The Digital Brain Tumour Atlas, an open histopathology resource.

Currently, approximately 150 different brain tumour types are defined by the WHO. Recent endeavours to exploit machine learning and deep learning methods for supporting more precise diagnostics based on the histological tumour appearance have been hampered by the relative paucity of accessible digital histopathological datasets. While freely available datasets are relatively common in many medical specialties such as radiology and genomic medicine, there is still an unmet need regarding histopathological data. Thus, we digitized a significant portion of a large dedicated brain tumour bank based at the Division of Neuropathology and Neurochemistry of the Medical University of Vienna, covering brain tumour cases from 1995-2019. A total of 3,115 slides of 126 brain tumour types (including 47 control tissue slides) have been scanned. Additionally, complementary clinical annotations have been collected for each case. In the present manuscript, we thoroughly discuss this unique dataset and make it publicly available for potential use cases in machine learning and digital image analysis, teaching and as a reference for external validation.

Background & Summary

Brain tumours account for a large fraction of years of potential life lost as compared with tumours from other sites[1], and have a significant negative impact on patients’ quality of life[2]. Overall, they are relatively uncommon neoplasms with an incidence of approximately 24 per 100.000 person-years[3]. Current diagnostic guidelines published by the WHO define approximately 150 distinct brain tumour types and assign grades I to IV, based on malignancy and potential to malignant transformation or progression. They are mainly differentiated by their histopathological phenotypes and molecular alterations[4]. While the majority of tumours is diagnosed solely based on histopathology, an integrated approach is mandatory for 19 tumour types.

Still, more accurate diagnostic distinctions are needed in order to i) better assess individual patients’ prognoses and ii) support more robust therapeutic decisions[4,5]. Recently, diagnostic algorithms trained on DNA methylation data have been shown to significantly increase diagnostic accuracy[6]. Similar advances focusing on histopathological data have been hampered, so far, by the lack of freely available histopathology datasets[7]. Most available histopathology data such as those available through TCGA[8], IvyGAP[9,10] or TCIA[11] focus on only a few diagnostic entities. They mostly consist of digitized fresh frozen tissue sections, which feature relatively poor tissue morphology as compared to formalin-fixed and paraffin-embedded tissues. Still, even with these limited data, computational algorithms have been successfully trained - amongst others - for survival prediction[12], detection of tumour-infiltrating lymphocytes[13], and assessments of tumour microvessels[14]. However, larger datasets encompassing an even wider range of brain tumours and featuring improved cellular and morphological characteristics are necessary to further develop these algorithms and extend their applicability to the entire spectrum of brain tumour types.

Thus, we set out to compile a comprehensive resource of digitized Haematoxylin-eosin(H&E)-stained brain tumour whole slide images (WSIs) with clinical annotations (Fig. 1). We aimed to capture the complete spectrum of brain tumours as encountered in day-to-day medical diagnostic practice. Importantly, we managed to specifically digitize slides of exceedingly rare pathologies, which are usually, if ever, seen only a few times in a pathologist’s lifetime. By performing a manual review of each slide, we ensure high scan quality and actuality of provided diagnoses. We envisage this dataset to be used for advancing digital pathology-based machine learning and for teaching purposes. Importantly, this dataset can be used for (1) inter-tumour comparisons thanks to the wide inclusion of distinct brain tumour types as well as (2) within-tumour-type investigations thanks to the inclusion of a large number of samples for the common tumour types.

Fig. 1

Overview of the data acquisition and publication process. First, histological slides and clinical records of brain tumour patients were retrieved from the biobank of the Division of Neuropathology and Neurochemistry, Medical University of Vienna. Then, slides were digitized using a Hamamatsu slidescanner. Clinical data were translated into standardized annotations. At least two experienced neuropathologists checked each slide scan to ensure conformity of the diagnosis with the current revised 4th edition of the “WHO Classification of Tumours of the Central Nervous System” and sufficient scan quality. Ambiguous cases were excluded and WSIs of inferior quality were re-scanned. Finally, data were made available via EBRAINS to the international research community. (Brain illustration adapted from Meaghan Hendricks from the Noun Project).

Methods

Sample acquisition

H&E stained tumour slides from FFPE tissues, which were collected for routine diagnostics in the time interval of 1995–2019 have been obtained from the biobank of the Division of Neuropathology & Neurochemistry, Medical University of Vienna. We digitized each slide in high magnification (40x objective, 228 nm/pixel) using a Hamamatsu NanoZoomer 2.0 HT slide-scanner. Each slide was manually reviewed to ensure high scan quality and sufficient diagnostic tumour tissue. Samples with equivocal diagnoses or missing molecular work-up otherwise needed to assign an integrated WHO 2016 diagnosis were excluded. A subset of glioblastoma scans (n = 381) has been published previously as part of the GBMatch study[15].

Basic clinical annotations consisting of patient age and sex as well as tumour location and recurrence were acquired from local electronic records where available. Tumour locations have been assigned to the following 19 categories: frontal; parietal; insular; occipital; temporal; cerebellar; brain stem; spinal; lateral ventricle; diencephalon; third ventricle; fourth ventricle; sellar region; cranial nerves; basal ganglia; cerebral, NOS (not otherwise specified); posterior fossa, NOS; cranial, NOS; and other.

This study complies with the relevant ethical, legal and institutional regulations and the study protocol has been approved by the Ethics Committee of the Medical University of Vienna (EK1691–2017). Participant informed consent has been obtained as by institutional guidelines, necessitating restrictions on commercial use of the obtained data.

Estimation of cell density and scanned tissue area

Additionally, the total tissue area and the average cellularities were estimated for each scan using a custom MATLAB script (MATLAB R2017b, MathWorks) with a similar approach as previously published[15,16]. In summary, H&E stained WSIs were first colour-deconvoluted into separate Haematoxylin and Eosin channels[17]. Then, global, Phansalkar and Otsu thresholding were applied to the Haematoxylin channel to identify nuclei[18,19]. Watershedding was used to separate densely clustered cells[20]. Only cells with a minimum size of 4 pixels were kept. The total tissue area was determined by averaging all colour channels, thresholding at a threshold of 220, followed by binary close and open operations.

Data Records

Data are provided via EBRAINS[21] as one ndpi-file per sample, sorted by diagnostic tumour type (in alphabetical order) for easier access. It is possible to download single files directly or all files of a specific tumour type or the whole dataset using a download manager (such as the Chrono Download Manager for the Google Chrome browser). Furthermore, supplementary clinical information, estimated cell densities and scanned tissue area is provided in a csv-spreadsheet with one row per tumour sample. An overview of all spreadsheet variables and descriptions is given in Table 1.

Table 1

Recorded clinical variables and corresponding descriptions.

Variable	Description
uuid	unique sample identifier
pat_id	unique patient identifier
diagnosis	primary diagnosis according to the WHO Classification of Tumours of the Central Nervous System (2016)
grade	WHO grade according to the WHO Classification of Tumours of the Central Nervous System (2016)
subtype	further specification of the histopathological subtype which is not a distinct entity as defined by the WHO, if applicable
secondary_diagnosis	secondary diagnosis in cases where two distinct diagnosis according to the WHO are applicable
control	1 if sample is a control sample without tumour tissue
age	patient age at the time of surgery
sex	biological patient sex
location	list (in square brackets) of all applicable tumour locations; empty if location is unknown
laterality	laterality of the tumour (left or right)
cellularity	estimated cell density of the tissue (given in 1/mm2)
tissue_area	estimated scanned tissue area (in mm2)
recurrence	0 if the entry corresponds to a primary tumour resection; if the entry corresponds to a tumour recurrence, the number of the recurrence is given (e.g., 2 corresponds to the second recurrence)
comment	notable findings that do not fit in other columns (e.g., important mutations not yet integrated in the WHO classification; other non-tumour pathologies in the control samples)

A total of 3,115 histological slides of 2,880 patients have been scanned. A total of 126 distinct diagnostic tumour types could be included. There are 1,395 female and 1,462 male patients in the dataset. The mean patient age at brain tumour surgery was 45 years, ranging from 9 days to 92 years. 2,530 of the scanned slides originated from primary operations and 538 from re-operations. See online-only Table 1 for descriptive properties broken down by tumour type. Descriptive visualizations of patient age, sex, tumour location, cellularity, and scanned tissue area are given in Fig. 2. Of note, we also scanned exceptionally rare tumour types such as melanotic schwannomas or liponeurocytomas (Fig. 3). A total of 47 non-tumour slides from different non-tumour CNS regions and with different pathologies were included as controls.

Online-only Table 1

Overview of the frequencies and descriptive statistics of all brain tumour types included in the DBTA.

	# of scans	Median age	Age range	Cellularity	Tissue Area	F:M ratio	Most frequent location
Adamantinomatous craniopharyngioma	85	30 years	2–79 years	5219 ± 254/mm2	81 ± 7 mm2	1.36	sellar region
Anaplastic astrocytoma, IDH-mutant	47	43 years	24–65 years	3765 ± 282/mm2	162 ± 17 mm2	0.74	frontal
Anaplastic astrocytoma, IDH-wildtype	47	54 years	12–79 years	4633 ± 361/mm2	85 ± 14 mm2	0.57	temporal
Anaplastic ependymoma	50	7 years	0.3–77 years	5571 ± 298/mm2	138 ± 16 mm2	0.92	fourth ventricle
Anaplastic ganglioglioma	5	36 years	14–70 years	3129 ± 801/mm2	96 ± 44 mm2	NA	temporal
Anaplastic meningioma	46	63 years	30–84 years	6291 ± 380/mm2	223 ± 17 mm2	1.19	frontal
Anaplastic oligodendroglioma, IDH-mutant and 1p/19q codeleted	91	49 years	30–78 years	4899 ± 217/mm2	162 ± 11 mm2	0.72	frontal
Anaplastic pleomorphic xanthoastrocytoma	1	48 years	NA	2933/mm2	245 mm2	NA	cerebral, NOS
Angiocentric glioma	3	8 years	3–35 years	3519/mm2	138 mm2	2	temporal
Angiomatous meningioma	32	60 years	43–79 years	4185 ± 330/mm2	201 ± 14 mm2	1.13	frontal
Angiosarcoma	2	66 years	53–79 years	2440/mm2	112 mm2	NA	cerebral, NOS
Astroblastoma	7	40 years	12–83 years	5417 ± 440/mm2	120 ± 24 mm2	1.33	frontal
Atypical choroid plexus papilloma	4	26 years	0.3–56 years	6555/mm2	98 mm2	1	lateral ventricle
Atypical meningioma	83	58 years	13–92 years	6901 ± 222/mm2	225 ± 11 mm2	1.13	frontal
Atypical teratoid/rhabdoid tumour	17	3 years	0.3–10 years	6846 ± 791/mm2	61 ± 21 mm2	1.29	temporal
CNS ganglioneuroblastoma	1	0 years	NA	7247/mm2	306 mm2	NA	frontal
Cellular schwannoma	25	54 years	27–79 years	7459 ± 526/mm2	128 ± 20 mm2	1.08	spinal
Central neurocytoma	20	28 years	6–41 years	8053 ± 650/mm2	110 ± 22 mm2	0.67	lateral ventricle
Cerebellar liponeurocytoma	4	50 years	43–57 years	6924/mm2	125 mm2	NA	cranial nerves
Chondrosarcoma	21	37 years	6–73 years	5642 ± 621/mm2	132 ± 28 mm2	1.33	cranial, NOS
Chordoid glioma of the third ventricle	4	34 years	26–42 years	5074/mm2	36 mm2	0.33	third ventricle
Chordoid meningioma	12	47 years	35–73 years	4408 ± 776/mm2	210 ± 17 mm2	3	cranial, NOS
Chordoma	28	61 years	4–85 years	2808 ± 342/mm2	114 ± 16 mm2	0.56	cranial, NOS
Choriocarcinoma	1	76 years	NA	8954/mm2	131 mm2	NA	sellar region
Choroid plexus carcinoma	7	3 years	0.5–46 years	6778 ± 691/mm2	100 ± 35 mm2	NA	cerebral, NOS
Choroid plexus papilloma	21	29 years	0.2–78 years	5454 ± 467/mm2	156 ± 26 mm2	0.91	fourth ventricle
Clear cell meningioma	13	39 years	8–74 years	5027 ± 648/mm2	182 ± 32 mm2	1.17	sellar region
Crystal-storing histiocytosis	1	62 years	NA	1435/mm2	211 mm2	NA	NA
Desmoplastic infantile astrocytoma and ganglioglioma	11	1 years	0.5–23 years	5206 ± 642/mm2	180 ± 37 mm2	1.5	parietal
Diffuse astrocytoma, IDH-mutant	70	37 years	18–60 years	3013 ± 171/mm2	105 ± 12 mm2	0.64	frontal
Diffuse astrocytoma, IDH-wildtype	19	58 years	20–77 years	2730 ± 315/mm2	90 ± 23 mm2	0.36	frontal
Diffuse large B-cell lymphoma of the CNS	59	68 years	9–84 years	6021 ± 450/mm2	90 ± 13 mm2	1.46	frontal
Diffuse leptomeningeal glioneuronal tumour	1	2 years	NA	8070/mm2	8 mm2	NA	frontal
Diffuse midline glioma, H3 K27M-mutant	21	19 years	3–64 years	4258 ± 460/mm2	26 ± 7 mm2	1.1	brain stem
Dysembryoplastic neuroepithelial tumour	25	31 years	8–57 years	2410 ± 196/mm2	76 ± 16 mm2	0.79	temporal
Dysplastic cerebellar gangliocytoma	1	38 years	NA	2345/mm2	196 mm2	NA	cerebellar
EBV-positive diffuse large B-cell lymphoma, NOS	1	34 years	NA	6595/mm2	5 mm2	NA	frontal
Embryonal carcinoma	1	39 years	NA	4888/mm2	291 mm2	NA	parietal
Embryonal tumour with multilayered rosettes, C19MC-altered	3	2 years	2–3 years	6087/mm2	231 mm2	2	parietal
Ependymoma	46	49 years	2–78 years	4813 ± 347/mm2	94 ± 12 mm2	0.96	spinal
Ependymoma, RELA fusion-positive	6	12 years	4–55 years	5814 ± 1401/mm2	138 ± 40 mm2	0.5	lateral ventricle
Epitheloid MPNST	1	50 years	NA	3003/mm2	70 mm2	NA	other
Erdheim-Chester disease	1	57 years	NA	2194/mm2	239 mm2	NA	NA
Ewing sarcoma	4	6 years	0.8–28 years	8370/mm2	91 mm2	1	spinal
Extraventricular neurocytoma	1	36 years	NA	1193/mm2	107 mm2	NA	spinal
Fibrosarcoma	2	27 years	20–34 years	4639/mm2	199 mm2	1	cerebral, NOS
Fibrous meningioma	57	58 years	12–84 years	6103 ± 237/mm2	228 ± 13 mm2	6.12	cranial, NOS
Follicular lymphoma	3	62 years	62–64 years	8741/mm2	276 mm2	2	occipital
Gangliocytoma	1	36 years	NA	1127/mm2	10 mm2	NA	occipital
Ganglioglioma	88	21 years	2–65 years	2932 ± 153/mm2	110 ± 9 mm2	0.6	temporal
Ganglioneuroma	2	33 years	27–39 years	4228/mm2	212 mm2	1	other
Gemistocytic astrocytoma, IDH-mutant	6	38 years	29–56 years	2036 ± 228/mm2	121 ± 25 mm2	0.5	temporal
Germinoma	20	16 years	9–33 years	7091 ± 686/mm2	21 ± 6 mm2	0.11	diencephalon
Giant cell glioblastoma	21	43 years	11–86 years	3170 ± 301/mm2	181 ± 20 mm2	0.62	temporal
Glioblastoma, IDH-mutant	34	38 years	25–73 years	4867 ± 296/mm2	172 ± 19 mm2	1	frontal
Glioblastoma, IDH-wildtype	474	62 years	17–87 years	4481 ± 96/mm2	151 ± 5 mm2	0.66	temporal
Gliosarcoma	59	57 years	9–86 years	4794 ± 276/mm2	221 ± 14 mm2	0.44	temporal
Granular cell tumour of the sellar region	1	46 years	NA	2172/mm2	97 mm2	NA	sellar region
Haemangioblastoma	88	50 years	16–81 years	5119 ± 185/mm2	109 ± 10 mm2	1	cerebellar
Haemangioma	30	51 years	0.2–76 years	2796 ± 292/mm2	133 ± 15 mm2	2	cranial, NOS
Haemangiopericytoma	34	39 years	25–83 years	9064 ± 489/mm2	186 ± 18 mm2	0.48	cranial, NOS
Hybrid nerve sheath tumours	3	58 years	32–72 years	3342/mm2	227 mm2	0.5	spinal
Immature teratoma	7	15 years	0.0–56 years	7927 ± 913/mm2	107 ± 28 mm2	0.4	third ventricle
Immunodeficiency-associated CNS lymphoma	5	53 years	31–73 years	5209 ± 1227/mm2	65 ± 34 mm2	0.67	cerebral, NOS
Inflammatory myofibroblastic tumour	1	26 years	NA	5226/mm2	298 mm2	NA	frontal
Intravascular large B-cell lymphoma	2	70 years	62–78 years	1242/mm2	427 mm2	NA	NA
Juvenile xanthogranuloma	1	23 years	NA	12519/mm2	75 mm2	NA	NA
Langerhans cell histiocytosis	32	13 years	1–53 years	6848 ± 565/mm2	104 ± 19 mm2	0.78	parietal
Leiomyoma	1	50 years	NA	1864/mm2	28 mm2	NA	cranial, NOS
Leiomyosarcoma	4	58 years	50–77 years	6955/mm2	213 mm2	1	occipital
Lipoma	38	10 years	0.3–76 years	757 ± 66/mm2	120 ± 16 mm2	0.9	spinal
Liposarcoma	1	52 years	NA	2697/mm2	107 mm2	NA	spinal
Low-grade B-cell lymphomas of the CNS	13	67 years	50–83 years	8087 ± 1088/mm2	86 ± 25 mm2	1.17	spinal
Lymphoplasmacyte-rich meningioma	2	46 years	37–55 years	9817/mm2	37 mm2	1	cranial, NOS
MALT lymphoma of the dura	5	68 years	39–79 years	9843 ± 1672/mm2	39 ± 23 mm2	1.5	cranial, NOS
Malignant peripheral nerve sheath tumour	15	61 years	17–81 years	5886 ± 794/mm2	136 ± 28 mm2	0.88	spinal
Mature teratoma	6	10 years	0.2–49 years	3259 ± 760/mm2	135 ± 45 mm2	0.2	spinal
Medulloblastoma, SHH-activated and TP53-mutant	3	16 years	0.5–52 years	9539/mm2	117 mm2	2	cerebellar
Medulloblastoma, SHH-activated and TP53-wildtype	9	30 years	1–75 years	10544 ± 581/mm2	185 ± 38 mm2	0.5	cerebellar
Medulloblastoma, WNT-activated	7	13 years	6–65 years	7641 ± 954/mm2	79 ± 20 mm2	0.4	fourth ventricle
Medulloblastoma, non-WNT/non-SHH	32	8 years	3–34 years	8799 ± 412/mm2	113 ± 13 mm2	0.33	fourth ventricle
Melanotic schwannoma	3	64 years	51–69 years	3110/mm2	55 mm2	0.5	cranial nerves
Meningeal melanocytoma	5	51 years	35–54 years	8152 ± 1898/mm2	172 ± 60 mm2	0.67	spinal
Meningeal melanoma	2	61 years	51–71 years	6763/mm2	160 mm2	NA	spinal
Meningothelial meningioma	104	55 years	25–88 years	5951 ± 212/mm2	162 ± 10 mm2	4.47	cranial, NOS
Metaplastic meningioma	4	75 years	56–85 years	4613/mm2	226 mm2	NA	frontal
Metastatic tumours	47	58 years	38–78 years	5092 ± 399/mm2	159 ± 14 mm2	0.88	spinal
Microcystic meningioma	23	48 years	33–75 years	4475 ± 382/mm2	213 ± 23 mm2	2.83	frontal
Mixed germ cell tumour	5	20 years	12–44 years	4379 ± 1084/mm2	142 ± 45 mm2	0.25	spinal
Myxopapillary ependymoma	23	35 years	11–71 years	3188 ± 360/mm2	154 ± 16 mm2	0.64	spinal
Neurofibroma	16	44 years	0.7–65 years	3640 ± 446/mm2	151 ± 29 mm2	0.6	spinal
Olfactory neuroblastoma	10	58 years	27–69 years	5053 ± 780/mm2	213 ± 34 mm2	0.25	cranial, NOS
Oligodendroglioma, IDH-mutant and 1p/19q codeleted	85	46 years	12–73 years	3587 ± 174/mm2	136 ± 12 mm2	0.7	frontal
Osteochondroma	1	14 years	NA	2388/mm2	39 mm2	NA	spinal
Osteoma	9	48 years	40–69 years	1570 ± 638/mm2	107 ± 33 mm2	1.25	frontal
Osteosarcoma	8	30 years	17–54 years	4328 ± 498/mm2	176 ± 20 mm2	3	cerebral, NOS
Papillary craniopharyngioma	13	61 years	44–82 years	4941 ± 427/mm2	65 ± 13 mm2	0.86	sellar region
Papillary ependymoma	2	35 years	35–35 years	5573/mm2	146 mm2	NA	spinal
Papillary glioneuronal tumour	2	12 years	12–13 years	5877/mm2	117 mm2	NA	cerebral, NOS
Papillary meningioma	3	38 years	20–61 years	7270/mm2	198 mm2	NA	temporal
Papillary tumour of the pineal region	11	17 years	4–48 years	6094 ± 807/mm2	82 ± 28 mm2	1.75	third ventricle
Paraganglioma	17	54 years	25–69 years	6734 ± 468/mm2	165 ± 24 mm2	0.89	spinal
Perineurioma	1	23 years	NA	3580/mm2	26 mm2	NA	other
Pilocytic astrocytoma	173	11 years	0.6–51 years	3327 ± 117/mm2	105 ± 8 mm2	0.66	cerebellar
Pilomyxoid astrocytoma	24	7 years	0.4–56 years	4073 ± 362/mm2	45 ± 11 mm2	1	cranial nerves
Pineal parenchymal tumour of intermediate differentiation	6	44 years	9–55 years	9287 ± 619/mm2	86 ± 44 mm2	0.2	diencephalon
Pineoblastoma	5	23 years	1–67 years	7436 ± 1352/mm2	87 ± 43 mm2	4	third ventricle
Pineocytoma	6	19 years	1–40 years	4086 ± 1245/mm2	60 ± 38 mm2	0.5	diencephalon
Pituicytoma	3	47 years	27–67 years	5231/mm2	33 mm2	2	sellar region
Pituitary adenoma	99	54 years	16–80 years	7842 ± 225/mm2	56 ± 6 mm2	0.98	sellar region
Pleomorphic xanthoastrocytoma	21	29 years	5–72 years	4215 ± 368/mm2	145 ± 23 mm2	1.38	temporal
Plexiform neurofibroma	1	12 years	NA	12310/mm2	66 mm2	NA	NA
Psammomatous meningioma	28	66 years	29–83 years	5201 ± 372/mm2	125 ± 15 mm2	8.33	spinal
Rhabdoid meningioma	5	63 years	52–89 years	5016 ± 475/mm2	235 ± 62 mm2	0.67	occipital
Rhabdomyosarcoma	3	51 years	49–62 years	2474/mm2	299 mm2	2	cranial, NOS
Rosette-forming glioneuronal tumour	11	25 years	13–47 years	3997 ± 822/mm2	56 ± 21 mm2	1.75	cerebellar
Schwannoma	81	53 years	14–78 years	5715 ± 229/mm2	124 ± 10 mm2	0.93	cranial nerves
Secretory meningioma	41	58 years	40–80 years	6112 ± 313/mm2	154 ± 16 mm2	12.67	cranial, NOS
Spindle cell oncocytoma	1	47 years	NA	4958/mm2	13 mm2	NA	NA
Subependymal giant cell astrocytoma	14	15 years	10–33 years	3336 ± 447/mm2	136 ± 31 mm2	0.56	lateral ventricle
Subependymoma	24	54 years	8–81 years	1829 ± 151/mm2	125 ± 20 mm2	1.18	lateral ventricle
T-cell and NK/T-cell lymphomas of the CNS	1	54 years	NA	7400/mm2	18 mm2	NA	cerebral, NOS
Tanycytic ependymoma	1	41 years	NA	5440/mm2	232 mm2	NA	spinal
Teratoma with malignant transformation	1	40 years	NA	7202/mm2	122 mm2	NA	frontal
Transitional meningioma	68	58 years	29–82 years	7221 ± 213/mm2	203 ± 13 mm2	2.09	frontal
Undifferentiated pleomorphic sarcoma	1	62 years	NA	5773/mm2	401 mm2	NA	cranial, NOS

Cellularity and Tissue area are given as [mean ± SEM].

Fig. 2

Descriptive statistics of the ‘Digital Brain Tumour Atlas’ patient cohort (not including control patients). (a) The age distribution by sex shows a bimodal distribution with most patients belonging to the higher-age categories. Since some uncommon tumour types like medulloblastoma occur mainly in children and have been strategically over-sampled, there is also a peak in younger patients. (b) The distribution of the different WHO grades shows a slight predominance of grade I and grade IV tumours. Of note, some tumour entities are not assigned WHO grades (‘NA’) and very few tumour types are assigned intermediate grades II-III (a total of five cases, not shown in the figure). (c) Tumour distribution with colour-coded locations and ratio-specific circle sizes. (Brain illustration adapted from Patrick J. Lynch, wikimedia) (d) Distribution of the cell densities of all included tumour samples by tumour grade. Note that lower-grade tumours are not necessarily less cell dense (e.g., in the case of cellular schwannoma). (e) The distribution of the scanned tissue areas (per slide).

Fig. 3

Exemplary images from exceedingly rare brain tumours, which are included in the DBTA. (a) Perineurioma component of a hybrid nerve sheath tumour. (b) Angiosarcoma. (c) Lymphoplasmacyte-rich meningioma. (d) Crystal-storing histiocytosis. (e) Embryonal tumour with multilayered rosettes. (f) Melanotic schwannoma. (g) Angiocentric glioma. (h) Cerebellar liponeurocytoma. (i) Pituicytoma.

Technical Validation

All cases were initially selected based on the given diagnosis in the diagnostic electronic records. To ensure conformity with the WHO 2016 diagnosis, all slides have been independently reviewed by two neuropathologists experienced in neuro-oncology. In disputed cases, a third senior neuropathologist was consulted. Older cases with missing necessary molecular analyses were not included in the dataset.

Inter- and intraobserver variability is one factor that contributes to misdiagnoses or discrepant diagnoses. We mitigated the risk by including only cases that had already undergone thorough routine diagnostic work-up and were additionally reviewed independently by at least two neuropathologists as described above. In this way, we also ensured excellent image quality and the presence of sufficient diagnostic tumour tissue on each WSI. Scans with suboptimal image quality were either re-scanned (if possible) or excluded.

Usage Notes

Data access

The data can be accessed via EBRAINS[21]. In order to download the data set, users have to register with EBRAINS and agree to the general terms of use, access policy as well as the data use agreement for pseudonymised human data (https://ebrains.eu/terms). The data are distributed under the conditions that users cite the respective DOI, adhere to EBRAINS’ Data Use Agreement and do not use the data for commercial purposes.

WSI processing

The ndp.view2 (© Hamamatsu) software can be freely used to view and annotate slide scans saved in the ndpi format[22]. Alternatively, most other WSI programs such as the open-source OMERO software platform[23] and the open-source QuPath software[24] can work directly on ndpi-files. However, most programming languages and non-specialized image processing software cannot handle ndpi-files out of the box. Thus, we also provide a toolbox of MATLAB scripts that depend on the openslide library[25] and can be used to

Automatically tile large slide scans and export multiple smaller image patches in a given magnification.

Convert annotation-files (.ndpa) to overlays, which can be used to extract specific regions of interest.

Estimate the total tissue area on a WSI.

Estimate the cell density on a WSI.

Of note, slide thickness and staining intensity vary to some degree, resulting in a slightly different histological appearance of each slide. Thus, for machine learning applications, we recommend astain normalization step such as WSICS[26], more recent methods employing generative adversarial networks[27] or style transfer learning[28]. Moreover, heavy stain colour augmentation should be performed[29]. Of note, the stain normalization step can be omitted with only a negligible drop in performance as has been shown by Tellez et al.[29].

Measurement(s)	Cancer Histology • Cellularity Measurement • Total Sample Tissue Area • brain neoplasm
Technology Type(s)	Hematoxylin and Eosin Staining Method • digital curation • Histology and Immunohistochemistry Shared Resource
Factor Type(s)	age • sex • location
Sample Characteristic - Organism	Homo sapiens