Biomedical Physics & Engineering Express最新文献

{"title":"External evaluation of a deep learning-based algorithm for automated lung tumor and esophagus segmentation on CBCT.","authors":"Zhehao Zhang, Dean Hobbis, Jue Jiang, Aneesh Rangnekar, Chloe Choi, Michael Waters, Harini Veeraraghavan, Geoffrey Hugo","doi":"10.1088/2057-1976/aeaf5c","DOIUrl":"10.1088/2057-1976/aeaf5c","url":null,"abstract":"<p><p>Online adaptive radiotherapy requires fast and accurate delineation of targets and organs at risk on cone-beam computed tomography (CBCT). Although deep learning (DL)-based CBCT auto-segmentation methods have been proposed, their clinical generalizability remains insufficiently validated. This study aimed to externally evaluate a pretrained DL model for lung tumor and esophagus segmentation on CBCT. Originally trained using C-arm linac CBCT images, the model was directly applied to an independent cohort of 12 lung cancer patients, each with 5 CBCT images acquired at a separate institution on an O-ring linac. Rigid propagation of planning CT contours served as a minimal baseline to determine whether the model matched or outperformed conventional rigid registration for gross tumor volume (GTV) and esophagus segmentation using geometric and dosimetric measures. Compared with rigid contour propagation, the model significantly improved GTV segmentation accuracy, with mean Dice similarity coefficient (DSC) increasing from 0.66 to 0.76. Significant differences in GTV D98% were observed between the model-generated contours and the reference contours (50.69 vs 53.58 Gy). The model showed a statistically non-significant decrease in performance compared with previously reported results on its original testing dataset (mean DSC = 0.84), which shared the same distribution as the training dataset. For esophagus segmentation, evaluation of 26 scans from 6 patients showed that the model achieved geometric accuracy within peri-target contour rings comparable to rigid propagation, with mean DSC of 0.72 and 0.70, respectively. Similar esophagus V32Gy values were observed between the model-generated and reference contours (0.06 vs 0.06 cc). Significantly reduced accuracy was observed compared with previously reported results, with mean DSC decreasing from 0.79 to 0.72. These findings highlight that DL-based CBCT auto-segmentation may suffer performance degradation under domain shifts across patient populations, clinical workflows, and imaging systems, emphasizing the necessity of external evaluation beyond the original development environment.</p>","PeriodicalId":8896,"journal":{"name":"Biomedical Physics & Engineering Express","volume":" ","pages":""},"PeriodicalIF":2.0,"publicationDate":"2026-10-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13662671/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149102852","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"When Bioinspiration Doesn't Help: Mechanical Behavior of Bio-inspired Curved Needles During Curvilinear Insertion.","authors":"Doyoung Kim, Parsaoran Hutapea","doi":"10.1088/2057-1976/aeb381","DOIUrl":"https://doi.org/10.1088/2057-1976/aeb381","url":null,"abstract":"<p><p>This study investigates the effectiveness of bio-inspired barbed geometries applied to curved needles for percutaneous procedures. The geometries of insect stingers, such as honeybees and mosquitoes, have been applied to straight needles to reduce insertion forces, tissue deformation, and tissue damage during linear insertion. However, whether these benefits extend to curved needles undergoing curvilinear base manipulation remains unclear. In this study, mosquito-inspired barb geometry was adapted onto bevel-tip curved needles with bent angles of 7.5°, 15°, 22.5°, and 30°. The effect on transverse and axial insertion forces, bending moments, and needle-tip trajectories was evaluated. The regular and bio-inspired curved needles were fabricated using high-resolution 3D printing and tested in tissue-mimicking gel under curvilinear insertion. Force and moment data were recorded at the needle base, and the needle tip paths were quantified using radius of curvature analysis from video tracking. No statistically significant reductions were found in insertion forces or bending moments under curvilinear insertion. These findings suggest that, for curved needles following predefined trajectories under the tested conditions, the measured forces did not change significantly with the addition of barbed surface features under curvilinear insertion. This suggests that the tested surface modification does not substantially alter the overall mechanical response under these conditions. This work highlights that friction-reducing surface features effective in linear insertion may not directly translate to curvilinear insertion, and that optimization strategies for curved needle applications may benefit from prioritizing geometric and trajectory-based design considerations.</p>","PeriodicalId":8896,"journal":{"name":"Biomedical Physics & Engineering Express","volume":" ","pages":""},"PeriodicalIF":2.0,"publicationDate":"2026-10-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149179226","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":"CT protocol optimisation for oncologic surveillance imaging: phantom-based comparison of multi-scan and single-scan acquisition strategies.","authors":"Desislava Kostova-Lefterova, Filip Simeonov, Desislava Hitova-Topkarova, Angel Yordanov, Virginia Payakova","doi":"10.1088/2057-1976/aeb383","DOIUrl":"https://doi.org/10.1088/2057-1976/aeb383","url":null,"abstract":"<p><p>Computed tomography (CT) plays an important role in oncologic surveillance after treatment of gynecological malignancies, including patients treated with vaginal electronic brachytherapy (eBT). This phantom study evaluated CT acquisition strategies for oncologic surveillance, focusing on radiation dose reduction and objective image quality in the cranial, thoracic, and upper abdominal regions assessed quantitatively. CT acquisition protocols were investigated using an anthropomorphic Kyoto Kagaku phantom. Two acquisition strategies were analysed: a multi-scan approach with separate acquisitions for the head, neck, thorax, and abdomen-pelvis, and a single-scan head-torso acquisition with region-specific reconstructions. Different exposure parameter settings and reconstruction kernels were evaluated. Objective image quality was assessed using semi-automated region-of-interest (ROI) measurements with a custom FIJI/ImageJ macro. Signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR), and Artifact Index (AI) were calculated for cranial, thoracic, and abdominal regions. For the single-scan protocols, CTDIvol decreased from 4.21 mGy in the standard protocol to 2.97-3.99 mGy in the optimized protocols, while DLP decreased from 410 mGy.cm to 287-388 mGy.cm. Across all anatomical regions, SNR and CNR remained comparable with the standard protocol. Thoracic CNR ranged from 2.68 to 3.01, abdominal CNR remained stable in both liver and aortic regions, and AI demonstrated only moderate variability across the optimized protocols. Optimized CT acquisition protocols reduced radiation dose while preserving the evaluated objective image quality metrics in the cranial, thoracic, and upper abdominal regions. Among the evaluated configurations, a reduced dose single scan head-torso acquisition using reduced tube voltage and tube current modulation showed a favourable balance between radiation dose reduction and objective image quality under the investigated phantom conditions and was identified as a candidate protocol for subsequent clinical and reader performance validation.</p>","PeriodicalId":8896,"journal":{"name":"Biomedical Physics & Engineering Express","volume":" ","pages":""},"PeriodicalIF":2.0,"publicationDate":"2026-10-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149178838","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":"Characterizing Altered Cardiac Autonomic Dynamics in Psychotic Disorders Through Multidomain Heart Rate Variability Analysis and Deep Learning.","authors":"Vivek Ranjan, Raghuwansh Singh, Anindita Ganguly, Suman Halder","doi":"10.1088/2057-1976/aeb384","DOIUrl":"https://doi.org/10.1088/2057-1976/aeb384","url":null,"abstract":"<p><p>Recent developments in wearable devices have opened the possibility of heart rate variability (HRV), a quantitative measure of autonomic nervous system (ANS) regulation, as a potential indicator for the early detection of psychiatric disorders. This study explores HRV for the detection of psychotic disorders by integrating multidomain HRV analysis with a hybrid deep learning technique. Polar H10 electrocardiogram (ECG) device was employed to collect RR interval data from 30 patients diagnosed with either schizophrenia (SZ) or bipolar disorder (BD), and 30 similar age healthy controls, with a duration of approximately 1.5-2 hours. HRV features, including time-domain, spectral, and nonlinear features such as the Poincaré plot and Recurrence Quantification Analysis (RQA), were extracted from this publicly available dataset. Group-wise statistical comparisons between the healthy and the psychotic disorders were performed using the Mann-Whitney U test with a significance level of p < 0.01. HRV indices such as Mean NN intervals, SDNN, RMSSD, triangular index, Cardiac Vagal Index, Poincaré plot measures (SD1 and SD2), Sample Entropy, and spectral power components (HF, LF, and Total Power), were significantly lower in the psychotic disorder group (p < 0.01), indicating attenuated autonomic modulation. All machine learning and deep learning models were conducted under a fully nested, subject-independent 5×4 cross-validation protocol in which Minimum Redundancy Maximum Relevance (mRMR) feature selection and hyper-parameter optimisation were performed exclusively inside each outer training fold. Gradient Boosting achieved the highest subject level accuracy of 88.33 ± 4.56, surpassing the best previously reported subject level accuracy of 83.33% on the same dataset. A hybrid CNN-BiLSTM-Attention network, applied to the same mRMR-selected features, achieved a competitive subject-level accuracy of 80.00±11.18%. These results demonstrate that a rigorously nested evaluation combined with multidomain HRV features yields reliable subject-level discrimination and offers both performance and interpretability.</p>","PeriodicalId":8896,"journal":{"name":"Biomedical Physics & Engineering Express","volume":" ","pages":""},"PeriodicalIF":2.0,"publicationDate":"2026-10-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149178840","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":"Pulsed Electromagnetic Field Driven Perturbation of Peripheral Arterial Hemodynamics: Integrated Evidence from Doppler Ultrasound and Photoplethysmographic Signal Analysis.","authors":"Ahmed Kiser, Tarikul Islam, Jubair Ibn Omar Hamza, Sumaiya Tabassum Dipty, Syed Salman Aziz, Md Mahbubur Rahman, Jamil Hasan Karami, Mohammad Rashal Chowdhury, Shahnaj Latif, Muhammad E H Chowdhury, Khondkar Siddique-E Rabbani, Zaid Mahbub","doi":"10.1088/2057-1976/aeb382","DOIUrl":"https://doi.org/10.1088/2057-1976/aeb382","url":null,"abstract":"<p><p>Pulsed electromagnetic fields (PEMF) have been widely studied for vascular and microcirculatory modulation, yet their effects on pulsatile hemodynamic organization remain poorly characterized. It is also unclear whether reversing field direction under otherwise comparable exposure conditions can alter vascular waveform behaviour. This study investigated acute direction-dependent responses to PEMF using Doppler ultrasound and photoplethysmography (PPG). Twenty healthy volunteers underwent PEMF exposure (10 Hz, 100 µs pulse width, 15 V) over the brachial artery with the coil in two opposing orientations, such that the local induced electric-field component along the arterial axis was directed either with (Forward) or against (Reverse) physiological blood flow. Downstream responses were assessed using radial artery Doppler and index-finger PPG. Participant-level feature summaries were analyzed using paired Wilcoxon signed-rank tests with Benjamini-Hochberg false-discovery-rate correction, including direct comparison of baseline-corrected Forward and Reverse responses, complemented by descriptive kernel-density L1 (KDE-L1) analysis. PPG showed the clearest participant-level responses. Six Forward features involving temporal, morphological, and pulse-area characteristics remained FDR-significant (q =0.018-0.037), whereas five Reverse features involving systolic amplitude, notch timing, and spectral organization remained significant (q = 0.003-0.012). Direct Forward-Reverse comparison identified significant differences in eight PPG features spanning amplitude, area, temporal, and spectral domains (q =0.005-0.029). Doppler showed a more selective response, with Forward Cardiac Cycle Duration remaining FDR-significant (q= 0.014), together with broader KDE-L1 redistribution of arterial velocity-waveform characteristics. These findings demonstrate orientation-sensitive peripheral pulse-wave responses together with complementary arterial waveform changes, supporting field direction as a potentially relevant exposure parameter in vascular PEMF research.</p>","PeriodicalId":8896,"journal":{"name":"Biomedical Physics & Engineering Express","volume":" ","pages":""},"PeriodicalIF":2.0,"publicationDate":"2026-10-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149179183","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":"Determination of argon metastable densities in an electrosurgical argon plasma by TDLAS.","authors":"Bastian Kogelheide, Alexander Böddecker, Ihor Korolov, Peter Awakowicz, Thomas Mussenbrock","doi":"10.1088/2057-1976/aeaeed","DOIUrl":"10.1088/2057-1976/aeaeed","url":null,"abstract":"<p><p>Electrosurgical argon plasma sources have recently shown promising clinical results for the treatment of cervical intraepithelial neoplasia without causing significant thermal tissue damage. Despite these findings, the underlying plasma-chemical mechanisms responsible for the therapeutic effects are still insufficiently understood, and quantitative data on key excited species under clinically relevant operating conditions remain scarce. In particular, argon metastable states are expected to play an important role in energy transfer and reactive species formation. In this work, tunable diode laser absorption spectroscopy is employed to determine the temporal evolution of the argon metastable state densities Ar(1s) and Ar(1s) in a clinically applied electrosurgical argon plasma with sub-microsecond temporal resolution. Metastable densities in the order offor Ar(1s) andfor Ar(1s) are observed. Both states exhibit pronounced temporal modulations within the discharge pulse, indicating the importance of stepwise excitation and de-excitation processes in the discharge kinetics. The measured densities are in good agreement with those reported in previous studies of similar systems. In addition, pressure-broadened absorption profiles are used to estimate characteristic electron densities and gas temperatures from Stark and van der Waals broadening, respectively. The results provide quantitative insight into the transient plasma chemistry of electrosurgical argon plasmas and support future modelling of reactive species generation relevant for biomedical applications.</p>","PeriodicalId":8896,"journal":{"name":"Biomedical Physics & Engineering Express","volume":" ","pages":""},"PeriodicalIF":2.0,"publicationDate":"2026-10-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149103858","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":"Repeat-consistency learning for improving repeatability of brain-age prediction from structural MRI.","authors":"Lulu Wang","doi":"10.1088/2057-1976/aeaeec","DOIUrl":"10.1088/2057-1976/aeaeec","url":null,"abstract":"<p><p>Structural magnetic resonance imaging (MRI)-based brain-age prediction is increasingly used as an imaging-derived marker of brain ageing, but good predictive accuracy does not guarantee stable estimates when the same individual is scanned repeatedly. This study proposes repeat-consistency learning (RCL), a training framework that uses same-subject, same-session repeated MRI acquisitions as paired training information. A compact three-dimensional convolutional neural network was trained using a chronological-age prediction loss together with output-level and latent feature-level consistency losses. An unweighted formulation treated all repeated-scan pairs equally, whereas a disagreement-weighted formulation reduced the influence of pairs with larger predicted-age differences. The framework was evaluated using the Open Access Series of Imaging Studies cross-sectional dataset with subject-level training, validation and held-out test partitions. Both RCL formulations showed numerically lower scan-rescan disagreement and narrower Bland-Altman limits of agreement than prediction-only training, while held-out age-prediction performance remained broadly comparable. The unweighted formulation provided the most favourable overall balance between repeatability and predictive accuracy. RCL also reduced prediction shifts under several controlled noise and affine perturbations, although substantial sensitivity to image blurring remained. These findings provide proof-of-concept evidence that repeated MRI acquisitions can be incorporated directly into model training to encourage more stable predictions and latent representations while maintaining broadly comparable chronological-age prediction performance. The observed improvements were modest and were demonstrated in a controlled, single-dataset, within-session setting. Validation in larger, multi-site, multi-scanner and longitudinal datasets is therefore required to determine the broader robustness, generalisability and practical utility of the approach.</p>","PeriodicalId":8896,"journal":{"name":"Biomedical Physics & Engineering Express","volume":" ","pages":""},"PeriodicalIF":2.0,"publicationDate":"2026-10-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149103589","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":"Integrating radiomic features and spatial semantic features from multiparametric MRI for IDH genotyping in glioma.","authors":"Yang Xi, Bingjie Shi, Qingzhu Wang, Xiaopeng Lv, Xuebing Jing","doi":"10.1088/2057-1976/aeaeef","DOIUrl":"10.1088/2057-1976/aeaeef","url":null,"abstract":"<p><p>Gliomas, one of the most prevalent primary tumors of the central nervous system, have IDH molecular subtypes that are strongly linked to patient survival, prognosis, and treatment plans. However, there are currently few noninvasive preoperative prediction techniques, and postoperative pathology and genetic testing are the main approaches used to assess IDH status. Radiomics features can quantitatively characterize intratumoral heterogeneity, whereas spatial semantic features may provide additional information related to tumor location and macroscopic growth patterns. When it comes to characteristics like glioma invasiveness, growth patterns, and angiogenesis, spatial semantic features can enhance radiomics features. In order to predict IDH molecular subtypes in gliomas, this study suggests a combined analytical approach that incorporates clinical data, spatial semantic features, and multiparametric magnetic resonance imaging (MRI) radiomic features. This technique uses the IBSI standard to extract radiomic data from raw multiparametric MRI images and combines morphological analysis, gradient statistics, and 3D geometric modeling to create spatial semantic information. To eliminate redundant features and account for multicollinearity, Pearson correlation analysis and the variance inflation factor were utilized. The best feature subset was then obtained through a multi-stage feature selection procedure that combined random forests (RFs), LASSO regression, and recursive feature elimination. Lastly, imaging characteristics and clinical data were combined using a RF model to identify IDH molecular subtypes in gliomas. According to experimental findings, the suggested model's area under the ROC curve (AUC) in a five-fold cross-validation was 96.08%. The model had the AUC of 95.62%, ACC of 91.67%, sensitivity of 88.89%, and specificity of 93.94% in the final test set. The limitations of radiomics in conveying tumor macroscopic development patterns and spatial invasion relationships can be somewhat compensated for by spatial semantic aspects, according to additional investigation. The source code at:.</p>","PeriodicalId":8896,"journal":{"name":"Biomedical Physics & Engineering Express","volume":" ","pages":""},"PeriodicalIF":2.0,"publicationDate":"2026-10-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149103434","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":"Optimization of selected SPECT system parameters for proton range verification: a Monte Carlo study.","authors":"S Gharib, P Taherparvar","doi":"10.1088/2057-1976/aeae43","DOIUrl":"10.1088/2057-1976/aeae43","url":null,"abstract":"<p><p>This Monte Carlo study investigated a dual-head SPECT-based system for prompt-gamma (PG) range verification in proton therapy, focusing on energy-window and collimator septal thickness effects on PG emission profiles and distal dose fall-off. Using GATE, correlations between PG emissions (4.44, 5.24, 6.13 MeV) and proton depth-dose distributions were evaluated for 90-180 MeV beams in a PMMA phantom. Narrow (4.2-4.6, 5.0-5.4, 5.9-6.3 MeV) and broad (4.2-6.3 MeV) windows were studied. A dual-head SPECT with parallel-hole collimator (1-5 mm septa) was simulated. Image reconstruction was performed with FBP and with OSEM including attenuation correction, using STIR 6.2. The 5.9-6.3 MeV window showed the highest spatial correlation with proton range (Δ ≈1 mm), but its low yield limited applicability. The 4.2-4.6 and 4.2-6.3 MeV windows balanced detection yield and accuracy (Δ ≈2-3 mm). Septal thicknesses ⩾4 mm produced shadowing artefacts, whereas 3 mm gave the most accurate PG localization for the 90 MeV proton beam studied. For the 4.2-4.6 and 4.2-6.3 MeV windows, the PG localization errors (relative to the true PG emission profile) were 1.5 ± 0.5 and 1.4 ± 0.4 mm with FBP, and 1.9 ± 0.3 and 1.7 ± 0.4 mm with OSEM, despite ∼20% lower PG yield than the 1 mm septum. Overall, appropriate window and septal thickness provided PG localization accuracy of 1-2 mm. A 3 mm septum with either the 4.2-4.6 or 4.2-6.3 MeV window offers a favourable trade-off between spatial accuracy and detection efficiency under the 90 MeV condition investigated here, highlighting the potential of optimized SPECT systems for proton range monitoring at this energy.</p>","PeriodicalId":8896,"journal":{"name":"Biomedical Physics & Engineering Express","volume":" ","pages":""},"PeriodicalIF":2.0,"publicationDate":"2026-10-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149103455","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":"Mamba-EdgeNet: learnable edge-guided state space model for skin lesion segmentation.","authors":"Jiaming Xu, Qi Mao, Lei Qiu, Yu Chen, Gongshuang Tao","doi":"10.1088/2057-1976/aeada8","DOIUrl":"10.1088/2057-1976/aeada8","url":null,"abstract":"<p><p>Skin lesion segmentation is one of the key tasks in computer-aided diagnosis. However, accurate boundary delineation remains challenging because of low contrast, irregular boundaries, and visual artifacts in skin lesion areas. To address these problems, we developed Mamba-EdgeNet. Mamba-EdgeNet is a hybrid encoder-decoder framework that selectively integrates ResNet50 with EfficientMamba2D state-space blocks. A resolution-adaptive scanning strategy dynamically alternates between multi-directional and sequential spatial scanning, enabling long-range contextual modeling with linear complexity relative to sequence length. To delineate fuzzy borders, a learnable edge-guidance branch is constructed. Structural priors are generated using a learnable edge enhancer and a shallow edge extractor and are subsequently refined through interactions with deep semantic features. A gated fusion decoder adaptively integrates these edge-aware cues with hierarchical semantic features to resolve feature ambiguity and restore fine-grained contours. Evaluations on the official ISIC 2016 and ISIC 2018 splits, together with an internal evaluation using a nonstandard custom partition of ISIC 2017, indicate that the proposed architecture achieves competitive segmentation performance relative to recent representative models across multiple evaluation metrics. Controlled ablation studies further support the contributions of the functional components and loss supervision strategies. Additionally, in-domain evaluation on PH2 and bidirectional cross-dataset transfer between PH2 and ISIC 2018 provide complementary assessments of segmentation performance in within-dataset and cross-dataset settings. These findings indicate that Mamba-EdgeNet is a competitive approach for automated skin lesion segmentation with improved boundary representation.</p>","PeriodicalId":8896,"journal":{"name":"Biomedical Physics & Engineering Express","volume":" ","pages":""},"PeriodicalIF":2.0,"publicationDate":"2026-10-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149079532","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}