Polymer gel dosimeters are fabricated from radiation sensitive chemicals which, upon irradiation, polymerize as a function of the absorbed radiation dose. These gel dosimeters, with the capacity to uniquely record the radiation dose distribution in three-dimensions (3D), have specific advantages when compared to one-dimensional dosimeters, such as ion chambers, and two-dimensional dosimeters, such as film. These advantages are particularly significant in dosimetry situations where steep dose gradients exist such as in intensity-modulated radiation therapy (IMRT) and stereotactic radiosurgery. Polymer gel dosimeters also have specific advantages for brachytherapy dosimetry. Potential dosimetry applications include those for low-energy x-rays, high-linear energy transfer (LET) and proton therapy, radionuclide and boron capture neutron therapy dosimetries. These 3D dosimeters are radiologically soft-tissue equivalent with properties that may be modified depending on the application. The 3D radiation dose distribution in polymer gel dosimeters may be imaged using magnetic resonance imaging (MRI), optical-computerized tomography (optical-CT), x-ray CT or ultrasound. The fundamental science underpinning polymer gel dosimetry is reviewed along with the various evaluation techniques. Clinical dosimetry applications of polymer gel dosimetry are also presented.

Although machine learning (ML) has shown promise across disciplines, out-of-sample generalizability is concerning. This is currently addressed by sharing multi-site data, but such centralization is challenging/infeasible to scale due to various limitations. Federated ML (FL) provides an alternative paradigm for accurate and generalizable ML, by only sharing numerical model updates. Here we present the largest FL study to-date, involving data from 71 sites across 6 continents, to generate an automatic tumor boundary detector for the rare disease of glioblastoma, reporting the largest such dataset in the literature (n = 6, 314). We demonstrate a 33% delineation improvement for the surgically targetable tumor, and 23% for the complete tumor extent, over a publicly trained model. We anticipate our study to: 1) enable more healthcare studies informed by large diverse data, ensuring meaningful results for rare diseases and underrepresented populations, 2) facilitate further analyses for glioblastoma by releasing our consensus model, and 3) demonstrate the FL effectiveness at such scale and task-complexity as a paradigm shift for multi-site collaborations, alleviating the need for data-sharing.