Enhancing the Clinical Utility of Radiomics: Addressing the Challenges of Repeatability and Reproducibility in CT and MRI DOI Creative Commons
Xinzhi Teng, Yongqiang Wang, Alexander James Nicol

et al.

Diagnostics, Journal Year: 2024, Volume and Issue: 14(16), P. 1835 - 1835

Published: Aug. 22, 2024

Radiomics, which integrates the comprehensive characterization of imaging phenotypes with machine learning algorithms, is increasingly recognized for its potential in diagnosis and prognosis oncological conditions. However, repeatability reproducibility radiomic features are critical challenges that hinder their widespread clinical adoption. This review aims to address paucity discussion regarding factors influence subsequent impact on application models. We provide a synthesis literature CT/MR-based features, examining sources variation, number reproducible availability individual feature indices. differentiate variation into random effects, challenging control but can be quantified through simulation methods such as perturbation, biases, arise from scanner variability inter-reader differences significantly affect generalizability model performance diverse settings. Four suggestions studies suggested: (1) detailed reporting sources, (2) transparent disclosure calculation parameters, (3) careful selection suitable reliability indices, (4) metrics. underscores importance effects harmonizing biases between development settings facilitate successful translation models research practice.

Language: Английский

CheckList for EvaluAtion of Radiomics research (CLEAR): a step-by-step reporting guideline for authors and reviewers endorsed by ESR and EuSoMII DOI Creative Commons
Burak Koçak, Bettina Baeßler, Spyridon Bakas

et al.

Insights into Imaging, Journal Year: 2023, Volume and Issue: 14(1)

Published: May 4, 2023

Even though radiomics can hold great potential for supporting clinical decision-making, its current use is mostly limited to academic research, without applications in routine practice. The workflow of complex due several methodological steps and nuances, which often leads inadequate reporting evaluation, poor reproducibility. Available guidelines checklists artificial intelligence predictive modeling include relevant good practices, but they are not tailored radiomic research. There a clear need complete checklist study planning, manuscript writing, evaluation during the review process facilitate repeatability reproducibility studies. We here present documentation standard research that guide authors reviewers. Our motivation improve quality reliability and, turn, name CLEAR (CheckList EvaluAtion Radiomics research), convey idea being more transparent. With 58 items, should be considered standardization tool providing minimum requirements presenting In addition dynamic online version checklist, public repository has also been set up allow community comment on items adapt future versions. Prepared revised by an international group experts using modified Delphi method, we hope will serve well as single scientific reviewers literature.

Language: Английский

Citations

222
METhodological RadiomICs Score (METRICS): a quality scoring tool for radiomics research endorsed by EuSoMII DOI Creative Commons
Burak Koçak, Tugba Akinci D’Antonoli, Nathaniel D. Mercaldo

et al.

Insights into Imaging, Journal Year: 2024, Volume and Issue: 15(1)

Published: Jan. 17, 2024

Abstract Purpose To propose a new quality scoring tool, METhodological RadiomICs Score (METRICS), to assess and improve research of radiomics studies. Methods We conducted an online modified Delphi study with group international experts. It was performed in three consecutive stages: Stage#1, item preparation; Stage#2, panel discussion among EuSoMII Auditing Group members identify the items be voted; Stage#3, four rounds exercise by panelists determine eligible for METRICS their weights. The consensus threshold 75%. Based on median ranks derived from expert opinion rank-sum based conversion importance scores, category weights were calculated. Result In total, 59 19 countries participated selection ranking categories. Final tool included 30 within 9 According weights, categories descending order importance: design, imaging data, image processing feature extraction, metrics comparison, testing, processing, preparation modeling, segmentation, open science. A web application repository developed streamline calculation score collect feedback community. Conclusion this work, we assessing methodological research, large protocol. With its conditional format cover variations, it provides well-constructed framework key concepts radiomic papers. Critical relevance statement assessment is made available domain experts, transparent methodology, aiming at evaluating improving machine learning. Key points • METRICS, proposed presents opinion-based methodology first time. accounts varying use cases, handcrafted entirely deep learning-based pipelines. has been help ( https://metricsscore.github.io/metrics/METRICS.html ) created community https://github.com/metricsscore/metrics ). Graphical

Language: Английский

Citations

117
Reporting guidelines in medical artificial intelligence: a systematic review and meta-analysis DOI Creative Commons
Fiona R. Kolbinger, Gregory P. Veldhuizen, Jiefu Zhu

et al.

Communications Medicine, Journal Year: 2024, Volume and Issue: 4(1)

Published: April 11, 2024

Abstract Background The field of Artificial Intelligence (AI) holds transformative potential in medicine. However, the lack universal reporting guidelines poses challenges ensuring validity and reproducibility published research studies this field. Methods Based on a systematic review academic publications standards demanded by both international consortia regulatory stakeholders as well leading journals fields medicine medical informatics, 26 between 2009 2023 were included analysis. Guidelines stratified breadth (general or specific to fields), underlying consensus quality, target phase (preclinical, translational, clinical) subsequently analyzed regarding overlap variations guideline items. Results AI for vary with respect quality process, breadth, phase. Some items such study design model performance recur across guidelines, whereas other are particular stages. Conclusions Our analysis highlights importance clinical underscores need common that address identified gaps current guidelines. Overall, comprehensive overview could help researchers public reinforce increased reliability, reproducibility, validity, trust healthcare. This facilitate safe, effective, ethical translation methods into applications will ultimately improve patient outcomes.

Language: Английский

Citations

24
Evaluating the impact of the Radiomics Quality Score: a systematic review and meta-analysis DOI Creative Commons
Nathaniel Barry, Jake Kendrick,

Kaylee Molin

et al.

European Radiology, Journal Year: 2025, Volume and Issue: unknown

Published: Jan. 10, 2025

Abstract Objectives Conduct a systematic review and meta-analysis on the application of Radiomics Quality Score (RQS). Materials methods A search was conducted from January 1, 2022, to December 31, 2023, for reviews which implemented RQS. Identification articles prior 2022 via previously published review. scores individual radiomics papers, their associated criteria scores, these all readers were extracted. Errors in RQS noted corrected. The papers matched with publication date, imaging modality, country, where available. Results total 130 included, quality 117/130 (90.0%), 98/130 (75.4%), multiple reader data 24/130 (18.5%) 3258 correlated study date publication. Criteria scoring errors discovered 39/98 (39.8%) articles. Overall mean 9.4 ± 6.4 (95% CI, 9.1–9.6) (26.1% 17.8% (25.3%–26.7%)). positively year (Pearson R = 0.32, p < 0.01) significantly higher after (year 2018, 5.6 6.1 (5.1–6.1); ≥ 10.1 (9.9–10.4); 0.01). Only 233/3258 (7.2%) 50% maximum different across modalities ( Ten year, one negatively correlated. Conclusion adherence is increasing time, although vast majority studies are developmental rarely provide high level evidence justify clinical translation proposed models. Key Points Question What have achieved has it increased sufficient? Findings extracted resulted score 6.4. time. Clinical relevance Although many not demonstrated sufficient translation. As new appraisal tools emerge, current role may change.

Language: Английский

Citations

2
Artificial intelligence-based radiomics in bone tumors: Technical advances and clinical application DOI
Yichen Meng, Yue Yang, Miao Hu

et al.

Seminars in Cancer Biology, Journal Year: 2023, Volume and Issue: 95, P. 75 - 87

Published: July 26, 2023

Language: Английский

Citations

23
Deep Semisupervised Transfer Learning for Fully Automated Whole-Body Tumor Quantification and Prognosis of Cancer on PET/CT DOI Creative Commons
Kevin Leung, Steven P. Rowe,

Moe S. Sadaghiani

et al.

Journal of Nuclear Medicine, Journal Year: 2024, Volume and Issue: 65(4), P. 643 - 650

Published: Feb. 29, 2024

Automatic detection and characterization of cancer are important clinical needs to optimize early treatment. We developed a deep, semisupervised transfer learning approach for fully automated, whole-body tumor segmentation prognosis on PET/CT. Methods: This retrospective study consisted 611 18F-FDG PET/CT scans patients with lung cancer, melanoma, lymphoma, head neck breast 408 prostate-specific membrane antigen (PSMA) prostate cancer. The had nnU-net backbone learned the task PSMA images using limited annotations radiomics analysis. True-positive rate Dice similarity coefficient were assessed evaluate performance. Prognostic models imaging measures extracted from predicted segmentations perform risk stratification based follow-up levels, survival estimation by Kaplan–Meier method Cox regression analysis, pathologic complete response prediction after neoadjuvant chemotherapy. Overall accuracy area under receiver-operating-characteristic (AUC) curve assessed. Results: Our yielded median true-positive rates 0.75, 0.85, 0.87, 0.75 coefficients 0.81, 0.76, 0.83, 0.73 respectively, task. model an overall 0.83 AUC 0.86. Patients classified as low- intermediate- high-risk mean levels 18.61 727.46 ng/mL, respectively (P < 0.05). score was significantly associated univariable multivariable analyses Predictive only pretherapy both pre- posttherapy accuracies 0.72 0.84 AUCs respectively. Conclusion: proposed demonstrated accurate in across 6 types scans.

Language: Английский

Citations

16
Conventional and novel [18F]FDG PET/CT features as predictors of CAR-T cell therapy outcome in large B-cell lymphoma DOI Creative Commons
Doris Leithner, Jessica Flynn, Sean M. Devlin

et al.

Journal of Hematology & Oncology, Journal Year: 2024, Volume and Issue: 17(1)

Published: April 23, 2024

Relapse and toxicity limit the effectiveness of chimeric antigen receptor T-cell (CAR-T) therapy for large B-cell lymphoma (LBCL), yet biomarkers that predict outcomes are lacking. We examined radiomic features extracted from pre-CAR-T

Language: Английский

Citations

12
Role of [68Ga]Ga-PSMA-11 PET radiomics to predict post-surgical ISUP grade in primary prostate cancer DOI
Samuele Ghezzo, Paola Mapelli, Carolina Bezzi

et al.

European Journal of Nuclear Medicine and Molecular Imaging, Journal Year: 2023, Volume and Issue: 50(8), P. 2548 - 2560

Published: March 18, 2023

Language: Английский

Citations

20
Impact of different reconstruction algorithms and setting parameters on radiomics features of PSMA PET images: A preliminary study DOI

Masoomeh Fooladi,

Yunus Soleymani, Arman Rahmim

et al.

European Journal of Radiology, Journal Year: 2024, Volume and Issue: 172, P. 111349 - 111349

Published: Feb. 1, 2024

Language: Английский

Citations

6
Automated multiclass segmentation, quantification, and visualization of the diseased aorta on hybrid PET/CT–SEQUOIA DOI Creative Commons
Gijs D. van Praagh, P. H. Nienhuis, Melanie Reijrink

et al.

Medical Physics, Journal Year: 2024, Volume and Issue: 51(6), P. 4297 - 4310

Published: Feb. 7, 2024

Abstract Background Cardiovascular disease is the most common cause of death worldwide, including infection and inflammation related conditions. Multiple studies have demonstrated potential advantages hybrid positron emission tomography combined with computed (PET/CT) as an adjunct to current clinical inflammatory infectious biochemical markers. To quantitatively analyze vascular diseases at PET/CT, robust segmentation aorta necessary. However, manual extremely time‐consuming labor‐intensive. Purpose investigate feasibility accuracy automated tool segment quantify multiple parts diseased on unenhanced low‐dose (LDCT) anatomical reference for PET‐assessed disease. Methods A software pipeline was developed using a 3D U‐Net, calcium scoring, PET uptake quantification, background measurement, radiomics feature extraction, 2D surface visualization vessel wall tracer distribution. train 352 non‐contrast LDCTs from (2‐[ 18 F]FDG Na[ F]F) PET/CTs performed in patients various pathologies ascending aorta, aortic arch, descending abdominal were used. The last 22 consecutive scans used hold‐out internal test set. remaining dataset randomly split into training ( n = 264; 80%) validation 66; 20%) sets. Further evaluation external set 49 PET/CTs. dice similarity coefficient (DSC) Hausdorff distance (HD) assess performance. Automatically obtained scores values compared scoring softwares syngo .via Affinity Viewer) six patient images. intraclass correlation coefficients (ICC) calculated validate values. Results Fully U‐Net feasible LDCT PET/CT scans. yielded DSC 0.867 ± 0.030 HD 1.0 [0.6–1.4] mm, similar open‐source model 0.864 0.023 1.4 [1.0–1.8] mm. Quantification excellent agreement (ICC: 1.00 [1.00–1.00] 0.99 [0.93–1.00] values, respectively). Conclusions We present accurately provide scores, features, visualization. call this algorithm SEQUOIA (SEgmentation, QUantification, visualizatiOn dIseased Aorta) available https://github.com/UMCG‐CVI/SEQUOIA . This could augment utility tremendously, irrespective tracer, potentially fast reliable quantification cardiovascular practice, both primary diagnosis monitoring.

Language: Английский

Citations

6