A5056339832- Radiomics and Machine Learning in Medical Imaging
- Cancer Immunotherapy and Biomarkers
- Ferroptosis and cancer prognosis
- AI in cancer detection
- Cancer Genomics and Diagnostics
- Advanced X-ray and CT Imaging
- Colorectal Cancer Treatments and Studies
- Medical Imaging Techniques and Applications
- COVID-19 diagnosis using AI
- Renal cell carcinoma treatment
- Colorectal and Anal Carcinomas
- MRI in cancer diagnosis
- Bladder and Urothelial Cancer Treatments
- Cancer, Stress, Anesthesia, and Immune Response
- Cancer Diagnosis and Treatment
- Digital Imaging for Blood Diseases
- CAR-T cell therapy research
- Immunotherapy and Immune Responses
- Glioma Diagnosis and Treatment
- Veterinary Equine Medical Research
- Lung Cancer Diagnosis and Treatment
- Gene expression and cancer classification
- Radiopharmaceutical Chemistry and Applications
- melanin and skin pigmentation
- Medical Imaging and Analysis
Vall d'Hebron Institute of Oncology
2019-2025
University Hospital Carl Gustav Carus
2024-2025
Technische Universität Dresden
2024-2025
Fresenius (Germany)
2024-2025
Universidad de Sevilla
2024
Vall d'Hebron Hospital Universitari
2020-2023
CIBBIM-Nanomedicine
2023
Abstract Objective To identify CT-acquisition parameters accounting for radiomics variability and to develop a post-acquisition CT-image correction method reduce improve classification in both phantom clinical applications. Methods protocols were prospectively tested phantom. The multi-centric retrospective study included CT scans of patients with colorectal/renal cancer liver metastases. Ninety-three features first order texture extracted. Intraclass correlation coefficients (ICCs) between...
The efficacy of immune checkpoint inhibitors (ICIs) depends on the tumor microenvironment (TIME), with a preference for T cell-inflamed TIME. However, challenges in tissue-based assessments via biopsies have triggered exploration non-invasive alternatives, such as radiomics, to comprehensively evaluate TIME across diverse cancers. To address these challenges, we develop an ICI response signature by integrating radiomics gene-expression profiles. We conducted pan-cancer investigation into...
Background Reliable predictive imaging markers of response to immune checkpoint inhibitors are needed. Purpose To develop and validate a pretreatment CT-based radiomics signature predict in advanced solid tumors. Materials Methods In this retrospective study, was developed patients with tumors (including breast, cervix, gastrointestinal) treated anti-programmed cell death-1 or programmed death ligand-1 monotherapy from August 2012 May 2018 (cohort 1). This tested bladder lung cancer (cohorts...
Immunotherapy by immune checkpoint inhibitors has become a standard treatment strategy for many types of solid tumors. However, the majority patients with cancer will not respond, and predicting response to this therapy is still challenge. Artificial intelligence (AI) methods can extract meaningful information from complex data, such as image data. In clinical routine, radiology or histopathology images are ubiquitously available. AI been used predict immunotherapy images, either directly...
Purpose To identify precise three-dimensional radiomics features in CT images that enable computation of stable and biologically meaningful habitats with machine learning for cancer heterogeneity assessment. Materials Methods This retrospective study included 2436 liver or lung lesions from 605 scans (November 2010–December 2021) 331 patients (mean age, 64.5 years ± 10.1 [SD]; 185 male patients). Three-dimensional were computed original perturbed (simulated retest) different combinations...
Abstract Medical image classification requires labeled, task-specific datasets which are used to train deep learning networks de novo, or fine-tune foundation models. However, this process is computationally and technically demanding. In language processing, in-context provides an alternative, where models learn from within prompts, bypassing the need for parameter updates. Yet, remains underexplored in medical analysis. Here, we systematically evaluate model Generative Pretrained...
Abstract Glioblastoma is the most common primary brain tumor. Standard therapy consists of maximum safe resection combined with adjuvant radiochemotherapy followed by chemotherapy temozolomide, however prognosis extremely poor. Assessment residual tumor after surgery and patient stratification into prognostic groups (i.e., volume) currently hindered subjective evaluation enhancement in medical images (magnetic resonance imaging [MRI]). Furthermore, objective evidence defining optimal time to...
Background Foundation models (FMs) show promise in medical AI by learning flexible features from large datasets, potentially surpassing handcrafted radiomics. Outcome prediction of head and neck squamous cell carcinomas (HNSCC) with FMs using routine imaging remains unexplored. Purpose To evaluate end-to-end FM-based multiple instance (MIL) for 2-year overall survival (OS), locoregional control (LRC), freedom distant metastasis (FFDM) risk group stratification pretreatment CT scans HNSCC....
Background Immune checkpoint inhibitors (ICIs) are the gold standard therapy in patients with deficient mismatch repair (dMMR)/microsatellite instability-high (MSI-H) metastatic colorectal cancer (mCRC). A significant proportion of show resistance, making identification determinants response crucial. Growing evidence supports role sex determining susceptibility to anticancer therapies, but data is lacking for MSI-H CRC. Methods In this real-world cohort comprising 624 mCRC receiving ICIs, we...
Abstract Multimodal vision-language models (VLM) have shown promising biomarker classification performance in individual data modalities such as radiology and pathology. By learning joint representations of visual textual information, these methods to provide better features from medical images. However, current are limited the multimodal landscape 2D images pathology, leaving a substantial gap for creation with 3D data. In this study, we introduce novel, VLM that can process both pathology...
Abstract Tumor heterogeneity has been postulated as a hallmark of treatment resistance and cure constraint in cancer patients. Conventional quantitative medical imaging (radiomics) can be extended to computing voxel-wise features aggregating tumor subregions with similar radiological phenotypes (imaging habitats) elucidate the distribution within among tumors. Despite promising applications habitats, they may affected by variability radiomics features, preventing comparison generalization...
Abstract Programmed death-ligand 1 (PD-L1) IHC is the most commonly used biomarker for immunotherapy response. However, quantification of PD-L1 status in pathology slides challenging. Neither manual nor a computer-based mimicking readouts perfectly reproducible, and predictive performance both approaches regarding response limited. In this study, we developed deep learning (DL) method to predict directly from raw image data, without explicit intermediary steps such as cell detection or...
Abstract The search for understanding immunotherapy response has sparked interest in diverse areas of oncology, with artificial intelligence (AI) and radiomics emerging as promising tools, capable gathering large amounts information to identify suitable patients treatment. application AI radiology grown, driven by the hypothesis that images capture tumor phenotypes thus could provide valuable insights into likelihood. However, despite rapid growth studies, no algorithms field have reached...
Abstract Chimeric antigen receptor (CAR) T‐cell therapy is a promising treatment option for relapsed or refractory (R/R) large B‐cell lymphoma (LBCL). However, only subset of patients will present long‐term benefit. In this study, we explored the potential PET‐based radiomics to predict outcomes with aim improving patient selection CAR therapy. We conducted single‐center study including 93 consecutive R/R LBCL who received infusion from 2018 2021, split in training set (73 patients) and test...
Dimensionality reduction is key to alleviate machine learning artifacts in clinical applications with Small Sample Size (SSS) unbalanced datasets. Existing methods rely on either the probabilistic distribution of training data or discriminant power reduced space, disregarding impact repeatability and uncertainty features.In present study proposed use reproducibility radiomics features select high inter-class correlation coefficient (ICC). The includes variability introduced image...
Medical image classification requires labeled, task-specific datasets which are used to train deep learning networks de novo, or fine-tune foundation models. However, this process is computationally and technically demanding. In language processing, in-context provides an alternative, where models learn from within prompts, bypassing the need for parameter updates. Yet, remains underexplored in medical analysis. Here, we systematically evaluate model Generative Pretrained Transformer 4 with...
White markings are characteristic of many equine breeds, being quite common in the Pura Raza Español horses (PRE). These white result a lack melanocytes skin and hair. In certain horse such as PRE, presence extension facial is penalised breed's patron morphological competitions, so it would be interesting to include their selection genetics programs select against this special feature. The aim study was calculate prevalence representative population sample PRE determine its depending on coat...
448 Background: Hyperprogressive disease (HPD) is a new pattern of progression in cancer patients (pts) treated with immune checkpoint inhibitors (ICI). The rate and outcome HDP pts metastatic renal carcinoma (RCC) urothelial (UC) are unknown. Here, we report the percentage HPD cohort GU malignancies at our center explore associations clinical variables. Methods: Medical records from phase I-III trials ICI alone or combination between July 2013 June 2018 were retrospectively reviewed. We...
Hematoxylin- and eosin (H&E) stained whole-slide images (WSIs) are the foundation of diagnosis cancer. In recent years, development deep learning-based methods in computational pathology enabled prediction biomarkers directly from WSIs. However, accurately linking tissue phenotype to at scale remains a crucial challenge for democratizing complex precision oncology. This protocol describes practical workflow solid tumor associative modeling (STAMP), enabling WSIs using learning. The STAMP is...
<p>Performance overview of the model for predicting PD-L1 status and response to immunotherapy in NSCLC-MSK pan-cancer-VHIO cohort. AUC curves predict (TPS ≥ 1%) training (NSCLC-MSK cohort; <b>A</b>) test cohort (pan-cancer-VHIO <b>B</b>) 5-folds cross-validation. All trained models were deployed Kaplan–Meier predicted (high/low) from differentiates patients with longer PFS shorter survival both (<b>C</b>) (<b>D</b>).</p>