SSiT: Saliency-guided Self-supervised Image Transformer for Diabetic Retinopathy Grading.

Self-supervised Learning (SSL) has been widely applied to learn image representations through exploiting unlabeled images. However, it has not been fully explored in the medical image analysis field. In this work, Saliency-guided Self-Supervised image Transformer (SSiT) is proposed for Diabetic Retinopathy (DR) grading from fundus images. We novelly introduce saliency maps into SSL, with a goal of guiding self-supervised pre-training with domain-specific prior knowledge. Specifically, two saliency-guided learning tasks are employed in SSiT: (1) Saliency-guided contrastive learning is conducted based on the momentum contrast, wherein fundus images' saliency maps are utilized to remove trivial patches from the input sequences of the momentum-updated key encoder. Thus, the key encoder is constrained to provide target representations focusing on salient regions, guiding the query encoder to capture salient features. (2) The query encoder is trained to predict the saliency segmentation, encouraging the preservation of fine-grained information in the learned representations. To assess our proposed method, four publicly-accessible fundus image datasets are adopted. One dataset is employed for pre-training, while the three others are used to evaluate the pre-trained models' performance on downstream DR grading. The proposed SSiT significantly outperforms other representative state-of-the-art SSL methods on all downstream datasets and under various evaluation settings. For example, SSiT achieves a Kappa score of 81.88% on the DDR dataset under fine-tuning evaluation, outperforming all other ViT-based SSL methods by at least 9.48%.

Regulation of neutrophil extracellular traps in cancer.

Neutrophil extracellular trap (NET) is one of the defense functions of neutrophils, which has a rapid ability to kill infections and is also crucial in a variety of immune-associated diseases including infections, tumors and autoimmune diseases. Recent studies have shown that NETs are closely related to the development of tumors. The regulatory role of NETs in tumors has been of interest to researchers. In addition to awakening latent tumor cells, NETs can also promote the proliferation and development of tumor cells and their metastasis to other sites. At the same time, NETs also have the effect of inhibiting tumors. At present, there are some new advances in the impact of NETs on tumor development, which will provide a more theoretical basis for developing NET-targeted drugs. Therefore, this review just summarized the formation process of NETs, the regulation of tumor development and the treatment methods based on NETs.

Deacetylase sirtuin 2 negatively regulates myeloid-derived suppressor cell functions in allograft rejection.

Although myeloid-derived suppressor cells (MDSCs) are critical for allograft survival, their regulatory mechanism remains unclear. Herein, our results showed that metabolism sensor sirtuin 2 (SIRT2) negatively regulates the functions of MDSCs in inducing allogeneic skin graft rejection. Genetic deletion of SIRT2 in myeloid cells (Sirt2Δmye) increased the number of CD11b+Gr1+ MDSCs in bone marrow, spleens, draining lymph nodes, and allografts, inhibited the production of proinflammatory cytokine tumor necrosis factor ɑ, enhanced the production of anti-inflammatory cytokine interleukin 10, and potentiated the suppressive activation of MDSCs in prolonging allograft skin survival. C-X-C motif chemokine receptor 2 is critical for mediating the recruitment and cytokine production of MDSCs induced by SIRT2. Mechanistically, Sirt2Δmye enhanced NAD+ levels, succinate dehydrogenase subunit A (SDHA) activities, and oxidative phosphorylation (OXPHOS) levels in MDSCs after transplantation. Pharmacologically blocking nicotinamide phosphoribosyltransferase effectively reverses the production of cytokines and suppressive activities of MDSC induced by Sirt2Δmye. Blocking OXPHOS with knockdown of SDHA or pharmacological blocking of SDHA significantly restores Sirt2Δmye-mediated stronger MDSC suppressive activity and inflammatory factor productions. Thus, our findings identify a previously unrecognized interplay between NAD+ and SDH-mediated OXPHOS metabolic pathways in regulating MDSC functions induced by the metabolic sensor SIRT2 in allogeneic transplantation.

Hippo Kinase MST1-Mediated Cell Metabolism Reprograms the Homeostasis and Differentiation of Granulocyte Progenitor Cells.

The mechanism of the development of granulocyte progenitor cells into neutrophils under steady-state and pathological conditions remains unclear. In this study, our results showed that with the development of neutrophils from hematopoietic stem cells to mature neutrophils, the expression level of the Hippo kinase MST1 gradually increased. Mst1-specific deficiency in myeloid cells caused neutrophilia, with an expanded granulocytic compartment resulting from a cell-autonomous increase in the number of granulocyte-macrophage progenitors under steady-state conditions and during Listeria monocytogenes infection. Mechanistically, mTOR and HIF1α signaling are required for regulating the balance between glycolysis and succinate dehydrogenase-mediated oxidative phosphorylation, which is crucial for Mst1-/--induced proliferation of granulocyte-monocyte progenitors, lineage-decision factor C/EBPα expression, and granulopoiesis. HIF1α directly regulated C/EBPα promoter activities. Blocking mTOR and HIF1α or adjusting the balance between glycolysis and succinate dehydrogenase-mediated oxidative phosphorylation reversed the granulopoiesis induced by Mst1-/- under steady-state conditions or infection in mice. Thus, our findings identify a previously unrecognized interplay between Hippo kinase MST1 signaling and mTOR-HIF1α metabolic reprogramming in granulocyte progenitor cells that underlies granulopoiesis.

Rootstock-scion interaction affects Malus transcriptome profiles in response to cadmium.

Apple production is threatened by cadmium contamination in orchards. Cd accumulation and tolerance in grafted Malus plants is affected by rootstock, scion, and their interaction. This dataset is part of an experiment investigating the molecular mechanism of Cd bioaccumulation and tolerance in different apple rootstock-scion combinations. We exposed four rootstock-scion combinations to Cd treatment consisting of Hanfu and Fuji apple (Malus domestica) scions grafted onto apple rootstocks of M. baccata or M. micromalus "qingzhoulinqin". RNA sequencing was conducted in roots and leaves of grafting combinations under 0 or 50 µM CdCl2 conditions. A comprehensive transcriptional dataset of affected rootstock, scion, and their interaction among different graft combinations was obtained. This dataset provides new insights in the transcriptional control of Cd bioaccumulation and tolerance in grafting plants regulated by rootstock and scion. Herein, we discuss the molecular mechanism underlying Cd absorption and bioaccumulation.

Identifying the Key Components in ResNet-50 for Diabetic Retinopathy Grading from Fundus Images: A Systematic Investigation.

Although deep learning-based diabetic retinopathy (DR) classification methods typically benefit from well-designed architectures of convolutional neural networks, the training setting also has a non-negligible impact on prediction performance. The training setting includes various interdependent components, such as an objective function, a data sampling strategy, and a data augmentation approach. To identify the key components in a standard deep learning framework (ResNet-50) for DR grading, we systematically analyze the impact of several major components. Extensive experiments are conducted on a publicly available dataset EyePACS. We demonstrate that (1) the DR grading framework is sensitive to input resolution, objective function, and composition of data augmentation; (2) using mean square error as the loss function can effectively improve the performance with respect to a task-specific evaluation metric, namely the quadratically weighted Kappa; (3) utilizing eye pairs boosts the performance of DR grading and; (4) using data resampling to address the problem of imbalanced data distribution in EyePACS hurts the performance. Based on these observations and an optimal combination of the investigated components, our framework, without any specialized network design, achieves a state-of-the-art result (0.8631 for Kappa) on the EyePACS test set (a total of 42,670 fundus images) with only image-level labels. We also examine the proposed training practices on other fundus datasets and other network architectures to evaluate their generalizability. Our codes and pre-trained model are available online.

New insights into follicular regulatory T cells in the intestinal and tumor microenvironments.

Follicular regulatory T (Tfr) cells are a novel and unique subset of effector regulatory T (Treg) cells that are located in germinal centers (GCs). Tfr cells express transcription profiles that are characteristic of both follicular helper T (Tfh) cells and Treg cells and negatively regulate GC reactions, including Tfh cell activation and cytokine production, class switch recombination and B cell activation. Evidence also shows that Tfr cells have specific characteristics in different local immune microenvironments. This review focuses on the regulation of Tfr cell differentiation and function in unique local immune microenvironments, including the intestine and tumor.

CD8+ T cell exhaustion in anti-tumour immunity: The new insights for cancer immunotherapy.

CD8+ T cells play a crucial role in anti-tumour immunity, but they often undergo exhaustion, which affects the anti-tumour activity of CD8+ T cells. The effect and mechanism of exhausted CD8+ T cells have become the focus of anti-tumour immunity research. Recently, a large number of studies have confirmed that long-term antigen exposure can induce exhaustion. Cytokines previously have identified their effects (such as IL-2 and IL-10) may play a dual role in the exhaustion process of CD8+ T cells, suggesting a new mechanism of inducing exhaustion. This review just focuses our current understanding of the biology of exhausted CD8+ T cells, including differentiation pathways, cellular characteristics and signalling pathways involved in inducing exhaustion, and summarizes how these can be applied to tumour immunotherapy.

Regulation of follicular T helper cell differentiation in antitumor immunity.

Follicular T helper (Tfh) cells are a subset of CD4+ T cells that play an important role in the formation of germinal centers and the maturation and differentiation of affinity-matured B cells. Recent studies have demonstrated important functions of Tfh cells in tertiary lymphoid structures of tumors, revealing great potential of Tfh cells in tumor immunity. However, Tfh development is incompletely understood. The differentiation of Tfh cells is a complex, multistage process regulated at the DNA, RNA and protein levels. This review just summarizes current research on the molecular mechanisms of Tfh cell differentiation to better understand the role of Tfh cells in antitumor immunity.

TMT proteomics analysis of a pseudocereal crop, quinoa (Chenopodium quinoa Willd.), during seed maturation.

Quinoa (Chenopodium quinoa Willd.), an Andean native crop, is increasingly popular around the world due to its high nutritional content and stress tolerance. The production and the popularity of this strategic global food are greatly restricted by many limiting factors, such as seed pre-harvest sprouting, bitter saponin, etc. To solve these problems, the underlying mechanism of seed maturation in quinoa needs to be investigated. In this study, based on the investigation of morphological characteristics, a quantitative analysis of its global proteome was conducted using the combinational proteomics of tandem mass tag (TMT) labeling and parallel reaction monitoring (PRM). The proteome changes related to quinoa seed maturation conversion were monitored to aid its genetic improvement. Typical changes of morphological characteristics were discovered during seed maturation, including mean grain diameter, mean grain thickness, mean hundred-grain weight, palea, episperm color, etc. With TMT proteomics analysis, 581 differentially accumulated proteins (DAPs) were identified. Functional classification analysis and Gene Ontology enrichment analysis showed that most DAPs involved in photosynthesis were downregulated, indicating low levels of photosynthesis. DAPs that participated in glycolysis, such as glyceraldehyde-3-phosphate dehydrogenase, pyruvate decarboxylase, and alcohol dehydrogenase, were upregulated to fulfill the increasing requirement of energy consumption during maturation conversion. The storage proteins, such as globulins, legumins, vicilins, and oleosin, were also increased significantly during maturation conversion. Protein-protein interaction analysis and function annotation revealed that the upregulation of oleosin, oil body-associated proteins, and acyl-coenzyme A oxidase 2 resulted in the accumulation of oil in quinoa seeds. The downregulation of ß-amyrin 28-oxidase was observed, indicating the decreasing saponin content, during maturation, which makes the quinoa "sweet". By the PRM and qRT-PCR analysis, the expression patterns of most selected DAPs were consistent with the result of TMT proteomics. Our study enhanced the understanding of the maturation conversion in quinoa. This might be the first and most important step toward the genetic improvement of quinoa.

Modulation of tumor-associated macrophages in colitis-associated colorectal cancer.

Intestinal macrophages are the most abundant immune cells in the small and large intestine, which maintain intestinal homeostasis by clearing invading bacteria and dead cells, secreting anti-inflammatory cytokines, and inducing tolerance to symbiotic bacteria and food particles. In addition, as antigen-presenting cells, they also participate in eliciting adaptive immune responses through bridging innate immune responses. After the intestinal homeostasis is disrupted, the damaged or apoptotic intestinal epithelial cells cannot be effectively cleared, and the infection of exogenous pathogens and leakage of endogenous antigens lead to persistent intestinal inflammation. Long-term chronic inflammation is one of the important causes of colitis-associated carcinogenesis (CAC). Tumor microenvironment (TME) is gradually formed around tumor cells, in which tumor associated macrophage (TAMs) is not only the builder, but also regulated by TME. This review just briefly summarized the role of intestinal macrophages under physiological and pathological inflammatory and cancerous conditions, and current therapeutic strategies for intestinal diseases targeting macrophages.

Multi-parametric Magnetic Resonance Imaging Fusion for Automatic Classification of Prostate Cancer.

Computer-aided diagnosis (CAD) of prostate cancer (PCa) using multi-parametric magnetic resonance imaging (mp-MRI) has recently gained great research interest. In this work, a fully automatic CAD pipeline of PCa using mp-MRI data is presented. In order to fully explore the mp-MRI data, we systematically investigate three multi-modal medical image fusion strategies in convolutional neural networks, namely input-level fusion, feature-level fusion, and decision-level fusion. Extensive experiments are conducted on two datasets with different PCa-related diagnostic tasks. We identify a pipeline that works relatively the best for both diagnostic tasks, two important components of which are stacking three adjacent slices as the input and performing decision-level fusion with specific loss weights. Clinical relevance- This work provides a practical method for automated diagnosis of PCa based on multi-parametric MRI.

AADG: Automatic Augmentation for Domain Generalization on Retinal Image Segmentation.

Convolutional neural networks have been widely applied to medical image segmentation and have achieved considerable performance. However, the performance may be significantly affected by the domain gap between training data (source domain) and testing data (target domain). To address this issue, we propose a data manipulation based domain generalization method, called Automated Augmentation for Domain Generalization (AADG). Our AADG framework can effectively sample data augmentation policies that generate novel domains and diversify the training set from an appropriate search space. Specifically, we introduce a novel proxy task maximizing the diversity among multiple augmented novel domains as measured by the Sinkhorn distance in a unit sphere space, making automated augmentation tractable. Adversarial training and deep reinforcement learning are employed to efficiently search the objectives. Quantitative and qualitative experiments on 11 publicly-accessible fundus image datasets (four for retinal vessel segmentation, four for optic disc and cup (OD/OC) segmentation and three for retinal lesion segmentation) are comprehensively performed. Two OCTA datasets for retinal vasculature segmentation are further involved to validate cross-modality generalization. Our proposed AADG exhibits state-of-the-art generalization performance and outperforms existing approaches by considerable margins on retinal vessel, OD/OC and lesion segmentation tasks. The learned policies are empirically validated to be model-agnostic and can transfer well to other models. The source code is available at https://github.com/CRazorback/AADG.

Imaging to intervention: Thoracic outlet syndrome.

Thoracic outlet syndrome (TOS) is a clinical disorder resulting from compression of the neurovascular bundle of the lower neck and upper chest. TOS can be categorized into neurogenic, venous, and arterial subtypes which result from anatomical compression of the brachial plexus, subclavian vein, and subclavian artery, respectively. This can lead to neurogenic pain as well as vascular injury with thrombosis and thromboembolism. Interventional and diagnostic radiologists play a critical role in the imaging diagnosis and treatment of vascular TOS. Prompt imaging and endovascular management with surgical collaboration has been shown to provide the most successful and long-lasting clinical outcomes, from vessel patency to symptom relief. In this article, we review the anatomy and clinical presentations of TOS as well as the initial imaging modalities used for diagnosis. Furthermore, we detail the role of the diagnostic and interventional radiologist in the management of TOS, including pre-procedure and endovascular interventions, along with medical and surgical treatments. PRECIS: Diagnostic and Interventional Radiologists play a key role in diagnosis and management of vascular thoracic outlet syndromes and are critical for timely and successful outcomes.

Lightweight deep neural networks for cholelithiasis and cholecystitis detection by point-of-care ultrasound.

BACKGROUND AND OBJECTIVE: Emergency physicians (EPs) frequently deal with abdominal pain, including that is caused by either gallstones or acute cholecystitis. Easy access and low cost justify point-of-care ultrasound (POCUS) use as a first-line test to detect these diseases; yet, the detection performance of POCUS by EPs is unreliable, causing misdiagnoses with serious impacts. This study aimed to develop a machine learning system to detect and localize gallstones and to detect acute cholecystitis by ultrasound (US) still images taken by physicians or technicians for preliminary diagnoses. METHODS: Abdominal US images (> 89,000) were collected from 2386 patients in a hospital database. We constructed training sets for gallstones with or without cholecystitis (N = 10,971) and cholecystitis with or without gallstones (N = 7348) as positives. Validation sets were also constructed for gallstones (N = 2664) and cholecystitis (N = 1919). We applied a single-shot multibox detector (SSD) and a feature pyramid network (FPN) to classify and localize objects using image features extracted by ResNet-50 for gallstones, and MobileNet V2 to classify cholecystitis. The deep learning models were pretrained using the COCO-2017 and ILSVRC-2012 datasets. RESULTS: Using the validation sets, the SSD-FPN-ResNet-50 and MobileNet V2 achieved areas under the receiver operating characteristics curve of 0.92 and 0.94, respectively. The inference speeds were 21 (47.6 frames per second, fps) and 7 ms (142.9 fps). CONCLUSIONS: A machine learning system was developed to detect and localize gallstones, and to detect cholecystitis, with acceptable discrimination and speed. This is the first study to develop this system for either gallstone or cholecystitis detection with absence or presence of each one. After clinical trials, this system may be used to assist EPs, including those in remote areas, for detecting these diseases.

Deep Learning-Based Detection of Early Renal Function Impairment Using Retinal Fundus Images: Model Development and Validation.

BACKGROUND: Retinal imaging has been applied for detecting eye diseases and cardiovascular risks using deep learning-based methods. Furthermore, retinal microvascular and structural changes were found in renal function impairments. However, a deep learning-based method using retinal images for detecting early renal function impairment has not yet been well studied. OBJECTIVE: This study aimed to develop and evaluate a deep learning model for detecting early renal function impairment using retinal fundus images. METHODS: This retrospective study enrolled patients who underwent renal function tests with color fundus images captured at any time between January 1, 2001, and August 31, 2019. A deep learning model was constructed to detect impaired renal function from the images. Early renal function impairment was defined as estimated glomerular filtration rate <90 mL/min/1.73 m2. Model performance was evaluated with respect to the receiver operating characteristic curve and area under the curve (AUC). RESULTS: In total, 25,706 retinal fundus images were obtained from 6212 patients for the study period. The images were divided at an 8:1:1 ratio. The training, validation, and testing data sets respectively contained 20,787, 2189, and 2730 images from 4970, 621, and 621 patients. There were 10,686 and 15,020 images determined to indicate normal and impaired renal function, respectively. The AUC of the model was 0.81 in the overall population. In subgroups stratified by serum hemoglobin A1c (HbA1c) level, the AUCs were 0.81, 0.84, 0.85, and 0.87 for the HbA1c levels of ≤6.5%, >6.5%, >7.5%, and >10%, respectively. CONCLUSIONS: The deep learning model in this study enables the detection of early renal function impairment using retinal fundus images. The model was more accurate for patients with elevated serum HbA1c levels.

The SUSTech-SYSU dataset for automated exudate detection and diabetic retinopathy grading.

Automated detection of exudates from fundus images plays an important role in diabetic retinopathy (DR) screening and evaluation, for which supervised or semi-supervised learning methods are typically preferred. However, a potential limitation of supervised and semi-supervised learning based detection algorithms is that they depend substantially on the sample size of training data and the quality of annotations, which is the fundamental motivation of this work. In this study, we construct a dataset containing 1219 fundus images (from DR patients and healthy controls) with annotations of exudate lesions. In addition to exudate annotations, we also provide four additional labels for each image: left-versus-right eye label, DR grade (severity scale) from three different grading protocols, the bounding box of the optic disc (OD), and fovea location. This dataset provides a great opportunity to analyze the accuracy and reliability of different exudate detection, OD detection, fovea localization, and DR classification algorithms. Moreover, it will facilitate the development of such algorithms in the realm of supervised and semi-supervised learning.

Probability distribution guided optic disc and cup segmentation from fundus images.

In this paper, we proposed and validated a probability distribution guided network for segmenting optic disc (OD) and optic cup (OC) from fundus images. Uncertainty is inevitable in deep learning, as induced by different sensors, insufficient samples, and inaccurate labeling. Since the input data and the corresponding ground truth label may be inaccurate, they may actually follow some potential distribution. In this study, a variational autoencoder (VAE) based network was proposed to estimate the joint distribution of the input image and the corresponding segmentation (both the ground truth segmentation and the predicted segmentation), making the segmentation network learn not only pixel-wise information but also semantic probability distribution. Moreover, we designed a building block, namely the Dilated Inception Block (DIB), for a better generalization of the model and a more effective extraction of multi-scale features. The proposed method was compared to several existing state-of-the-art methods. Superior segmentation performance has been observed over two datasets (ORIGA and REFUGE), with the mean Dice overlap coefficients being 96.57% and 95.81% for OD and 88.46% and 88.91% for OC.

Analysis of factors related to recurrence of paediatric hepatoblastoma - a single Centre retrospective study.

BACKGROUND: This study was performed to identify risk factors associated with recurrence of hepatoblastoma. METHODS: A retrospective study was conducted on 56 patients with hepatoblastoma from 2012 to 2015 in Beijing Children's Hospital. Pretreatment extension stage (PRETEXT), serum alpha fetoprotein (AFP) value, change trend of tumors after treatment and some other clinical characteristics were collected and analyzed. The comparison of independent variables that were not distributed normally was performed with the log-rank test. RESULTS: Twenty-eight patients with tumour recurrence and 28 patients without recurrence were included in this study, and the median age at presentation was 46.5 (26, 71.5) months. There was a significant difference in the 3-year recurrence-free survival (RFS) probability between patients aged over 54 months and those younger than 54 months (p = 0.007). After neoadjuvant chemotherapy, the chance of recurrence in partial response (PR) patients was significantly lower than that in stable disease (SD) patients (p = 0.004). The 3-year RFS rate of patients with a reduction in AFP of more than 60% after neoadjuvant chemotherapy was significantly higher than that of patients with a reduction of less than 60% (p = 0.005). The postoperative follow-up revealed that patients whose postoperative AFP fell to normal levels within 6 months of the start of treatment had a 3-year RFS rate of 68.6%, which is higher than that of patients whose AFP fell below the normal range after 6 months (p = 0.0005). Finally, the multivariate analysis by Cox regression showed that AFP decreased by less than 60% and tumour size decreased by less than 50% after neoadjuvant chemotherapy were significant independent prognostic risk factors for the 3-year RFS rate. The other clinical features were not significantly associated with tumour recurrence in this study. CONCLUSIONS: Through this study, we concluded that the prognosis of childhood HB is related to the age at presentation and the response of chemotherapy. The results of the multivariate analysis showed that AFP decreased by less than 60% and tumour size decreased by less than 50% after neoadjuvant chemotherapy were significant independent prognostic risk factors. These findings can be helpful to evaluate therapeutic effects and predict prognosis.

Diagnosis and treatment of pancreatoblastoma in children: a retrospective study in a single pediatric center.

BACKGROUND: Pancreatoblastoma is a very rare malignant pancreatic tumor in children. Pancreatoblastoma is the most common pancreatic tumor in children less than 10 years of age, accounting for 25% of the pancreatic neoplasm. There were only a few published literatures about the standardized diagnostic and management protocol for PB in the last decade. OBJECTIVE: To summarize our experience in the management of pancreatoblastoma in children and adolescents with emphasis on the presentation, diagnosis, treatment, and outcomes. A management strategy will also be discussed. METHODS: This was a retrospective case-series study of all pancreatoblastoma in patients < 18 years of age who were treated at Beijing children's hospital (BCH) from January 2002-January 2015. The diagnoses of PB were confirmed by histopathology analysis of the resected specimen. The variables being analyzed included patient demographics, age at diagnosis, clinical presentation, tumor size, metastasis if present, tumor markers (AFP), type of surgery, length of follow-up, and outcome. The assessment of the tumor location, size, extent of the tumor, and distant metastasis was made by ultrasound (US), computed tomography (CT), and/or magnetic resonance imaging (MRI). RESULT: 21 patients with pancreatoblastoma were diagnosed at a median age of 4 years, 7 girls, and 14 boys. The diagnosis of pancreatoblastoma was identified by the histology examination. The most common syndrome was abdominal mass (n = 11), followed by abdominal pain (N = 10), elevated serum AFP levels were noted in almost all cases (17/18), 17 patients with disease initially unresectable on diagnosis accepted neo-adjuvant chemotherapy consisting of CDV, OPEC, PLADO, IEV, and AVCP. All patients underwent surgery, including pancreaticoduodenectomy (Whipple's procedure), the Pylorus-preserving pancreaticoduodenectomy (traverse-Longmire procedure), Spleen-preserving distal pancreatectomy, and distal pancreatectomy with en bloc splenectomy, Roux-en-Y end-to-end pancreatojejunostomy. In all, 13 children were disease free with a median follow-up of 53 months (range 11-156 months). CONCLUSIONS: The pancreatoblastoma in children and adolescents is a malignant tumor. Complete resection combined with chemotherapy is associated with long-term survival. For the unresectable tumor at diagnosis, preoperative chemotherapy was recommended to reduce tumor volume. AFP is critical for diagnosis and monitoring the disease as a tumors marker.