Iulian Emil Tampu

Refined detection methods, more detailed tumor characterization, and adequate distinction between different pediatric tumor subtypes are necessary to improve diagnosis and treatment, enable precision medicine, and advance patient prognosis. However, the application of computational approaches to pediatric brain tumors remains limited, largely due to the lack of accessible datasets. To address part of this gap, we provide whole slide images (WSIs) of hematoxylin and eosin (H&E)-stained tissue sections from all pediatric central nervous system (CNS) samples collected in Sweden between 2013 and 2023. These data represent a population-based national cohort encompassing all six pediatric oncology centers in Sweden and are available through the Swedish Childhood Tumor Biobank (BTB). The dataset includes 1,446 WSIs of sufficient image quality with confirmed CNS tumor diagnoses, derived from 537 unique subjects (562 cases). In addition, diagnosticrelevant clinical information is included. Corresponding whole-genome sequencing (WGS), wholetranscriptome sequencing (WTS), and methylation array data are available for most tumor samples through separate resources. This H&E dataset has been specifically curated to support artificial intelligence-based analyses, while also serving broader applications in medical research and education. When combined with matched molecular data, it provides a valuable resource for advancing multimodal and precision diagnostic approaches in the pediatric population.

A bstract: The classification of pediatric brain tumors is investigated using deep learning on hematoxylin and eosin (H&E) and antigen Ki-67 (Ki-67) whole slide images (WSIs) from the Children’s Brain Tumor Network (CBTN) dataset. A total of 1,662 unregistered WSIs (1,047 H&E and 615 Ki-67 images) were analyzed, including low-grade glioma/astrocytoma (grades 1, 2) (LGG), high-grade glioma/astrocytoma (grades 3, 4) (HGG), medulloblastoma (MB), ependymoma (EP) and ganglioglioma. The The aim of this study was to effectively classify pediatric brain tumors using H&E and Ki-67 WSIs individually, and to investigate whether early, intermediate, and late fusion could improve the predictive performance. From each WSI, 224× 224 pixel patches were extracted, and the instance (patch)-level features were obtained using the histology foundation model CONCHv1_5. The instances were aggregated using clustering-constrained attention multiple instance learning (CLAM) for patient-level classification. Model interpretability and explainability was assessed through attention heatmaps, cell density and Ki-67 labelling index (LI) maps. In the binary grade classification between LGG and HGG, the intermediate concatenation fusion achieved the best performance with a balanced accuracy of 0.88 ± 0.05, ( p < 0.005) compared to the single-stain models (H&E: 0.84 ± 0.05, Ki-67: 0.86 ± 0.05). For the 5-class tumor type classification, the one-hidden layer late fusion learning model achieved the highest balanced accuracy of 0.83 ± 0.04 ( p < 0.005), outperforming the single-stain models (H&E: 0.77 ± 0.05, Ki-67: 0.74 ± 0.05). Overall, most of the fusion approaches outperformed the single-stain models in both classification tasks ( p < 0.005). The Ki-67 attention maps demonstrated moderate to strong Spearman correlation ( ρ = 0.576 − 0.823) with the cell density and Ki-67 LI maps, suggesting that these features are associated with the model’s predictions, although additional features may contribute. The results show that H&E and Ki-67 images provide complementary information, and most of the multi-stain fusion approaches using deep learning improve pediatric brain tumor diagnosis.

Brain tumors are the most common solid tumors in children and young adults, but the scarcity of large histopathology datasets has limited the application of computational pathology in this group. This study implements two weakly supervised multiple‐instance learning (MIL) approaches on patch features obtained from state‐of‐the‐art histology‐specific foundation models to classify pediatric brain tumors in hematoxylin and eosin whole slide images (WSIs) from a multi‐center Swedish cohort. WSIs from 540 subjects (age 8.5 ± 4.9 years) diagnosed with brain tumors were gathered from the six Swedish university hospitals. Instance (patch)‐level features were obtained from WSIs using three pre‐trained feature extractors: ResNet50, UNI, and CONCH. Instances were aggregated using attention‐based MIL (ABMIL) or clustering‐constrained attention MIL (CLAM) for patient‐level classification. Models were evaluated on three classification tasks based on the hierarchical classification of pediatric brain tumors: tumor category, family, and type. Model generalization was assessed by training on data from two of the centers and testing on data from four other centers. Model interpretability was evaluated through attention mapping. The highest classification performance was achieved using UNI features and ABMIL aggregation, with Matthew's correlation coefficient of 0.76 ± 0.04, 0.63 ± 0.04, and 0.60 ± 0.05 for tumor category, family, and type classification, respectively. When evaluating generalization, models utilizing UNI and CONCH features outperformed those using ResNet50. However, the drop in performance from the in‐site to out‐of‐site testing was similar across feature extractors. These results show the potential of state‐of‐the‐art computational pathology methods in diagnosing pediatric brain tumors at different hierarchical levels with fair generalizability on a multi‐center national dataset.

Purpose: To implement and evaluate deep learning-based methods for the classification of pediatric brain tumors (PBT) in magnetic resonance (MR) data. Methods: A subset of the “Children’s Brain Tumor Network” dataset was retrospectively used (n = 178 subjects, female = 72, male = 102, NA = 4, age range [0.01, 36.49] years) with tumor types being low-grade astrocytoma (n = 84), ependymoma (n = 32), and medulloblastoma (n = 62). T1w post-contrast (n = 94 subjects), T2w (n = 160 subjects), and apparent diffusion coefficient (ADC: n = 66 subjects) MR sequences were used separately. Two deep learning models were trained on transversal slices showing tumor. Joint fusion was implemented to combine image and age data, and 2 pre-training paradigms were utilized. Model explainability was investigated using gradient-weighted class-activation mapping (Grad-CAM), and the learned feature space was visualized using principal component analysis (PCA). Results: The highest tumor-type classification performance was achieved when using a vision transformer model pre-trained on ImageNet and fine-tuned on ADC images with age fusion (Matthews correlation coefficient [MCC]: 0.77 ± 0.14, Accuracy: 0.87 ± 0.08), followed by models trained on T2w (MCC: 0.58 ± 0.11, Accuracy: 0.73 ± 0.08) and T1w post-contrast (MCC: 0.41 ± 0.11, Accuracy: 0.62 ± 0.08) data. Age fusion marginally improved the model’s performance. Both model architectures performed similarly across the experiments, with no differences between the pre-training strategies. Grad-CAMs showed that the models’ attention focused on the brain region. PCA of the feature space showed greater separation of the tumor-type clusters when using contrastive pre-training. Conclusion: Classification of PBT on MR images could be accomplished using deep learning, with the top-performing model being trained on ADC data, which radiologists use for the clinical classification of these tumors.

Medical imaging is one of the cornerstones of clinical diagnosis, providing insights into the anatomy and physiology of organs and tissues for screening, initial diagnosis, treatment planning, and ...

Purpose To implement and evaluate deep learning-based methods for the classification of pediatric brain tumors in MR data. Materials and methods A subset of the “Children’s Brain Tumor Network” dataset was retrospectively used (n=178 subjects, female=72, male=102, NA=4, age-range [0.01, 36.49] years) with tumor types being low-grade astrocytoma (n=84), ependymoma (n=32), and medulloblastoma (n=62). T1w post-contrast (n=94 subjects), T2w (n=160 subjects), and ADC (n=66 subjects) MR sequences were used separately. Two deep-learning models were trained on transversal slices showing tumor. Joint fusion was implemented to combine image and age data, and two pre-training paradigms were utilized. Model explainability was investigated using gradient-weighted class activation mapping (Grad-CAM), and the learned feature space was visualized using principal component analysis (PCA). Results The highest tumor-type classification performance was achieved when using a vision transformer model pre-trained on ImageNet and fine-tuned on ADC images with age fusion (MCC: 0.77 ± 0.14 Accuracy: 0.87 ± 0.08), followed by models trained on T2w (MCC: 0.58 ± 0.11, Accuracy: 0.73 ± 0.08) and T1w post-contrast (MCC: 0.41 ± 0.11, Accuracy: 0.62 ± 0.08) data. Age fusion marginally improved the model’s performance. Both model architectures performed similarly across the experiments, with no differences between the pre-training strategies. Grad-CAMs showed that the models’ attention focused on the brain region. PCA of the feature space showed greater separation of the tumor-type clusters when using contrastive pre-training. Conclusion Classification of pediatric brain tumors on MR-images could be accomplished using deep learning, with the top-performing model being trained on ADC data, which is used by radiologists for the clinical classification of these tumors. Key points The vision transformer model pre-trained on ImageNet and fine-tuned on ADC data with age fusion achieved the highest performance, which was significantly better than models trained on T2w (second-best) and T1w-Gd data. Fusion of age information with the image data marginally improved classification, and model architecture (ResNet50 -vs -ViT) and pre-training strategies (supervised -vs -self-supervised) did not show to significantly impact models’ performance. Model explainability, by means of class activation mapping and principal component analysis of the learned feature space, show that the models use the tumor region information for classification and that the tumor type clusters are better separated when using age information. Summary Deep learning-based classification of pediatric brain tumors can be achieved using single-sequence pre-operative MR data, showing the potential of automated decision support tools that can aid radiologists in the primary diagnosis of these tumors.

The infiltrative nature of malignant gliomas results in active tumor spreading into the peritumoral edema, which is not visible in conventional magnetic resonance imaging (cMRI) even after contrast injection. MR relaxometry (qMRI) measures relaxation rates dependent on tissue properties and can offer additional contrast mechanisms to highlight the non-enhancing infiltrative tumor. To investigate if qMRI data provides additional information compared to cMRI sequences when considering deep learning-based brain tumor detection and segmentation, preoperative conventional (T1w per- and post-contrast, T2w and FLAIR) and quantitative (pre- and post-contrast R₁, R₂ and proton density) MR data was obtained from 23 patients with typical radiological findings suggestive of a high-grade glioma. 2D deep learning models were trained on transversal slices (n = 528) for tumor detection and segmentation using either cMRI or qMRI. Moreover, trends in quantitative R₁ and R₂ rates of regions identified as relevant for tumor detection by model explainability methods were qualitatively analyzed. Tumor detection and segmentation performance for models trained with a combination of qMRI pre- and post-contrast was the highest (detection Matthews correlation coefficient (MCC) = 0.72, segmentation dice similarity coefficient (DSC) = 0.90), however, the difference compared to cMRI was not statistically significant. Overall analysis of the relevant regions identified using model explainability showed no differences between models trained on cMRI or qMRI. When looking at the individual cases, relaxation rates of brain regions outside the annotation and identified as relevant for tumor detection exhibited changes after contrast injection similar to region inside the annotation in the majority of cases. In conclusion, models trained on qMRI data obtained similar detection and segmentation performance to those trained on cMRI data, with the advantage of quantitatively measuring brain tissue properties within a similar scan time. When considering individual patients, the analysis of relaxation rates of regions identified by model explainability suggests the presence of infiltrative tumor outside the cMRI-based tumor annotation.

Background and purpose: Brain tumors are among the leading causes of cancer deaths in children. Initial diagnosis based on MR images can be a challenging task for radiologists, depending on the tumor type and location. Deep learning methods could support the diagnosis by predicting the tumor type. Materials and methods: A subset (181 subjects) of the data from “Children’s Brain Tumor Network” (CBTN) was used, including infratentorial and supratentorial tumors, with the main tumor types being low-grade astrocytomas, ependymomas, and medulloblastomas. T1w-Gd, T2-w, and ADC MR sequences were used separately. Classification was performed on 2D MR images using four different off-the-shelf deep learning models and a custom-designed shallow network all pre-trained on adult MR images. Joint fusion was implemented to combine image and age data, and tumor type prediction was computed volume-wise. Matthew’s correlation coefficient (MCC), accuracy, and F1 scores were used to assess the models’ performance. Model explainability, using gradient-weighted class activation mapping (Grad-CAM), was implemented and the network’s attention on the tumor region was quantified. Results: The shallow custom network resulted in the highest classification performance when trained on T2-w or ADC MR images fused with age information, when considering infratentorial tumors only (MCC: 0.71 for ADC and 0.64 for T2-w), and both infra- and supratentorial tumors (MCC: 0.70 for ADC and 0.57 for T2-w). Conclusion: Classification of pediatric brain tumors on MR images could be accomplished using deep learning, and the fusion of age information improved model performance.

The infiltrative nature of malignant gliomas results in active tumor spreading into the peritumoral edema, which is not visible in conventional magnetic resonance imaging (cMRI) even after contrast injection. MR relaxometry (qMRI) measures relaxation rates dependent on tissue properties, and can offer additional contrast mechanisms to highlight the non-enhancing infiltrative tumor. The aim of this study is to investigate if qMRI data provides additional information compared to cMRI sequences (T1w, T1wGd, T2w, FLAIR), when considering deep learning-based brain tumor (1) detection and (2) segmentation. A total of 23 patients with histologically confirmed malignant glioma were retrospectively included in the study. Quantitative MR imaging was used to obtain R ₁ (1/T1), R ₂ (1/T2) and proton density maps pre- and post-gadolinium contrast injection. Conventional MR imaging was also performed. A 2D CNN detection model and a 2D U-Net were trained on transversal slices (n=528) using either cMRI or a combination of qMRI pre- and post-contrast data for tumor detection and segmentation, respectively. Moreover, trends in quantitative R ₁ and R ₂ rates of regions identified as relevant for tumor detection by model explainability methods were qualitatively analyzed. Tumor detection and segmentation performance for models trained with a combination of qMRI pre- and post-contrast was the highest (detection MCC=0.72, segmentation Dice=0.90), however, improvements were not statistically significant compared to cMRI (detection MCC=0.67, segmentation Dice=0.90). The analysis of the relaxation rates of the relevant regions identified using model explainability methods showed no differences between models trained on cMRI or qMRI. Relevant regions which fell outside the annotation showed changes in relaxation rates after contrast injection similar to those within the annotation, when looking at majority of the individual cases. A similar trend could not be seen when looking at relaxation trends over all the dataset. In conclusion, models trained on qMRI data obtain similar performance to those trained on cMRI data, with the advantage of quantitatively measuring brain tissue properties within the scan time (11.8 minutes for qMRI with and without contrast, and 12.2 minutes for cMRI). Moreover, when considering individual patients, regions identified by model explainability methods as relevant for tumor detection outside the manual annotation of the tumor showed changes in quantitative relaxation rates after contrast injection similar to regions within the annotation, suggestive of infiltrative tumor in the peritumoral edema.

Intraoperative guidance tools for thyroid surgery based on optical coherence tomography (OCT) could aid distinguish between normal and diseased tissue. However, OCT images are difficult to interpret, thus, real‐time automatic analysis could support the clinical decision‐making. In this study, several deep learning models were investigated for thyroid disease classification on 2D and 3D OCT data obtained from ex vivo specimens of 22 patients undergoing surgery and diagnosed with several thyroid pathologies. Additionally, two open‐access datasets were used to evaluate the custom models. On the thyroid dataset, the best performance was achieved by the 3D vision transformer model with a Matthew's correlation coefficient (MCC) of 0.79 (accuracy = 0.90) for the normal‐versus‐abnormal classification. On the open‐access datasets, the custom models achieved the best performance (MCC > 0.88, accuracy > 0.96). Results obtained for the normal‐versus‐abnormal classification suggest OCT, complemented with deep learning‐based analysis, as a tool for real‐time automatic diseased tissue identification in thyroid surgery.