A5057938768- Radiomics and Machine Learning in Medical Imaging
- Artificial Intelligence in Healthcare and Education
- AI in cancer detection
- Head and Neck Cancer Studies
- Lung Cancer Diagnosis and Treatment
- Advanced X-ray and CT Imaging
- Medical Imaging Techniques and Applications
- Advanced Radiotherapy Techniques
- Sarcoma Diagnosis and Treatment
- MRI in cancer diagnosis
- Colorectal Cancer Surgical Treatments
- Colorectal Cancer Screening and Detection
- Medical Imaging and Analysis
- Colorectal and Anal Carcinomas
- Cancer Diagnosis and Treatment
- Pancreatic and Hepatic Oncology Research
- Advances in Oncology and Radiotherapy
- Gastric Cancer Management and Outcomes
- Brain Tumor Detection and Classification
- Dental Radiography and Imaging
- COVID-19 diagnosis using AI
University of Toronto
2020-2025
Princess Margaret Cancer Centre
2021-2025
University Health Network
2021-2025
Maastro Clinic
2018-2020
Maastricht University Medical Centre
2019-2020
Maastricht University
2018-2019
Aarhus University Hospital
2018
Artificial intelligence (AI) and machine learning (ML) are becoming critical in developing deploying personalized medicine targeted clinical trials. Recent advances ML have enabled the integration of wider ranges data including both medical records imaging (radiomics). However, development prognostic models is complex as no modeling strategy universally superior to others validation developed requires large diverse datasets demonstrate that (regardless method) from one dataset applicable...
The aims of this study are to evaluate the stability radiomic features from Apparent Diffusion Coefficient (ADC) maps cervical cancer with respect to: (1) reproducibility in inter-observer delineation, and (2) image pre-processing (normalization/quantization) prior feature extraction.Two observers manually delineated tumor on ADC derived pre-treatment diffusion-weighted Magnetic Resonance imaging 81 patients FIGO stage IB-IVA cancer. First-order, shape, texture were extracted original...
PurposeHighlighting the risk of biases in radiomics-based models will help improve their quality and increase usage as decision support systems clinic. In this study we use machine learning-based methods to identify presence volume-confounding effects radiomics features.Methods841 features were extracted from two retrospective publicly available datasets lung head neck cancers using open source software. Unsupervised hierarchical clustering principal component analysis (PCA) identified...
<ns3:p>Background Machine learning and AI promise to revolutionize the way we leverage medical imaging data for improving care but require large datasets train computational models that can be implemented in clinical practice. However, processing complex remains an open challenge. Methods To address this issue, developed Med-ImageTools, a new Python open-source software package automate curation while allowing researchers share their configurations more easily, lowering barrier other...
Abstract Purpose This manuscript presents RADCURE, one of the most extensive head and neck cancer (HNC) imaging datasets accessible to public. Initially collected for clinical radiation therapy (RT) treatment planning, this dataset has been retrospectively reconstructed use in research. Acquisition Validation Methods RADCURE encompasses data from 3346 patients, featuring computed tomography (CT) RT simulation images with corresponding target organ‐at‐risk contours. These CT scans were using...
Background and purposeComputed tomography (CT) is one of the most common medical imaging modalities in radiation oncology radiomics research, computational voxel-level analysis images. Radiomics vulnerable to effects dental artifacts (DA) caused by metal implants or fillings can hamper future reproducibility on new datasets. In this study we seek better understand robustness quantitative radiomic features DAs. Furthermore, propose a novel method detecting DAs order safeguard studies improve...
Predicting outcomes, such as survival or metastasis for individual cancer patients is a crucial component of precision oncology. Machine learning (ML) offers promising way to exploit rich multi-modal data, including clinical information and imaging learn predictors disease trajectory help inform decision making. In this paper, we present novel graph-based approach incorporate characteristics existing spread local lymph nodes (LNs) well their connectivity patterns in prognostic ML model. We...
<ns3:p><ns3:bold>Background: </ns3:bold>Machine learning and AI promise to revolutionize the way we leverage medical imaging data for improving care but require large datasets train computational models that can be implemented in clinical practice. However, processing complex remains an open challenge.</ns3:p><ns3:p> <ns3:bold>Methods: </ns3:bold>To address this issue, developed Med-ImageTools, a new Python open-source software package automate curation while allowing researchers share their...
Radical treatment of patients diagnosed with inoperable and locally advanced head neck cancers (LAHNC) is still a challenge for clinicians. Prediction incomplete response (IR) primary tumour would be value to the optimization LAHNC. Aim this study was develop evaluate models based on clinical radiomics features prediction IR in LAHNC treated definitive chemoradiation or radiotherapy. Clinical imaging data 290 were included into retrospective study. model built patient related features....
<ns4:p>Background Machine learning and AI promise to revolutionize the way we leverage medical imaging data for improving care but require large datasets train computational models that can be implemented in clinical practice. However, processing complex remains an open challenge. Methods To address this issue, developed Med-ImageTools, a new Python open-source software package automate curation while allowing researchers share their configurations more easily, lowering barrier other...
Abstract Accurate prognosis for an individual patient is a key component of precision oncology. Recent advances in machine learning have enabled the development models using wider range data, including imaging. Radiomics aims to extract quantitative predictive and prognostic biomarkers from routine medical imaging, but evidence computed tomography radiomics remains inconclusive. We conducted institutional challenge develop accurate model overall survival prediction head neck cancer clinical...
Abstract Purpose We developed QUANNOTATE , a new web-application for rapid review of radiotherapy (RT) target volumes, and used it to evaluate the relationship between delineation compliance with international guidelines treatment outcomes in nasopharyngeal carcinoma (NPC) patients undergoing definitive RT. Methods Materials The dataset this study consists anonymized CT simulation scans, RT structures, clinical data 332 pathologically confirmed NPC treated intensity-modulated July 2005...
6545 Background: Immunotherapy has become a standard of care in the treatment R/M HNSCC, however only subset patients respond, highlighting need for predictive and prognostic biomarkers. Radiomics is non-invasive method to quantitatively analyze tumors through conventional imaging. Methods: The pre-treatment first-on-treatment (after 8 weeks) computed tomography (CT) scans from 132 HNSCC treated with single-agent Pembrolizumab (10mg/kg Q2W or 200mg Q3W IV) on KEYNOTE-012 study were analyzed....
ABSTRACT Computed tomography (CT) is one of the most common medical imaging modalities and main technology used in radiomics research, computational voxel-level analysis images. Analysis CT images vulnerable to effects dental artifacts (DA) caused by metal implants or fillings. Running automated pipelines with uncurated datasets can reduce performance hamper future reproducibility on new datasets. This work introduces a tool detect location magnitude DAs based combination deep learning...
Accurate survival prediction is crucial for development of precision cancer medicine, creating the need new sources prognostic information. Recently, there has been significant interest in exploiting routinely collected clinical and medical imaging data to discover markers multiple types. However, most previous studies focus on individual modalities alone do not make use recent advances machine learning prediction. We present Deep-CR MTLR -- a novel approach accurate from multi-modal...
Background and Purpose: Auto-segmentation of organs at risk (OAR) in cancer patients is essential for enhancing radiotherapy planning efficacy reducing inter-observer variability. Deep learning auto-segmentation models have shown promise, but their lack transparency reproducibility hinders generalizability clinical acceptability, limiting use settings. Materials Methods: This study introduces SCARF (auto-Segmentation Clinical Acceptability & Reproducibility Framework), a comprehensive...
<p>Model performances across permutations by clinical and imaging variables.</p>
<p>Model performances across permutations by clinical and imaging variables.</p>
<div>Abstract<p>Artificial intelligence (AI) and machine learning (ML) are becoming critical in developing deploying personalized medicine targeted clinical trials. Recent advances ML have enabled the integration of wider ranges data including both medical records imaging ({radiomics}). However, development prognostic models is complex as no modelling strategy universally superior to others validation developed requires large diverse datasets demonstrate that (regardless method)...
<p>Calibration of predicted 2-year event probabilities for the best performing model in each category and ensemble all models.</p>
<p>Supplementary Data of figures/tables</p>
<p>Calibration of predicted 2-year event probabilities for the best performing model in each category and ensemble all models.</p>