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

Position dependence of recovery coefficients in 177Lu-SPECT/CT reconstructions – phantom simulations and measurements DOI Creative Commons
Julian Leube,

Wies Claeys,

Johan Gustafsson

et al.

EJNMMI Physics, Journal Year: 2024, Volume and Issue: 11(1)

Published: June 28, 2024

Although the importance of quantitative SPECT has increased tremendously due to newly developed therapeutic radiopharmaceuticals, there are still no accreditation programs harmonize imaging. Work is currently underway develop an for

Language: Английский

Citations

6
Assessment of RadiomIcS rEsearch (ARISE): a brief guide for authors, reviewers, and readers from the Scientific Editorial Board of European Radiology DOI Open Access
Burak Koçak, Leonid Chepelev, Linda C. Chu

et al.

European Radiology, Journal Year: 2023, Volume and Issue: 33(11), P. 7556 - 7560

Published: June 26, 2023

Language: Английский

Citations

15
Deep learning and radiomics framework for PSMA-RADS classification of prostate cancer on PSMA PET DOI Creative Commons
Kevin Leung, Steven P. Rowe,

Jeffrey P. Leal

et al.

EJNMMI Research, Journal Year: 2022, Volume and Issue: 12(1)

Published: Dec. 29, 2022

Accurate classification of sites interest on prostate-specific membrane antigen (PSMA) positron emission tomography (PET) images is an important diagnostic requirement for the differentiation prostate cancer (PCa) from foci physiologic uptake. We developed a deep learning and radiomics framework to perform lesion-level patient-level PSMA PET patients with PCa. This was IRB-approved, HIPAA-compliant, retrospective study. Lesions [18F]DCFPyL PET/CT scans were assigned reporting data system (PSMA-RADS) categories randomly partitioned into training, validation, test sets. The extracted image features, radiomic tissue type information cropped slice containing lesion performed PSMA-RADS PCa classification. Performance evaluated by assessing area under receiver operating characteristic curve (AUROC). A t-distributed stochastic neighbor embedding (t-SNE) analysis performed. Confidence probability scores measured. Statistical significance determined using two-tailed t test. 267 men had 3794 lesions categories. yielded AUROC values 0.87 0.90 classification, respectively, set. 0.92 0.85 t-SNE revealed learned relationships between disease findings. Mean confidence reflected expected accuracy significantly higher correct predictions than incorrect (P < 0.05). Measured likelihood consistent framework. provided images. interpretable that may assist physicians in making more informed clinical decisions.

Language: Английский

Citations

22
The Promise and Future of Radiomics for Personalized Radiotherapy Dosing and Adaptation DOI Creative Commons
Rachel Ger, Lise Wei, Issam El Naqa

et al.

Seminars in Radiation Oncology, Journal Year: 2023, Volume and Issue: 33(3), P. 252 - 261

Published: June 16, 2023

Language: Английский

Citations

12
Prediction of pathological response after neoadjuvant chemotherapy using baseline FDG PET heterogeneity features in breast cancer DOI Creative Commons
Carla Oliveira, Francisco P. M. Oliveira, Sofia C. Vaz

et al.

British Journal of Radiology, Journal Year: 2023, Volume and Issue: 96(1146)

Published: March 3, 2023

Complete pathological response to neoadjuvant systemic treatment (NAST) in some subtypes of breast cancer (BC) has been used as a surrogate long-term outcome. The possibility predicting BC NAST based on the baseline 18F-Fluorodeoxyglucose positron emission tomography (FDG PET), without need an interim study, is focus recent discussion. This review summarises characteristics and results available studies regarding potential impact heterogeneity features primary tumour burden FDG PET patients. Literature search was conducted PubMed database relevant data from each selected study were collected. A total 13 eligible for inclusion, all them published over last 5 years. Eight out analysed indicated association between PET-based uptake prediction NAST. When associated with derived, these varied studies. Therefore, definitive reproducible findings across series difficult establish. lack consensus may reflect low number included series. clinical relevance this topic justifies further investigation about predictive role PET.

Language: Английский

Citations

11
Application of PET imaging delta radiomics for predicting progression-free survival in rare high-grade glioma DOI Creative Commons
Shamimeh Ahrari, Timothée Zaragori,

Adeline Zinsz

et al.

Scientific Reports, Journal Year: 2024, Volume and Issue: 14(1)

Published: Feb. 8, 2024

Abstract This study assesses the feasibility of using a sample-efficient model to investigate radiomics changes over time for predicting progression-free survival in rare diseases. Eighteen high-grade glioma patients underwent two L-3,4-dihydroxy-6-[ 18 F]-fluoro-phenylalanine positron emission tomography (PET) dynamic scans: first during treatment and second at temozolomide chemotherapy discontinuation. Radiomics features from static/dynamic parametric images, alongside conventional features, were extracted. After excluding highly correlated 16 different models trained by combining various feature selection methods time-to-event algorithms. Performance was assessed cross-validation. To evaluate robustness, an additional dataset including 35 with single PET scan therapy discontinuation used. Model performance compared strategy extracting informative set applying them 2 scans. Delta-absolute achieved highest when pipeline directly applied 18-patient subset (support vector machine (SVM) recursive elimination (RFE): C-index = 0.783 [0.744–0.818]). result remained consistent transferring (SVM + RFE: 0.751 [0.716–0.784], p 0.06). In addition, it significantly outperformed delta-absolute (C-index 0.584 [0.548–0.620], < 0.001) single-time-point 0.546 [0.512–0.580], 0.001), highlighting considerable potential delta cancer cohorts.

Language: Английский

Citations

4
PET radiomics for histologic subtype classification of non-small cell lung cancer: a systematic review and meta-analysis DOI
Jucheng Zhang, Xiaohui Zhang, Yan Zhong

et al.

European Journal of Nuclear Medicine and Molecular Imaging, Journal Year: 2025, Volume and Issue: unknown

Published: Jan. 11, 2025

Language: Английский

Citations

0
Robust vs. Non-robust radiomic features: the quest for optimal machine learning models using phantom and clinical studies DOI Creative Commons
Seyyed Ali Hosseini, Ghasem Hajianfar, Brandon Hall

et al.

Cancer Imaging, Journal Year: 2025, Volume and Issue: 25(1)

Published: March 12, 2025

Abstract Purpose This study aimed to select robust features against lung motion in a phantom and use them as input feature selection algorithms machine learning classifiers clinical predict the lymphovascular invasion (LVI) of non-small cell cancer (NSCLC). The results were also compared with conventional techniques without considering robustness radiomic features. Methods An in-house developed was two 22mm lesion sizes based on study. A specific motor built simulate orthogonal directions. Lesions both studies segmented using Fuzzy C-means-based segmentation algorithm. After inducing extracting 105 4 sets, including shape, first-, second-, higher-order statistics from each region interest (ROI) image, statistical analyses performed motion. Subsequently, these total extracted 126 data. Various (FS) multiple (ML) implemented LVI NSCLC, followed by comparing predicting common not Results Our demonstrated that selecting FS ML surges sensitivity, which has gentle negative effect accuracy area under curve (AUC) predictions commonly used methods 12 15 outcomes. top performance prediction achieved NB classifier RFE 95% AUC, 67% accuracy, 100% sensitivity. Moreover, belonged Boruta 92% 86% Conclusion Robustness over various influential factors is critical should be considered Selecting solution overcome low reproducibility Although setting minor impact AUC prediction, it boosts sensitivity large extent.

Language: Английский

Citations

0
Radiomics prognostic analysis of PET/CT images in a multicenter head and neck cancer cohort: investigating ComBat strategies, sub-volume characterization, and automatic segmentation DOI
Hui Xu, Nassib Abdallah, Jean-Marie Marion

et al.

European Journal of Nuclear Medicine and Molecular Imaging, Journal Year: 2023, Volume and Issue: 50(6), P. 1720 - 1734

Published: Jan. 24, 2023

Language: Английский

Citations

10
A radiomic‐ and dosiomic‐based machine learning regression model for pretreatment planning in 177Lu‐DOTATATE therapy DOI Creative Commons
Dimitris Plachouris,

Vassilios Eleftheriadis,

Thomas Nanos

et al.

Medical Physics, Journal Year: 2023, Volume and Issue: 50(11), P. 7222 - 7235

Published: Sept. 18, 2023

Standardized patient-specific pretreatment dosimetry planning is mandatory in the modern era of nuclear molecular radiotherapy, which may eventually lead to improvements final therapeutic outcome. Only a comprehensive definition dosage window encompassing range absorbed doses, that is, helpful without being detrimental can therapy individualization and improved outcomes. As result, setting dose safety limits for organs at risk (OARs) requires knowledge dose-effect relationship. Data sets consistent reliable inter-center findings are required characterize this relationship.We developed standardized new model consisting predictive procedure OARs patients with neuroendocrine tumors (NETs) treated 177 Lu-DOTATATE (Lutathera). In retrospective study described herein, we used machine learning (ML) regression algorithms predict doses by exploiting combination radiomic dosiomic features extracted from patients' imaging data.Pretreatment posttreatment data 20 NETs were collected two clinical centers. A total 3412 computed tomography (CT) scans maps, respectively. All maps generated using Monte Carlo simulations. An ML was designed based on predicting every OAR (liver, left kidney, right spleen) before after between each session, thus any possible radiotoxic effects.We evaluated nine algorithms. Our achieved mean absolute error (MAE, Gy) 0.61 liver, 1.58 spleen, 1.30 1.35 kidney pretherapy 68 Ga-DOTATOC positron emission (PET)/CT posttherapy single photon (SPECT)/CT scans. Τhe best performance observed gradient boost extra tree regressor spleen. Evaluation model's according its ability PET/CT treatment cycle SPECT/CT as well consequent revealed differences ranges -0.55 0.68 Gy. Incorporating radiodosiomics first further precision minimized standard deviation predictions out 12 instances. average improvement 57.34% (range: 17.53%-96.12%). However, it's important note three instances (i.e., Ga,C.1 → C3 spleen C2 kidney) did not observe an (absolute 0.17, 0.08, 0.05 Gy, respectively). Wavelet-based proved have high correlated value, whereas non-linear-based be more capable than linear-based producing precise prediction our case.The radiomics dosiomics has potential utility personalized radiotherapy (PMR) response evaluation prediction. These radiodosiomic potentially provide information disease recurrence highly useful decision-making, especially regarding escalation issues.

Language: Английский

Citations

10