Synergistic effects of organic carbon and silica in preserving structural stability of drying soils.

Predicted climate warming and prolonged droughts pose a threat to the soil structure as organic carbon losses weaken the stability of soil aggregates. Well-structured soils are important for storage and movement of water, solutes, and air, the development of plant roots, as habitat for soil organisms, and the microbial activity. Structural stability is measured in terms of hydro-mechanical properties. This study compares effects of amorphous silica with those of organic carbon on stability parameters during drying of aggregates from relatively finer- and coarser-textured soils. Silica amendment enhanced the positive effect of organic carbon on structural stability in terms of the tensile strength. Synergistic effects between silica and organic carbon in soil colloids appear to dynamically alter aggregate density and friability (i.e., ability to crumble) during drying. Silica together with organic carbon could help soil management to reduce negative effects of predicted prolonged droughts on soil structure and stability.

Hydrochlorothiazide and increased risk of atypical fibroxanthoma and pleomorphic dermal sarcoma.

BACKGROUND AND OBJECTIVES: Previous work has demonstrated that hydrochlorothiazide (HCTZ) is a risk factor for squamous cell carcinomas (SCC) and basal cell carcinomas (BCC) due to pro-photocarcinogenic effects. Atypical fibroxanthoma (AFX) and pleomorphic sarcoma (PDS), both ultraviolet-induced cancers, display a rare but rising cutaneous tumor entity. This study aimed to evaluate if the use of HCTZ is higher in patients with AFX/PDS than in patients with SCC/BCC and subsequently may be a risk factor for AFX/PDS-development. PATIENTS AND METHODS: In a retrospective study of four German skin cancer centers, AFX/PDS cases and SCC/BCC controls were sex and age matched (1:3) over a time-period of 7 years (2013-2019) to evaluate the use of HCTZ, immunosuppressive medication, second malignancies, and presence of diabetes mellitus. RESULTS: Overall, 146 AFX/PDS and 438 controls (SCC/BCC) were included in the study. The use of HCTZ was significantly higher in patients with AFX/PDS (44.5%) compared to patients with SCC/BCC (25.3%). Additionally, the presence of diabetes mellitus was significantly higher in AFX/PDS patients. CONCLUSIONS: This study demonstrates a significantly higher use of HCTZ in patients with AFX/PDS compared to SCC/BCC. This result suggests that HCTZ may be a risk factor for AFX/PDS. Additionally, diabetes mellitus or its comorbidities may be associated with an increased risk for AFX/PDS.

Effects of leaf silicon on drought performance of tropical tree seedlings.

Elevated leaf silicon (Si) concentrations improve drought resistance in cultivated plants, suggesting Si might also improve drought performance of wild species. Tropical tree species, for instance, take up substantial amounts of Si, and leaf Si varies markedly at local and regional scales, suggesting consequences for seedling drought resistance. Yet, whether elevated leaf Si improves seedling drought performance in tropical forests is unknown. To manipulate leaf Si concentrations, seedlings of seven tropical tree species were grown in Si-rich and -poor soil, before exposing them to drought in the forest understorey. Survival, growth and wilting were monitored. Elevated leaf Si did not improve drought survival and growth in any of the species. In one species, drought survival was reduced in seedlings previously grown in Si-rich soil, contrary to our expectation. Our results suggest that elevated leaf Si does not improve drought resistance of wild tropical tree species. Elevated leaf Si may even reduce drought performance, suggesting differences in soil conditions influencing leaf Si may contribute to soil-related variation of tropical seedling performance. Furthermore, our results are at odds with most studies on cultivated species and show that alleviative effects of Si in crops cannot be generalized to wild plants in natural systems.

Silicon-based anti-herbivore defense in tropical tree seedlings.

Silicon-based defenses deter insect herbivores in many cultivated and wild grass species. Furthermore, in some of these species, silicon (Si) uptake and defense can be induced by herbivory. Tropical trees also take up Si and leaf Si concentrations vary greatly across and within species. As herbivory is a major driver of seedling mortality and niche differentiation of tropical tree species, understanding anti-herbivore defenses is pivotal. Yet, whether silicon is a constitutive and inducible herbivory defense in tropical forest tree species remains unknown. We grew seedlings of eight tropical tree species in a full factorial experiment, including two levels of plant-available soil Si concentrations (-Si/+Si) and a simulated herbivory treatment (-H/+H). The simulated herbivory treatment was a combination of clipping and application of methyl jasmonate. We then carried out multiple-choice feeding trials, separately for each tree species, in which leaves of each treatment combination were offered to a generalist caterpillar (Spodoptera frugiperda). Leaf damage was assessed. Three species showed a significant decrease in leaf damage under high compared to low Si conditions (by up to 72%), consistent with our expectation of Si-based defenses acting in tropical tree species. In one species, leaf damage was increased by increasing soil Si and in four species, no effect of soil Si on leaf damage was observed. Opposite to our expectation of Si uptake and defense being inducible by herbivory damage, simulated herbivory increased leaf damage in two species. Furthermore, simulated herbivory reduced Si concentrations in one species. Our results showed that tropical tree seedlings can be better defended when growing in Si-rich compared to Si-poor soils, and that the effects of Si on plant defense vary strongly across species. Furthermore, Si-based defenses may not be inducible in tropical tree species. Overall, constitutive Si-based defense should be considered part of the vast array of anti-herbivore defenses of tropical tree species. Our finding that Si-based defenses are highly species-specific combined with the fact that herbivory is a major driver of mortality in tropical tree seedling, suggests that variation in soil Si concentrations may have pervasive consequences for regeneration and performance across tropical tree species.

Applying an artificial intelligence deep learning approach to routine dermatopathological diagnosis of basal cell carcinoma.

BACKGROUND: Institutes of dermatopathology are faced with considerable challenges including a continuously rising numbers of submitted specimens and a shortage of specialized health care practitioners. Basal cell carcinoma (BCC) is one of the most common tumors in the fair-skinned western population and represents a major part of samples submitted for histological evaluation. Digitalizing glass slides has enabled the application of artificial intelligence (AI)-based procedures. To date, these methods have found only limited application in routine diagnostics. The aim of this study was to establish an AI-based model for automated BCC detection. PATIENTS AND METHODS: In three dermatopathological centers, daily routine practice BCC cases were digitalized. The diagnosis was made both conventionally by analog microscope and digitally through an AI-supported algorithm based on a U-Net architecture neural network. RESULTS: In routine practice, the model achieved a sensitivity of 98.23% (center 1) and a specificity of 98.51%. The model generalized successfully without additional training to samples from the other centers, achieving similarly high accuracies in BCC detection (sensitivities of 97.67% and 98.57% and specificities of 96.77% and 98.73% in centers 2 and 3, respectively). In addition, automated AI-based basal cell carcinoma subtyping and tumor thickness measurement were established. CONCLUSIONS: AI-based methods can detect BCC with high accuracy in a routine clinical setting and significantly support dermatopathological work.

Exploring the potential: Can arsenic (As) resistant silicate-solubilizing bacteria manage the dual effects of silicon on As accumulation in rice?

Rice (Oryza sativa L.) cultivation in regions marked by elevated arsenic (As) concentrations poses significant health concerns due to As uptake by the plant and its subsequent entry into the human food chain. With rice serving as a staple crop for a substantial share of the global population, addressing this issue is critical for food security. In flooded paddy soils, where As availability is pronounced, innovative strategies to reduce As uptake and enhance agricultural sustainability are mandatory. Silicon (Si) and Si nanoparticles have emerged as potential candidates to mitigate As accumulation in rice. However, their effects on As uptake exhibit complexity, influenced by initial Si levels in the soil and the amount of Si introduced through fertilization. While low Si additions may inadvertently increase As uptake, higher Si concentrations may alleviate As uptake and toxicity. The interplay among existing Si and As availability, Si supplementation, and soil biogeochemistry collectively shapes the outcome. Adding water-soluble Si fertilizers (e.g., Na2SiO3 and K2SiO3) has demonstrated efficacy in mitigating As toxicity stress in rice. Nonetheless, the expense associated with these fertilizers underscores the necessity for low cost innovative solutions. Silicate-solubilizing bacteria (SSB) resilient to As hold promise by enhancing Si availability by accelerating mineral dissolution within the rhizosphere, thereby regulating the Si biogeochemical cycle in paddy soils. Promoting SSB could make cost-effective Si sources more soluble and, consequently, managing the intricate interplay of Si's dual effects on As accumulation in rice. This review paper offers a comprehensive exploration of Si's nuanced role in modulating As uptake by rice, emphasizing the potential synergy between As-resistant SSB and Si availability enhancement. By shedding light on this interplay, we aspire to shed light on an innovative attempt for reducing As accumulation in rice while advancing agricultural sustainability.

Silicon in Plants: Alleviation of Metal(loid) Toxicity and Consequential Perspectives for Phytoremediation.

For the majority of higher plants, silicon (Si) is considered a beneficial element because of the various favorable effects of Si accumulation in plants that have been revealed, including the alleviation of metal(loid) toxicity. The accumulation of non-degradable metal(loid)s in the environment strongly increased in the last decades by intensified industrial and agricultural production with negative consequences for the environment and human health. Phytoremediation, i.e., the use of plants to extract and remove elemental pollutants from contaminated soils, has been commonly used for the restoration of metal(loid)-contaminated sites. In our viewpoint article, we briefly summarize the current knowledge of Si-mediated alleviation of metal(loid) toxicity in plants and the potential role of Si in the phytoremediation of soils contaminated with metal(loid)s. In this context, a special focus is on metal(loid) accumulation in (soil) phytoliths, i.e., relatively stable silica structures formed in plants. The accumulation of metal(loid)s in phytoliths might offer a promising pathway for the long-term sequestration of metal(loid)s in soils. As specific phytoliths might also represent an important carbon sink in soils, phytoliths might be a silver bullet in the mitigation of global change. Thus, the time is now to combine Si/phytolith and phytoremediation research. This will help us to merge the positive effects of Si accumulation in plants with the advantages of phytoremediation, which represents an economically feasible and environmentally friendly way to restore metal(loid)-contaminated sites.

Increased wheat yield and soil C stocks after silica fertilization at the field scale.

Increased crop production is a main goal to feed the predicted human population in future. The current management practice is, however, not sustainable as it depends on high amounts of fertilizer application and is highly vulnerable to decreased soil water availability. At the same time it becomes more and more crucial to reduce or even mitigate anthropogenic greenhouse gas (GHG) emissions. A possible way to enable this, might be the increase of the soil C sequestration and thus the C sink function of arable lands. A recent and potentially more sustainable idea is the single time fertilization with amorphous silicon (ASi) which is known to increase both nutrient and water availability. Here we show for the first time on the basis of a field plot experiment how a fertilization with ASi is affecting both, crop yield and the C sequestration of the soils in an agricultural system cultivating wheat. We found a strong increase in wheat yield and biomass production after ASi fertilization by increasing soil moisture during the whole growing season. Additionally, despite a relatively short growing season, Si fertilization increased the net C uptake by soils, i.e., C sequestration with both Si fertilized treatments showing a negative net ecosystem C balance (soil C gain) during the measurement period, while the control showed a small positive net ecosystem C balance (soil C loss). To our best knowledge, this is the first time such effect has ever been observed. In summary, our study demonstrates a new management strategy for crop production increasing yield and biomass production as well as soil C uptake on a more sustainable basis, by a single time fertilization with ASi.

Deep learning detection of melanoma metastases in lymph nodes.

BACKGROUND: In melanoma patients, surgical excision of the first draining lymph node, the sentinel lymph node (SLN), is a routine procedure to evaluate lymphogenic metastases. Metastasis detection by histopathological analysis assesses multiple tissue levels with hematoxylin and eosin and immunohistochemically stained glass slides. Considering the amount of tissue to analyze, the detection of metastasis can be highly time-consuming for pathologists. The application of artificial intelligence in the clinical routine has constantly increased over the past few years. METHODS: In this multi-center study, a deep learning method was established on histological tissue sections of sentinel lymph nodes collected from the clinical routine. The algorithm was trained to highlight potential melanoma metastases for further review by pathologists, without relying on supplementary immunohistochemical stainings (e.g. anti-S100, anti-MelanA). RESULTS: The established method was able to detect the existence of metastasis on individual tissue cuts with an area under the curve of 0.9630 and 0.9856 respectively on two test cohorts from different laboratories. The method was able to accurately identify tumour deposits>0.1 mm and, by automatic tumour diameter measurement, classify these into 0.1 mm to -1.0 mm and>1.0 mm groups, thus identifying and classifying metastasis currently relevant for assessing prognosis and stratifying treatment. CONCLUSIONS: Our results demonstrate that AI-based SLN melanoma metastasis detection has great potential and could become a routinely applied aid for pathologists. Our current study focused on assessing established parameters; however, larger future AI-based studies could identify novel biomarkers potentially further improving SLN-based prognostic and therapeutic predictions for affected patients.

Arctic soil CO2 release during freeze-thaw cycles modulated by silicon and calcium.

Arctic soils are the largest pool of soil organic carbon worldwide. Temperatures in the Arctic have risen faster than the global average during the last decades, decreasing annual freezing days and increasing the number of freeze-thaw cycles (temperature oscillations passing through zero degrees) per year as the temperature is expected to fluctuate more around 0 °C. At the same time, proceeding deepening of seasonal thaw may increase silicon (Si) and calcium (Ca) concentrations in the active layer of Arctic soils as the concentrations in the thawing permafrost layer might be higher depending on location. We analyzed the importance of freeze-thaw cycles for Arctic soil CO2 fluxes. Furthermore, we tested how Si (mobilizing organic C) and Ca (immobilizing organic C) interfere with the soil CO2 fluxes in the context of freeze-thaw cycles. Our results show that with each freeze-thaw cycle the CO2 fluxes from the Arctic soils decreased. Our data revealed a considerable CO2 emission below 0 °C. We also show that pronounced differences emerge in Arctic soil CO2 fluxes with Si increasing and Ca decreasing CO2 fluxes. Furthermore, we show that both Si and Ca concentrations in Arctic soils are central controls on Arctic soil CO2 release, with Si increasing Arctic soil CO2 release especially when temperatures are just below 0 °C. Our findings could provide an important constraint on soil CO2 emissions upon soil thaw, as well as on the greenhouse gas budget of high latitudes. Thus we call for work improving understanding of freeze-thaw cycles as well as the effect of Ca and Si on carbon fluxes, as well as for increased consideration of those factors in wide-scale assessments of carbon fluxes in the high latitudes.

Why do plants silicify?

Despite seminal papers that stress the significance of silicon (Si) in plant biology and ecology, most studies focus on manipulations of Si supply and mitigation of stresses. The ecological significance of Si varies with different levels of biological organization, and remains hard to capture. We show that the costs of Si accumulation are greater than is currently acknowledged, and discuss potential links between Si and fitness components (growth, survival, reproduction), environment, and ecosystem functioning. We suggest that Si is more important in trait-based ecology than is currently recognized. Si potentially plays a significant role in many aspects of plant ecology, but knowledge gaps prevent us from understanding its possible contribution to the success of some clades and the expansion of specific biomes.

Zunahme des atypischen Fibroxanthoms und pleomorphen dermalen Sarkoms: eine retrospektive Analyse vier deutscher Hauttumorzentren.

HINTERGRUND UND ZIELE: In den letzten Jahren konnten umfassende Erkenntnisse über die Pathogenese, Diagnostik und Behandlung von kutanen Sarkomen, insbesondere des atypischen Fibroxanthoms (AFX) und pleomorphen dermalen Sarkoms (PDS) gesammelt werden. Beide Entitäten zeigten innerhalb der letzten Dekade steigende Inzidenzraten. Die vorliegende Studie diente der Untersuchung, welchen Einfluss die neuen Erkenntnisse auf die Fallzahlen von AFX/PDS im Vergleich zu anderen Sarkom-Entitäten haben. PATIENTEN UND METHODIK: Diese retrospektive Studie wurde an vier deutschen Hauttumorzentren durchgeführt und alle von zertifizierten Dermatopathologen bestätigten histopathologischen Befunde von kutanen Sarkomen (AFX, PDS, Dermatofibrosarcoma protuberans, kutanes Leiomyosarkom, Angiosarkom und Kaposi-Sarkom) in einem Zeitraum von sieben Jahren (2013-2019) evaluiert. Zusätzlich wurde der Einsatz von immunhistochemischen Markern als diagnostische Hilfe (Panzytokeratin, S100, Desmin, CD34, CD10, Prokollagen-1, CD99, CD14 und CD68) erfasst. ERGEBNISSE: Insgesamt konnten 255 kutane Sarkome in die vorliegende Studie eingeschlossen werden. Die Zahl der kutanen Sarkome nahm kontinuierlich von 2013 bis 2019 zu (von 16 auf 52 Fälle im Jahr). Die Diagnose eines AFX/PDS konnte in 2019 4,6-mal häufiger als in 2013 gestellt werden. Das AFX stellte mit 49,3 % aller kutanen Sarkome den häufigsten Sarkom-Subtypen dar. Zusätzlich war der Anstieg von AFX/PDS mit dem Einsatz von Immunhistochemie assoziiert. Der Einsatz von spezifischen Immunhistochemischen Markern stieg von 57,1 % im Jahr 2013 auf 100 % in 2019. SCHLUSSFOLGERUNGEN: Diese retrospektive Studie von vier deutschen Hauttumorzentren demonstriert eine substanzielle Zunahme von AFX/PDS, wahrscheinlich infolge kürzlich etablierter beziehungsweise verbesserter diagnostischer und terminologischer Standards. Dieser Anstieg ist vermutlich mit dem vermehrten Einsatz von bestimmten immunhistochemischen Markern assoziiert. AFX/PDS treten wahrscheinlich häufiger auf als bisher vermutet und repräsentieren möglicherweise den häufigsten kutanen Sarkom-Subtyp.

Increase of atypical fibroxanthoma and pleomorphic dermal sarcoma: a retrospective analysis of four German skin cancer centers.

BACKGROUND AND OBJECTIVES: In recent years, considerable insight has been gained into the pathogenesis, diagnosis and treatment of cutaneous sarcomas, including atypical fibroxanthoma (AFX) and pleomorphic dermal sarcoma (PDS). Both entities have shown increasing incidence rates in the last decade. This study was initiated to evaluate how these new insights impact the number of diagnoses of AFX/PDS compared to other cutaneous sarcoma entities. PATIENTS AND METHODS: In a retrospective study of four German skin cancer centers, all histopathological reports of cutaneous sarcomas (AFX, PDS, dermatofibrosarcoma protuberans, cutaneous leiomyosarcoma, angiosarcoma, and Kaposi sarcoma) confirmed by board-certified dermatopathologists were analyzed during a time-period of seven years (2013-2019). Additionally, utilization of immunohistochemical markers (including pan-cytokeratin, S100, desmin, CD34, CD10, procollagen-1, CD99, CD14, and CD68) as an adjunct to diagnose AFX/PDS was recorded. RESULTS: Overall, 255 cutaneous sarcomas were included in the present study. The diagnosis of a cutaneous sarcoma has consequently risen from 2013 to 2019 (from 16 to 52 annual cases). The results of AFX/PDS revealed 4.6 times more diagnoses in 2019 than in 2013. Atypical fibroxanthoma represented the most common subtype, displaying 49.3 % of all diagnosed cutaneous sarcomas. Additionally, the increase of AFX/PDS was linked to the use of immunohistochemistry, with specific immunohistochemical markers used in 57.1 % of cases in 2013 compared to 100 % in 2019. CONCLUSIONS: This retrospective study of four German skin cancer centers demonstrates a substantial rise of AFX/PDS, possibly due to recently established diagnostic and terminology standards. This rise is probably linked to increased utilization of specific immunohistochemical markers. Atypical fibroxanthoma/PDS may be more common than previously thought and seems to represent the most frequent cutaneous sarcoma subtype.

Silicon as a potential limiting factor for phosphorus availability in paddy soils.

Rice cultivation requires high amounts of phosphorus (P). However, significant amounts of P fertilizer additions may be retained by iron (Fe) oxides and are thus unavailable for plants. At the same time, rice cultivation has a high demand for silicic acid (Si), reducing Si availability after short duration of rice cultivation. By studying a paddy chronosequence with rice cultivation up to 2000 years, we show that Si limitation, observed as early as a few decades of rice cultivation, is limiting P availability along the paddy soils chronosequence. Using near edge X-ray absorption fine structure spectroscopy (NEXAFS) in a scanning transmission (soft) X-ray microscope (STXM) we show release of available P was linked to a Si-induced change in speciation of Fe-phases in soil particles and competition of Si with P for binding sites. Hence, low Si availability is limiting P availability in paddy soils. We propose that proper management of Si availability is a promising tool to improve the P supply of paddy plants.

Deep Learning Assisted Diagnosis of Onychomycosis on Whole-Slide Images.

BACKGROUND: Onychomycosis numbers among the most common fungal infections in humans affecting finger- or toenails. Histology remains a frequently applied screening technique to diagnose onychomycosis. Screening slides for fungal elements can be time-consuming for pathologists, and sensitivity in cases with low amounts of fungi remains a concern. Convolutional neural networks (CNNs) have revolutionized image classification in recent years. The goal of our project was to evaluate if a U-NET-based segmentation approach as a subcategory of CNNs can be applied to detect fungal elements on digitized histologic sections of human nail specimens and to compare it with the performance of 11 board-certified dermatopathologists. METHODS: In total, 664 corresponding H&E- and PAS-stained histologic whole-slide images (WSIs) of human nail plates from four different laboratories were digitized. Histologic structures were manually annotated. A U-NET image segmentation model was trained for binary segmentation on the dataset generated by annotated slides. RESULTS: The U-NET algorithm detected 90.5% of WSIs with fungi, demonstrating a comparable sensitivity with that of the 11 board-certified dermatopathologists (sensitivity of 89.2%). CONCLUSIONS: Our results demonstrate that machine-learning-based algorithms applied to real-world clinical cases can produce comparable sensitivities to human pathologists. Our established U-NET may be used as a supportive diagnostic tool to preselect possible slides with fungal elements. Slides where fungal elements are indicated by our U-NET should be reevaluated by the pathologist to confirm or refute the diagnosis of onychomycosis.

Mismatch of N release from the permafrost and vegetative uptake opens pathways of increasing nitrous oxide emissions in the high Arctic.

Biogeochemical cycling in permafrost-affected ecosystems remains associated with large uncertainties, which could impact the Earth's greenhouse gas budget and future climate policies. In particular, increased nutrient availability following permafrost thaw could perturb the greenhouse gas exchange in these systems, an effect largely unexplored until now. Here, we enhance the terrestrial ecosystem model QUINCY (QUantifying Interactions between terrestrial Nutrient CYcles and the climate system), which simulates fully coupled carbon (C), nitrogen (N) and phosphorus (P) cycles in vegetation and soil, with processes relevant in high latitudes (e.g., soil freezing and snow dynamics). In combination with site-level and satellite-based observations, we use the model to investigate impacts of increased nutrient availability from permafrost thawing in comparison to other climate-induced effects and CO2 fertilization over 1960 to 2018 across the high Arctic. Our simulations show that enhanced availability of nutrients following permafrost thaw account for less than 15% of the total Gross primary productivity increase over the time period, despite simulated N limitation over the high Arctic scale. As an explanation for this weak fertilization effect, observational and model data indicate a mismatch between the timing of peak vegetative growth (week 26-27 of the year, corresponding to the beginning of July) and peak thaw depth (week 32-35, mid-to-late August), resulting in incomplete plant use of nutrients near the permafrost table. The resulting increasing N availability approaching the permafrost table enhances N loss pathways, which leads to rising nitrous oxide (N2 O) emissions in our model. Site-level emission trends of 2 mg N m-2  year-1 on average over the historical time period could therefore predict an emerging increasing source of N2 O emissions following future permafrost thaw in the high Arctic.

Evaluation of a Deep Learning Approach to Differentiate Bowen's Disease and Seborrheic Keratosis.

Background: Some of the most common cutaneous neoplasms are Bowen's disease and seborrheic keratosis, a malignant and a benign proliferation, respectively. These entities represent a significant fraction of a dermatopathologists' workload, and in some cases, histological differentiation may be challenging. The potential of deep learning networks to distinguish these diseases is assessed. Methods: In total, 1935 whole-slide images from three institutions were scanned on two different slide scanners. A U-Net-based segmentation deep learning algorithm was trained on data from one of the centers to differentiate Bowen's disease, seborrheic keratosis, and normal tissue, learning from annotations performed by dermatopathologists. Optimal thresholds for the class distinction of diagnoses were extracted and assessed on a test set with data from all three institutions. Results: We aimed to diagnose Bowen's diseases with the highest sensitivity. A good performance was observed across all three centers, underlining the model's robustness. In one of the centers, the distinction between Bowen's disease and all other diagnoses was achieved with an AUC of 0.9858 and a sensitivity of 0.9511. Seborrheic keratosis was detected with an AUC of 0.9764 and a sensitivity of 0.9394. Nevertheless, distinguishing irritated seborrheic keratosis from Bowen's disease remained challenging. Conclusions: Bowen's disease and seborrheic keratosis could be correctly identified by the evaluated deep learning model on test sets from three different centers, two of which were not involved in training, and AUC scores > 0.97 were obtained. The method proved robust to changes in the staining solution and scanner model. We believe this demonstrates that deep learning algorithms can aid in clinical routine; however, the results should be confirmed by qualified histopathologists.