Geometric and dosimetric impact of inter-professional contour variability in the upper abdomen for MR-linac

Journal of medical imaging and radiation sciences, Год журнала: 2025, Номер 56(5), С. 101980 - 101980

Опубликована: Май 13, 2025

Язык: Английский

---

ESR Essentials: a step-by-step guide of segmentation for radiologists—practice recommendations by the European Society of Medical Imaging Informatics

European Radiology, Год журнала: 2025, Номер unknown

Опубликована: Май 22, 2025

Abstract High-quality segmentation is important for AI-driven radiological research and clinical practice, with the potential to play an even more prominent role in future. As medical imaging advances, accurately segmenting anatomical pathological structures increasingly used obtain quantitative data valuable insights. Segmentation volumetric analysis could enable precise diagnosis, treatment planning, patient monitoring. These guidelines aim improve accuracy consistency, allowing better decision-making both environments. Practical advice on planning organization provided, focusing quality, precision, communication among teams. Additionally, tips strategies improving practices radiology radiation oncology are discussed, as pitfalls avoid. Key Points AI continues advance, volumetry will become integrated into making it essential radiologists stay informed about its applications diagnosis . There a significant lack of practical resources tailored specifically technical topics like Establishing clear rules best can streamline assessment settings, easier manage leading accurate care

Язык: Английский

---

Recommendations for radiotherapy quality assurance in clinical trials

Radiotherapy and Oncology, Год журнала: 2025, Номер unknown, С. 110950 - 110950

Опубликована: Май 1, 2025

Robust quality assurance (QA) of clinical trials in radiotherapy (RT) is paramount for minimising uncertainties treatment delivery, thereby strengthening the statistical power study and increasing likelihood accurately answering research question. As RT techniques evolve become more complex, establishing an appropriate QA program a specific trial becomes increasingly challenging, highlighting importance clear standardised recommendations. This provide such recommendations Principal Investigators (PIs) to consider when planning conducting Quality Assurance (RTQA) trials. They arise from experiences with RTQA conducted Danish Multidisciplinary Cancer Groups (DMCGs). The include checklist guide PIs developing effective program.

Язык: Английский

---

Diagnosis of Acute Aortic Syndromes on Non-Contrast CT Images with Radiomics-Based Machine Learning

Biology, Год журнала: 2023, Номер 12(3), С. 337 - 337

Опубликована: Фев. 21, 2023

We aimed to detect acute aortic syndromes (AAS) on non-contrast computed tomography (NCCT) images using a radiomics-based machine learning model. A total of 325 patients who underwent CT angiography (CTA) were enrolled retrospectively from 2 medical centers in China form the internal cohort (230 patients, 60 with AAS) and external testing (95 AAS). The was divided into training (n = 135), validation 49), 46). mask manually delineated NCCT by radiologist. Least Absolute Shrinkage Selection Operator regression (LASSO) used filter out nine feature parameters; Support Vector Machine (SVM) model showed best performance. In cohorts, SVM had an area under curve (AUC) 0.993 (95% CI, 0.965-1); accuracy (ACC), 0.946 0.877-1); sensitivity, 0.9 0.696-1); specificity, 0.964 0.903-1). cohort, AUC 0.997 0.992-1); ACC, 0.957 0.945-0.988); 0.889 0.888-0.889); 0.973 0.959-1). ACC 0.991 0.937-1). This can AAS NCCT, reducing misdiagnosis improving examinations prognosis.

Язык: Английский

---

Interobserver variation in tumor delineation of liver metastases using Magnetic Resonance Imaging

Physics and Imaging in Radiation Oncology, Год журнала: 2024, Номер 30, С. 100592 - 100592

Опубликована: Апрель 1, 2024

Язык: Английский

---

Impact of motion management strategies on abdominal organ at risk delineation for magnetic resonance-guided radiotherapy

Physics and Imaging in Radiation Oncology, Год журнала: 2024, Номер 32, С. 100650 - 100650

Опубликована: Сен. 17, 2024

Язык: Английский

---

Artificial intelligence-assisted delineation for postoperative radiotherapy in patients with lung cancer: a prospective, multi-center, cohort study

Frontiers in Oncology, Год журнала: 2024, Номер 14

Опубликована: Окт. 22, 2024

Background Postoperative radiotherapy (PORT) is an important treatment for lung cancer patients with poor prognostic features, but accurate delineation of the clinical target volume (CTV) and organs at risk (OARs) challenging time-consuming. Recently, deep learning-based artificial intelligent (AI) algorithms have shown promise in automating this process. Objective To evaluate utility a auto-segmentation model AI-assisted delineating CTV OARs undergoing PORT, to compare its accuracy efficiency manual by radiation oncology residents from different levels medical institutions. Methods We previously developed AI 664 validated contouring performance 149 patients. In multi-center, validation trial, we prospectively involved 55 compared 3 methods: (i) unmodified auto-segmentation, (ii) fully junior centers, (iii) modifications based on segmentation (AI-assisted delineation). The ground truth was delineated senior oncologists. Contouring evaluated Dice similarity coefficient (DSC), Hausdorff distance (HD), mean agreement (MDA). Inter-observer consistency assessed variation (CV). Results achieved significantly higher auto-contouring oncologists, median HD, MDA, DCS values 20.03 vs. 21.55 mm, 2.57 3.06 0.745 0.703 (all P<0.05) CTV, respectively. results contours were similar. CV reduced approximately 50%. addition better accuracy, decreased consuming time improved efficiency. Conclusion PORT improves real-world setting, pure or approach has promising potential enhance quality planning further improve outcomes cancer.

Язык: Английский

---

Clinical adoption of deep learning target auto-segmentation for radiation therapy: challenges, clinical risks, and mitigation strategies

Deleted Journal, Год журнала: 2024, Номер 1(1)

Опубликована: Янв. 1, 2024

Abstract Radiation therapy is a localized cancer treatment that relies on precise delineation of the target to be treated and healthy tissues guarantee optimal effect. This step, known as contouring or segmentation, involves identifying both volumes organs at risk imaging modalities like CT, PET, MRI guide radiation delivery. Manual however, time-consuming highly subjective, despite presence guidelines. In recent years, automated segmentation methods, particularly deep learning models, have shown promise in addressing this task. However, challenges persist their clinical use, including need for robust quality assurance (QA) processes risks associated with use models. review examines considerations adoption auto-segmentation radiotherapy, focused volume. We discuss potential (eg, over- under-segmentation, automation bias, appropriate trust), mitigation strategies human oversight, uncertainty quantification, education professionals), we highlight importance expanding QA include geometric, dose-volume, outcome-based performance monitoring. While offers significant benefits, careful attention rigorous measures are essential its successful integration practice.

Язык: Английский

---

Framework for Radiation Oncology Department-wide Evaluation and Implementation of Commercial Artificial Intelligence Autocontouring

Practical Radiation Oncology, Год журнала: 2023, Номер 14(2), С. e150 - e158

Опубликована: Ноя. 5, 2023

Язык: Английский

---

A multi-task deep learning model for EGFR genotyping prediction and GTV segmentation of brain metastasis

Journal of Translational Medicine, Год журнала: 2023, Номер 21(1)

Опубликована: Ноя. 7, 2023

The precise prediction of epidermal growth factor receptor (EGFR) mutation status and gross tumor volume (GTV) segmentation are crucial goals in computer-aided lung adenocarcinoma brain metastasis diagnosis. However, these two tasks present continuous difficulties due to the nonuniform intensity distributions, ambiguous boundaries, variable shapes (BM) MR images.The existing approaches for tackling challenges mainly rely on single-task algorithms, which overlook interdependence between tasks.To comprehensively address challenges, we propose a multi-task deep learning model that simultaneously enables GTV EGFR subtype classification. Specifically, multi-scale self-attention encoder consists convolutional module is designed extract shared spatial global information decoder an genotype classifier. Then, hybrid CNN-Transformer classifier consisting block Transformer combine local information. Furthermore, task correlation heterogeneity issues solved with loss function, aiming balance above by incorporating classification functions learnable weights.The experimental results demonstrate our proposed achieves excellent performance, surpassing approaches. Our mean Dice score 0.89 genotyping accuracy 0.88 internal testing set, attains 0.81 average 0.85 external set. This shows method has outstanding performance generalization.With introduction efficient feature extraction module, classifier, network significantly enhances achieved both tasks. Thus, can serve as noninvasive tool facilitating clinical treatment.

Язык: Английский

---