In this article the Tomographic Iterative GPU-based Reconstruction (TIGRE) Toolbox, a MATLAB/CUDA toolbox for fast and accurate 3D x-ray image reconstruction, is presented. One of the key features is the implementation of a wide variety of iterative algorithms as well as FDK, including a range of algorithms in the SART family, the Krylov subspace family and a range of methods using total variation regularization. Additionally, the toolbox has GPU-accelerated projection and back projection using the latest techniques and it has a modular design that facilitates the implementation of new algorithms. We present an overview of the structure and techniques used in the creation of the toolbox, together with two usage examples. The TIGRE Toolbox is released under an open source licence, encouraging people to contribute.
Institute of Physics and Engineering in Medicine
IPEM's aim is to promote the advancement of physics and engineering applied to medicine and biology for the public benefit. Its members are professionals working in healthcare, education, industry and research.
IPEM publishes scientific journals and books and organises conferences to disseminate knowledge and support members in their development. It sets and advises on standards for the practice, education and training of scientists and engineers working in healthcare to secure an effective and appropriate workforce.
Purpose-led Publishing is a coalition of three not-for-profit publishers in the field of physical sciences: AIP Publishing, the American Physical Society and IOP Publishing.
Together, as publishers that will always put purpose above profit, we have defined a set of industry standards that underpin high-quality, ethical scholarly communications.
We are proudly declaring that science is our only shareholder.
ISSN: 2057-1976
A broad, inclusive, rapid review journal devoted to publishing new research in all areas of biomedical engineering, biophysics and medical physics, with a special emphasis on interdisciplinary work between these fields.
Open all abstracts, in this tab
Ander Biguri et al 2016 Biomed. Phys. Eng. Express 2 055010
Christiana Subaar et al 2024 Biomed. Phys. Eng. Express 10 035029
A lot of underdeveloped nations particularly in Africa struggle with cancer-related, deadly diseases. Particularly in women, the incidence of breast cancer is rising daily because of ignorance and delayed diagnosis. Only by correctly identifying and diagnosing cancer in its very early stages of development can be effectively treated. The classification of cancer can be accelerated and automated with the aid of computer-aided diagnosis and medical image analysis techniques. This research provides the use of transfer learning from a Residual Network 18 (ResNet18) and Residual Network 34 (ResNet34) architectures to detect breast cancer. The study examined how breast cancer can be identified in breast mammography pictures using transfer learning from ResNet18 and ResNet34, and developed a demo app for radiologists using the trained models with the best validation accuracy. 1, 200 datasets of breast x-ray mammography images from the National Radiological Society's (NRS) archives were employed in the study. The dataset was categorised as implant cancer negative, implant cancer positive, cancer negative and cancer positive in order to increase the consistency of x-ray mammography images classification and produce better features. For the multi-class classification of the images, the study gave an average accuracy for binary classification of benign or malignant cancer cases of 86.7% validation accuracy for ResNet34 and 92% validation accuracy for ResNet18. A prototype web application showcasing ResNet18 performance has been created. The acquired results show how transfer learning can improve the accuracy of breast cancer detection, providing invaluable assistance to medical professionals, particularly in an African scenario.
Benita S Mackay et al 2021 Biomed. Phys. Eng. Express 7 052002
Tissue engineering is a branch of regenerative medicine that harnesses biomaterial and stem cell research to utilise the body's natural healing responses to regenerate tissue and organs. There remain many unanswered questions in tissue engineering, with optimal biomaterial designs still to be developed and a lack of adequate stem cell knowledge limiting successful application. Advances in artificial intelligence (AI), and deep learning specifically, offer the potential to improve both scientific understanding and clinical outcomes in regenerative medicine. With enhanced perception of how to integrate artificial intelligence into current research and clinical practice, AI offers an invaluable tool to improve patient outcome.
Fan Peng et al 2024 Biomed. Phys. Eng. Express 10 035038
Objective. Ultrasound-assisted orthopaedic navigation held promise due to its non-ionizing feature, portability, low cost, and real-time performance. To facilitate the applications, it was critical to have accurate and real-time bone surface segmentation. Nevertheless, the imaging artifacts and low signal-to-noise ratios in the tomographical B-mode ultrasound (B-US) images created substantial challenges in bone surface detection. In this study, we presented an end-to-end lightweight US bone segmentation network (UBS-Net) for bone surface detection. Approach. We presented an end-to-end lightweight UBS-Net for bone surface detection, using the U-Net structure as the base framework and a level set loss function for improved sensitivity to bone surface detectability. A dual attention (DA) mechanism was introduced at the end of the encoder, which considered both position and channel information to obtain the correlation between the position and channel dimensions of the feature map, where axial attention (AA) replaced the traditional self-attention (SA) mechanism in the position attention module for better computational efficiency. The position attention and channel attention (CA) were combined with a two-class fusion module for the DA map. The decoding module finally completed the bone surface detection. Main Results. As a result, a frame rate of 21 frames per second (fps) in detection were achieved. It outperformed the state-of-the-art method with higher segmentation accuracy (Dice similarity coefficient: 88.76% versus 87.22%) when applied the retrospective ultrasound (US) data from 11 volunteers. Significance. The proposed UBS-Net for bone surface detection in ultrasound achieved outstanding accuracy and real-time performance. The new method out-performed the state-of-the-art methods. It had potential in US-guided orthopaedic surgery applications.
Sotiris Raptis et al 2024 Biomed. Phys. Eng. Express 10 035016
Radiomics-based prediction models have shown promise in predicting Radiation Pneumonitis (RP), a common adverse outcome of chest irradiation. Τhis study looks into more than just RP: it also investigates a bigger shift in the way radiomics-based models work. By integrating multi-modal radiomic data, which includes a wide range of variables collected from medical images including cutting-edge PET/CT imaging, we have developed predictive models that capture the intricate nature of illness progression. Radiomic features were extracted using PyRadiomics, encompassing intensity, texture, and shape measures. The high-dimensional dataset formed the basis for our predictive models, primarily Gradient Boosting Machines (GBM)—XGBoost, LightGBM, and CatBoost. Performance evaluation metrics, including Multi-Modal AUC-ROC, Sensitivity, Specificity, and F1-Score, underscore the superiority of the Deep Neural Network (DNN) model. The DNN achieved a remarkable Multi-Modal AUC-ROC of 0.90, indicating superior discriminatory power. Sensitivity and specificity values of 0.85 and 0.91, respectively, highlight its effectiveness in detecting positive occurrences while accurately identifying negatives. External validation datasets, comprising retrospective patient data and a heterogeneous patient population, validate the robustness and generalizability of our models. The focus of our study is the application of sophisticated model interpretability methods, namely SHAP (SHapley Additive exPlanations) and LIME (Local Interpretable Model-Agnostic Explanations), to improve the clarity and understanding of predictions. These methods allow clinicians to visualize the effects of features and provide localized explanations for every prediction, enhancing the comprehensibility of the model. This strengthens trust and collaboration between computational technologies and medical competence. The integration of data-driven analytics and medical domain expertise represents a significant shift in the profession, advancing us from analyzing pixel-level information to gaining valuable prognostic insights.
Sami Ullah Bangash et al 2024 Biomed. Phys. Eng. Express 10 035032
We have previously reported the design of a portable 109Cd x-ray fluorescence (XRF) system to measure iron levels in the skin of patients with either iron overload disease, such as thalassemia, or iron deficiency disease, such as anemia. In phantom studies, the system was found to have a detection limit of 1.35 μg Fe per g of tissue for a dose of 1.1 mSv. However, the system must provide accurate as well as precise measurements of iron levels in the skin in order to be suitable for human studies. The accuracy of the system has been explored using several methods. First, the iron concentrations of ten pigskin samples were assessed using both the portable XRF system and ICP-MS, and the results were compared. Overall, it was found that XRF and ICP-MS reported average values for iron in skin that were comparable to within uncertainties. The mean difference between the two methodologies was not significant, 2.5 ± 4.6 μg Fe per g. On this basis, the system could be considered accurate. However, ICP-MS measurements reported a wider range of values than XRF, with two individual samples having ICP-MS results that were significantly elevated (p < 0.05) compared to XRF. Synchrotron μXRF maps of iron levels in pigskin were acquired on the BioXAS beam line of the Canadian Light Source. The μXRF maps indicated two important features in the distribution of iron in pigskin. First, there were small areas of high iron concentration in the pigskin samples, that were predominantly located in the dermis and hypodermis at depths greater than 0.5 mm. Monte Carlo modelling using the EGS 5 code determined that if these iron 'hot spots' were located towards the back of the skin at depths greater than 0.5 mm, they would not be observed by XRF, but would be measured by ICP-MS. These results support a hypothesis that iron levels in the two samples that reported significantly elevated ICP-MS results compared to XRF may have had small blood vessels at the back of the skin. Second, the synchrotron μXRF maps also showed a narrow (approximately 100μm thick) layer of elevated iron at the surface of the skin. Monte Carlo models determined that, as expected, the XRF system was most sensitive to these skin layers. However, the simulations found that the XRF system, when calibrated against homogenous water-based phantoms, was found to accurately measure average iron levels in the skin of normal pigs despite the greater sensitivity to the surface layer. The Monte Carlo results further indicated that with highly elevated skin surface iron levels, the XRF system would not provide a good estimate of average skin iron levels. The XRF estimate could, with correction factors, provide a good estimate of the iron levels in the surface layers of skin. There is limited data on iron distribution in skin, especially under conditions of disease. If iron levels are elevated at the skin surface by diseases including thalassemia and hemochromatosis, this XRF device may prove to be an accurate clinical tool. However, further data are required on skin iron distributions in healthy and iron overload disease before this system can be verified to provide accurate measurements.
Muhammad Suhaib Shahid et al 2024 Biomed. Phys. Eng. Express 10 032001
The paper aims to explore the current state of understanding surrounding in silico oral modelling. This involves exploring methodologies, technologies and approaches pertaining to the modelling of the whole oral cavity; both internally and externally visible structures that may be relevant or appropriate to oral actions. Such a model could be referred to as a 'complete model' which includes consideration of a full set of facial features (i.e. not only mouth) as well as synergistic stimuli such as audio and facial thermal data. 3D modelling technologies capable of accurately and efficiently capturing a complete representation of the mouth for an individual have broad applications in the study of oral actions, due to their cost-effectiveness and time efficiency. This review delves into the field of clinical phonetics to classify oral actions pertaining to both speech and non-speech movements, identifying how the various vocal organs play a role in the articulatory and masticatory process. Vitaly, it provides a summation of 12 articulatory recording methods, forming a tool to be used by researchers in identifying which method of recording is appropriate for their work. After addressing the cost and resource-intensive limitations of existing methods, a new system of modelling is proposed that leverages external to internal correlation modelling techniques to create a more efficient models of the oral cavity. The vision is that the outcomes will be applicable to a broad spectrum of oral functions related to physiology, health and wellbeing, including speech, oral processing of foods as well as dental health. The applications may span from speech correction, designing foods for the aging population, whilst in the dental field we would be able to gain information about patient's oral actions that would become part of creating a personalised dental treatment plan.
Li Liu et al 2018 Biomed. Phys. Eng. Express 4 015004
Adhesives that involve adhesion to the skin have been of great technological importance in medical or pharmaceutical fields, including recently emerging wearable sensors and electronics. The objective of this work was to evaluate the performances of silicone-based adhesives with skin, using a peel adhesion test. Specifically, we explored the effect of adhesive cleansing, which is an inevitable daily event for patients' comfort in long-term applications. Firstly, three medical grade silicone gels, Silbione® RT 4717, Silbione® RT 4642, and Silpuran® 2130, were used to fabricate adhesive pads. Their peel strength values were subsequently measured and compared, among which Silbione® RT gel 4717 possessed the highest peel strength. Therefore, it was selected as the raw material to fabricate the pads with a thickness range of 640–740 μm. Secondly, the peel adhesion of Silbione® RT 4717 adhesive pad was further compared with a series of commercial products that employ various medical-grade adhesives. The peel strength results indicated that our custom-made adhesive pad had an adequately strong adhesion for clinical use. Thirdly, in order to observe and predict the long-term performance of the adhesives, an aging test was performed in an ambient environment, revealing that Silbione® RT 4717 adhesive remained highly sticky for 5 days. Lastly, adequate cleansing protocols were established by monitoring the changes in peel strength after washing and wiping events. The reusability analysis showed that Silbione® 4717 adhesive pad was reusable in a one-week period for the washing method and 3 days for the wiping method.
Reshma H et al 2024 Biomed. Phys. Eng. Express 10 035036
To localize the unusual cardiac activities non-invasively, one has to build a prior forward model that relates the heart, torso, and detectors. This model has to be constructed to mathematically relate the geometrical and functional activities of the heart. Several methods are available to model the prior sources in the forward problem, which results in the lead field matrix generation. In the conventional technique, the lead field assumed the fixed prior sources, and the source vector orientations were presumed to be parallel to the detector plane with the unit strength in all directions. However, the anomalies cannot always be expected to occur in the same location and orientation, leading to misinterpretation and misdiagnosis. To overcome this, the work proposes a new forward model constructed using the VCG signals of the same subject. Furthermore, three transformation methods were used to extract VCG in constructing the time-varying lead fields to steer to the orientation of the source rather than just reconstructing its activities in the inverse problem. In addition, the unit VCG loop of the acute ischemia patient was extracted to observe the changes compared to the normal subject. The abnormality condition was achieved by delaying the depolarization time by 15ms. The results involving the unit vectors of VCG demonstrated the anisotropic nature of cardiac source orientations, providing information about the heart's electrical activity.
Taisa Higino and Rodrigo França 2022 Biomed. Phys. Eng. Express 8 042001
The use of nanoparticles as biomaterials with applications in the biomedical field is growing every day. These nanomaterials can be used as contrast imaging agents, combination therapy agents, and targeted delivery systems in medicine and dentistry. Usually, nanoparticles are found as synthetic or natural organic materials, such as hydroxyapatite, polymers, and lipids. Besides that, they are could also be inorganic, for instance, metallic or metal-oxide-based particles. These inorganic nanoparticles could additionally present magnetic properties, such as superparamagnetic iron oxide nanoparticles. The use of nanoparticles as drug delivery agents has many advantages, for they help diminish toxicity effects in the body since the drug dose reduces significantly, increases drugs biocompatibility, and helps target drugs to specific organs. As targeted-delivery agents, one of the applications uses nanoparticles as drug delivery particles for bone-tissue to treat cancer, osteoporosis, bone diseases, and dental treatments such as periodontitis. Their application as drug delivery agents requires a good comprehension of the nanoparticle properties and composition, alongside their synthesis and drug attachment characteristics. Properties such as size, shape, core-shell designs, and magnetic characteristics can influence their behavior inside the human body and modify magnetic properties in the case of magnetic nanoparticles. Based on that, many different studies have modified the synthesis methods for these nanoparticles and developed composite systems for therapeutics delivery, adapting, and improving magnetic properties, shell-core designs, and particle size and nanosystems characteristics. This review presents the most recent studies that have been presented with different nanoparticle types and structures for bone and dental drug delivery.
Open all abstracts, in this tab
H Ara et al 2024 Biomed. Phys. Eng. Express 10 045029
The Z-scan technique is a nonlinear optical method that has found applications in characterizing various materials, particularly those exhibiting nonlinear optical response (NLOR). This study applies the continuous wave (CW) Z-scan technique to examine the NLOR in terms of the nonlinear optical phase shifts ) exhibited by the ccfDNA extracted from blood plasma samples collected from a group constituting 30 cancer-diagnosed patients and another group constituting 30 non-diagnosed individuals. The cancer group exhibited significantly higher versus incident power slopes compared to the non-cancer group (0.34 versus 0.12) providing a clear distinction between the two groups. The receiver operating characteristic (ROC) curve analysis of the results indicates a clear separation between cancer and non-cancer groups, along with a 94% accuracy rate of the data. The Z-scan results are corroborated by spectrophotometric analysis, revealing a consistent trend in the concentration values of ccfDNA samples extracted from both cancerous and non-cancerous samples, measuring 3.24 and 1.41 respectively. Additionally, more sensitive fluorometric analyses of the respective samples demonstrate significantly higher concentrations of ccfDNA in the cancer group, further affirming the correlation with the Z-scan results. The study suggests that the Z-scan technique holds promise as an effective method for cancer detection, potentially contributing to improved oncology diagnosis and prognosis in the future.
Shunsuke Suzuki et al 2024 Biomed. Phys. Eng. Express 10 045028
Evaluating neutron output is important to ensure proper dose delivery for patients in boron neutron capture therapy (BNCT). It requires efficient quality assurance (QA) and quality control (QC) while maintaining measurement accuracy. This study investigated the optimal measurement conditions for QA/QC of activation measurements using a high-purity germanium (HP-Ge) detector in an accelerator-based boron neutron capture therapy (AB-BNCT) system employing a lithium target. The QA/QC uncertainty of the activation measurement was evaluated based on counts, reproducibility, and standard radiation source uncertainties. Measurements in a polymethyl methacrylate (PMMA) cylindrical phantom using aluminum-manganese (Al–Mn) foils and aluminum-gold (Al–Au) foils and measurements in a water phantom using gold wire with and without cadmium cover were performed to determine the optimal measurement conditions. The QA/QC uncertainties of the activation measurements were 4.5% for Au and 4.6% for Mn. The optimum irradiation proton charge and measurement time were determined to be 36 C and 900 s for measurements in a PMMA cylindrical phantom, 7.0 C and 900 s for gold wire measurements in a water phantom, and 54 C and 900 s at 0–2.2 cm depth and 3,600 s at deeper depths for gold wire measurements with cadmium cover. Our results serve as a reference for determining measurement conditions when performing QA/QC of activation measurements using HP-Ge detectors at an AB-BNCT employing a lithium target.
Choirul Anam et al 2024 Biomed. Phys. Eng. Express 10 045027
Purpose. To develop a method to extract statistical low-contrast detectability (LCD) and contrast-detail (C-D) curves from clinical patient images. Method. We used the region of air surrounding the patient as an alternative for a homogeneous region within a patient. A simple graphical user interface (GUI) was created to set the initial configuration for region of interest (ROI), ROI size, and minimum detectable contrast (MDC). The process was started by segmenting the air surrounding the patient with a threshold between −980 HU (Hounsfield units) and −1024 HU to get an air mask. The mask was trimmed using the patient center coordinates to avoid distortion from the patient table. It was used to automatically place square ROIs of a predetermined size. The mean pixel values in HU within each ROI were calculated, and the standard deviation (SD) from all the means was obtained. The MDC for a particular target size was generated by multiplying the SD by 3.29. A C-D curve was obtained by iterating this process for the other ROI sizes. This method was applied to the homogeneous area from the uniformity module of an ACR CT phantom to find the correlation between the parameters inside and outside the phantom, for 30 thoracic, 26 abdominal, and 23 head images. Results. The phantom images showed a significant linear correlation between the LCDs obtained from outside and inside the phantom, with R2 values of 0.67 and 0.99 for variations in tube currents and tube voltages. This indicated that the air region outside the phantom can act as a surrogate for the homogenous region inside the phantom to obtain the LCD and C-D curves. Conclusion. The C-D curves obtained from outside the ACR CT phantom show a strong linear correlation with those from inside the phantom. The proposed method can also be used to extract the LCD from patient images by using the region of air outside as a surrogate for a region inside the patient.
Jingyuan Wu et al 2024 Biomed. Phys. Eng. Express 10 045026
Evaluation of skin recovery is an important step in the treatment of burns. However, conventional methods only observe the surface of the skin and cannot quantify the injury volume. Optical coherence tomography (OCT) is a non-invasive, non-contact, real-time technique. Swept source OCT uses near infrared light and analyzes the intensity of light echo at different depths to generate images from optical interference signals. To quantify the dynamic recovery of skin burns over time, laser induced skin burns in mice were evaluated using deep learning of Swept source OCT images. A laser-induced mouse skin thermal injury model was established in thirty Kunming mice, and OCT images of normal and burned areas of mouse skin were acquired at day 0, day 1, day 3, day 7, and day 14 after laser irradiation. This resulted in 7000 normal and 1400 burn B-scan images which were divided into training, validation, and test sets at 8:1.5:0.5 ratio for the normal data and 8:1:1 for the burn data. Normal images were manually annotated, and the deep learning U-Net model (verified with PSPNe and HRNet models) was used to segment the skin into three layers: the dermal epidermal layer, subcutaneous fat layer, and muscle layer. For the burn images, the models were trained to segment just the damaged area. Three-dimensional reconstruction technology was then used to reconstruct the damaged tissue and calculate the damaged tissue volume. The average IoU value and f-score of the normal tissue layer U-Net segmentation model were 0.876 and 0.934 respectively. The IoU value of the burn area segmentation model reached 0.907 and f-score value reached 0.951. Compared with manual labeling, the U-Net model was faster with higher accuracy for skin stratification. OCT and U-Net segmentation can provide rapid and accurate analysis of tissue changes and clinical guidance in the treatment of burns.
Ratheesh K E and Mayakannan Krishnan 2024 Biomed. Phys. Eng. Express 10 045025
Radiotherapy (RT) is one of the major treatment modalities among surgery and chemotherapy for carcinoma breast. The surface dose study of modified reconstructive constructive Mastectomy (MRM) breast is important due to the heterogeneity in the body contour and the conventional treatment angle to save the lungs and heart from the radiation. These angular entries of radiation beam cause an unpredictable dose deposition on the body surface, which has to be monitored. Thermoluminescent dosimeter (TLD) or optically stimulated luminescent dosimeter (nano OSLD) are commonly preferable dosimeters for this purpose. The surface dose response of TLD and nano OSLD during MRM irradiation has been compared with the predicted dose from the treatment planning system (TPS). The study monitored 100 MRM patients by employing a total 500 dosimeters consisting of TLD (n = 250) and nano OSLD (n = 250), during irradiation from an Elekta Versa HD 6 MV Linear accelerator. The study observed a variance of 3.9% in the dose measurements for TLD and 3.2% for nano OSLD from the planned surface dose, with a median percentage dose of 44.02 for nano OSLD and 40.30 for TLD (p value 0.01). There was no discernible evidence of variation in dose measurements attributable to differences in field size or from patient to patient. Additionally, no variation was observed in dose measurements when comparing the placement of the dosimeter from central to off-centre positions. In comparison, a minor difference in dose measurements were noted between TLD and nano OSLD, The study's outcomes support the applicability of both TLD and nano OSLD as effective dosimeters during MRM breast irradiation for surface dose evaluation.
Open all abstracts, in this tab
Mohammed Ali et al 2024 Biomed. Phys. Eng. Express 10 032003
Guided tissue/bone regeneration (GTR/GBR) is a widely used technique in dentistry to facilitate the regeneration of damaged bone and tissue, which involves guiding materials that eventually degrade, allowing newly created tissue to take its place. This comprehensive review the evolution of biomaterials for guided bone regeneration that showcases a progressive shift from non-resorbable to highly biocompatible and bioactive materials, allowing for more effective and predictable bone regeneration. The evolution of biomaterials for guided bone regeneration GTR/GBR has marked a significant progression in regenerative dentistry and maxillofacial surgery. Biomaterials used in GBR have evolved over time to enhance biocompatibility, bioactivity, and efficacy in promoting bone growth and integration. This review also probes into several promising fabrication techniques like electrospinning and latest 3D printing fabrication techniques, which have shown potential in enhancing tissue and bone regeneration processes. Further, the challenges and future direction of GTR/GBR are explored and discussed.
Muhammad Suhaib Shahid et al 2024 Biomed. Phys. Eng. Express 10 032001
The paper aims to explore the current state of understanding surrounding in silico oral modelling. This involves exploring methodologies, technologies and approaches pertaining to the modelling of the whole oral cavity; both internally and externally visible structures that may be relevant or appropriate to oral actions. Such a model could be referred to as a 'complete model' which includes consideration of a full set of facial features (i.e. not only mouth) as well as synergistic stimuli such as audio and facial thermal data. 3D modelling technologies capable of accurately and efficiently capturing a complete representation of the mouth for an individual have broad applications in the study of oral actions, due to their cost-effectiveness and time efficiency. This review delves into the field of clinical phonetics to classify oral actions pertaining to both speech and non-speech movements, identifying how the various vocal organs play a role in the articulatory and masticatory process. Vitaly, it provides a summation of 12 articulatory recording methods, forming a tool to be used by researchers in identifying which method of recording is appropriate for their work. After addressing the cost and resource-intensive limitations of existing methods, a new system of modelling is proposed that leverages external to internal correlation modelling techniques to create a more efficient models of the oral cavity. The vision is that the outcomes will be applicable to a broad spectrum of oral functions related to physiology, health and wellbeing, including speech, oral processing of foods as well as dental health. The applications may span from speech correction, designing foods for the aging population, whilst in the dental field we would be able to gain information about patient's oral actions that would become part of creating a personalised dental treatment plan.
Abdallah El Ouaridi et al 2024 Biomed. Phys. Eng. Express 10 032002
Positron emission tomography (PET) is a powerful medical imaging modality used in nuclear medicine to diagnose and monitor various clinical diseases in patients. It is more sensitive and produces a highly quantitative mapping of the three-dimensional biodistribution of positron-emitting radiotracers inside the human body. The underlying technology is constantly evolving, and recent advances in detection instrumentation and PET scanner design have significantly improved the medical diagnosis capabilities of this imaging modality, making it more efficient and opening the way to broader, innovative, and promising clinical applications. Some significant achievements related to detection instrumentation include introducing new scintillators and photodetectors as well as developing innovative detector designs and coupling configurations. Other advances in scanner design include moving towards a cylindrical geometry, 3D acquisition mode, and the trend towards a wider axial field of view and a shorter diameter. Further research on PET camera instrumentation and design will be required to advance this technology by improving its performance and extending its clinical applications while optimising radiation dose, image acquisition time, and manufacturing cost. This article comprehensively reviews the various parameters of detection instrumentation and PET system design. Firstly, an overview of the historical innovation of the PET system has been presented, focusing on instrumental technology. Secondly, we have characterised the main performance parameters of current clinical PET and detailed recent instrumental innovations and trends that affect these performances and clinical practice. Finally, prospects for this medical imaging modality are presented and discussed. This overview of the PET system's instrumental parameters enables us to draw solid conclusions on achieving the best possible performance for the different needs of different clinical applications.
Nadia Muhammad Hussain et al 2024 Biomed. Phys. Eng. Express 10 022002
Intrapartum fetal hypoxia is related to long-term morbidity and mortality of the fetus and the mother. Fetal surveillance is extremely important to minimize the adverse outcomes arising from fetal hypoxia during labour. Several methods have been used in current clinical practice to monitor fetal well-being. For instance, biophysical technologies including cardiotocography, ST-analysis adjunct to cardiotocography, and Doppler ultrasound are used for intrapartum fetal monitoring. However, these technologies result in a high false-positive rate and increased obstetric interventions during labour. Alternatively, biochemical-based technologies including fetal scalp blood sampling and fetal pulse oximetry are used to identify metabolic acidosis and oxygen deprivation resulting from fetal hypoxia. These technologies neither improve clinical outcomes nor reduce unnecessary interventions during labour. Also, there is a need to link the physiological changes during fetal hypoxia to fetal monitoring technologies. The objective of this article is to assess the clinical background of fetal hypoxia and to review existing monitoring technologies for the detection and monitoring of fetal hypoxia. A comprehensive review has been made to predict fetal hypoxia using computational and machine-learning algorithms. The detection of more specific biomarkers or new sensing technologies is also reviewed which may help in the enhancement of the reliability of continuous fetal monitoring and may result in the accurate detection of intrapartum fetal hypoxia.
Wei-Jen Chan and Huatian Li 2024 Biomed. Phys. Eng. Express 10 022001
In recent years, nanoparticles (NPs) have been extensively developed as drug carriers to overcome the limitations of cancer therapeutics. However, there are several biological barriers to nanomedicines, which include the lack of stability in circulation, limited target specificity, low penetration into tumors and insufficient cellular uptake, restricting the active targeting toward tumors of nanomedicines. To address these challenges, a variety of promising strategies were developed recently, as they can be designed to improve NP accumulation and penetration in tumor tissues, circulation stability, tumor targeting, and intracellular uptake. In this Review, we summarized nanomaterials developed in recent three years that could be utilized to improve drug delivery for cancer treatments.
Open all abstracts, in this tab
Manfred Köller et al 2018 Biomed. Phys. Eng. Express 4 055002
The role of bacterial cell division on the damage of adherent bacteria to titanium (Ti) nano-pillar cicada wing like surface was analyzed. Therefore nano-pillar Ti thin films were fabricated by glancing angle sputter deposition (GLAD) on silicon substrates. Gram-negative E. coli bacteria were allowed to adhere and to proliferate on these nanostructured samples for 3 h at 37 °C either under optimal cell growth conditions (brain heart infusion medium, BHI) or limited growth conditions (RPMI1640 medium). The bacteria adhered to the samples in both media. Compared to BHI medium the growth of E. coli in RPMI1640 medium was significantly inhibited. Concomitantly, the ratio of dead/living adherent bacteria on the nano-pillar surface was significantly decreased after the incubation period in RPMI1640. In addition, when the bacterial proliferation was biochemically halted using DL-serine-hydroxamate a comparable decrease in the ratio of dead/living adherent bacteria was also obtained in BHI medium. These results indicate that cell growth of adherent E. coli which is accompanied by cell elongations of the rod structure is involved in the damage induced by the titanium nano-pillar surface.
James Archer et al 2018 Biomed. Phys. Eng. Express 4 044003
Cherenkov radiation is the primary source of unwanted light in a scintillator dosimetry system. In this work we compare two techniques for temporally separating Cherenkov radiation from a slow scintillator signal. These techniques are applicable to a pulsed radiation beam. We found that by analysing the rising edge of the light pulse to identify the fast Cherenkov light only removed 74% of the Cherenkov light. By integrating the tail of the signal where only scintillation light is present a more accurate result is achieved. The average of the results of the two methods provides up to a 90% improvement in the accuracy of the relative dose when compared to ionisation chamber, in certain measurements. This work demonstrates an alternative methodology for the removal of Cherenkov light using signal analysis, while preserving all the scintillation light signal and minimising the bulk of the experimental equipment.
Natasha Maurmann et al 2017 Biomed. Phys. Eng. Express 3 045005
Materials, such as biopolymers, can be applied to produce scaffolds as mechanical support for cell growth in regenerative medicine. Two examples are polycaprolactone (PCL) and poly (lactic-co-glycolic acid) (PLGA), both used in this study to evaluate the behavior of umbilical cord-derived mesenchymal stem cells. The scaffolds were produced by the 3D printing technique using PCL as a polymer covered with PLGA fibers obtained by electrospinning. The cells were seeded in three concentrations: 8.5 × 103; 25.5 × 103 and 51.0 × 103 on the two surfaces of the scaffolds. With scanning electron microscopy (SEM), it was observed that the electrospun fibers were integrated into the 3D printed matrices. Confocal laser scanning microscopy and SEM confirmed the presence of attached cells and the lactate dehydrogenase release test showed the scaffolds were not cytotoxic. The cells were able to differentiate into osteogenic and chondrogenic lineages on the scaffolds. Mechanical test showed that the cells seeded on the 3D printed PCL matrices coated with PLGA electrospun nanofibers (3D + ES + SC) did not show significant difference in tensile modulus than the pure PCL matrix (3D) or PCL matrices coated with PLGA electrospun nanofibers (3D + ES). The combination of the two polymers facilitated the production of a support with greater mechanical stability due to the presence of the 3D printed PCL matrices fabricated by melted filaments and greater cell adhesion due to the PLGA fibers. The scaffolds are suitable for use in cell therapy and also for tissue regeneration purposes.
Open all abstracts, in this tab
Correia et al
Introduction. The positioning of γ ray interactions in positron emission tomography (PET) detectors is commonly made through the evaluation of the Anger logic flood histograms. Machine learning techniques, leveraging features extracted from signal waveform, have demonstrated successful applications in addressing various challenges in PET instrumentation. Aim. This paper evaluates the use of artificial neural networks (NN) for γ ray interaction positioning in pixelated scintillators coupled to a multiplexed array of silicon photomultipliers (SiPM). Methods. An array of 16 Cerium doped Lutetium-based (LYSO) crystal pixels (cross-section 2x2 mm2) coupled to 16 SiPM (S13360-1350) were used for the experimental setup. Data from each of the 16 LYSO pixels was recorded, a total of 160 000 events. The detectors were irradiated by 511 keV annihilation γ rays from a Sodium-22 (22Na) source. Another LYSO crystal was used for electronic collimation. Features extracted from the signal waveform were used to train the model. Two models were tested: i) single multiple-class neural network (mcNN), with 16 possible outputs followed by a softmax and ii) 16 binary classification neural networks (bNN), each one specialized in identifying events occurred in each position. Results. Both NN models showed a mean positioning accuracy above 85% on the evaluation dataset, although the mcNN is faster to train. Discussion. The method's accuracy is affected by the introduction of misclassified events that interacted in the neighbour's crystals and were misclassified during the dataset acquisition. Electronic collimation reduces this effect, however results could be improved using a more complex acquisition setup, such as a light-sharing configuration. Conclusions. The methods comparison showed that mcNN and bNN can surpass the Anger logic, showing the feasibility of using these models in positioning procedures of future multiplexed detector systems in a linear configuration.
Suppiah et al
Physiological Signals like Electromography (EMG) and Electroencephalography (EEG) can be analysed and decoded to provide vital information that can be used in a range of applications like rehabilitative robotics and remote device control. The process of acquiring and using these signals requires many compute-intensive tasks like signal acquisition, signal processing, feature extraction and machine learning. Performing these activities on a PC-based system with well-established software tools like Python and Matlab is the first step in designing solutions based upon these signals. In the application domain of rehabilitative robotics, one of the main goals is to develop solutions that can be deployed for the use of individuals who need them in improving their Acitivities-for-Daily Living (ADL). To achieve this objective, the final solution must be deployed onto an embedded solution that allows high portability and ease-of-use. Porting a solution from a PC-based environment onto a resource-constraint one such as a microcontroller poses many challenges. In this research paper, we propose the use of a ARM-based Corex M-4 processor. We explore the various stages of the design from the initial testing and validation, to the deployment of the proposed algorithm on the controller, and further investigate the use of Cepstrum features to obtain a high classification accuracy with minimal input features. The proposed solution is able to achieve an average classification accuracy of 95.34% for all five classes in the EMG domain and 96.16% in the EEG domain on the embedded board.
R et al
Deep learning models have demonstrated remarkable performance in the classification of motor imagery BCI systems. However, these models exhibit sensitivity to challenging trials, often called hard trials, leading to performance degradation. In this paper, we address this issue by proposing two novel methods for identifying and mitigating the impact of hard trials on model performance. The first method leverages model prediction scores to discern hard trials. The second approach employs a quantitative explainable artificial intelligence (XAI) approach, enabling a more transparent and interpretable means of hard trials identification.
The identified hard trials are removed from the entire motor imagery training and validation dataset, and the deep learning model is further re-trained using the dataset without hard trials.
To evaluate the efficacy of these proposed methods, experiments were conducted on the Open
BMI dataset. The results for hold-out analysis show that, the proposed quantitative XAI based hard trial removal method has statistically improved the average classification accuracy of the baseline deep CNN model from 63.77 % to 68.70 %, with p-value = 7.66 −11 for the subject specific MI classification. Additionally, analyzing the scalp map depicting the average relevance scores of correctly classified trials compared to a misclassified trial provides a deeper insight into identifying hard trials. The results indicates that the proposed quantitative based XAI approach outperforms the prediction-score based approach in hard trial identification.
Iytha Sridhar et al
Accurate lung segmentation in chest X-ray images plays a pivotal role in early disease detection and clinical decision-making. In this study, we introduce an innovative approach to enhance the precision of lung segmentation using the Segment Anything Model (SAM). Despite its versatility, SAM faces the challenge of prompt decoupling, often resulting in misclassifications, especially with intricate structures like the clavicle. Our research focuses on the integration of spatial at- tention mechanisms within SAM. This approach enables the model to concentrate specifically on the lung region, fostering adaptability to image variations and reducing the likelihood of false positives. This work has the potential to significantly advance lung segmentation, improving the identification and quantification of lung anomalies across diverse clinical contexts.
Sokol et al
Noise activity is known to affect neural networks, enhance the system response to weak external signals, and lead to stochastic resonance phenomenon that can effectively amplify signals in nonlinear systems. In most treatments, channel noise has been modeled based on multi-state Markov descriptions or the use stochastic differential equation models. Here we probe a computationally simple approach based on a minor modification of the traditional Hodgkin-Huxley approach to embed noise in neural response. Results obtained from numerous simulations with different excitation frequencies and noise amplitudes for the action potential firing show very good agreement with output obtained from well-established models. Furthermore, results from the Mann-Whitney U test reveal a statistically insignificant difference. The distribution of the time interval between successive potential spikes obtained from this simple approach compared very well with the results of complicated Pu-Thomas type methods at much reduced computational cost. This present method could also possibly be applied to the analysis of spatial variations and/or differences in characteristics of random incident electromagnetic signals.
Trending on Altmetric
Open all abstracts, in this tab
Pedro Manuel Mendes Correia et al 2024 Biomed. Phys. Eng. Express
Introduction. The positioning of γ ray interactions in positron emission tomography (PET) detectors is commonly made through the evaluation of the Anger logic flood histograms. Machine learning techniques, leveraging features extracted from signal waveform, have demonstrated successful applications in addressing various challenges in PET instrumentation. Aim. This paper evaluates the use of artificial neural networks (NN) for γ ray interaction positioning in pixelated scintillators coupled to a multiplexed array of silicon photomultipliers (SiPM). Methods. An array of 16 Cerium doped Lutetium-based (LYSO) crystal pixels (cross-section 2x2 mm2) coupled to 16 SiPM (S13360-1350) were used for the experimental setup. Data from each of the 16 LYSO pixels was recorded, a total of 160 000 events. The detectors were irradiated by 511 keV annihilation γ rays from a Sodium-22 (22Na) source. Another LYSO crystal was used for electronic collimation. Features extracted from the signal waveform were used to train the model. Two models were tested: i) single multiple-class neural network (mcNN), with 16 possible outputs followed by a softmax and ii) 16 binary classification neural networks (bNN), each one specialized in identifying events occurred in each position. Results. Both NN models showed a mean positioning accuracy above 85% on the evaluation dataset, although the mcNN is faster to train. Discussion. The method's accuracy is affected by the introduction of misclassified events that interacted in the neighbour's crystals and were misclassified during the dataset acquisition. Electronic collimation reduces this effect, however results could be improved using a more complex acquisition setup, such as a light-sharing configuration. Conclusions. The methods comparison showed that mcNN and bNN can surpass the Anger logic, showing the feasibility of using these models in positioning procedures of future multiplexed detector systems in a linear configuration.
Ravi Suppiah et al 2024 Biomed. Phys. Eng. Express
Physiological Signals like Electromography (EMG) and Electroencephalography (EEG) can be analysed and decoded to provide vital information that can be used in a range of applications like rehabilitative robotics and remote device control. The process of acquiring and using these signals requires many compute-intensive tasks like signal acquisition, signal processing, feature extraction and machine learning. Performing these activities on a PC-based system with well-established software tools like Python and Matlab is the first step in designing solutions based upon these signals. In the application domain of rehabilitative robotics, one of the main goals is to develop solutions that can be deployed for the use of individuals who need them in improving their Acitivities-for-Daily Living (ADL). To achieve this objective, the final solution must be deployed onto an embedded solution that allows high portability and ease-of-use. Porting a solution from a PC-based environment onto a resource-constraint one such as a microcontroller poses many challenges. In this research paper, we propose the use of a ARM-based Corex M-4 processor. We explore the various stages of the design from the initial testing and validation, to the deployment of the proposed algorithm on the controller, and further investigate the use of Cepstrum features to obtain a high classification accuracy with minimal input features. The proposed solution is able to achieve an average classification accuracy of 95.34% for all five classes in the EMG domain and 96.16% in the EEG domain on the embedded board.
Rishika Iytha Sridhar and Rishikesan Kamaleswaran 2024 Biomed. Phys. Eng. Express
Accurate lung segmentation in chest X-ray images plays a pivotal role in early disease detection and clinical decision-making. In this study, we introduce an innovative approach to enhance the precision of lung segmentation using the Segment Anything Model (SAM). Despite its versatility, SAM faces the challenge of prompt decoupling, often resulting in misclassifications, especially with intricate structures like the clavicle. Our research focuses on the integration of spatial at- tention mechanisms within SAM. This approach enables the model to concentrate specifically on the lung region, fostering adaptability to image variations and reducing the likelihood of false positives. This work has the potential to significantly advance lung segmentation, improving the identification and quantification of lung anomalies across diverse clinical contexts.
Matthew Sokol et al 2024 Biomed. Phys. Eng. Express
Noise activity is known to affect neural networks, enhance the system response to weak external signals, and lead to stochastic resonance phenomenon that can effectively amplify signals in nonlinear systems. In most treatments, channel noise has been modeled based on multi-state Markov descriptions or the use stochastic differential equation models. Here we probe a computationally simple approach based on a minor modification of the traditional Hodgkin-Huxley approach to embed noise in neural response. Results obtained from numerous simulations with different excitation frequencies and noise amplitudes for the action potential firing show very good agreement with output obtained from well-established models. Furthermore, results from the Mann-Whitney U test reveal a statistically insignificant difference. The distribution of the time interval between successive potential spikes obtained from this simple approach compared very well with the results of complicated Pu-Thomas type methods at much reduced computational cost. This present method could also possibly be applied to the analysis of spatial variations and/or differences in characteristics of random incident electromagnetic signals.
Anna Chiara Giovannelli et al 2024 Biomed. Phys. Eng. Express
Purpose
4D computed tomography (4DCT) is the clinical standard to image organ motion in radiotherapy, although it is limited in imaging breathing variability. We propose a method to transfer breathing motion across longitudinal imaging datasets to include intra-patient variability and verify its performance in lung cancer patients. 
Methods 
Five repeated control 4DCTs for 6 non-small cell lung cancer patients were combined into multi-breath datasets (m4DCT) by merging stages of deformable image registration to isolate respiratory motion. The displacement of the centre of mass of the primary tumour and its volume changes were evaluated to quantify intra-patient differences. Internal target volumes defined on the m4DCT were compared with those conventionally drawn on the 4DCT.
Results 
Motion analysis suggests no discontinuity at the junction between successive breaths, confirming the method's ability to merge repeated imaging into a continuum. Motion (variability) is primarily in superior-inferior direction and goes from 14.4 mm (8.7 mm) down to 0.1 mm (0.6 mm), respectively for tumours located in the lower lobes or most apical ones. On average, up to 65% and 74% of the tumour volume was subject to expansion or contraction in the inhalation and exhalation phases. These variations lead to an enlargement of the ITV up to 8% of its volume in our dataset.
Conclusion 
4DCT can be extended to model variable breathing motion by adding synthetic phases from multiple time-resolved images. The inclusion of this improved knowledge of patients' breathing allows better definition of treatment volumes and their margins for radiation therapy.
James J Sohn and Indra J Das 2024 Biomed. Phys. Eng. Express
Small field dosimetry presents unique challenges with source occlusion, lateral charged particle equilibrium and detector size. As detector volume decreases, signal strength declines while noise increases, deteriorating the signal-to-noise ratio (SNR). This issue may be compounded by triaxial cables connecting detectors to electrometers. However, effects of cables, critical for precision dosimetry, are often overlooked. There is a need to evaluate triaxial cable and detector impacts on SNR in small fields. To evaluate the influence of triaxial cables and microdetectors on signal-to-noise ratios in small-field dosimetry. This study also aims to establish the importance of cable quality assurance for measurement accuracy.Six 9.1 m length triaxial cables from different manufacturers were tested with six microdetectors . A 6 MV photon beam (TrueBeam) was used, with a water phantom at 5 cm depth with 0.5×0.5 cm2 to 10×10 cm2 fields at 600 MU/min. Readings were acquired using cable-detector permutations with a dedicated electrometer. Cables had differing connector types, conductor materials, insulation, and diameter. Detectors had various sensitive volumes, materials, typical signals, and bias voltages.
Normalized FOFs showed 13.4% and 4.6% variation across cables for 0.5×0.5 cm2 and 1×1 cm2 fields, respectively. The maximum estimated error between any cable-detector combinations was 0.2%. No consistent FOF trend was observed with increasing cable diameter, likely due to different types of detectors used. However, absolute FOF differences of 0.9% and 0.3% were noted between cables for 0.5×0.5 cm2 and 1×1 cm2 fields, respectively.
Regular triaxial cable quality assurance is critical for precision small field dosimetry. A national protocol is needed to standardize cable evaluations/calibrations, particularly for small signals from modern detectors. This could enhance measurement accuracy and treatment delivery with advanced small-field radiotherapy techniques that promise improved patient outcomes. Further studies should expand detector and cable models tested across institutions to establish robust quality control guidelines.
J E Parker et al 2024 Biomed. Phys. Eng. Express 10 045024
A study of burn thresholds from superficially penetrating radio-frequency (RF) energy at 8.2 and 95 GHz for swine skin was conducted. The study determined the thresholds for superficial, partial-thickness, and full-thickness burn severities after 5 seconds of exposure at power densities of 4–30 W/cm2 and 2–15 W/cm2 at 8.2 and 95 GHz, respectively. There were significant differences in he burn thresholds at the different severities between the two frequencies due to the large difference in energy penetration depths. Biopsies were collected from each burn site at 1, 24, 72, and 168 hr post exposure. Each sample was assessed by a burn pathologist against 20 histological factors to characterize the damage resulting from these RF overexposures. A one-dimensional, layered digital phantom that utilized realistic values for dielectric and thermal properties was used to explain some observed thresholds. The results of the heating and cooling response of the animal model and histology scores of each exposure are provided to enhance future efforts at simulation of RF overexposures and to establish damage thresholds.
Steven Squires et al 2024 Biomed. Phys. Eng. Express 10 045021
Purpose. To improve breast cancer risk prediction for young women, we have developed deep learning methods to estimate mammographic density from low dose mammograms taken at approximately 1/10th of the usual dose. We investigate the quality and reliability of the density scores produced on low dose mammograms focussing on how image resolution and levels of training affect the low dose predictions. Methods. Deep learning models are developed and tested, with two feature extraction methods and an end-to-end trained method, on five different resolutions of 15,290 standard dose and simulated low dose mammograms with known labels. The models are further tested on a dataset with 296 matching standard and real low dose images allowing performance on the low dose images to be ascertained. Results. Prediction quality on standard and simulated low dose images compared to labels is similar for all equivalent model training and image resolution versions. Increasing resolution results in improved performance of both feature extraction methods for standard and simulated low dose images, while the trained models show high performance across the resolutions. For the trained models the Spearman rank correlation coefficient between predictions of standard and low dose images at low resolution is 0.951 (0.937 to 0.960) and at the highest resolution 0.956 (0.942 to 0.965). If pairs of model predictions are averaged, similarity increases. Conclusions. Deep learning mammographic density predictions on low dose mammograms are highly correlated with standard dose equivalents for feature extraction and end-to-end approaches across multiple image resolutions. Deep learning models can reliably make high quality mammographic density predictions on low dose mammograms.
Esmaeil Mehrara 2024 Biomed. Phys. Eng. Express 10 045020
Thermoluminescent dosimeters (TLDs) serve as compact and user-friendly tools for various applications, including personal radiation dosimetry and radiation therapy. This study explores the potential of utilizing TLD-100 personal dosimetry, conventionally applied in PET/CT (positron emission tomography/computed tomography) settings, in the PET/MRI (magnetic resonance imaging) environment. The integration of MRI into conventional radiotherapy and PET systems necessitates ionizing radiation dosimetry in the presence of static magnetic fields. In this study, TLD-100 dosimeters were exposed on the surface of a water-filled cylindrical phantom containing PET-radioisotope and positioned on the patient table of a 3 T PET/MRI, where the magnetic field strength is around 0.2 T, aiming to replicate real-world scenarios experienced by personnel in PET/MRI environments. Results indicate that the modified MR-safe TLD-100 personal dosimeters exhibit no significant impact from the static magnetic field of the 3 T PET/MRI, supporting their suitability for personal dosimetry in PET/MRI settings. This study addresses a notable gap in existing literature on the effect of MRI static magnetic field on TLDs.
Wondesen T Gebreamlak and Hassaan H Alkhatib 2024 Biomed. Phys. Eng. Express 10 045017
Purpose. The aim of this study is to determine the planar dose distribution of irregularly-shaped electron beams at their maximum dose depth (zmax) using the modied lateral build-up ratio (LBR) and curve-fitting methods. Methods. Circular and irregular cutouts were created using Cerrobend alloy for a 14 × 14 cm2 applicator. Percentage depth dose (PDD) at the standard source-surface-distance (SSD = 100 cm) and point dose at different SSD were measured for each cutout. Orthogonal profiles of the cutouts were measured at zmax. Data were collected for 6, 9, 12, and 15 MeV electron beam energies on a VERSA HDTM LINAC using the IBA Blue Phantom2 3D water phantom system. The planar dose distributions of the cutouts were also measured at zmax in solid water using EDR2 films. Results. The measured PDD curves were normalized to a normalization depth (d0) of 1 mm. The lateral-buildup-ratio (LBR), lateral spread parameter (σR(z)), and effective SSD (SSDeff) for each cutout were calculated using the PDD of the open applicator as the reference field. The modified LBR method was then employed to calculate the planar dose distribution of the irregular cutouts within the field at least 5 mm from the edge. A simple curve-fitting model was developed based on the profile shapes of the circular cutouts around the field edge. This model was used to calculate the planar dose distribution of the irregular cutouts in the region from 3 mm outside to 5 mm inside the field edge. Finally, the calculated planar dose distribution was compared with the film measurement. Conclusions. The planar dose distribution of electron therapy for irregular cutouts at zmax was calculated using the improved LBR method and a simple curve-fitting model. The calculated profiles were within 3% of the measured values. The gamma passing rate with a 3%/3 mm and 10% dose threshold was more than 96%.