Evolution of Personalized Dosimetry for Radioembolization of Hepatocellular Carcinoma

Yttrium-90 transarterial radioembolization (TARE) has progressed from a salvage or palliative lobar or sequential bilobar regional liver therapy for patients with advanced disease to a versatile, potentially curative, and often highly selective local treatment for patients across Barcelona Clinic Liver Cancer stages. With this shift, radiation dosimetry has evolved to become more tailored to patients and target lesion(s), with treatment dose and distributions adapted for specific clinical goals (ie, palliation, bridging or downstaging to liver transplantation, converting to surgical resection candidacy, or ablative/curative intent). Data have confirmed that “personalizing” dosimetry yields real-world improvements in tumor response and overall survival while maintaining a favorable adverse event profile. In this review, imaging techniques used before, during, and after TARE have been reviewed. Historical algorithms and contemporary image-based dosimetry methods have been reviewed and compared. Finally, recent and upcoming developments in TARE methodologies and tools have been discussed.     

Fully automated radiomics-based machine learning models for multiclass classification of single brain tumors: Glioblastoma,lymphoma, and metastasis

Background and purposeTo investigate the diagnostic performance of fully automated radiomics-based models for multiclass classification of a single enhancing brain tumor among glioblastoma, central nervous system lymphoma, and metastasis.Materials and methodsThe training and test sets were comprised of 538 cases (300 glioblastomas, 73 lymphomas, and 165 metastases) and 169 cases (101 glioblastomas, 29 lymphomas, and 39 metastases), respectively. After fully automated segmentation, radiomic features were extracted. Three conventional machine learning classifiers, including least absolute shrinkage and selection operator (LASSO), adaptive boosting (Adaboost), and support vector machine with the linear kernel (SVC), combined with one of four feature selection methods, including forward sequential feature selection, F score, mutual information, and LASSO, were trained. Additionally, one ensemble classifier based on the three classifiers was used. The diagnostic performance of the optimized models was tested in the test set using the accuracy, F1-macro score, and the area under the receiver operating characteristic curve (AUCROC).ResultsThe best performance was achieved when the LASSO was used as a feature selection method. In the test set, the best performance was achieved by the ensemble classifier, showing an accuracy of 76.3% (95% CI, 70.0–82.7), a F1-macro score of 0.704, and an AUCROC of 0.878.ConclusionOur fully automated radiomics-based models for multiclass classification might be useful for differential diagnosis of a single enhancing brain tumor with a good diagnostic performance and generalizability.