Medical Oncology, Journal Year: 2020, Volume and Issue: 37(5)
Published: March 31, 2020
Language: Английский
Nature Reviews Clinical Oncology, Journal Year: 2022, Volume and Issue: 20(2), P. 69 - 82
Published: Nov. 28, 2022
Language: Английский
Citations
135
Frontiers in Oncology, Journal Year: 2021, Volume and Issue: 11
Published: March 29, 2021
Radiomics is the method of choice for investigating association between cancer imaging phenotype, genotype and clinical outcome prediction in era precision medicine. The fast dispersal this new methodology has benefited from existing advances core technologies involved radiomics workflow: image acquisition, tumor segmentation, feature extraction machine learning. However, despite rapidly increasing body publications, there no real use a developed signature so far. Reasons are multifaceted. One major challenges lack reproducibility generalizability reported signatures (features models). Sources variation exist each step workflow; some controllable or can be controlled to certain degrees, while others uncontrollable even unknown. Insufficient transparency reporting studies further prevents translation bench bedside. This review article first addresses sources variation, which illustrated using demonstrative examples. Then, it reviews number published progresses made date investigation improvement model performance. Lastly, discusses potential strategies practical considerations reduce variability improve quality study. focuses on CT quantitative extraction, disease lung cancer.
Language: Английский
Citations
119
Radiotherapy and Oncology, Journal Year: 2019, Volume and Issue: 135, P. 107 - 114
Published: March 19, 2019
Language: Английский
Citations
138
Investigative Radiology, Journal Year: 2020, Volume and Issue: 55(9), P. 601 - 616
Published: March 24, 2020
Radiological images have been assessed qualitatively in most clinical settings by the expert eyes of radiologists and other clinicians. On hand, quantification radiological has potential to detect early disease that may be difficult with human eyes, complement or replace biopsy, provide clear differentiation stage. Further, objective assessment is a prerequisite personalized/precision medicine. This review article aims summarize discuss how variability quantitative values derived from are induced number factors these variabilities mitigated standardization achieved. We specific biomarkers magnetic resonance imaging computed tomography, focus on diffusion-weighted imaging, relaxometry, lung density evaluation, computer-aided tomography volumetry. also sources current efforts rapidly evolving techniques, which include radiomics artificial intelligence.
Language: Английский
Citations
118
EBioMedicine, Journal Year: 2019, Volume and Issue: 44, P. 289 - 297
Published: May 23, 2019
BackgroundEvaluating clinical outcome prior to concurrent chemoradiotherapy remains challenging for oesophageal squamous cell carcinoma (OSCC) as traditional prognostic markers are assessed at the completion of treatment. Herein, we investigated potential using sub-region radiomics a novel tumour biomarker in predicting overall survival OSCC patients treated by chemoradiotherapy.MethodsIndependent patient cohorts from two hospitals were included training (n = 87) and validation 46). Radiomics features extracted sub-regions clustered patients' regions K-means method. The LASSO regression 'Cox' method was used feature selection. prediction model constructed based on Cox proportional hazards model. biological significance correlation analysis characteristics copy number alterations(CNAs) dataset.FindingsThe combining with seven sub-regional constructed. C-indexes proposed 0.729 (0.656–0.801, 95% CI) 0.705 (0.628–0.782, 95%CI) cohorts, respectively. 3-year receiver operating characteristic (ROC) curve showed an area under ROC 0.811 (0.670–0.952, 0.805 (0.638–0.973, validation. significant between CNAs.InterpretationThe could predict risk definitive chemoradiotherapy.FundThis work supported Zhejiang Provincial Foundation Natural Sciences, National Science China.
Language: Английский
Citations
113
Scientific Reports, Journal Year: 2018, Volume and Issue: 8(1)
Published: Aug. 23, 2018
Abstract Radiomics has shown promise in improving models for predicting patient outcomes. However, to maximize the information gain of radiomics features, especially larger cohorts, variability features owing differences between scanners and scanning protocols must be accounted for. To this aim, imaging feature values was evaluated on 100 computed tomography at 35 clinics by a phantom using controlled protocol commonly used chest head local clinic. We linear mixed-effects model determine degree which manufacturer individual contribute overall variability. Using reduced 57% 52% compared respectively. The also relative contribution total For almost all variabilities (manufacturer, scanner, residual with different preprocesssing), scans had significantly smaller than did. most small inter-patient non–small cell lung cancer neck squamous carcinoma cohorts. From study, we conclude that can reduce our results demonstrate importance prospective studies.
Language: Английский
Citations
101
Lung Cancer, Journal Year: 2020, Volume and Issue: 146, P. 197 - 208
Published: June 2, 2020
Radiomics has become a popular image analysis method in the last few years. Its key hypothesis is that medical images harbor biological, prognostic and predictive information not revealed upon visual inspection. In contrast to previous work with priori defined imaging biomarkers, radiomics instead calculates features at scale uses statistical methods identify those most strongly associated outcome. This builds on years of research into computer aided diagnosis pattern recognition. While potential aid personalized medicine widely recognized, several technical limitations exist which hinder biomarker translation. Aspects radiomic workflow lack repeatability or reproducibility under particular circumstances, requirement for translation biomarkers clinical practice. One commonly studied applications Non-Small Cell Lung Cancer (NSCLC). this review, we summarize reported methodological CT based analyses together suggested solutions. We then evaluate current NSCLC literature assess risk accepting published conclusions respect these limitations. review different complementary scoring systems initiatives can be used critically appraise data from studies. Wider awareness should improve quality ongoing future studies advance their as clinically relevant patients NSCLC.
Language: Английский
Citations
91
Scientific Reports, Journal Year: 2019, Volume and Issue: 9(1)
Published: March 13, 2019
Radiomics extracts high-throughput quantitative data from medical images to contribute precision medicine. Radiomic shape features have been shown correlate with patient outcomes. However, how radiomic vary in function of tumor complexity and volume, as well method used for meshing voxel resampling, remains unknown. The aims this study are create models varying degrees complexity, or spiculatedness, evaluate their relationship quantitatively extracted features. Twenty-eight were mathematically created using spherical harmonics the spiculatedness degree d being increased by increments 3 (d = 11 92). Models 3D printed identical bases 5 cm, imaged a CT scanner two different slice thicknesses, semi-automatically delineated. Resampling resulting masks on 1 × mm
Language: Английский
Citations
89
Physica Medica, Journal Year: 2018, Volume and Issue: 48, P. 27 - 36
Published: March 27, 2018
Language: Английский
Citations
88
Computer Methods and Programs in Biomedicine, Journal Year: 2021, Volume and Issue: 215, P. 106609 - 106609
Published: Dec. 27, 2021
Language: Английский
Citations
88