with improved target biological effective dose (BED) to brain lesions.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Methods</h3>\u0000 \u0000 <p>Thirty-nine high-quality clinical three-fraction HyperArc brain SRT plans (24–27 Gy) were used to train the HARP model, with a separate 10 plans used to validate its effectiveness. Fifty-eight five-fraction HyperArc brain SRT plans (30–40 Gy) attempted to be retrospectively replanned for three fractions scheme using the HARP model. All planning was done within the Eclipse treatment planning system for a TrueBeam LINAC with a 6 MV-FFF beam and Millenium 120 MLCs and dosimetric parameters were analyzed per brain SRT protocol.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Results</h3>\u0000 \u0000 <p>The HyperArc RapidPlan model was successfully trained and tested, with the validation set demonstrating a statistically significant (<i>p</i> = 0.01) increase in GTV D_100% from 28.5 ± 0.7 Gy to 29.4 ± 0.6 Gy from the original to RapidPlan plans. No statistically significant differences were found for the OAR metrics (<i>p</i> &gt; 0.05). The five-fraction replans were successful for 20 of the 58 five-fraction brain SRT plans. For those 20 successful brain SRT plans, the maximum doses to OAR were clinically acceptable with a three-fraction scheme including an average V_18Gy to Brain-PTV of 9.9 ± 5.9 cc. Additionally, the replanned five-fraction brain SRT plans achieved a higher BED to the tumors, increasing from a GTV D_100% of 52.9 ± 4.5 Gy for the original five-fraction plans to 57.3 ± 3.1 Gy for the three-fraction RapidPlan plans. All RapidPlan plans were generated automatically, without manual input, in under 20 min.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusions</h3>\u0000 \u0000 <p>The HARP model developed in this research was used to successfully identify select five-fraction plans that were able to be reduced to three-fraction SRT treatments while achieving clinically acceptable OAR doses and improved target BED. This tool encourages a fast and standardized way to provide physicians with more options when choosing the necessary fractionation scheme(s) for HyperArc SRT to single- and multiple brain lesions.</p>\u0000 </section>\u0000 </div>","PeriodicalId":14989,"journal":{"name":"Journal of Applied Clinical Medical Physics","volume":"26 4","pages":""},"PeriodicalIF":2.0,"publicationDate":"2025-02-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/acm2.70055","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143476609","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"The effect of kV imaging dose on PTV and OAR planning constraints in lung SBRT using stereoscopic/monoscopic real-time tumor-monitoring system.","authors":"Ruwan Abeywardhana, Mike Sattarivand","doi":"10.1002/acm2.70019","DOIUrl":"https://doi.org/10.1002/acm2.70019","url":null,"abstract":"<p><strong>Purpose: </strong>Quantify the impact of additional imaging doses on clinical dose constraints during lung stereotactic body radiotherapy (SBRT) treatment utilizing stereoscopic/monoscopic real-time tumor monitoring.</p><p><strong>Materials and methods: </strong>Thirty lung SBRT patients treated with the volumetric arc therapy technique were randomly selected from the institutional clinical database. Contours of patients' and computed tomography data were extracted from the Eclipse treatment planning system, along with information regarding the treatment dose. Subsequently, patient-specific three-dimensional real-time imaging dose distributions were computed using a validated Monte Carlo simulation of the ExacTrac imaging. The 3D imaging dose was added to the treatment dose, and the influence of the imaging dose on clinical dose constraints was analyzed for planning target volume (PTV) and various organs at risk (OARs).</p><p><strong>Results: </strong>Among the 30 patients, 14 patients exhibited one or more failed OAR constraints based solely on the treatment dose, resulting in a total of 24 constraint failures. The addition of the real-time imaging dose altered the pass/fail criteria for one OAR constraint and two PTV constraints. The change in constraint due to additional imaging dose relative to the prescription dose was less than 1% for all patients, except for one case, where it reached 1.9%, which had remained below the threshold of 5% recommended by AAPM TG-180 guidelines. Furthermore, the additional imaging dose relative to the treatment dose resulted in an increase in OAR constraints ranging from 0 to 27% (mean of 0.8%), with nine cases exceeding 5%.</p><p><strong>Conclusion: </strong>The current study represents the first attempt to investigate the impact of additional imaging doses on clinical planning constraints in real-time tumor monitoring during lung SBRT utilizing ExacTrac imaging system. The addition of an imaging dose will likely have minimal clinical impact.</p>","PeriodicalId":14989,"journal":{"name":"Journal of Applied Clinical Medical Physics","volume":" ","pages":"e70019"},"PeriodicalIF":2.0,"publicationDate":"2025-02-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143476647","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"The implications of the American Board of Radiology's decision to relinquish its specialty board designation on prospective authorized medical physicists (AMPs) and radiation safety officers (RSOs)","authors":"Christopher J. Tien,&nbsp;Samantha J. Simiele,&nbsp;Joann I. Prisciandaro,&nbsp;Jacqueline E. Zoberi,&nbsp;Y. Jessica Huang,&nbsp;William A. Hinchcliffe,&nbsp;Hania A. Al-Hallaq","doi":"10.1002/acm2.70001","DOIUrl":"10.1002/acm2.70001","url":null,"abstract":"<p>In order to independently supervise the medical use of byproduct material, physicists in the United States (US) must legally meet the qualifications defined by the Nuclear Regulatory Commission (NRC) in the 35th part of the tenth title of the Code of Federal Regulations (§ 10 CFR Part 35). The American Board of Radiology (ABR) relinquished its NRC-recognized specialty board (NSB) status at the end of 2023, which eliminated the NSB application pathway for those who earn ABR certification in 2024 and beyond. While these changes in NSB status are not retroactive and will not affect eligibility for diplomates who already possess certificates, these changes will nonetheless have repercussions for those individuals who regularly provide training and experience (T&amp;E) attestations to the NRC, such as residency program directors, brachytherapy rotation preceptors, or radiation safety officers. This article will focus on the repercussions for new authorized medical physicist and radiation safety officer applicants with ABR certificates to be conferred in 2024 and later.</p>","PeriodicalId":14989,"journal":{"name":"Journal of Applied Clinical Medical Physics","volume":"26 4","pages":""},"PeriodicalIF":2.0,"publicationDate":"2025-02-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/acm2.70001","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143483272","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Automatic skin flash optimization in breast and chestwall VMAT with static angle modulated ports: Effect of HU and flash margin size on plan quality and robustness.","authors":"Emily Hubley, Brandon Koger, Taoran Li, Michael Salerno, Ryan M Scheuermann, Lei Dong, Boon-Keng Kevin Teo","doi":"10.1002/acm2.70036","DOIUrl":"https://doi.org/10.1002/acm2.70036","url":null,"abstract":"<p><p>Skin flash is typically added to breast and chestwall plans to ensure robust target coverage in the presence of respiratory motion, anatomic changes, and small setup uncertainties. Adding skin flash in volumetric modulated arc therapy (VMAT) plans is an iterative and manual process. RapidArc dynamic (RAD) is a new solution that integrates a dynamic collimator and static-gantry angle modulated ports directly into arc delivery. The automatic skin flash tool (ASF) allows users to automatically add skin flash directly within the optimizer. The user must select the thickness and Hounsfield Units (HUs) of the flash region, but the optimal values are not currently known. For 13 left- and right-sided breast and chestwall patients, RAD plans were created with no skin flash and with ASF with thickness of 5 to 20 mm and HU of -500 to 0 HU. To assess plan quality, DVH metrics for planning target volume (PTV), heart, ipsilateral lung, contralateral lung, and contralateral breast were recorded. To assess plan robustness, the isocenter was shifted 5 mm, moving the target 5 mm anteriorly into the flash region. The changes in clinical target volume (CTV) D95% and D99% were recorded. A paired t-test was used to determine if changes in plan quality or robustness were statistically significant (p ≤ 0.05). The addition of ≥ 7 mm of skin flash resulted in robust plans. Varying the HU did not affect robustness. Increasing the skin flash beyond 10 mm increased PTV V105%. This increase was much larger in the 0 HU plans than in the -350 HU plans. We therefore recommend using -350 HU and 7-10 mm of skin flash for anticipated inter- and intra-fraction motion of 5 mm.</p>","PeriodicalId":14989,"journal":{"name":"Journal of Applied Clinical Medical Physics","volume":" ","pages":"e70036"},"PeriodicalIF":2.0,"publicationDate":"2025-02-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143476611","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"T2-weighted imaging of rectal cancer using a 3D fast spin echo sequence with and without deep learning reconstruction: A reader study.","authors":"Dan Nguyen, Sarah Palmquist, Ken-Pin Hwang, Jingfei Ma, Usama Salem, Jia Sun, Xinzeng Wang, Jong Bum Son, Randy Ernst, Peng Wei, Harmeet Kaur, Nir Stanietzky","doi":"10.1002/acm2.70031","DOIUrl":"https://doi.org/10.1002/acm2.70031","url":null,"abstract":"<p><strong>Purpose: </strong>To compare image quality and clinical utility of a T2-weighted (T2W) 3-dimensional (3D) fast spin echo (FSE) sequence using deep learning reconstruction (DLR) versus conventional reconstruction for rectal magnetic resonance imaging (MRI).</p><p><strong>Methods: </strong>The study included 50 patients with rectal cancer who underwent rectal MRI consecutively between July 7, 2020 and January 20, 2021 using a T2W 3D FSE sequence with DLR and conventional reconstruction. Three radiologists reviewed the two sets of images, scoring overall SNR, motion artifacts, and overall image quality on a 3-point scale and indicating clinical preference for DLR or conventional reconstruction based on those three criteria as well as image characterization of bowel wall layer definition, tumor invasion of muscularis propria, residual disease, fibrosis, nodal margin, and extramural venous invasion.</p><p><strong>Results: </strong>Image quality was rated as moderate or good for both DLR and conventional reconstruction for most cases. DLR was preferred over conventional reconstruction in all of the categories except for bowel wall layer definition.</p><p><strong>Conclusion: </strong>Both conventional reconstruction and DLR provide acceptable image quality for T2W 3D FSE imaging of rectal cancer. DLR was clinically preferred over conventional reconstruction in almost all categories.</p>","PeriodicalId":14989,"journal":{"name":"Journal of Applied Clinical Medical Physics","volume":" ","pages":"e70031"},"PeriodicalIF":2.0,"publicationDate":"2025-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143458003","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Assessing the rigidity of thermoplastic masks for head immobilization effectiveness in stereotactic radiosurgery.","authors":"Iris Pasion Apale, Adam Agnew, Daniel Foley","doi":"10.1002/acm2.70058","DOIUrl":"https://doi.org/10.1002/acm2.70058","url":null,"abstract":"<p><p>This study compared three Brainlab thermoplastic masks-Cranial 4pi basic, stereotactic (Close Mask V2), and open-face-to see how well they limited head movement during Stereotactic Radiosurgery (SRS). Using a head phantom, we tested rotational movements (pitch, yaw, and roll) and measured displacements with the ExacTrac system. The open-face mask had the smallest mean displacements (pitch: 0.14 ± 0.03°, yaw: 0.11 ± 0.02°, roll: 0.16 ± 0.03°) and performed slightly better than the stereotactic mask in pitch (0.20 ± 0.04°, p = 0.0173). The stereotactic mask performed similarly in yaw (0.09 ± 0.02°) and roll (0.16 ± 0.04°). The basic mask showed much more movement (pitch: 0.44 ± 0.13°, yaw: 0.28 ± 0.07°, roll: 0.26 ± 0.07°), making it less suitable for SRS. These results apply to the solid two-piece masks tested here and show that both the open-face and stereotactic masks provide reliable immobilization for accurate SRS treatments.</p>","PeriodicalId":14989,"journal":{"name":"Journal of Applied Clinical Medical Physics","volume":" ","pages":"e70058"},"PeriodicalIF":2.0,"publicationDate":"2025-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143468248","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Impact of acceleration treatment on treatment plan and delivery qualities in tomotherapy for lung cancer.","authors":"Ryosuke Shirata, Tatsuya Inoue, Yugo Ebinuma, Akihiro Yamano, Takayuki Yagihashi, Hironori Nagata, Yumiko Minagawa, Yuki Mukai, Akiko Sato, Motoko Omura","doi":"10.1002/acm2.70049","DOIUrl":"https://doi.org/10.1002/acm2.70049","url":null,"abstract":"<p><strong>Background: </strong>Acceleration treatment (AT) is a novel treatment planning parameter introduced in the tomotherapy-dedicated treatment planning system, Precision. This study explores the effects of AT on tomotherapy plans using helical (TomoHelical) and direct (TomoDirect) irradiation techniques.</p><p><strong>Methods: </strong>This study enrolled 20 patients with lung cancer. Initially, 10 TomoHelical and 10 TomoDirect treatment plans were created for each patient, utilizing patient-specific field width and pitch with an AT setting of 0. These original plans were subsequently reoptimized by changing only the AT values to 1, 4, 7, and 10 without changing other calculation parameters to assess the impact of AT on dosimetric and delivery parameters. Additionally, the deliverability of all plans was evaluated through patient-specific quality assurance using gamma analysis.</p><p><strong>Results: </strong>Increasing the AT from 0 to 10 led to a slight increase in maximum doses and a decrease in minimum doses within the target volume, thereby impairing dose homogeneity. Dose conformity to the target also deteriorated. Conversely, target coverage and delivery time improved considerably with higher AT values. Moreover, doses to organs at risk, including the lung, spinal cord, heart, and esophagus, remained clinically acceptable across all plans. Changes in these doses and the gamma pass rate in patient-specific quality assurance were negligible with variations in AT. This trend was consistent across both delivery techniques.</p><p><strong>Conclusion: </strong>AT is a crucial parameter in tomotherapy planning for modulating plan and delivery qualities. Higher AT values can enhance target coverage and delivery time efficiency.</p>","PeriodicalId":14989,"journal":{"name":"Journal of Applied Clinical Medical Physics","volume":" ","pages":"e70049"},"PeriodicalIF":2.0,"publicationDate":"2025-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143457998","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A comparative analysis of deep learning architectures with data augmentation and multichannel input for locoregional breast cancer radiotherapy.","authors":"Rosalie Klarenberg, Nienke L M Bakx, Coen W Hurkmans","doi":"10.1002/acm2.70047","DOIUrl":"https://doi.org/10.1002/acm2.70047","url":null,"abstract":"<p><strong>Purpose: </strong>Studies on deep learning dose prediction increasingly focus on 3D models with multiple input channels and data augmentation, which increases the training time and thus also the environmental burden and hampers the ease of re-training. Here we compare 2D and 3D U-Net models with clinical accepted plans to evaluate the appropriateness of using less computationally heavy models.</p><p><strong>Methods: </strong>A 2D Attention U-Net, a 2D HD U-Net, and a 3D U-Net were trained using 1 or 5 input channels with or without data augmentation and data from 89 locoregional breast cancer patients. Results were compared to clinically accepted plans. The significance of inclusion of more channels or data augmentation was compared to the base models and the HD U-Net and Attention U-Net were compared to their respective identically trained counterparts.</p><p><strong>Results: </strong>The Attention U-Net reached fewest PTV clinical goals (28%, mostly due to a too high average breast PTV dose) and improved using significantly using five channels and augmentation (49%). The HD U-Net already fulfilled 70% of the PTV goals, which did not improve much by adding more channels or augmentation. The 3D U-Net with five channels and augmentation reached 76%, compared to 81% in the clinically accepted plans. The lower rates for the HD U-Net compared to the 3D U-Net and clinical plans were mainly caused by a lower PTVn1n2 D98%, which was still on average 93%. Organ-at-risk goals were met in most cases for all models. Training time was approximately 8 fold for the 3D model.</p><p><strong>Conclusions: </strong>Comparable results to a 3D U-Net and clinical plans can be reached with a 2D HD U-net for a dataset size commonly seen in clinical practice. The Attention U-Net did profit from adding extra channels and data augmentation but did not reach the same level of accuracy as the other models.</p>","PeriodicalId":14989,"journal":{"name":"Journal of Applied Clinical Medical Physics","volume":" ","pages":"e70047"},"PeriodicalIF":2.0,"publicationDate":"2025-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143468244","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}