*2.2. GB Histological Samples*

The specimens investigated in this research work consist of human biopsies extracted during brain tumor resection procedures. The pathological slides in this study were processed and analyzed by the Pathological Anatomy Department of the University Hospital Doctor Negrín at Las Palmas of Gran Canaria (Spain). The study protocol and consent procedures were approved by the Comité Ético de Investigación Clínica-Comité de Ética en la Investigación (CEIC/CEI) of the same hospital. After the resection, the samples were dehydrated and embedded in paraffin blocks. The blocks were then mounted in microtomes and sliced in 4 μm thick slices. Finally, the slices were rehydrated and stained with H&E. After routine examination of the samples, every sample was diagnosed by pathologists as GB, according to the World Health Organization (WHO) classification of tumors of the nervous system [25]. After the pathologist confirmed the GB diagnosis, macroscopic annotations of the GB locations were made on the physical glass slides using a marker-pen. Non-tumor areas are defined as areas in the pathological slide where there is no discrete presence of tumor cells. Within the areas annotated by a pathologist, we selected regions of interest (ROI) that were subsequently digitized using HS instrumentation. Within each ROI, different numbers of HS images were acquired for analysis. Figure 2 shows an example of the annotations within the pathological slide, the selection of different ROIs (shown at 5×), and the HS images (imaged at 20×) that are used in this study for classification. In

this case, red color annotations indicate areas diagnosed as GB, while non-tumor areas were annotated in blue marker. In this feasibility study, a total of 13 patients were analyzed.

**Figure 2.** Pathological samples used in this study. (**a**) Macroscopic annotations performed in pathological slides after diagnosis. Blue squares denote regions of interest (ROIs) within annotations; (**b**) ROIs from (**a**) shown at 5×; (**c**) Examples of HS images used in this study for classification (imaged at 20×).
