共找到 20 条结果
The development of modern digital technologies in the field of three-dimensional scanning and the construction of three-dimensional models of human body parts are relevant tasks in conducting forensic medical (medical-criminalistic) examinations for identifying the identity of deceased persons and for monitoring the long-term preservation of research objects, particularly the face of a deceased person. The purpose of the study is to experimentally evaluate the reproducibility of the results of the 3D face scanning methodology for determining some criteria for preserving the volume of soft tissues and the relief of skin of a model biological object. In accordance with the previously developed method of 3D scanning of body parts of an embalmed human body, three-dimensional images of the face part of the head of a model biological object, which is on long-term preservation with ensuring the preservation of the appearance, were obtained. To obtain three-dimensional images of the surface of the object, a 3D scanner Stereoscan 5MP (Breuckmann GmbH, Germany) was used, and for subsequent processing, comparison and analysis of images, specialized computer programs Optocat 2007 (Breuckmann GmbH, Germany) and GOM Inspect V8 (GOM, Germany) were used. The evaluation of the obtained three-dimensional images and the three-dimensional models based on them showed high reproducibility of the method's results, which allowed us to determine a set of criteria for preserving the volume of soft tissues and the skin relief of the face part of the embalmed human body during its long-term preservation, namely, the acceptable average and maximum deviations in the geometry of the three-dimensional images and 3D-models of the face. Развитие современных цифровых технологий в области трехмерного сканирования, построение трехмерных моделей частей тела человека являются актуальными задачами при проведении судебно-медицинских (медико-криминалистических) экспертиз при идентификации личности погибших и для контроля длительной сохранности объектов исследования, в частности лица умершего человека. Экспериментальная оценка воспроизводимости результатов методики 3D-сканирования лица для определения некоторых критериев сохранности объема мягких тканей и рельефа кожного покрова модельного биологического объекта. В соответствии с ранее разработанной методикой 3D-сканирования частей тела бальзамированного тела человека получали трехмерные изображения лицевого отдела головы модельного биологического объекта, находящегося на длительном сохранении с обеспечением сохранности облика. Для получения трехмерных изображений поверхности объекта использовали 3D-сканер Stereoscan 5МП (Breuckmann GmbH, Германия), для последующей обработки, сравнения и анализа изображений применяли специализированные компьютерные программы Optocat 2007 (Breuckmann GmbH, Германия) и GOM Inspect V8 (GOM, Германия). Оценка полученных трехмерных снимков и разработанных на их основе трехмерных моделей показала высокую воспроизводимость результатов методики, что позволило определить ряд критериев сохранности объема мягких тканей и рельефа кожного покрова лицевого отдела головы бальзамированного тела человека при его длительном сохранении, а именно: допустимые среднее и максимальное отклонения геометрии трехмерных снимков и 3D-моделей лица.
Early and accurate detection of Tetralogy of Fallot (TOF), along with proper risk stratification management, is critical for improving patient survival and prognosis. We developed an end-to-end automated framework for TOF, aimed at supporting decisions from preoperative diagnosis through postoperative risk prediction. A total of 1986 filtered participants (1018 healthy controls, 480 TOF mimics, and 488 patients with TOF) from four centres were recruited for the development and validation of DynaTOF, an integrated diagnostic and predictive system. The DynaTOF system comprises: (1) an echocardiographic view classification module built on ResNet-18; (2) a key diameter localisation and calculation module constructed with HRNet and a custom composite loss combining heatmap loss with geometric constraint loss; (3) a multimodal TOF diagnostic module that integrates a ResNet-LSTM-based video feature extractor for echocardiographic videos and a Transformer-based feature extractor for key diameters; (4) a time-aware postoperative prediction module, implemented with a GNN (Graph Neural Network), which estimates postoperative abnormal score dynamics based on preoperative video data, key diameters, surgical type, and specific postoperative time; and (5) a risk-stratification module that employs a Random Forest classifier to differentiate between high- and low-risk patients using the predicted abnormal score series. The view classification module achieved AUC values of 0.999, 0.999, 0.998, and 0.998 for classifying the Apical Four-Chamber (A4C), Apical Five-Chamber (A5C), Parasternal Short-Axis (PSAX), and Parasternal Long-Axis (PLAX) views, respectively. The key diameter localisation and calculation module demonstrated R2 values of 0.98, 0.76, and 0.97 for the prediction of LVD (left ventricular diameter), RVD (right ventricular diameter), and MPAD (main pulmonary artery diameter). The multimodal diagnostic module exhibited excellent performance in identifying TOF, with an accuracy of 0.910 (95% CI 0.881-0.938), an AUC of 0.989 (95% CI 0.977-0.992), a precision of 0.893 (95% CI 0.860-0.927), and a recall of 0.892 (95% CI 0.856-0.927), surpassing all single-modality approaches. The time-aware prediction module showed a high correlation (R2 = 0.852) between predicted and observed postoperative abnormal scores. Finally, the risk stratification module achieved an AUC of 0.904 for distinguishing between high-risk and low-risk patients. DynaTOF enables efficient diagnosis of TOF and provides personalised abnormal dynamics after operation, facilitating early screening and longitudinal monitoring. This system holds promise for improving comprehensive clinical care for infants with TOF. This work was supported by Shanghai Municipal Education Commission (No. 2024AIYB010), Fundamental Research Funds for the Central Universities (YG2025LC03), Shanghai Special Fund for Promoting High-Quality Industrial Development - Pilot Industry Innovation Development (AI Special Topic) Project (No. 2025-GZL-RGZN-02078), National Key Research and Development Program of China (2025YFC2511603), Shenzhen Medical Research Special Project clinical multi-center study (No. C2405001), the Science and Technology Commission of Shanghai Municipality (STCSM) (Grant No. 23JS1400700; 24JS2840200; 25JS2850100), the Innovative Research Team of High-Level Local Universities in Shanghai, and the Sanya Science and Technology Special Fund (No. 2022KJCX41).
Accurate assessment of the Ki-67 proliferation index (PI) is essential for grading and prognostication of gastrointestinal well-differentiated neuroendocrine tumours (NETs). While manual counting (MC) of 500-2000 tumour cells remains the standard, digital image analysis (DIA) offers potential advantages in efficiency and reproducibility. We evaluated the comparability of open-source DIA platforms on camera-captured (CC) images and whole-slide images (WSI) for Ki-67 quantification. Ki-67 hotspot areas of 70 NETs were photographed using a microscope-mounted camera. PI was determined by MC (gold standard) and compared with automated counts in 68 cases (two excluded owing to high background staining) using QuPath (V.0.4.4). In a randomly selected subset of 20 cases, the same hotspot areas were analysed using ImageJ, ChatGPT V.4.0 (colour-based segmentation) and the IHCexpert.com platform. Additionally, WSI files of these 20 cases were imported into QuPath for DIA; PI of identical areas were compared against static images. DIA using QuPath (on CC images) demonstrated excellent agreement with MC (intraclass correlation coefficient). Only one case showed grade reclassification (manual G1, 2.92%; DIA G2, 3.38%). In the subset analysis (n=20), comparable Ki-67 indices were observed across all digital platforms and between CC images and WSI. Grade switches from changes in Ki-67 PI were observed in two additional cases (G2 to G1 in IHCexpert.com group and G1 to G2 in ChatGPT group). Our findings offer the prospect of eliminating variability in the analysis of PI estimation. Of note, CC images yield results similar to WSI, supporting broader applicability in resource-limited practice settings.
To explore the perspectives of patients and healthcare professionals (HCP) in a specialist oncology centre regarding independent patient access to imaging results through Digital Health Records (DHR). A single-centre mixed methods service evaluation (SE) was developed and conducted at a tertiary oncology centre. The service evaluation comprised online questionnaires and semi-structured interviews with oncology patients and three HCP groups involved in cancer care. Qualitative data was analysed using an inductive thematic analysis, with quantitative data analysis consisting of descriptive summary statistics and tests for group differences. A total of 131 questionnaires (Patients n = 41; Radiographers n = 35; Radiologists n = 14; Oncology team n = 41) and 32 semi-structured interviews were completed. Patient qualitative analysis highlighted three main themes, i) information access - oncology patients want direct access to their imaging information; ii) autonomy - access to information as per their preferences, preferably prior to clinic appointments and iii) accessibility - imaging reports and information should be clear and understandable. HCP's overall qualitative analysis underlines a support towards independent patient access, however, with concerns raised, particularly among radiologists, about potential misinterpretation leading to increased patient anxiety. Quantitative analysis demonstrated that most patients (98%) favoured independent access to reports and images, with 76% preferring immediate availability. In contrast, 80% of HCP recommended access only after clinical discussion. Lay summary report formats were viewed positively by 85% of patients and 63% of HCP. Patients and HCP agreed that sharing imaging reports is desirable, however they differed in views regarding timing of release, highlighting the need to balance patient preferences with clinical context. As DHR adoption expands alongside independent patient access, these insights can inform the design of accessible, safe, and patient-centred imaging data sharing in oncology.
The 2023 iteration of the Global Burden of Diseases, Injuries, and Risk Factors Study (GBD) estimated prevalence, incidence, and health burden for 375 diseases and injuries, including 12 mental disorders. We assess past, current, and emerging trends in the prevalence and burden of mental disorders across sexes and age groups, for 21 regions, 204 countries and territories, and by Socio-demographic Index (SDI) quintile, from 1990 to 2023. Mental disorders included in GBD 2023 were anxiety disorders, major depressive disorder, dysthymia, bipolar disorder, schizophrenia, autism spectrum disorders, conduct disorder, attention-deficit hyperactivity disorder, anorexia nervosa, bulimia nervosa, idiopathic developmental intellectual disability, and a residual category of other mental disorders. A literature review identified epidemiological data for each disorder. These were analysed via a Bayesian meta-regression to estimate prevalence by disorder, sex, age, location, and year. Disorder-specific prevalence was multiplied by disability weights representing the severity of health loss associated with each disorder to estimate years lived with disability (YLDs). Deaths due to anorexia nervosa were assessed with a Cause of Death Ensemble modelling strategy to estimate deaths by sex, age, location, and year, and then multiplied by the standard life expectancy at age of death to estimate years of life lost (YLLs). YLDs equalled disability-adjusted life-years (DALYs) for all mental disorders except anorexia nervosa (the only mental disorder considered as an underlying cause of death in GBD), for which DALYs represented the sum of YLDs and YLLs. We presented prevalence, deaths, YLDs, YLLs, and DALYs as counts, age-specific rates per 100 000 population, and age-standardised rates per 100 000 population. We estimated 1·17 billion (95% uncertainty interval 1·06-1·31) prevalent cases of mental disorders globally in 2023, equivalent to an age-standardised prevalence rate of 14 210·7 cases (12 849·5-15 940·1) per 100 000 population. These estimates represented a 95·5% (75·0-121·2) increase in prevalent cases and 24·2% (11·4-41·4) increase in age-standardised prevalence rate between 1990 and 2023. All mental disorders showed increases in prevalent cases between 1990 and 2023, while notable increases were seen in age-standardised prevalence rates for anxiety disorders, major depressive disorder, dysthymia, anorexia nervosa, bulimia nervosa, schizophrenia, and conduct disorder. There were an estimated 171 million (127-228) DALYs due to mental disorders globally across sex and age in 2023, equivalent to an age-standardised DALY rate of 2070·5 DALYs (1519·1-2750·5) per 100 000 population. Mental disorders contributed to 6·1% (4·8-7·6) of all-cause DALYs in 2023, making them the fifth leading cause of global DALYs (up from 12th in 1990). DALYs were almost entirely composed of YLDs. Mental disorders were the leading cause of YLDs in 2023 (up from second in 1990), explaining 17·3% (14·8-20·6) of all-cause global YLDs. Leading causes of mental disorder DALYs were anxiety disorders (ranked 11th among the 304 diseases and injuries at Level 4 of the GBD cause hierarchy), major depressive disorder (15th), and schizophrenia (41st). Globally in 2023, mental disorder age-standardised DALY rates were higher among females (2239·6 [1643·7-3014·1] per 100 000) than among males (1900·2 [1399·8-2510·8] per 100 000), and peaked in the 15-19 years age group (2617·3 [1850·6-3696·8] per 100 000). All locations showed increased mental disorder DALY rates in 2023 compared with 1990, ranging across countries and territories from 1302·4 (952·7-1683·7) per 100 000 in Viet Nam to 3555·8 (2661·9-4715·0) per 100 000 in the Netherlands. Across SDI quintiles, DALY rates ranged from 1853·0 (1352·1-2469·3) per 100 000 for middle SDI to 2184·1 (1606·1-2890·3) per 100 000 for high SDI. A significant health burden was imposed by mental disorders in all countries and territories in 2023, irrespective of the health resources available. In some instances, this burden has increased over time and is unevenly distributed across populations. Stronger surveillance systems, particularly in low-income and middle-income countries, are required. Additionally, we need more coordinated and inclusive policies to reduce the burden through early treatment and prevention, tailored to sex and age differences across locations. Responding to the mental health needs of our global population, especially those most vulnerable, is an obligation, not a choice. Gates Foundation, Queensland Health, and University of Queensland.
Obtaining negative margins at the time of segmental mastectomy (SM) is important to reduce risk of local recurrence and avoid the need for a second surgery due to positive margins. Our current institutional standard practice (ISP) includes intraoperative assessment of the gross specimen and specimen radiographs by a multidisciplinary team including radiologists and pathologists. In this study, we utilized digital breast tomosynthesis (DBT) images of SM specimens in the operating room and recorded when additional (selective) shave margins were proposed by the surgeon. A breast radiologist later reviewed the archived DBT images. The timing and accuracy of surgeon and breast radiologists' margin assessment using DBT was then compared to our ISP. There were 193 patients enrolled and included in the analysis. Of 196 SM specimens, 9.7% (n = 19) had positive margins prior to excision of selective shave margins. Of these, 16 were identified by ISP with a sensitivity of 84%, specificity of 29%, false-negative rate (FNR) of 16%, positive predictive value (PPV) of 11%, and negative predictive value (NPV) of 95%. Surgeon assessment of DBT images identified 16/19 specimens with positive margins, with a sensitivity of 84% (p > 0.05), specificity of 49% (p < 0.001), FNR of 16% (> 0.05), PPV of 15% (p > 0.05), and NPV of 97% (p > 0.05). The median time for surgeon interpretation was 6 (range 2-35) minutes vs. 33 (range 15-88) for ISP. Surgeon interpretation of DBT images offers an alternative to a more time- and labor-intensive ISP for detecting positive margins during breast-conserving surgery, with comparable accuracy and higher sensitivity.
Background Radiologist expertise plays a critical role in breast cancer screening outcomes; yet approximately 70% of screening mammography interpretations in the United States are performed by general radiologists (generalists) rather than breast imaging specialists (specialists). Purpose To evaluate the impact of a multistage artificial intelligence (AI)-driven workflow on the clinical performance of generalists and fellowship-trained specialists. Materials and Methods This prospective study included screening mammogram interpretations from radiologists across 109 U.S. imaging facilities performed between September 2021 and December 2022. Only radiologists who interpreted digital breast tomosynthesis screening examinations during both study periods were included, and only bilateral examinations from women aged 35 years or older were included. The multistage AI-driven workflow integrated a computer-aided detection and diagnosis device and a separate "Safeguard Review" that routes AI-identified suspicious screening mammogram examinations that were not recalled for additional expert review. Adjusted cancer detection rate (CDR), positive predictive value (PPV) of recalls, and recall rate (RR) were compared before and after introduction of the multistage AI-driven workflow using logistic regression with generalized estimating equations. Results A total of 95 radiologists (60 generalists with a median of 19 years of experience [IQR, 14-31 years; range, 3-50 years] and 35 specialists with a median of 11 years of experience [IQR, 6-21 years; range, 1-32 years]) interpreting 577 742 examinations were included. Adjusted results show that the CDR of generalists increased from 3.76 (95% CI: 3.46, 4.08) to 4.99 (95% CI: 4.51, 5.53; P < .001) cancers per 1000 examinations with the AI workflow. The CDR of specialists was similar before and after adding the AI workflow (from 4.47 to 4.76 per 1000 examinations, respectively; P = .33) and was similar to that of generalists with AI (P = .53). Among generalists, PPV of recalls increased by 15.09% (from 3.38% [95% CI: 3.03, 3.75] to 3.89% [95% CI: 3.39, 4.46]; P = .02), indicating more efficient cancer detection (more cancers detected per recall) despite a 14.79% relative increase in RR (from 9.06% to 10.40%, P = .007). Specialists showed no change in PPV of recalls with the AI workflow (P = .86). Conclusion A multistage AI-driven workflow was associated with substantially improved CDR and PPV of recalls for generalists, which were on par with those of specialists. © RSNA, 2026 Supplemental material is available for this article. See also the editorial by Schiaffino and Cozzi in this issue.
This article focuses on the series of viewpoints on the prevention and treatment of medication-related osteonecrosis of the jaw issued by the American Association of Oral and Maxillofacial Surgery (AAOMS), and puts forward several uncertain issues from a clinical perspective. The points of doubt in diagnosis include the definition of drug categories and the threshold of bone exposure duration, as well as the significance of imaging features of bone lesions for diagnostic and treatment decisions. The concerns in terms of prevention include drug risks and disease incidence, tooth extraction and drug holiday, intervention measures promoting bone wound healing and their effectiveness. In terms of treatment, with integration of the author's practical experience, the discussion points focused on the dominant principles of non-surgical treatment, the identification and confirmation of the surgical bone incision boundary, the application of buccal fat pads, chin flaps and submandibular gland transposition, as well as the applicable conditions for permanent repair of fibular transplantation, temporary repair of reconstruction plate bridging, compromise mandibular resection and maxillary sinus opening, etc. The article suggests seeking evidence-based research on the above issues. 本文围绕美国口腔颌面外科学会（AAOMS）发布的关于药物相关性颌骨坏死防治的系列意见书，从临床角度对其带有不确定性的若干问题提出商榷意见。在诊断方面的质疑点包括对药物类别和骨暴露持续时间阈值的界定、骨病变影像学特征对诊断和治疗决策的意义；在预防方面的关注点包括药物风险与疾病发生率、拔牙与药物假期、促进骨创愈合的干预措施及其有效性；在治疗方面融入了笔者的实践体会，讨论点集中于非手术治疗的主导原则，手术中对手术切骨边界的确认，颊脂垫、颏瓣和颌下腺转位的应用，以及腓骨移植永久性修复、重建板桥接暂时性修复、姑息性下颌骨切除和上颌窦开放适用条件等。文章建议对上述问题寻求循证研究。.
Architectural distortion (AD) is a subtle but clinically important mammographic finding. With the adoption of synthesized 2D mammography (s2D) and digital breast tomosynthesis (DBT), AD detection has increased, often with a reduced positive predictive value for malignancy. To evaluate AD visibility on s2D versus DBT, assess its predictive value for malignancy compared with benign or high-risk lesions, assess interreader agreement, and analyze biopsy outcomes based on visibility. This retrospective study included 297 BI-RADS 4 or 5 AD cases detected on combined s2D + DBT mammography (2017-2018). Three breast radiologists independently assessed AD visibility on s2D and DBT. Pathologic outcomes were categorized as malignant, high-risk, or benign. Visibility patterns were correlated with pathology, positive predictive values (PPVs) were calculated by subgroup, and interreader agreement was assessed using Cohen's kappa. Of 297 architectural distortion (AD) cases, 33% were malignant (99/297), 30% high-risk (89/297), and 37% benign (109/297), yielding an overall PPV of 33%. When analyzed by visibility group, 23.1% of DBT-only ADs were malignant compared with 50.5% of ADs visible on at least one synthesized 2D (s2D) view (p < 0.001). Benign outcomes were significantly associated with decreasing s2D visibility (p = 0.03), whereas malignancy showed no significant association with increasing s2D + DBT visibility (p = 0.93). s2D-visible ADs demonstrated higher and more consistent PPVs for malignancy (approximately 48-54%) compared with DBT-only ADs (approximately 21-25%). Interreader agreement was moderate for s2D (κ = 0.46) and fair for DBT-only ADs (κ = 0.38). Architectural distortion visible on s2D and DBT is associated with a higher likelihood of malignancy than DBT-only distortion; however, the nontrivial malignancy rate of DBT-only lesions limits the reliability of pre-biopsy visibility as a standalone discriminator. While DBT enhances detection, it also increases biopsies of benign and high-risk lesions, highlighting the essential role of radiologic-pathologic correlation in management decisions.
To describe the requirements for modeling the tele-ICU platform, its development, the process, software, and hardware technologies employed. We also present its performance based on a proof-of-concept in a remote intensive care unit environment. This multicenter, prospective implementation study was conducted in three Level III Intensive Care Units in distinct Brazilian regions between June 2021 and December 2022. The INTEGRARE®, an Internet of Medical Things-based hardware/software architecture enabling agnostic integration of multiparameter monitors and mechanical ventilators, was deployed to provide continuous high-frequency telemetry, a unified analytical dashboard, and synchronous audiovisual communication for tele-round sessions. The proof-of-concept comprised continuous multimodal monitoring, collaborative tele-round discussions, and structured knowledge transfer. Nineteen months of multiparameter monitors and mechanical ventilators data were processed through a four-step data workflow (edge server, cloud storage, preprocessing, and analytical layer). All clinical, operational, and performance metrics were automatically generated by the platform. The INTEGRARE® enabled integration of multiparameter monitors and mechanical ventilator devices across 30 intensive care unit beds, generating over 2 billion data points with a median acquisition frequency of ~1 second and cross-device synchronization under 5 seconds. A total of 361 patients were monitored (7,235 intensive care unit-days), with a median intensive care unit stay of 14 days and 11 days on mechanical ventilators among ventilated patients. Tele-round sessions completed 484 hours, with a median of 3 hours and 39 minutes per intensive care unit per week. The platform supported real-time visualization and retrospective review of physiological curves, ventilator mechanics, laboratory results, and imaging within a unified dashboard. High adherence to tele-round routines was observed across multidisciplinary teams, who spontaneously incorporated multimodal telemetry into case discussions. Immediate bedside impact was common, including rapid identification and correction of patients' ventilator asynchrony. The Tele-UTI Conectada model provides a technical infrastructure that supports standardization, situation awareness, and collaborative clinical decision-making by integrating real-time telemetry with synchronous audiovisual interaction across heterogeneous intensive care units. Multidisciplinary teams successfully adopted the platform and routinely used it to conduct structured case discussions.
This study aimed to determine whether flagship smartphones can approach the performance of professional digital single-lens reflex (DSLR) cameras using a standardized workflow incorporating color calibration and optical zoom. Three DSLR cameras (Canon EOS 5D Mark IV, Canon EOS 80D, Nikon D610) and two smartphones (iPhone 17 Pro Max, Galaxy S24 Ultra) were used to capture nine standardized extraoral and intraoral views for each of 25 volunteers. Images were evaluated for color accuracy, dimensional accuracy, and image quality. Statistical analyses were conducted using one-way repeated-measures analysis of variance and paired t-tests, with Bonferroni correction applied for multiple comparisons (α = 0.05). Gray-card calibration significantly reduced smartphone image ΔE values (P < 0.001), resulting in lower ΔE values than those of the DSLR group with standardized white balance (P < 0.001). Regarding dimensional accuracy, images captured with the iPhone 17 Pro Max at 4× optical zoom showed no significant difference from DSLR cameras (P = 0.178), whereas the Samsung device significantly underestimated arch width (P = 0.041). Samsung achieved the most favorable BRISQUE score. Under a standardized workflow incorporating color calibration and appropriate optical zoom, smartphone photography achieved gray-card-based color accuracy and relative dimensional consistency comparable to those of DSLR cameras, providing a more convenient and feasible imaging option. However, DSLR cameras still demonstrated advantages in clinically demanding aesthetic cases. Using a standardized workflow that includes appropriate optical zoom, professional dental lighting, and gray-card-based color calibration, smartphone photography can achieve relatively satisfactory reproduction of dental color and dimensional consistency, representing a potentially reliable and cost-effective option for clinical documentation.
Background Digital breast tomosynthesis (DBT) uses 1-mm slices, resulting in a larger number of images and longer interpretation times than conventional digital two-dimensional mammography. Slab reconstruction technologies address this challenge by generating thicker slices, thereby reducing the number of images requiring review, improving efficiency, and lowering storage demand. Purpose To compare the diagnostic accuracy of screening DBT before and after the implementation of artificial intelligence (AI)-based slab reconstruction technology. Materials and Methods Consecutive screening DBT examinations obtained before and after the implementation of a slab reconstruction technology at an academic medical center were retrospectively reviewed. The slab reconstruction technology uses AI to generate 6-mm synthetic slices with 3-mm overlap and 70-μm pixel resolution. The preimplementation period was between January 2018 and December 2019, and the postimplementation period was between October 2021 and September 2022. Multivariable logistic regression models were used to compare screening performance metrics in both periods, and a noninferiority analysis was performed. Results A total of 119 662 screening DBT examinations in 64 949 women were analyzed: 77 577 (52 649 women; mean age, 60 years ± 11 [SD]) during the preimplementation period and 42 085 (42 059 women; mean age, 60 years ± 11) during the postimplementation period. The cancer detection rate (CDR) (5.8 vs 6.5 per 1000 examinations; adjusted odds ratio [OR], 1.1; P = .49), sensitivity (82.3% vs 85.9%; adjusted OR, 1.3; P = .27), and false-negative rate (1.2 vs 1.1 per 1000 examinations; adjusted OR, 0.8; P = .39) did not differ between periods, and all three metrics met the noninferiority criteria. The abnormal interpretation rate (AIR) was lower (6.2% vs 5.8%; adjusted OR, 0.9; P < .001) and the specificity was higher (94.4% vs 94.9%; adjusted OR, 1.1; P < .001) during the postimplementation period. Conclusion The implementation of AI-based slab reconstruction technology was associated with noninferior CDR and sensitivity, improved specificity, and reduced AIR. © The Author(s) 2026. Published by the Radiological Society of North America under a CC BY 4.0 license. Supplemental material is available for this article. See also the editorial by Grimm in this issue.
Conventional imaging techniques lack the ability to quantify localized brain tissue displacements and strains associated with Glioblastoma (GBM) growth and treatment response. In this proof-of-concept study, a non-invasive approach is presented to map displacement and strain fields using Digital Volume Correlation (DVC) on serial T1-Gadolinium MRI scans of a GBM patient over 63 days. A comprehensive MRI pre-processing pipeline was applied, followed by the generation of meshes segmenting healthy brain tissue, tumor, and ventricles. Three distinct regularization scenarios were implemented to capture localized tissue deformations. DVC results were qualitatively compared to MRI anatomical changes and quantitatively validated against symmetric normalization (SyN) registration. DVC outperforms conventional SyN registration in capturing heterogeneous tissue deformations. These preliminary findings suggest greater sensitivity to biomechanical alterations induced by tumor progression and demonstrate the potential of DVC-augmented imaging to serve as a quantitative biomarker for assessing GBM-induced brain mechanics. By generating maps of ventricle deformation, the method provides new opportunities for early detection and monitoring of elevated intracranial pressure (ICP) in GBM patients.
Background: Diagnosing unilateral digital nerve injuries in small or already closed wounds is challenging, as clinical tests may be unreliable. Routine surgical exploration, although traditionally recommended, can represent an unnecessarily invasive procedure in cases without nerve rupture. High-resolution ultrasonography (US) offers a potential minimally invasive alternative. This study aimed to assess the reliability of US for evaluating unilateral digital nerve injuries in cases with small or previously sutured wounds. Methods: We retrospectively reviewed 19 digits from 19 patients with suspected unilateral digital nerve injury between April 2016 and December 2024. All underwent US evaluation within a mean of 8.9 days after injury. Injuries were classified as complete, partial or no tear based on long-axis US findings. Treatment decisions were guided by the US. Surgical findings were compared with preoperative US results. Clinical outcomes were assessed by pain resolution, numbness and sensory recovery. Results: US evaluation was feasible in all cases without local anaesthesia or complications. Nine patients had complete tears, six had partial tears and four had no tears. Twelve underwent surgery, and intraoperative findings matched US findings in all cases (100% accuracy). The remaining seven patients were managed conservatively. At follow-up (mean 6.6 months), all patients achieved improvement of numbness, hypaesthesia and local wound pain. Near-normal sensory recovery was achieved in 6/10 surgical patients and 4/6 conservative patients. No patients required additional treatment at the final follow-up. Conclusions: Early US evaluation is a reliable, safe and minimally invasive method for diagnosing unilateral digital nerve injuries in small or previously sutured wounds. This approach may prevent unnecessary surgical exploration in intact nerves and provide accurate guidance for treatment selection. Level of Evidence: Level III (Diagnostic).
Artificial Intelligence (AI) is increasingly proposed to enhance population-based cancer screening. While several applications are under evaluation, there is limited evidence on professional and organisational perspectives regarding their value, risks, and implementation challenges. We conducted a web-based survey targeting coordinators and reference professionals of Italian cancer screening programmes. The survey comprised 39 questions and explored prior experience with AI, perceived benefits and risks, organisational and legal barriers, and willingness to invest in AI applications. Recruitment combined a cascade procedure through regional and sub-regional coordinators with distribution via the mailing list of the Federation of Oncology Screening Associations (FASO). Descriptive analyses summarised responses, while multiple linear regression assessed determinants of willingness to invest in AI applications. A total of 305 professionals responded, representing all but one Italian region. About one-third reported AI experience, most frequently in colorectal polyp detection and breast imaging support. Overall, 74.8% expressed a strong interest in adopting AI applications, particularly for risk prediction and diagnostic support. The most valued benefits were improved diagnostic accuracy, workflow optimisation, and shorter reporting times. Concerns focused on legal liability (72.8%), lack of transparency and algorithmic bias (67.2%), and professional de-skilling (46.2%). Organisational barriers included inadequate IT infrastructure (77.0%), absence of guidelines (75.7%), and limited financial resources (77.4%). Regression confirmed that higher perceived benefits were positively associated with willingness to invest in AI applications, whereas risks were negatively associated. Italian screening professionals generally support the adoption of AI but emphasise the need for robust evidence, clear regulatory frameworks, and sustained investment in digital infrastructure to ensure its safe and effective implementation.
This study examines physicians' perceptions of privacy and data protection in Turkey's national personal health record system, e-Nabız. While e-Nabız enhances continuity of care through centralized access to prescriptions, laboratory results, imaging, and clinical records, its centralized governance model raises concerns regarding unauthorized access, accountability, and alignment with professional privacy norms. To empirically investigate these concerns, we conducted a survey with 309 healthcare professionals. The results reveal a pronounced usage-trust paradox: although system usage is high (87%), only 56% of respondents consider existing data-protection mechanisms adequate, and a substantial proportion express concerns about potential misuse or leakage of health data. Importantly, physicians' concerns are not primarily directed at the absence of role-based access control (RBAC), but at the lack of verifiable enforcement, transparent oversight, and tamper-evident auditability of access decisions. This indicates a perceived misalignment between expected information flows grounded in the physician-patient confidentiality context and the opaque governance of access practices within the system. Based on these findings, the study derives a governance-oriented design implication: a supplementary layer in which RBAC decisions and access events are recorded through blockchain-supported immutable logs and smart contracts to enhance accountability and auditability. The proposed approach does not store medical data on the blockchain; rather, it aims to make authorization and access histories verifiable and resistant to manipulation. The study contributes a physician-centered empirical assessment of privacy governance in a nationwide digital health system and highlights the importance of transparent, enforceable access governance for sustaining professional trust.
To evaluate the validity of digital manual and automated structural superimposition on the anterior cranial base using Björk's two-dimensional (2D) cephalometric method. Secondary aims included assessing repeatability and reproducibility, comparing manual and automated methods, and testing a color blending superimposition mode. Fourteen dry human skulls were radiographed twice. A gold standard superimposition was established using metallic markers. Three observers performed digital manual superimpositions twice on each skull; once using grayscale images and once using a color blending mode. Superimpositions were repeated one week later. An automated mutual information-based superimposition was also performed. Validity was assessed using root mean square error (RMSE) at 14 standardized cephalometric landmarks from a reference template. The RMSE of the automated method was compared with the average RMSE of each examiner across both trials using General Linear Models. Alignment time was recorded for the manual methods. The automated method demonstrated high validity (RMSE 0.48 mm, SD 0.35), closely matching the performance of the most accurate observer (RMSE 0.38 mm, SD 0.22), and showed no statistically significant difference when compared with the digital manual methods. A statistically significant difference in accuracy was observed between the observers (F = 4.76, P = 0.017). Color blending did not improve accuracy, but significantly reduced alignment time for two observers. No significant differences were found between repeats (P > 0.5). The digital manual and automated methods showed comparable validity. While color blending did not improve accuracy, it may reduce alignment time for some clinicians. Both approaches appear reliable and efficient.
Magnetic resonance imaging (MRI) is a complex-valued technique incorporating magnitude and phase information, with phase images critical for susceptibility-weighted imaging and quantitative susceptibility mapping yet often absent in reconstruction. This study aimed to synthesize phase images from magnitude-only data via innovative phase modulation to reconstruct complete MR images. Synthetic phase information was generated using irregular phase modulation based on two-dimensional sinusoidal functions. Comprehensive assessments were performed to validate the synthetic phase performance. First, SVG (Synthetic Variable Gradient) and PSE (Phase-Shift Estimation) experiments were performed to evaluate the spatial heterogeneity of the synthesized phase structures. Second, five metrics, including signal-to-noise ratio (SNR), contrast, correlation, homogeneity, energy, evaluated synthetic phase image similarity to true phase image, the Wilcoxon rank-sum test and Bland-Altman analyses further validated the consistency and approximation between synthetic and true phase images. Final, compressed sensing (CS) and Dense-U-Dense Net (DUD-Net) were utilized to verify the feasibility of synthetic phase images for MRI reconstruction, with peak signal-to-noise ratio (PSNR), root mean square error (RMSE) and structural similarity index (SSIM) used for comparative evaluation. The SVG and PSE results confirmed that the irregular phase provides more effective spatial heterogeneity than regular modulation. Statistical analyses demonstrated that the irregular phase achieved higher consistency with true phase images across five metrics including SNR, contrast, correlation, homogeneity, energy, with the Wilcoxon rank sum test (P < 0.05) and Bland Altman analyses further confirming its significant approximation to real physical phase. Regarding MRI reconstruction, the proposed irregular method demonstrated superior fidelity compared to regular methods. Quantitative validation showed that the irregular method achieved a DUD-Net PSNR of 34.90 ± 5.50 for magnitude reconstruction, which significantly outperformed the regular method and demonstrated comparable performance to the magnitude only reconstruction baseline. While irregular phase metrics were lower in high noise background areas, these have a competitive performance with the true phase when background noise was removed. Both CS and DUD-Net effectively reconstructed phase information, and phase modulation can provide reference phase data for establishing MR image reconstruction models without direct phase information. Notably, this heuristic synthetic phase serves as an approximation for reconstruction studies and cannot substitute true phase data for advanced physics-driven applications such as quantitative susceptibility mapping or susceptibility-weighted imaging.
Multiplexed and sensitive detection of pathogens is crucial for controlling risks of public health. Here, we present a mesophilic Clostridium butyricum Argonaute (CbAgo)-based guide DNA (gDNA)-target DNA-circuit (GTC) platform based on encoded microsphere microscopic imaging for multiplexed and ultrasensitive detection of pathogens. A user-friendly interactive interface enables rapid customized screening of orthogonal gDNAs and rational design of cleavage circuits from pathogen genome sequences, reducing cross-hybridization to below 0.032% and enhancing the cleavage efficiency of CbAgo by up to 157%. Encoded polystyrene microspheres by particle size and color serve as programmable multiprobes in microimaging bioassay, allowing simultaneous and ultrasensitive detection of multiple pathogens through spatial confinement and facilitating straightforward signal decoding via computer vision. The platform demonstrated excellent performance in multiplexed detection of three pathogenic bacteria (101 to 107 CFU/mL) across 60 real-world samples within 70 min without DNA amplification. Compared with quantitative polymerase chain reaction, the GTC platform shows a 10-fold improvement in sensitivity with a limit of detection down to 1 CFU/mL and exhibits higher accuracy in the detection of real samples. This work establishes an intelligent biosensing strategy for multiplexed pathogen detection and provides a promising next-generation digital platform for public health monitoring.
After radiological diagnostics, the patient journey does not end with image acquisition and reporting. Patients often face questions regarding report access, image sharing, follow-up, second opinions and care coordination. Digital patient portals may structure this transition, but their functionality varies substantially. To assess the role of digital patient portals after radiological diagnostics and to clinically classify their relevance within the "After the examination: what comes next?" phase. Narrative review focusing on radiology patient portals, direct release of imaging results, image access and sharing, patient-controlled image exchange, notification of follow-up recommendations and emergency department contexts. After imaging, patient portals mainly serve six functions: providing reports and images, contextualizing findings, enabling secure communication, supporting transfer to external physicians, tracking follow-up recommendations and guiding downstream care. Studies indicate that many patients actively access radiology results online, with reports being viewed more frequently than images. Portals create most value when reports, images and clear action options are delivered together. After radiological diagnostics, the patient portal becomes a digital bridge between imaging and subsequent care. Therefore, radiological quality should not be assessed solely by image acquisition and report quality, but also by whether patients know what to do next after receiving the results. HINTERGRUND: Nach radiologischer Diagnostik endet die Versorgung nicht mit der Befundung. Häufig entstehen Anschlussfragen zu Befundzugang, Bildmitnahme, Nachsorge, Zweitmeinung und Terminsteuerung. Digitale Patientenportale können diesen Übergang strukturieren, sind jedoch funktional sehr unterschiedlich ausgeprägt. Welche Funktion haben digitale Patientenportale nach radiologischer Diagnostik, und wie lassen sie sich klinisch einordnen? Narrative Übersichtsarbeit mit Fokus auf radiologische Patientenportale, direkten Ergebniszugang, Bildfreigabe, patientenkontrollierten Bildaustausch, Benachrichtigung bei Folgeempfehlungen sowie Notaufnahme- und Notfallkontexte. Nach radiologischer Diagnostik übernehmen Patientenportale vor allem 6 Aufgaben: Bereitstellung von Bericht und Bild, Kontextualisierung des Befunds, sichere Kommunikation, Weitergabe an externe Behandler, Nachverfolgung von Folgeempfehlungen und Navigation in die Anschlussversorgung. Studien zeigen, dass viele Patientinnen und Patienten radiologische Ergebnisse aktiv online aufrufen, Berichte häufiger als Bilder. Der größte Nutzen entsteht, wenn Bericht, Bilder und konkrete Handlungsoptionen gemeinsam angeboten werden. Nach radiologischer Diagnostik wird das Patientenportal zur digitalen Brücke zwischen Bildgebung und weiterer Versorgung. Radiologische Qualität bemisst sich deshalb nicht nur an Untersuchungs- und Befundqualität, sondern zunehmend auch daran, ob nach dem Befund klar ist, was der nächste sinnvolle Schritt ist.