Demystifying Supervised Learning in Healthcare 4.0: A New Reality of Transforming Diagnostic Medicine

Diagnostics, Год журнала: 2022, Номер 12(10), С. 2549 - 2549

Опубликована: Окт. 20, 2022

The global healthcare sector continues to grow rapidly and is reflected as one of the fastest-growing sectors in fourth industrial revolution (4.0). majority industry still uses labor-intensive, time-consuming, error-prone traditional, manual, manpower-based methods. This review addresses current paradigm, potential for new scientific discoveries, technological state preparation, supervised machine learning (SML) prospects various sectors, ethical issues. effectiveness innovation disease diagnosis, personalized medicine, clinical trials, non-invasive image analysis, drug discovery, patient care services, remote monitoring, hospital data, nanotechnology learning-based automation along with requirement explainable artificial intelligence (AI) are evaluated. In order understand architecture treatment, a thorough study medical imaging analysis from technical point view presented. also represents thinking developments that will push boundaries increase opportunity through AI SML near future. Nowadays, SML-based applications require lot data quality awareness data-heavy, knowledge management paramount. biomedical needs skills, consciousness data-intensive study, knowledge-centric health system. As result, merits, demerits, precautions need take ethics other effects into consideration. overall insight this paper help researchers academia address future research be discussed on sectors.

Язык: Английский

---

CT-based machine learning radiomics predicts CCR5 expression level and survival in ovarian cancer

Journal of Ovarian Research, Год журнала: 2023, Номер 16(1)

Опубликована: Янв. 3, 2023

Abstract Objective We aimed to evaluate the prognostic value of C-C motif chemokine receptor type 5 (CCR5) expression level for patients with ovarian cancer and establish a radiomics model that can predict CCR5 using The Cancer Imaging Archive (TCIA) Genome Atlas (TCGA) database. Methods A total 343 cases from TCGA were used gene-based analysis. Fifty seven had preoperative computed tomography (CT) images stored in TCIA genomic data feature extraction construction. 89 both clinical evaluation. After extraction, signature was constructed least absolute shrinkage selection operator (LASSO) regression scoring system incorporating based on clinicopathologic risk factors proposed survival prediction. Results identified as differentially expressed prognosis-related gene tumor normal sample, which involved regulation immune response invasion metastasis. Four optimal features selected overall survival. performance predicting 10-fold cross- validation achieved Area Under Curve (AUCs) 0.770 0.726, respectively, training sets. predictive nomogram generated score each patient, AUCs time-dependent receiver operating characteristic (ROC) curve 0.8, 0.673 0.792 1-year, 3-year 5-year, respectively. Along features, important imaging biomarkers could improve accuracy prediction model. Conclusion levels affect prognosis cancer. CT-based serve new tool

Язык: Английский

---

Shearlet Transform Applied to a Prostate Cancer Radiomics Analysis on MR Images

Mathematics, Год журнала: 2024, Номер 12(9), С. 1296 - 1296

Опубликована: Апрель 25, 2024

For decades, wavelet theory has attracted interest in several fields dealing with signals. Nowadays, it is acknowledged that not very suitable to face aspects of multidimensional data like singularities and this led the development other mathematical tools. A recent application radiomics, an emerging field aiming improve diagnostic, prognostic predictive analysis various cancer types through features extracted from medical images. In paper, for a radiomics study prostate magnetic resonance (MR) images, we apply similar but more sophisticated tool, namely shearlet transform which, contrast transform, allows us examine variations along orientations. particular, conduct parallel based on two different transformations highlight better performance (evaluated terms statistical measures) use (in absolute value). The results achieved suggest taking into consideration studies contexts.

Язык: Английский

---

[18F]PSMA-1007 PET/CT-based radiomics may help enhance the interpretation of bone focal uptakes in hormone-sensitive prostate cancer patients

European Journal of Nuclear Medicine and Molecular Imaging, Год журнала: 2025, Номер unknown

Опубликована: Янв. 28, 2025

Abstract Purpose We hypothesised that applying radiomics to [ 18 F]PSMA-1007 PET/CT images could help distinguish Unspecific Bone Uptakes (UBUs) from bone metastases in prostate cancer (PCa) patients. compared the performance of radiomic features human visual interpretation. Materials and methods retrospectively analysed 102 hormone-sensitive PCa patients who underwent exhibited at least one focal uptake with known clinical follow-up (reference standard). Using matRadiomics, we extracted PET CT each identified best predictor model for using a machine-learning approach generate score. Blinded readers low ( n = 2) high experience rated as either UBU or metastasis. The same performed second read three months later, access Results Of 178 uptakes, 74 (41.5%) were classified by reference standard. A combining achieved an accuracy 84.69%, though it did not surpass expert round. Less-experienced had significantly lower diagnostic baseline p < 0.05) but improved addition scores 0.05 first round). Conclusion Radiomics might differentiate UBUs. While exceed assessments, has potential enhance less-experienced evaluating uptakes.

Язык: Английский

---

Phenotyping the Histopathological Subtypes of Non-Small-Cell Lung Carcinoma: How Beneficial Is Radiomics?

Diagnostics, Год журнала: 2023, Номер 13(6), С. 1167 - 1167

Опубликована: Март 18, 2023

The aim of this study was to investigate the usefulness radiomics in absence well-defined standard guidelines. Specifically, we extracted features from multicenter computed tomography (CT) images differentiate between four histopathological subtypes non-small-cell lung carcinoma (NSCLC). In addition, results that varied with model were compared. We investigated presence batch effects and impact feature harmonization on models’ performance. Moreover, question how training dataset composition influenced selected subsets and, consequently, model’s performance also investigated. Therefore, through combining data two publicly available datasets, involves a total 152 squamous cell (SCC), 106 large (LCC), 150 adenocarcinoma (ADC), 58 no other specified (NOS). Through matRadiomics tool, which is an example Image Biomarker Standardization Initiative (IBSI) compliant software, 1781 each malignant lesions identified CT images. After analysis harmonization, based ComBat tool integrated matRadiomics, datasets (the harmonized non-harmonized) given as input machine learning modeling pipeline. following steps articulated: (i) training-set/test-set splitting (80/20); (ii) Kruskal–Wallis LASSO linear regression for selection; (iii) training; (iv) validation hyperparameter optimization; (v) testing. Model optimization consisted 5-fold cross-validated Bayesian optimization, repeated ten times (inner loop). whole pipeline 10 (outer loop) six different classification algorithms. stability selection evaluated. Results showed present even if voxels resampled isotropic form whether correctly removed them, though performances decreased. low accuracy (61.41%) reached when differentiating subtypes, high average area under curve (AUC) (0.831). Further, NOS subtype classified almost completely correct (true positive rate ~90%). increased (77.25%) only SCC ADC considered, well AUC (0.821) obtained—although decreased 58%. contributed most those wavelet decomposed Laplacian Gaussian (LoG) filtered they belonged texture class.. conclusion, our affected by effects, could significantly alter performance, them. Although seemed be informative features, absolute subset not since it changed depending training/testing splitting. chosen methods, reach binary tasks, but underperform multiclass problems. It is, therefore, essential scientific community propose more systematic approach, focusing studies, clear solid guidelines facilitate translation clinical practice.

Язык: Английский

---

Cancer Radiomic and Perfusion Imaging Automated Framework: Validation on Musculoskeletal Tumors

JCO Clinical Cancer Informatics, Год журнала: 2024, Номер 8

Опубликована: Янв. 5, 2024

Limitations from commercial software applications prevent the implementation of a robust and cost-efficient high-throughput cancer imaging radiomic feature extraction perfusion analysis workflow. This study aimed to develop validate research computational solution using open-source for vendor- sequence-neutral image processing extraction.

Язык: Английский

---

Comparative Evaluation of Machine Learning-Based Radiomics and Deep Learning for Breast Lesion Classification in Mammography

Diagnostics, Год журнала: 2025, Номер 15(8), С. 953 - 953

Опубликована: Апрель 9, 2025

Background: Breast cancer is the second leading cause of cancer-related mortality among women, accounting for 12% cases. Early diagnosis, based on identification radiological features, such as masses and microcalcifications in mammograms, crucial reducing rates. However, manual interpretation by radiologists complex subject to variability, emphasizing need automated diagnostic tools enhance accuracy efficiency. This study compares a radiomics workflow machine learning (ML) with deep (DL) approach classifying breast lesions benign or malignant. Methods: matRadiomics was used extract features from mammographic images 1219 patients CBIS-DDSM public database, including 581 cases 638 masses. Among ML models, linear discriminant analysis (LDA) demonstrated best performance both lesion types. External validation conducted private dataset 222 evaluate generalizability an independent cohort. Additionally, EfficientNetB6 model employed comparison. Results: The LDA achieved mean AUC 68.28% 61.53% In external validation, values 66.9% 61.5% were obtained, respectively. contrast, superior performance, achieving 81.52% 76.24% masses, highlighting potential DL improved accuracy. Conclusions: underscores limitations ML-based diagnosis. Deep proves be more effective approach, offering enhanced supporting clinicians improving patient management.

Язык: Английский

---

A Critical Analysis of the Robustness of Radiomics to Variations in Segmentation Methods in 18F-PSMA-1007 PET Images of Patients Affected by Prostate Cancer

Diagnostics, Год журнала: 2023, Номер 13(24), С. 3640 - 3640

Опубликована: Дек. 11, 2023

Radiomics shows promising results in supporting the clinical decision process, and much effort has been put into its standardization, thus leading to Imaging Biomarker Standardization Initiative (IBSI), that established how radiomics features should be computed. However, still lacks standardization many factors, such as segmentation methods, limit study reproducibility robustness.We investigated impact three different methods (manual, thresholding region growing) have on extracted from 18F-PSMA-1007 Positron Emission Tomography (PET) images of 78 patients (43 Low Risk, 35 High Risk). Segmentation was repeated for each patient, datasets segmentations. Then, feature extraction performed dataset, 1781 (107 original, 930 Laplacian Gaussian (LoG) features, 744 wavelet features) were extracted. Feature robustness assessed through intra class correlation coefficient (ICC) measure agreement between methods. To assess had machine learning models, selection a hybrid descriptive-inferential method, selected given input classifiers, K-Nearest Neighbors (KNN), Support Vector Machines (SVM), Linear Discriminant Analysis (LDA), Random Forest (RF), AdaBoost Neural Networks (NN), whose performance discriminating low-risk high-risk validated 30 times five-fold cross validation.Our showed influence Shape least reproducible (average ICC: 0.27), while GLCM most reproducible. Moreover, changed depending type, resulting 51.18% LoG exhibiting excellent (range average 0.68-0.87) 47.85% poor varied sub-bands 0.34-0.80) resulted LLL band showing highest ICC (0.80). Finally, model growing led accuracy (74.49%), improved sensitivity (84.38%) AUC (79.20%) contrast with manual segmentation.

Язык: Английский

---

Machine learning models combining computed tomography semantic features and selected clinical variables for accurate prediction of the pathological grade of bladder cancer

Frontiers in Oncology, Год журнала: 2023, Номер 13

Опубликована: Май 8, 2023

The purpose of this research was to develop a radiomics model that combines several clinical features for preoperative prediction the pathological grade bladder cancer (BCa) using non-enhanced computed tomography (NE-CT) scanning images.The (CT), clinical, and data 105 BCa patients attending our hospital between January 2017 August 2022 were retrospectively evaluated. study cohort comprised 44 low-grade 61 high-grade patients. subjects randomly divided into training (n = 73) validation 32) cohorts at ratio 7:3. Radiomic extracted from NE-CT images. A total 15 representative screened least absolute shrinkage selection operator (LASSO) algorithm. Based on these characteristics, six models predicting grade, including support vector machine (SVM), k-nearest neighbor (KNN), gradient boosting decision tree (GBDT), logical regression (LR), random forest (RF), extreme (XGBOOST) constructed. combining score factors further predictive performance evaluated based area under receiver operating characteristic (ROC) curve, DeLong test, curve analysis (DCA).The selected included age tumor size. LASSO identified most linked which in learning model. SVM revealed highest AUC 0.842. nomogram signature variables showed accurate preoperatively. 0.919, whereas 0.854. value combined validated calibration DCA.Machine CT semantic can accurately predict BCa, offering non-invasive approach

Язык: Английский

---

QuantImage v2: a comprehensive and integrated physician-centered cloud platform for radiomics and machine learning research

European Radiology Experimental, Год журнала: 2023, Номер 7(1)

Опубликована: Март 22, 2023

Radiomics, the field of image-based computational medical biomarker research, has experienced rapid growth over past decade due to its potential revolutionize development personalized decision support models. However, despite research momentum and important advances toward methodological standardization, translation radiomics prediction models into clinical practice only progresses slowly. The lack physicians leading insufficient integration tools in workflow contributes this slow uptake.We propose a physician-centered vision derive minimal functional requirements for software vision. Free-to-access frameworks were reviewed identify best practices reveal shortcomings existing solutions optimally physician-driven environment.Support user-friendly evaluation via machine learning was found be missing most tools. QuantImage v2 (QI2) designed implemented address these shortcomings. QI2 relies on well-established open-source libraries realize concretely demonstrate one-stop tool research. It provides web-based access cohort management, feature extraction, visualization supports "no-code" against patient-specific outcome data.QI2 fills gap landscape by enabling including model validation, environment. Further information about QI2, public instance system, source code is available at https://medgift.github.io/quantimage-v2-info/ . Key points As domain experts, play key role Existing do not optimally. implements web-based, platform.

Язык: Английский

---