PET radiomics for histologic subtype classification of non-small cell lung cancer: a systematic review and meta-analysis

European Journal of Nuclear Medicine and Molecular Imaging, Journal Year: 2025, Volume and Issue: unknown

Published: Jan. 11, 2025

Language: Английский

---

Artificial Intelligence Assisted 18F-FDG PET Radiomics in Classifying Histological Subtypes of Lung Cancer: Systematic Review and Meta-analysis

Nuclear Medicine and Molecular Imaging, Journal Year: 2025, Volume and Issue: unknown

Published: May 17, 2025

Language: Английский

---

Fully Automated Region-Specific Human-Perceptive-Equivalent Image Quality Assessment: Application to 18F-FDG PET Scans

Clinical Nuclear Medicine, Journal Year: 2024, Volume and Issue: 49(12), P. 1079 - 1090

Published: Oct. 21, 2024

Introduction We propose a fully automated framework to conduct region-wise image quality assessment (IQA) on whole-body 18 F-FDG PET scans. This (1) can be valuable in daily clinical acquisition procedures instantly recognize low-quality scans for potential rescanning and/or reconstruction, and (2) make significant impact dataset collection the development of artificial intelligence–driven analysis models by rejecting images those presenting with artifacts, toward building clean datasets. Patients Methods Two experienced nuclear medicine physicians separately evaluated 174 from 87 patients, each body region, based 5-point Likert scale. The regisons included following: head neck, including brain, chest, (3) chest-abdomen interval (diaphragmatic region), (4) abdomen, (5) pelvis. Intrareader interreader reproducibility scores were calculated using 39 randomly selected dataset. Utilizing binarized classification, dichotomized into versus high-quality physician ≤3 >3, respectively. Inputting PET/CT scans, our proposed applies 2 deep learning (DL) CT perform region identification contour extraction (excluding extremities), then classifies regions as low high quality. For mainstream approaches, machine (ML) radiomic features DL, investigated. All trained attributed physician, average reported. DL radiomics-ML same test performance evaluation was carried out area under curve, accuracy, sensitivity, specificity compared Delong P values <0.05 regarded statistically significant. Results In interval, pelvis regions, best achieved [0.97, 0.95, 0.96, 0.95], [0.85, 0.82, 0.87, 0.76], [0.83, 0.76, 0.68, 0.80], [0.73, 0.72, 0.64, 0.77], [0.72, 0.70, 0.67], all revealed highest performance, when developed higher intrareader reproducibility. Comparison did not show any differences, though showed overall improved trends. Conclusions human-perceptive equivalent model IQA over images. Our emphasizes necessity developing separate performing data annotation multiple experts’ consensus studies.

Language: Английский

---

Ablation for Single Pulmonary Nodules, Primary or Metastatic. Εndobronchial Ablation Systems or Percutaneous

Journal of Cancer, Journal Year: 2024, Volume and Issue: 15(4), P. 880 - 888

Published: Jan. 1, 2024

Single pulmonary nodules are a difficult to diagnose imagining artifact.Currently novel diagnostic tools such as Radial-EBUS with or not C-ARM flouroscopy, electromagnetic navigation systems, robotic bronchoscopy and cone beam-compuer tomography (CBCT) can assist in the optimal guidance of biopsy equipment.After diagnosis lung cancer metastatic disease nodule, then surgery ablation methods local treatment be applied.The percutaneous systems under computed radiofrequency, microwave, cryo thermosphere have been used for several years.In past 10 years extensive research has made endobronchial methods.We will present comment on two different up date data.

Language: Английский

---

MRI-based radiomics for predicting histology in malignant salivary gland tumors: methodology and “proof of principle”

Scientific Reports, Journal Year: 2024, Volume and Issue: 14(1)

Published: April 30, 2024

Defining the exact histological features of salivary gland malignancies before treatment remains an unsolved problem that compromises ability to tailor further therapeutic steps individually. Radiomics, a new methodology extract quantitative information from medical images, could contribute characterizing individual cancer phenotype already in fast and non-invasive way. Consequently, standardization implementation radiomic analysis clinical routine work predict histology (SGC) also provide improvements decision-making. In this study, we aimed investigate potential as imaging biomarker distinguish between high grade low-grade malignancies. We have investigated effect image feature level harmonization on performance models. For our dual center cohort consisted 126 patients, with histologically proven SGC, who underwent curative-intent two tertiary oncology centers. extracted analyzed radiomics 120 pre-therapeutic MRI images gadolinium (T1 sequences), correlated those definitive post-operative histology. study best model achieved average AUC 0.66 balanced accuracy 0.63. According results, there is significant difference models based intensity normalized + harmonized other (p value < 0.05) which indicates case dealing heterogeneous dataset, applying methods beneficial. Among minimum first order, gray level-variance texture category were frequently selected during multivariate indicate these being used biomarker. The present bicentric presents for time feasibility implementing MR-based, handcrafted radiomics, T1 contrast-enhanced sequences ComBat method effort formal grading carcinoma satisfactory performance.

Language: Английский

---

CT-based radiomics analysis for prediction of pathological subtypes of lung adenocarcinoma

Journal of Radiation Research and Applied Sciences, Journal Year: 2024, Volume and Issue: 17(4), P. 101174 - 101174

Published: Nov. 3, 2024

Language: Английский

---

Differential diagnosis of radiation encephalopathy and post-radiation brain tumor recurrence by machine learning models based on contrast-enhanced MRI

Published: Jan. 26, 2024

Language: Английский

---