Comparative Characterization of Human Meibomian Glands, Free Sebaceous Glands, and Hair-Associated Sebaceous Glands Based on Biomarkers, Analysis of Secretion Composition, and Gland Morphology.

Meibomian gland dysfunction (MGD) is one of the main causes of dry eye disease. To better understand the physiological functions of human meibomian glands (MGs), the present study compared MGs with free sebaceous glands (SGs) and hair-associated SGs of humans using morphological, immunohistochemical, and liquid chromatography-mass spectrometry (LCMS)-based lipidomic approaches. Eyelids with MGs, nostrils, lips, and external auditory canals with free SGs, and scalp with hair-associated SGs of body donors were probed with antibodies against cytokeratins (CK) 1, 8, 10, and 14, stem cell markers keratin 15 and N-cadherin, cell-cell contact markers desmoglein 1 (Dsg1), desmocollin 3 (Dsc3), desmoplakin (Dp), plakoglobin (Pg), and E-cadherin, and the tight junction protein claudin 5. In addition, Oil Red O staining (ORO) was performed in cryosections. Secretions of MGs as well as of SGs of nostrils, external auditory canals, and scalps were collected from healthy volunteers, analyzed by LCMS, and the data were processed using various multivariate statistical analysis approaches. Serial sections of MGs, free SGs, and hair-associated SGs were 3D reconstructed and compared. CK1 was expressed differently in hair-associated SGs than in MGs and other free SGs. The expression levels of CK8, CK10, and CK14 in MGs were different from those in hair-associated SGs and other free SGs. KRT15 was expressed differently in hair-associated SGs, whereas N-cadherin was expressed equally in all types of glands. The cell-cell contact markers Dsg1, Dp, Dsc3, Pg, and E-cadherin revealed no differences. ORO staining showed that lipids in MGs were more highly dispersed and had larger lipid droplets than lipids in other free SGs. Hair-associated SGs had a smaller number of lipid droplets. LCMS revealed that the lipid composition of meibum was distinctively different from that of the sebum of the nostrils, external auditory canals, and scalp. The 3D reconstructions of the different glands revealed different morphologies of the SGs compared with MGs which are by far the largest type of glands. In humans, MGs differ in their morphology and secretory composition and show major differences from free and hair-associated SGs. The composition of meibum differs significantly from that of sebum from free SGs and from hair-associated SGs. Therefore, the MG can be considered as a highly specialized type of holocrine gland that exhibits all the histological characteristics of SGs, but is significantly different from them in terms of morphology and lipid composition.

Domain generalization across tumor types, laboratories, and species - Insights from the 2022 edition of the Mitosis Domain Generalization Challenge.

Recognition of mitotic figures in histologic tumor specimens is highly relevant to patient outcome assessment. This task is challenging for algorithms and human experts alike, with deterioration of algorithmic performance under shifts in image representations. Considerable covariate shifts occur when assessment is performed on different tumor types, images are acquired using different digitization devices, or specimens are produced in different laboratories. This observation motivated the inception of the 2022 challenge on MItosis Domain Generalization (MIDOG 2022). The challenge provided annotated histologic tumor images from six different domains and evaluated the algorithmic approaches for mitotic figure detection provided by nine challenge participants on ten independent domains. Ground truth for mitotic figure detection was established in two ways: a three-expert majority vote and an independent, immunohistochemistry-assisted set of labels. This work represents an overview of the challenge tasks, the algorithmic strategies employed by the participants, and potential factors contributing to their success. With an F1 score of 0.764 for the top-performing team, we summarize that domain generalization across various tumor domains is possible with today's deep learning-based recognition pipelines. However, we also found that domain characteristics not present in the training set (feline as new species, spindle cell shape as new morphology and a new scanner) led to small but significant decreases in performance. When assessed against the immunohistochemistry-assisted reference standard, all methods resulted in reduced recall scores, with only minor changes in the order of participants in the ranking.

Tracking cell turnover in human brain using 15N-thymidine imaging mass spectrometry.

Microcephaly is often caused by an impairment of the generation of neurons in the brain, a process referred to as neurogenesis. While most neurogenesis in mammals occurs during brain development, it thought to continue to take place through adulthood in selected regions of the mammalian brain, notably the hippocampus. However, the generality of neurogenesis in the adult brain has been controversial. While studies in mice and rats have provided compelling evidence for neurogenesis occurring in the adult rodent hippocampus, the lack of applicability in humans of key methods to demonstrate neurogenesis has led to an intense debate about the existence and, in particular, the magnitude of neurogenesis in the adult human brain. Here, we demonstrate the applicability of a powerful method to address this debate, that is, the in vivo labeling of adult human patients with 15N-thymidine, a non-hazardous form of thymidine, an approach without any clinical harm or ethical concerns. 15N-thymidine incorporation into newly synthesized DNA of specific cells was quantified at the single-cell level with subcellular resolution by Multiple-isotype imaging mass spectrometry (MIMS) of brain tissue resected for medical reasons. Two adult human patients, a glioblastoma patient and a patient with drug-refractory right temporal lobe epilepsy, were infused for 24 h with 15N-thymidine. Detection of 15N-positive leukocyte nuclei in blood samples from these patients confirmed previous findings by others and demonstrated the appropriateness of this approach to search for the generation of new cells in the adult human brain. 15N-positive neural cells were easily identified in the glioblastoma tissue sample, and the range of the 15N signal suggested that cells that underwent S-phase fully or partially during the 24 h in vivo labeling period, as well as cells generated therefrom, were detected. In contrast, within the hippocampus tissue resected from the epilepsy patient, none of the 2,000 dentate gyrus neurons analyzed was positive for 15N-thymidine uptake, consistent with the notion that the rate of neurogenesis in the adult human hippocampus is rather low. Of note, the likelihood of detecting neurogenesis was reduced because of (i) the low number of cells analyzed, (ii) the fact that hippocampal tissue was explored that may have had reduced neurogenesis due to epilepsy, and (iii) the labeling period of 24 h which may have been too short to capture quiescent neural stem cells. Yet, overall, our approach to enrich NeuN-labeled neuronal nuclei by FACS prior to MIMS analysis provides a promising strategy to quantify even low rates of neurogenesis in the adult human hippocampus after in vivo15N-thymidine infusion. From a general point of view and regarding future perspectives, the in vivo labeling of humans with 15N-thymidine followed by MIMS analysis of brain tissue constitutes a novel approach to study mitotically active cells and their progeny in the brain, and thus allows a broad spectrum of studies of brain physiology and pathology, including microcephaly.

The specific DNA methylation landscape in focal cortical dysplasia ILAE type 3D.

Focal Cortical Dysplasia (FCD) is a frequent cause of drug-resistant focal epilepsy in children and young adults. The international FCD classifications of 2011 and 2022 have identified several clinico-pathological subtypes, either occurring isolated, i.e., FCD ILAE Type 1 or 2, or in association with a principal cortical lesion, i.e., FCD Type 3. Here, we addressed the DNA methylation signature of a previously described new subtype of FCD 3D occurring in the occipital lobe of very young children and microscopically defined by neuronal cell loss in cortical layer 4. We studied the DNA methylation profile using 850 K BeadChip arrays in a retrospective cohort of 104 patients with FCD 1 A, 2 A, 2B, 3D, TLE without FCD, and 16 postmortem specimens without neurological disorders as controls, operated in China or Germany. DNA was extracted from formalin-fixed paraffin-embedded tissue blocks with microscopically confirmed lesions, and DNA methylation profiles were bioinformatically analyzed with a recently developed deep learning algorithm. Our results revealed a distinct position of FCD 3D in the DNA methylation map of common FCD subtypes, also different from non-FCD epilepsy surgery controls or non-epileptic postmortem controls. Within the FCD 3D cohort, the DNA methylation signature separated three histopathology subtypes, i.e., glial scarring around porencephalic cysts, loss of layer 4, and Rasmussen encephalitis. Differential methylation in FCD 3D with loss of layer 4 mapped explicitly to biological pathways related to neurodegeneration, biogenesis of the extracellular matrix (ECM) components, axon guidance, and regulation of the actin cytoskeleton. Our data suggest that DNA methylation signatures in cortical malformations are not only of diagnostic value but also phenotypically relevant, providing the molecular underpinnings of structural and histopathological features associated with epilepsy. Further studies will be necessary to confirm these results and clarify their functional relevance and epileptogenic potential in these difficult-to-treat children.

A comprehensive multi-domain dataset for mitotic figure detection.

The prognostic value of mitotic figures in tumor tissue is well-established for many tumor types and automating this task is of high research interest. However, especially deep learning-based methods face performance deterioration in the presence of domain shifts, which may arise from different tumor types, slide preparation and digitization devices. We introduce the MIDOG++ dataset, an extension of the MIDOG 2021 and 2022 challenge datasets. We provide region of interest images from 503 histological specimens of seven different tumor types with variable morphology with in total labels for 11,937 mitotic figures: breast carcinoma, lung carcinoma, lymphosarcoma, neuroendocrine tumor, cutaneous mast cell tumor, cutaneous melanoma, and (sub)cutaneous soft tissue sarcoma. The specimens were processed in several laboratories utilizing diverse scanners. We evaluated the extent of the domain shift by using state-of-the-art approaches, observing notable differences in single-domain training. In a leave-one-domain-out setting, generalizability improved considerably. This mitotic figure dataset is the first that incorporates a wide domain shift based on different tumor types, laboratories, whole slide image scanners, and species.

Ganglioglioma with adverse clinical outcome and atypical histopathological features were defined by alterations in PTPN11/KRAS/NF1 and other RAS-/MAP-Kinase pathway genes.

Exome-wide sequencing studies recently described PTPN11 as a novel brain somatic epilepsy gene. In contrast, germline mutations of PTPN11 are known to cause Noonan syndrome, a multisystem disorder characterized by abnormal facial features, developmental delay, and sporadically, also brain tumors. Herein, we performed a deep phenotype-genotype analysis of a comprehensive series of ganglioglioma (GG) with brain somatic alterations of the PTPN11/KRAS/NF1 genes compared to GG with common MAP-Kinase signaling pathway alterations, i.e., BRAFV600E. Seventy-two GG were submitted to whole exome sequencing and genotyping and 84 low grade epilepsy associated tumors (LEAT) to DNA-methylation analysis. In 28 tumours, both analyses were available from the same sample. Clinical data were retrieved from hospital files including disease onset, age at surgery, brain localization, and seizure outcome. A comprehensive histopathology staining panel was available in all cases. We identified eight GG with PTPN11 alterations, copy number variant (CNV) gains of chromosome 12, and the commonality of additional CNV gains in NF1, KRAS, FGFR4 and RHEB, as well as BRAFV600E alterations. Histopathology revealed an atypical glio-neuronal phenotype with subarachnoidal tumor spread and large, pleomorphic, and multinuclear cellular features. Only three out of eight patients with GG and PTPN11/KRAS/NF1 alterations were free of disabling-seizures 2 years after surgery (38% had Engel I). This was remarkably different from our series of GG with only BRAFV600E mutations (85% had Engel I). Unsupervised cluster analysis of DNA methylation arrays separated these tumours from well-established LEAT categories. Our data point to a subgroup of GG with cellular atypia in glial and neuronal cell components, adverse postsurgical outcome, and genetically characterized by complex alterations in PTPN11 and other RAS-/MAP-Kinase and/or mTOR signaling pathways. These findings need prospective validation in clinical practice as they argue for an adaptation of the WHO grading system in developmental, glio-neuronal tumors associated with early onset focal epilepsy.

Identification of a novel role for matrix metalloproteinase-3 in the modulation of B cell responses in multiple sclerosis.

There has been a growing interest in the presence and role of B cell aggregates within the central nervous system of multiple sclerosis patients. However, very little is known about the expression profile of molecules associated with these aggregates and how they might be influencing aggregate development or persistence in the brain. The current study focuses on the effect of matrix metalloproteinase-3, which is associated with B cell aggregates in autopsied multiple sclerosis brain tissue, on B cells. Autopsied brain sections from multiple sclerosis cases and controls were screened for the presence of CD20+ B cell aggregates and expression of matrix metalloproteinase-3. Using flow cytometry, enzyme-linked immunosorbent assay and gene array as methods, in vitro studies were conducted using peripheral blood of healthy volunteers to demonstrate the effect of matrix metalloproteinase-3 on B cells. Autopsied brain sections from multiple sclerosis patients containing aggregates of B cells expressed a significantly higher amount of matrix metalloproteinase-3 compared to controls. In vitro experiments demonstrated that matrix metalloproteinase-3 dampened the overall activation status of B cells by downregulating CD69, CD80 and CD86. Furthermore, matrix metalloproteinase-3-treated B cells produced significantly lower amounts of interleukin-6. Gene array data confirmed that matrix metalloproteinase-3 altered the proliferation and survival profiles of B cells. Taken together, out data indicate a role for B cell modulatory properties of matrix metalloproteinase-3.

AI Prediction of Neuropathic Pain after Lumbar Disc Herniation-Machine Learning Reveals Influencing Factors.

The treatment options for neuropathic pain caused by lumbar disc herniation have been debated controversially in the literature. Whether surgical or conservative therapy makes more sense in individual cases can hardly be answered. We have investigated whether a machine learning-based prediction of outcome, regarding neuropathic pain development, after lumbar disc herniation treatment is possible. The extensive datasets of 123 consecutive patients were used to predict the development of neuropathic pain, measured by a visual analogue scale (VAS) for leg pain and the Oswestry Disability Index (ODI), at 6 weeks, 6 months and 1 year after treatment of lumbar disc herniation in a machine learning approach. Using a decision tree regressor algorithm, a prediction quality within the limits of the minimum clinically important difference for the VAS and ODI value could be achieved. An analysis of the influencing factors of the algorithm reveals the important role of psychological factors as well as body weight and age with pre-existing conditions for an accurate prediction of neuropathic pain. The machine learning algorithm developed here can enable an assessment of the course of treatment after lumbar disc herniation. The early, comparative individual prediction of a therapy outcome is important to avoid unnecessary surgical therapies as well as insufficient conservative therapies and prevent the chronification of neuropathic pain.

An Evolution Gaining Momentum-The Growing Role of Artificial Intelligence in the Diagnosis and Treatment of Spinal Diseases.

In recent years, applications using artificial intelligence have been gaining importance in the diagnosis and treatment of spinal diseases. In our review, we describe the basic features of artificial intelligence which are currently applied in the field of spine diagnosis and treatment, and we provide an orientation of the recent technical developments and their applications. Furthermore, we point out the possible limitations and challenges in dealing with such technological advances. Despite the momentary limitations in practical application, artificial intelligence is gaining ground in the field of spine treatment. As an applying physician, it is therefore necessary to engage with it in order to benefit from those advances in the interest of the patient and to prevent these applications being misused by non-medical partners.

Multilobar unilateral hypoplasia with emphasis on the posterior quadrant and severe epilepsy in children with FCD ILAE Type 1A.

OBJECTIVE: Focal cortical dysplasia (FCD) Type 1 and its three subtypes have yet not been fully characterized at the clinical, anatomopathological, and molecular level (International League Against Epilepsy [ILAE] FCD classification from 2011). We aimed to describe the clinical phenotype of patients with histopathologically confirmed FCD1A obtained from a single epilepsy center between 2002 and 2016. METHODS: Medical records were retrieved from the hospital's archive. Results from electroencephalography (EEG) video recordings, neuroimaging, and histopathology were reevaluated. Magnetic resonance imaging (MRI) post-processing was retrospectively performed in nine patients. DNA methylation studies were carried out from archival surgical brain tissue in 11 patients. RESULTS: Nineteen children with a histopathological diagnosis of FCD1A were included. The average onset of epilepsy was 0.9 years (range 0.2-10 years). All children had severe cognitive impairment and one third had mild motor deficits, yet fine finger movements were preserved in all patients. All patients had daily seizures, being drug resistant from disease onset. Interictal electroencephalography revealed bilateral multi-regional epileptiform discharges. Interictal status epilepticus was observed in 8 and countless subclinical seizures in 11 patients. Regional continuous irregular slow waves were of higher lateralizing and localizing yield than spikes. Posterior background rhythms were normal in 16 of 19 children. Neuroimaging showed unilateral multilobar hypoplasia and increased T2-FLAIR signals of the white matter in 18 of 19 patients. All children underwent tailored multilobar resections, with seizure freedom achieved in 47% (Engel class I). There was no case with frontal involvement without involvement of the posterior quadrant by MRI and histopathology. DNA methylation profiling distinguished FCD1A samples from all other epilepsy specimens and controls. SIGNIFICANCE: We identified a cohort of young children with drug resistance from seizure onset, bad EEG with posterior emphasis, lack of any focal neurological deficits but severe cognitive impairment, subtle hypoplasia of the epileptogenic area on MRI, and histopathologically defined and molecularly confirmed by DNA methylation analysis as FCD ILAE Type 1A.

DNA methylation-based classification of malformations of cortical development in the human brain.

Malformations of cortical development (MCD) comprise a broad spectrum of structural brain lesions frequently associated with epilepsy. Disease definition and diagnosis remain challenging and are often prone to arbitrary judgment. Molecular classification of histopathological entities may help rationalize the diagnostic process. We present a retrospective, multi-center analysis of genome-wide DNA methylation from human brain specimens obtained from epilepsy surgery using EPIC 850 K BeadChip arrays. A total of 308 samples were included in the study. In the reference cohort, 239 formalin-fixed and paraffin-embedded (FFPE) tissue samples were histopathologically classified as MCD, including 12 major subtype pathologies. They were compared to 15 FFPE samples from surgical non-MCD cortices and 11 FFPE samples from post-mortem non-epilepsy controls. We applied three different statistical approaches to decipher the DNA methylation pattern of histopathological MCD entities, i.e., pairwise comparison, machine learning, and deep learning algorithms. Our deep learning model, which represented a shallow neuronal network, achieved the highest level of accuracy. A test cohort of 43 independent surgical samples from different epilepsy centers was used to test the precision of our DNA methylation-based MCD classifier. All samples from the test cohort were accurately assigned to their disease classes by the algorithm. These data demonstrate DNA methylation-based MCD classification suitability across major histopathological entities amenable to epilepsy surgery and age groups and will help establish an integrated diagnostic classification scheme for epilepsy-associated MCD.

Combined Artificial Intelligence Approaches Analyzing 1000 Conservative Patients with Back Pain-A Methodological Pathway to Predicting Treatment Efficacy and Diagnostic Groups.

Patients with back pain are common and present a challenge in everyday medical practice due to the multitude of possible causes and the individual effects of treatments. Predicting causes and therapy efficien cy with the help of artificial intelligence could improve and simplify the treatment. In an exemplary collective of 1000 conservatively treated back pain patients, it was investigated whether the prediction of therapy efficiency and the underlying diagnosis is possible by combining different artificial intelligence approaches. For this purpose, supervised and unsupervised artificial intelligence methods were analyzed and a methodology for combining the predictions was developed. Supervised AI is suitable for predicting therapy efficiency at the borderline of minimal clinical difference. Non-supervised AI can show patterns in the dataset. We can show that the identification of the underlying diagnostic groups only becomes possible through a combination of different AI approaches and the baseline data. The presented methodology for the combined application of artificial intelligence algorithms shows a transferable path to establish correlations in heterogeneous data sets when individual AI approaches only provide weak results.

Artificial intelligence facilitates decision-making in the treatment of lumbar disc herniations.

PURPOSE: Apart from patients with severe neurological deficits, it is not clear whether surgical or conservative treatment of lumbar disc herniations is superior for the individual patient. We investigated whether deep learning techniques can predict the outcome of patients with lumbar disc herniation after 6 months of treatment. METHODS: The data of 60 patients were used to train and test a deep learning algorithm with the aim to achieve an accurate prediction of the ODI 6 months after surgery or the start of conservative therapy. We developed an algorithm that predicts the ODI of 6 randomly selected test patients in tenfold cross-validation. RESULTS: A 100% accurate prediction of an ODI range could be achieved by dividing the ODI scale into 12% sections. A maximum absolute difference of only 3.4% between individually predicted and actual ODI after 6 months of a given therapy was achieved with our most powerful model. The application of artificial intelligence as shown in this work also allowed to compare the actual patient values after 6 months with the prediction for the alternative therapy, showing deviations up to 18.8%. CONCLUSION: Deep learning in the supervised form applied here can identify patients at an early stage who would benefit from conservative therapy, and on the contrary avoid painful and unnecessary delays for patients who would profit from surgical therapy. In addition, this approach can be used in many other areas of medicine as an effective tool for decision-making when choosing between opposing treatment options, despite small patient groups.

Modulation of Pancreatic Neuroendocrine Neoplastic Cell Fate by Autophagy-Mediated Death.

INTRODUCTION: Autophagic cell death in cancer cells can be mediated by inhibition of deacetylases. Although extensive studies have focused on the autophagic process in cancer, little is known about the role of autophagy in degrading cytosolic and nuclear components of pancreatic neuroendocrine neoplastic (pNEN) cells leading to cell death, thus improving the therapy of patients affected by pNEN. METHODS: 2D and 3D human pNEN and pancreatic stellate cells were treated with panobinostat and bafilomycin. Autophagy markers were detected by RT-qPCR, immunofluorescence, and Western blot. Autophagosomes were detected by electron microscopy and their maturation by real-time fluorescence of LC3B stable transfected cells. ChIP was performed at the cAMP responsive element. Immunofluorescence was performed in murine pancreatic tissue. RESULTS: We observed that pan-deacetylase inhibitor panobinostat treatment causes autophagic cell death in pNEN cells. We also found that although AMPK-α phosphorylation is counterbalanced by phosphorylated AKT, it is not capable to inhibiting autophagic cell death. However, the binding activity of the cAMP responsive element is prompted by panobinostat. Although autophagy inhibition prevented autophagosome synthesis, maturation, and cell death, panobinostat treatment induced the accumulation of mature autophagosomes in the cytosol and the nucleus, leading to disruption of the organelles, cellular digestion, and decay. Observation of autophagosome membrane proteins Beclin1 and LC3B aggregation in murine pancreatic islets indicates that autophagy restoration may also lead to autophagosome aggregation in murine insulinoma cells. A basal low expression of autophagy markers was detectable in patients affected by pNEN, and, interestingly, the expression of these markers was significantly lower in metastatic pNEN. DISCUSSION/CONCLUSION: Our study highlights that the autophagy functional restoration and prolongation of this catabolic process, mediated by inhibition of deacetylase, is responsible for the reduction of pNEN cells. Prompting of autophagy cell death could be a promising strategy for the therapy of pNEN.

Disseminated Multifocal Intracerebral Bleeding Events in Three Coronavirus Disease 2019 Patients on Extracorporeal Membrane Oxygenation As Rescue Therapy.

OBJECTIVES: To describe three coronavirus disease 2019 patients suffering from acute respiratory distress syndrome under venovenous extracorporeal membrane oxygenation therapy and tight anticoagulation monitoring presenting a novel pattern of multifocal brain hemorrhage in various degrees in all cerebral and cerebellar lobes. DESIGN: Clinical observation of three patients. Post mortem examinations. SETTING: Two ICUs at the University Hospital Erlangen. PATIENTS: Three patients (medium age 56.6 yr, two male with hypertension and diabetes, one female with no medical history) developed severe acute respiratory distress syndrome on the basis of a severe acute respiratory syndrome coronavirus 2 infection. All required mechanical ventilation and venovenous extracorporeal membrane oxygenation support. INTERVENTIONS: Clinical observation, CT, data extraction from electronic medical records, and post mortem examinations. MAIN RESULTS: We report on an unusual multifocal bleeding pattern in the white matter in three cases with severe acute respiratory distress syndrome due to coronavirus disease 2019 undergoing venovenous extracorporeal membrane oxygenation therapy. Bleeding pattern with consecutive herniation was found in CT scans as well as in neuropathologic post mortem examinations. Frequency for this unusual brain hemorrhage in coronavirus disease 2019 patients with extracorporeal membrane oxygenation therapy at our hospital is currently 50%, whereas bleeding events in extracorporeal membrane oxygenation patients generally occur at 10-15%. CONCLUSIONS: Multifocality and high frequency of the unusual white matter hemorrhage pattern suggest a coherence to coronavirus disease 2019. Neuropathological analyses showed circumscribed thrombotic cerebrovascular occlusions, which eventually led to microvascular and later on macrovascular disseminated bleeding events. However, signs of cerebrovascular inflammation could not be detected. Polymerase chain reaction analyses of brain tissue or cerebrospinal fluid remained negative. Increased susceptibility for fatal bleeding events should be taken into consideration in terms of systemic anticoagulation strategies in coronavirus disease 2019.

Mosaic trisomy of chromosome 1q in human brain tissue associates with unilateral polymicrogyria, very early-onset focal epilepsy, and severe developmental delay.

Polymicrogyria (PMG) is a developmental cortical malformation characterized by an excess of small and frustrane gyration and abnormal cortical lamination. PMG frequently associates with seizures. The molecular pathomechanisms underlying PMG development are not yet understood. About 40 genes have been associated with PMG, and small copy number variations have also been described in selected patients. We recently provided evidence that epilepsy-associated structural brain lesions can be classified based on genomic DNA methylation patterns. Here, we analyzed 26 PMG patients employing array-based DNA methylation profiling on formalin-fixed paraffin-embedded material. A series of 62 well-characterized non-PMG cortical malformations (focal cortical dysplasia type 2a/b and hemimegalencephaly), temporal lobe epilepsy, and non-epilepsy autopsy controls was used as reference cohort. Unsupervised dimensionality reduction and hierarchical cluster analysis of DNA methylation profiles showed that PMG formed a distinct DNA methylation class. Copy number profiling from DNA methylation data identified a uniform duplication spanning the entire long arm of chromosome 1 in 7 out of 26 PMG patients, which was verified by additional fluorescence in situ hybridization analysis. In respective cases, about 50% of nuclei in the center of the PMG lesion were 1q triploid. No chromosomal imbalance was seen in adjacent, architecturally normal-appearing tissue indicating mosaicism. Clinically, PMG 1q patients presented with a unilateral frontal or hemispheric PMG without hemimegalencephaly, a severe form of intractable epilepsy with seizure onset in the first months of life, and severe developmental delay. Our results show that PMG can be classified among other structural brain lesions according to their DNA methylation profile. One subset of PMG with distinct clinical features exhibits a duplication of chromosomal arm 1q.

Deep Learning-Based Quantification of Pulmonary Hemosiderophages in Cytology Slides.

Exercise-induced pulmonary hemorrhage (EIPH) is a common condition in sport horses with negative impact on performance. Cytology of bronchoalveolar lavage fluid by use of a scoring system is considered the most sensitive diagnostic method. Macrophages are classified depending on the degree of cytoplasmic hemosiderin content. The current gold standard is manual grading, which is however monotonous and time-consuming. We evaluated state-of-the-art deep learning-based methods for single cell macrophage classification and compared them against the performance of nine cytology experts and evaluated inter- and intra-observer variability. Additionally, we evaluated object detection methods on a novel data set of 17 completely annotated cytology whole slide images (WSI) containing 78,047 hemosiderophages. Our deep learning-based approach reached a concordance of 0.85, partially exceeding human expert concordance (0.68 to 0.86, mean of 0.73, SD of 0.04). Intra-observer variability was high (0.68 to 0.88) and inter-observer concordance was moderate (Fleiss' kappa = 0.67). Our object detection approach has a mean average precision of 0.66 over the five classes from the whole slide gigapixel image and a computation time of below two minutes. To mitigate the high inter- and intra-rater variability, we propose our automated object detection pipeline, enabling accurate, reproducible and quick EIPH scoring in WSI.