Quality of science and reporting of radiomics in oncologic studies: room for improvement according to radiomics quality score and TRIPOD statement

European Radiology, Год журнала: 2019, Номер 30(1), С. 523 - 536

Опубликована: Июль 26, 2019

Язык: Английский

---

Radiomics and Deep Learning: Hepatic Applications

Korean Journal of Radiology, Год журнала: 2020, Номер 21(4), С. 387 - 387

Опубликована: Янв. 1, 2020

Radiomics and deep learning have recently gained attention in the imaging assessment of various liver diseases.Recent research has demonstrated potential utility radiomics staging fibroses, detecting portal hypertension, characterizing focal hepatic lesions, prognosticating malignant tumors, segmenting tumors.In this review, we outline basic technical aspects summarize recent investigations application these techniques disease.

Язык: Английский

---

Predicting Breast Cancer in Breast Imaging Reporting and Data System (BI-RADS) Ultrasound Category 4 or 5 Lesions: A Nomogram Combining Radiomics and BI-RADS

Scientific Reports, Год журнала: 2019, Номер 9(1)

Опубликована: Авг. 15, 2019

Abstract Radiomics reflects the texture and morphological features of tumours by quantitatively analysing grey values medical images. We aim to develop a nomogram incorporating radiomics Breast Imaging Reporting Data System (BI-RADS) for predicting breast cancer in BI-RADS ultrasound (US) category 4 or 5 lesions. From January 2017 August 2018, total 315 pathologically proven lesions were included. Patients from study population divided into training group (n = 211) validation 104) according cut-off date March 1 st , 2018. Each lesion was assigned (4A, 4B, 4C 5) second edition American College Radiology (ACR) US. A score generated US image. developed based on results multivariate regression analysis group. Discrimination, calibration clinical usefulness assessed The included 9 selected features. independently associated with malignancy. showed better discrimination (area under receiver operating characteristic curve [AUC]: 0.928; 95% confidence interval [CI]: 0.876, 0.980) between malignant benign than either ( P 0.029) 0.011). demonstrated good usefulness. In conclusion, combining is potentially useful malignancy

Язык: Английский

---

Ultrasound-based radiomics nomogram: A potential biomarker to predict axillary lymph node metastasis in early-stage invasive breast cancer

European Journal of Radiology, Год журнала: 2019, Номер 119, С. 108658 - 108658

Опубликована: Сен. 7, 2019

Язык: Английский

---

Radiomics in liver diseases: Current progress and future opportunities

Liver International, Год журнала: 2020, Номер 40(9), С. 2050 - 2063

Опубликована: Июнь 9, 2020

Abstract Liver diseases, a wide spectrum of pathologies from inflammation to neoplasm, have become an increasingly significant health problem worldwide. Noninvasive imaging plays critical role in the clinical workflow liver but conventional assessment may provide limited information. Accurate detection, characterization and monitoring remain challenging. With progress quantitative analysis techniques, radiomics emerged as efficient tool that shows promise aid personalized diagnosis treatment decision‐making. Radiomics could reflect heterogeneity lesions via extracting high‐throughput high‐dimensional features multi‐modality imaging. Machine learning algorithms are then used construct target‐oriented biomarkers assist disease management. Here, we review methodological process studies stepwise fashion data acquisition curation, region interest segmentation, liver‐specific feature extraction, task‐oriented modelling. Furthermore, applications diseases outlined aspects staging, evaluation tumour biological behaviours, prognosis according different type. Finally, discuss current limitations explore its future opportunities.

Язык: Английский

---

Radiomics analysis of [18F]FDG PET/CT for microvascular invasion and prognosis prediction in very-early- and early-stage hepatocellular carcinoma

European Journal of Nuclear Medicine and Molecular Imaging, Год журнала: 2021, Номер 48(8), С. 2599 - 2614

Опубликована: Янв. 8, 2021

Язык: Английский

---

Contrast‐enhanced CT radiomics for preoperative evaluation of microvascular invasion in hepatocellular carcinoma: A two‐center study

Clinical and Translational Medicine, Год журнала: 2020, Номер 10(2)

Опубликована: Июнь 1, 2020

Abstract Background The present study constructed and validated the use of contrast‐enhanced computed tomography (CT)‐based radiomics to preoperatively predict microvascular invasion (MVI) status (positive vs negative) risk (low high) in patients with hepatocellular carcinoma (HCC). Methods We enrolled 637 from two independent institutions. Patients Institution I were randomly divided into a training cohort 451 test 111 patients. II served as an validation set. LASSO algorithm was used for selection 798 features. Two classifiers predicting MVI developed using multivariable logistic regression. also performed survival analysis investigate potentially prognostic value proposed classifiers. Results signature predicted area under receiver operating characteristic curve (AUC) .780, .776, .743 training, test, cohorts, respectively. final classifier that integrated clinical factors (age α‐fetoprotein level) achieved AUC .806, .803, .796 For stratification, AUCs .746, .664, .700 respectively, classifier‐integrated stage .783, .778, .740, Survival showed our significantly stratified short overall or early tumor recurrence. Conclusions Our CT radiomics‐based models able HCC might serve reliable preoperative evaluation tool.

Язык: Английский

---

Artificial intelligence in precision medicine in hepatology

Journal of Gastroenterology and Hepatology, Год журнала: 2021, Номер 36(3), С. 569 - 580

Опубликована: Март 1, 2021

Abstract The advancement of investigation tools and electronic health records (EHR) enables a paradigm shift from guideline‐specific therapy toward patient‐specific precision medicine. multiparametric large detailed information necessitates novel analyses to explore the insight diseases aid diagnosis, monitoring, outcome prediction. Artificial intelligence (AI), machine learning, deep learning (DL) provide various models supervised, or unsupervised algorithms, sophisticated neural networks generate predictive more precisely than conventional ones. data, application tasks, algorithms are three key components in AI. Various data formats available daily clinical practice hepatology, including radiological imaging, EHR, liver pathology, wearable devices, multi‐omics measurements. images abdominal ultrasonography, computed tomography, magnetic resonance imaging can be used predict fibrosis, cirrhosis, non‐alcoholic fatty disease (NAFLD), differentiation benign tumors hepatocellular carcinoma (HCC). Using AI help diagnosis outcomes HCC, NAFLD, portal hypertension, varices, transplantation, acute failure. helps severity patterns steatosis, activity survival HCC by using pathological data. Despite these high potentials application, preparation, collection, quality, labeling, sampling biases major concerns. selection, evaluation, validation as well real‐world models, also challenging. Nevertheless, opens new era medicine which will change our future practice.

Язык: Английский

---

Deep Learning With 3D Convolutional Neural Network for Noninvasive Prediction of Microvascular Invasion in Hepatocellular Carcinoma

Journal of Magnetic Resonance Imaging, Год журнала: 2021, Номер 54(1), С. 134 - 143

Опубликована: Фев. 8, 2021

Background Microvascular invasion (MVI) is a critical prognostic factor of hepatocellular carcinoma (HCC). However, it could only be obtained by postoperative histological examination. Purpose To develop an end‐to‐end deep‐learning models based on MRI images for preoperative prediction MVI in HCC patients who underwent surgical resection. Study type Retrospective. Population Two hundred and thirty‐seven with histologically confirmed HCC. Field strength 1.5 T 3.0 T. Sequence Axial 2 ‐weighted (T ‐w) turbo spin echo sequence, ‐Spectral Presaturation Inversion Recovery ‐SPIR), dynamic contrast‐enhanced (DCE) imaging fat suppressed enhanced 1 high‐resolution isotropic volume Assessment The were randomly divided into training ( N = 158) validation 79) sets. Data augmentation random rotation was performed the set sample size increased to 1940 each MR sequence. A three‐dimensional convolutional neural network (3D CNN) used four models, including three single‐layer single‐sequence, fusion model combining sequences. status from pathology reports. Statistical Tests dice similarity coefficient (DSC) Hausdorff distance (HD) applied assess reproducibility between manual segmentations tumor two radiologists. Receiver operating characteristic curve analysis evaluate performance. identified 92 (38.8%) patients. Good interobserver DSCs 0.90, 0.89, 0.89 HDs 4.09, 3.67, 3.60 observed PVP, WI, ‐SPIR, respectively. achieved area under (AUC) 0.81, sensitivity 69%, specificity 79% 0.72, 55%, 81% set. Conclusion 3D CNN may serve as noninvasive tool predict HCC, whereas its accuracy needs larger cohort. Level Evidence 3 Technical Efficacy Stage

Язык: Английский

---

The progress of multimodal imaging combination and subregion based radiomics research of cancers

International Journal of Biological Sciences, Год журнала: 2022, Номер 18(8), С. 3458 - 3469

Опубликована: Янв. 1, 2022

In recent years, with the standardization of radiomics methods; development tools; and popularization concept, has been widely used in all aspects tumor diagnosis; treatment; prognosis.As study cancer become more advanced, currently methods have revealed their shortcomings.The performance based on single-modality medical images, which imaging principles, only partially reflects information, necessarily compromised.Using whole as a region interest to extract radiomic features inevitably leads loss intra-tumoral heterogeneity of, also affects radiomics.Radiomics multimodal images extracts various information from each modality then integrates them together for model construction; thus, avoiding missing information.Subregional segmentation image combinations allows acquired subregions retain heterogeneity, further improving radiomics.In this review, we provide detailed summary current research subregion-based radiomics, raised some problems thorough discussion these issues.

Язык: Английский

---