Abstract:Background: Retinoblastoma (RB) is the most common malignant tumor in children under the age of 3 years and is associated with a high disability and mortality rate. The aim of this study was, first, to evaluate the clinical efficacy of multimodal imaging in differentially diagnosing RB in children and in predicting the efficacy of selective ophthalmic artery infusion (SOAI) and, second, to identify the factors associated with this efficacy. Methods: This study retrospectively collected the data from 256 children with unilateral RB and intraocular involvement, including multimodal imaging magnetic resonance imaging (MRI), computed tomography (CT), and clinical characteristics. Among the cases, 33 with both CT and MRI data available were used to evaluate the diagnostic accuracy in distinguishing RB, with histopathological results serving as the gold standard. Additionally, a retrospective analysis was conducted on the MRI and clinical characteristics of 256 cases of unilateral RB with intraocular involvement before SOAI treatment. The predictive ability of imaging features and clinical characteristics for the treatment efficacy of children was analyzed, and the differences in globe salvage rates and visual preservation based on different tumor stages were evaluated. Results: The diagnostic accuracy of CT imaging for RB was 96.96% while that of MRI was 84.84%, with both showing high consistency with the histopathological results. CT images demonstrated a posterior intraocular mass with a high-density appearance, with spots, patches, or clustered calcifications visible within the tumor. The CT values were mostly above 100 Hounsfield units (HU), and enhanced scanning showed varying degrees of enhancement in noncalcified masses. MRI showed low or moderate signal intensity on T1-weighted images and moderate-to-high signal intensity on T2-weighted images, with significant enhancement after contrast administration. Tumors with more calcifications showed long T1 and short T2 signals. Patients with better prognosis had a higher delta signal increase (ΔSI), a greater distance from the optic disc, smaller tumor diameter, absence of implantation nodules or smaller implantation range, endogenous growth pattern, smaller extent of retinal detachment, absence of clinical high-risk factors, no vitreous hemorrhage, no globe shrinkage, and smaller calcification volume. The distance between the tumor and optic disc, clinical high-risk factors, and tumor growth pattern were found to be independent factors associated with prognosis. The rate of successful globe salvage and visual acuity decreased with increasing tumor stage. Conclusions: CT and MRI are highly valuable for the comprehensive assessment of tumors in pediatric RB. MRI alone can complete a comprehensive assessment of patients with RB and thus allow for the reduction radiation dose in children. Calcification of the tumor is crucial for diagnosis, and imaging findings can serve to inform patient prognosis and treatment planning. The distance between the tumor and optic disc, clinical high-risk factors, and tumor growth pattern are closely related to the prognosis of children.

Nanostructure analysis in polymeric materials with ion beam based Raman spectroscopy for retinoblastoma tumor imaging using ensemble machine learning technique

Early Tumor Detection Afforded by in Vivo Imaging of Near-Infrared Ii Fluorescence

Screening Ovarian Cancers with Raman Spectroscopy of Blood Plasma Coupled with Machine Learning Data Processing.

Nimg-18. Rapid Intraoperative Differentiation Of Primary Cns Lymphoma From Common Cns Tumors Using Stimulated Raman Histology And Deep Learning

Rapid Automated Analysis of Skull Base Tumor Specimens Using Intraoperative Optical Imaging and Artificial Intelligence

Early Prediction and Diagnosis of Retinoblastoma Using Deep Learning Techniques

Monitoring Retinoblastoma by Machine Learning of Aqueous Humor Metabolic Fingerprinting (Small Methods 1/2022)

Raman optical identification of renal cell carcinoma via machine learning

Machine Learning-Assisted Classification of Paraffin-Embedded Brain Tumors with Raman Spectroscopy

Machine learning approaches in non-contact autofluorescence spectrum classification

Surface-Enhanced Raman Spectroscopy and Transfer Learning Toward Accurate Reconstruction of the Surgical Zone

Multi photon micro material analysis based on Raman spectroscopy biosensor for cancer detection using biomarker with deep learning techniques

Deep learning models for benign and malign ocular tumor growth estimation

Raman spectroscopy and artificial intelligence to predict the Bayesian probability of breast cancer

Low coherence quantitative phase microscopy with machine learning model and Raman spectroscopy for the study of breast cancer cells and their classification

Improving the Accuracy of Brain Tumor Surgery Via Raman-based Technology

Early Detection of Pre-Cancerous and Cancerous Cells Using Raman Spectroscopy-Based Machine Learning

IMG-05. USING CONVOLUTIONAL NEURAL NETWORKS TO IDENTIFY COMMON MOLECULAR ALTERATIONS IN PEDIATRIC BRAIN TUMORS

Multimodal imaging for the differential diagnosis and efficacy evaluation of intraocular retinoblastoma in children with selective ophthalmic artery infusion

A Two-Stage Detection Methodology for Thyroid Cancer Using Photonic Crystal: Logistic Regression and Artificial Neural Networks

Label-Free Neurosurgical Pathology with Stimulated Raman Imaging.