Gross failure rates and failure modes for a commercial AI‐based auto‐segmentation algorithm in head and neck cancer patients DOI Creative Commons
Simon W. P. Temple, C. Rowbottom

Journal of Applied Clinical Medical Physics, Journal Year: 2024, Volume and Issue: 25(6)

Published: Jan. 23, 2024

Abstract Purpose Artificial intelligence (AI) based commercial software can be used to automatically delineate organs at risk (OAR), with potential for efficiency savings in the radiotherapy treatment planning pathway, and reduction of inter‐ intra‐observer variability. There has been little research investigating gross failure rates modes such systems. Method 50 head neck (H&N) patient data sets “gold standard” contours were compared AI‐generated produce expected mean standard deviation values Dice Similarity Coefficient (DSC), four common H&N OARs (brainstem, mandible, left right parotid). An AI‐based system was applied 500 patients. manual contours, outlined by an expert human, a set three deviations below DSC. Failures inspected assess reason failures relating suboptimal contouring censored. True classified into 4 sub‐types (setup position, anatomy, image artefacts unknown). Results 24 true software, rate 1.2%. Fifteen due dental artefacts, position two unknown. OAR 0.4% (brainstem), 2.2% (mandible), 1.4% (left parotid) 0.8% (right Conclusion predominantly associated non‐standard element within CT scan. It is likely that these elements failure, suggests datasets train AI model did not contain sufficient heterogeneity data. Regardless reasons region investigated low (∼1%).

Language: Английский

Revolutionizing radiation therapy: the role of AI in clinical practice DOI Creative Commons

Mariko Kawamura,

Takeshi Kamomae, Masahiro Yanagawa

et al.

Journal of Radiation Research, Journal Year: 2023, Volume and Issue: 65(1), P. 1 - 9

Published: Oct. 19, 2023

This review provides an overview of the application artificial intelligence (AI) in radiation therapy (RT) from a oncologist's perspective. Over years, advances diagnostic imaging have significantly improved efficiency and effectiveness radiotherapy. The introduction AI has further optimized segmentation tumors organs at risk, thereby saving considerable time for oncologists. also been utilized treatment planning optimization, reducing several days to minutes or even seconds. Knowledge-based deep learning techniques employed produce plans comparable those generated by humans. Additionally, potential applications quality control assurance plans, optimization image-guided RT monitoring mobile during treatment. Prognostic evaluation prediction using increasingly explored, with radiomics being prominent area research. future oncology offers establish standardization minimizing inter-observer differences improving dose adequacy evaluation. through may global implications, providing world-standard resource-limited settings. However, there are challenges accumulating big data, including patient background information correlating disease outcomes. Although remain, ongoing research integration technology hold promise advancements oncology.

Language: Английский

Citations

34
HaN‐Seg: The head and neck organ‐at‐risk CT and MR segmentation dataset DOI Creative Commons
Gašper Podobnik, Primož Strojan, Primož Peterlin

et al.

Medical Physics, Journal Year: 2023, Volume and Issue: 50(3), P. 1917 - 1927

Published: Jan. 3, 2023

For the cancer in head and neck (HaN), radiotherapy (RT) represents an important treatment modality. Segmentation of organs-at-risk (OARs) is starting point RT planning, however, existing approaches are focused on either computed tomography (CT) or magnetic resonance (MR) images, while multimodal segmentation has not been thoroughly explored yet. We present a dataset CT MR images same patients with curated reference HaN OAR segmentations for objective evaluation methods.The cohort consists 56 that underwent both T1-weighted imaging image-guided RT. each patient, up to 30 OARs were obtained by experts performing manual pixel-wise image annotation. By maintaining distribution patient age gender, annotation type, randomly split into training Set 1 (42 cases 75%) test 2 (14 25%). Baseline auto-segmentation results also provided publicly available deep nnU-Net architecture 1, evaluating its performance 2.The data through open-access repository under name HaN-Seg: The Head Neck Organ-at-Risk & Dataset. Images stored NRRD file format, where filenames correspond nomenclature recommended American Association Physicists Medicine, demographics information separate comma-separated value files.The Challenge launched parallel release promote development automated techniques HaN. Other potential applications include out-of-challenge algorithm benchmarking, as well external validation developed algorithms.

Language: Английский

Citations

32
Recent advances in artificial intelligence for cardiac CT: Enhancing diagnosis and prognosis prediction DOI Creative Commons
Fuminari Tatsugami, Takeshi Nakaura, Masahiro Yanagawa

et al.

Diagnostic and Interventional Imaging, Journal Year: 2023, Volume and Issue: 104(11), P. 521 - 528

Published: July 4, 2023

Language: Английский

Citations

32
Deep learning-based accurate delineation of primary gross tumor volume of nasopharyngeal carcinoma on heterogeneous magnetic resonance imaging: A large-scale and multi-center study DOI
Xiangde Luo, Wenjun Liao, Yuan He

et al.

Radiotherapy and Oncology, Journal Year: 2023, Volume and Issue: 180, P. 109480 - 109480

Published: Jan. 16, 2023

Language: Английский

Citations

27
Review and recommendations on deformable image registration uncertainties for radiotherapy applications DOI Creative Commons
Lena Nenoff, Florian Amstutz, Martina Murr

et al.

Physics in Medicine and Biology, Journal Year: 2023, Volume and Issue: 68(24), P. 24TR01 - 24TR01

Published: Nov. 16, 2023

Deformable image registration (DIR) is a versatile tool used in many applications radiotherapy (RT). DIR algorithms have been implemented commercial treatment planning systems providing accessible and easy-to-use solutions. However, the geometric uncertainty of can be large difficult to quantify, resulting barriers clinical practice. Currently, there no agreement RT community on how quantify these uncertainties determine thresholds that distinguish good result from poor one. This review summarises current literature sources their impact applications. Recommendations are provided handle for patient-specific use, commissioning, research. also developers vendors help users understand make application safer more reliable.

Language: Английский

Citations

26
The role of Artificial intelligence in the assessment of the spine and spinal cord DOI
Teodoro Martín‐Noguerol,

Marta Oñate Miranda,

Timothy J. Amrhein

et al.

European Journal of Radiology, Journal Year: 2023, Volume and Issue: 161, P. 110726 - 110726

Published: Feb. 3, 2023

Language: Английский

Citations

25
Deep learning for autosegmentation for radiotherapy treatment planning: State-of-the-art and novel perspectives DOI Creative Commons
Ayhan Can Erdur,

Daniel Rusche,

Daniel Scholz

et al.

Strahlentherapie und Onkologie, Journal Year: 2024, Volume and Issue: unknown

Published: Aug. 6, 2024

Abstract The rapid development of artificial intelligence (AI) has gained importance, with many tools already entering our daily lives. medical field radiation oncology is also subject to this development, AI all steps the patient journey. In review article, we summarize contemporary techniques and explore clinical applications AI-based automated segmentation models in radiotherapy planning, focusing on delineation organs at risk (OARs), gross tumor volume (GTV), target (CTV). Emphasizing need for precise individualized plans, various commercial freeware state-of-the-art approaches. Through own findings based literature, demonstrate improved efficiency consistency as well time savings different scenarios. Despite challenges implementation such domain shifts, potential benefits personalized treatment planning are substantial. integration mathematical growth detection further enhances possibilities refining volumes. As advancements continue, prospect one-stop-shop represents an exciting frontier radiotherapy, potentially enabling fast enhanced precision individualization.

Language: Английский

Citations

13
CT radiomics based on different machine learning models for classifying gross tumor volume and normal liver tissue in hepatocellular carcinoma DOI Creative Commons

Huai-wen Zhang,

Delong Huang, Yiren Wang

et al.

Cancer Imaging, Journal Year: 2024, Volume and Issue: 24(1)

Published: Jan. 26, 2024

Abstract Background & aims The present study utilized extracted computed tomography radiomics features to classify the gross tumor volume and normal liver tissue in hepatocellular carcinoma by mainstream machine learning methods, aiming establish an automatic classification model. Methods We recruited 104 pathologically confirmed patients for this study. GTV samples were manually segmented into regions of interest randomly divided five-fold cross-validation groups. Dimensionality reduction using LASSO regression. Radiomics models constructed via logistic regression, support vector (SVM), random forest, Xgboost, Adaboost algorithms. diagnostic efficacy, discrimination, calibration algorithms verified area under receiver operating characteristic curve (AUC) analyses plot comparison. Results Seven screened excelled at distinguishing area. Xgboost algorithm had best discrimination comprehensive performance with AUC 0.9975 [95% confidence interval (CI): 0.9973–0.9978] mean MCC 0.9369. SVM second 0.9846 (95% CI: 0.9835– 0.9857), Matthews correlation coefficient (MCC)of 0.9105, a better calibration. All other showed excellent ability distinguish between (mean 0.9825, 0.9861,0.9727,0.9644 Adaboost, naivem Bayes respectively). Conclusion CT based on can accurately tissue, while served as complementary

Language: Английский

Citations

11
HaN-Seg: The head and neck organ-at-risk CT and MR segmentation challenge DOI Creative Commons
Gašper Podobnik, Bulat Ibragimov, Elias Tappeiner

et al.

Radiotherapy and Oncology, Journal Year: 2024, Volume and Issue: 198, P. 110410 - 110410

Published: June 24, 2024

To promote the development of auto-segmentation methods for head and neck (HaN) radiation treatment (RT) planning that exploit information computed tomography (CT) magnetic resonance (MR) imaging modalities, we organized HaN-Seg: The Head Neck Organ-at-Risk CT MR Segmentation Challenge. challenge task was to automatically segment 30 organs-at-risk (OARs) HaN region in 14 withheld test cases given availability 42 publicly available training cases. Each case consisted one contrast-enhanced T1-weighted image same patient, with up corresponding reference OAR delineation masks. performance evaluated terms Dice similarity coefficient (DSC) 95-percentile Hausdorff distance (HD95), statistical ranking applied each metric by pairwise comparison submitted using Wilcoxon signed-rank test. While 23 teams registered challenge, only seven their final phase. top-performing team achieved a DSC 76.9 % HD95 3.5 mm. All participating utilized architectures based on U-Net, winning leveraging rigid registration combined network entry-level concatenation both modalities. This simulated real-world clinical scenario providing non-registered images varying fields-of-view voxel sizes. Remarkably, segmentation surpassing inter-observer agreement dataset. These results set benchmark future research this dataset paired multi-modal general.

Language: Английский

Citations

10
Towards more precise automatic analysis: a systematic review of deep learning-based multi-organ segmentation DOI Creative Commons

Xiaoyu Liu,

Linhao Qu, Ziyue Xie

et al.

BioMedical Engineering OnLine, Journal Year: 2024, Volume and Issue: 23(1)

Published: June 8, 2024

Abstract Accurate segmentation of multiple organs in the head, neck, chest, and abdomen from medical images is an essential step computer-aided diagnosis, surgical navigation, radiation therapy. In past few years, with a data-driven feature extraction approach end-to-end training, automatic deep learning-based multi-organ methods have far outperformed traditional become new research topic. This review systematically summarizes latest this field. We searched Google Scholar for papers published January 1, 2016 to December 31, 2023, using keywords “multi-organ segmentation” “deep learning”, resulting 327 papers. followed PRISMA guidelines paper selection, 195 studies were deemed be within scope review. summarized two main aspects involved segmentation: datasets methods. Regarding datasets, we provided overview existing public conducted in-depth analysis. Concerning methods, categorized approaches into three major classes: fully supervised, weakly supervised semi-supervised, based on whether they require complete label information. achievements these terms accuracy. discussion conclusion section, outlined current trends segmentation.

Language: Английский

Citations

9