Skip to main content

Table 1 Distribution of WSI in validation cohort

From: Visualizing histopathologic deep learning classification and anomaly detection using nonlinear feature space dimensionality reduction

Diagnosis Unique Slides Unique Cases >‚ÄČ15 Lesion Tiles) Misclassified by Prediction score Misclassified by t-SNE Misclassified by both
Trained Classes
Glioma 47 15
12 Glioblastoma, WHO IV, IDH-wt
1 Anaplastic Astrocytoma, WHO III
1 Anaplastic Oligo, WHO III
1 Gliosarcoma, WHO IV
43/47 8/43 4/25 2/23
Meningioma 57 18
15 Meningioma, WHO I (Meningothelial, Angiomatous, Fibrous, Transitional)
3 Atypical Meningioma, WHO II
55/57 1/55 1/47 1/47
Schwannoma 23 7
Conventional Type, WHO I
23/23 0/23 0/22 0/22
Metastasis 12 7
3 Lung Adenocarcinoma
1 Lung Squamous Cell Carcinoma
1 Breast Adenocarcinoma
1 Esophageal Adenocarcinoma
1 Squamous Cell Carcinoma, NYD
12/12 3/12 0/9 0/9
Lymphoma 2 1 Diffuse Large B-Cell Lymphoma 2/2 0/2 0/1 0/1
Hematoma 15 3 3/15 2/3 1/1 N/A
Brain Tissue 3 2 1/3 1/1 N/A N/A
Dura 1 1 1/1 1/1 0/1 N/A
Novel Classes
Hemangioblastoma 5 2 5/5 5/5 N/A N/A
Radiation necrosis 3 1 2/3 2/2 N/A N/A
Focal Cortical Dysplasia 5 1 0 N/A N/A N/A
Cartilaginous Material 5 2 0 N/A N/A N/A
Epidermoid Cyst 2 1 0 N/A N/A N/A
Total: 13 Class 180 61 82 23 6 3
  1. WSI were randomly selected from prospective cases from our local surgical neuropathology service. All cases selected had diagnostic consensus amongst 3 board certified neuropathologists and had confirmatory immunohistochemical staining patterns. Up to 5 slides of the same cases were used when available