[Guillain-Barre syndrome after allogeneic hematopoietic stem cell transplantation: a case report and literature review].

Guillain-Barre syndrome rarely develops after allogeneic hematopoietic stem cell transplantation (allo-HSCT), and only a few reports exist in China. Guillain-Barre syndrome is an acute and life-threatening condition that requires early diagnosis and treatment. A patient with acute myeloid leukemia underwent allogeneic HSCT for >5 months and gradually developed limb muscle weakness and limited eye movement after coexisting with delayed acute intestinal graft-versus-host disease. After the examination of cerebrospinal fluid and electromyography, the diagnosis of Guillain-Barre syndrome was confirmed. After a high-dose intravenous immunoglobulin (IVIg) treatment, muscle strength gradually recovered, and the prognosis was good.

[Nomogram prediction model of cervical anastomotic leakage after esophageal cancer surgery].

Objective: To retrospectively analyze the risk factors of anastomotic leakage in the neck after esophageal cancer and establish a nomogram prediction model that can accurately predict the occurrence of anastomotic leakage in the neck of the patient. Methods: The study retrospectively analyzed 702 patients who underwent radical esophageal cancer surgery between January 2010 and May 2015 at Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College. A multivariate logistic regression model was used to determine the risk factors for neck anastomotic leak, and a nomogram model was constructed, internal validation methods were used to evaluate and verify the predictive effectiveness of the nomogram. Results: There were 702 patients in the whole group, 492 in the training group and 210 in the validation group. The incidence of postoperative cervical anastomotic leak was 16.1% (79/492) in 492 patients with esophageal cancer in the training group. Multifactorial analysis revealed calcification of the descending aorta (OR=2.12, 95% CI: 1.14, 3.94, P=0.018), calcification of the celiac artery (OR=2.29, 95% CI: 1.13, 4.64, P=0.022), peripheral vascular disease (OR=5.50, 95% CI: 1.64, 18.40, P=0.006), postoperative ventilator-assisted breathing (OR=5.33, 95% CI: 1.83, 15.56, P=0.002), pleural effusion or septic chest (OR=3.08, 95% CI: 1.11, 8.55, P=0.031), incisional fat liquefaction and infection (OR=3.49, 95% CI: 1.68, 7.27, P=0.001) were independent risk factors for the development of cervical anastomotic leak after esophageal cancer surgery. The results of the nomogram prediction model showed that the consistency indices of the training and external validation groups were 0.73 and 0.74, respectively (P<0.001), suggesting that the prediction model has good predictive efficacy. Conclusion: The nomogram prediction model can intuitively predict the incidence of postoperative cervical anastomotic leakage in patients with high prediction accuracy, which can help provide a clinical basis for preventing cervical anastomotic leak and individualized treatment of patients.

[A clinicopathological classification of space-occupying lesions of the orbit in 1 913 patients from 2000 to 2021].

Objective: To investigate the histopathological classification of orbital space-occupying lesions. Methods: This is a retrospective case series study. The clinical and pathological data of 1 913 tissue specimens from 1 913 patients with space-occupying lesions of the orbit which were examined in the Second Affiliated Hospital, Zhejiang University School of Medicine from January 2000 to December 2021 were collected. The mass lesions were classified based on histogenesis, pathological nature and age. Results: There were 913 males (47.7%) and 1 000 females (52.3%). The lesions were benign in 1 489 patients (77.8%) and malignant in 424 patients (22.2%). Based on histogenesis, there were 521 vasculogenic lesions (27.2%), which rancked first, 407 cystoid lesions (21.3%), 277 lymphoproliferative lesions (14.5%), 182 lacrimal gland lesions (9.5%) and 121 inflammatory lesions (6.3%). By pathological nature, there were 1 489 benign lesions, including cavernous hemangioma (275, 14.4%), dermoid cyst (225, 11.8%), other hemangiomas (199, 10.4%), epidermoid cyst (136, 7.1%) and benign mixed tumor of the lacrimal gland (134, 7.0%), and 257 malignant lesions, including lymphoma (210, 11.0%) and sebaceous gland carcinoma (47, 2.5%). The age of all patients ranged from 0 to 90 years, while 247 lesions (12.9%) occurred in patients aged 0 to18 years, 1 270 lesions (66.4%) in patients aged 19 to 59 years, and 396 lesions (20.7%) in patients aged 60 to 90 years. Conclusions: In 22 years, almost 2/3 benign orbital lesions in the Second Affiliated Hospital, Zhejiang University School of Medicine occurred in young and middle-aged patients, and males were fewer than females. The most common benign orbital tumors was cavernous hemangioma, followed by dermoid cyst and epidermoid cyst. And the most common malignant orbital tumor was lymphoma, which occurred more frequently in older patients.

Association between CYP2E1 C-1054T and 96-bp I/D genetic variations and the risk of polycystic ovary syndrome in Chinese women.

PURPOSE: To investigate the association of cytochrome P450 2E1 (CYP2E1) C-1054T (rs2031920) and 96-bp I/D genetic variations with the risk of polycystic ovary syndrome (PCOS), and to estimate the effects of genotypes on the clinical, metabolic, hormonal, and oxidative stress indicators. METHODS: This case-control study included 762 control women and 1034 patients with PCOS. Genotypes were determined using polymerase chain reaction and/or restriction fragment length polymorphism analysis. Clinical and biochemical parameters were also analyzed. RESULTS: Frequencies of the TT + CT genotype (35.4 vs. 28.9%) and T allele (19.6 vs. 16.0%) of the CYP2E1 C-1054T polymorphism were significantly higher in the PCOS group than in the control group (OR = 1.350, 95% CI 1.103-1.652, P = 0.004 for the dominant model). Genotype TT + CT remained a significant predictor of PCOS in a logistic regression model including age, body mass index (BMI), and recruitment year of participants (OR = 1.345, 95% CI 1.071-1.688, P = 0.011). No statistical differences were found in the genotype and allele frequencies of CYP2E1 96-bp I/D polymorphism. However, the combined genotype DD/TT + CT was related to an increased risk of PCOS when the DD/CC wild-type combined genotype was used as a reference. Patients with the I allele of 96-bp I/D polymorphism had a lower BMI but higher plasma apolipoprotein B and oxidized low-density lipoprotein cholesterol levels than those with the DD genotype. CONCLUSION: CYP2E1 C-1054T, but not 96-bp I/D, genetic polymorphism is associated with an increased risk of PCOS in Chinese women.

[Influence of age on advanced neoplasia detection in colorectal cancer screening in population at high risk].

Objective: To compare the detection rate of advanced neoplasia and the number of people needing endoscopy in colorectal cancer screening giving at different starting age in population at high risk. Methods: Based on the screening project of early diagnosis and treatment of colorectal cancer in Jiashan county, Zhejiang province, two rounds of colorectal cancer screening were conducted between January 2007 and December 2020. After excluding participants who were not at high risk or had incomplete information, 27 130 participants and 31 205 participants were finally enrolled in round one and in round two, respectively. The spline analysis based on the generalized additive model was used to describe the trend of detection rate of advanced neoplasia with age. The detection rate and number of people needing endoscopy for the groups with starting age at 50, 45 and 40 years were calculated, and the differences in the detection rate were tested by χ2 goodness of fit test. Results: A total of 21 077 (77.69%) participants in round one and 25 249 (80.91%) participants in round two received endoscopy, in whom 1 097 (detection rate=52.05‰) and 1 151 (detection rate=45.59‰) had advanced neoplasia (cancers and advanced adenomas), respectively. The detection rate increased significantly with age, and the detection rate in round one were significantly higher than that in round two (P<0.05). The overall detection rates of advanced neoplasia for the groups with starting age at 50, 45 and 40 years were 61.11‰, 56.14‰ and 52.05‰ in round one, and 49.10‰, 46.75‰ and 45.59‰ in round two, respectively. The rates were significantly higher for the group with starting age at 50 years than that with starting age at 40 years in both round one and round two (P<0.05). The numbers of people needing endoscopy of advanced neoplasia for the groups with starting age at 50, 45 and 40 years were 17, 18, and 20 in round one, and 21, 22 and 22 in round two. Conclusions: The detection rate of advanced neoplasia increased with age. Starting screening at lower age might contribute to decreased detection rate and increased number of people needing endoscopy. However, the difference was limited.

Radiomics-Based Machine Learning for Outcome Prediction in a Multicenter Phase II Study of Programmed Death-Ligand 1 Inhibition Immunotherapy for Glioblastoma.

BACKGROUND AND PURPOSE: Imaging assessment of an immunotherapy response in glioblastoma is challenging due to overlap in the appearance of treatment-related changes with tumor progression. Our purpose was to determine whether MR imaging radiomics-based machine learning can predict progression-free survival and overall survival in patients with glioblastoma on programmed death-ligand 1 inhibition immunotherapy. MATERIALS AND METHODS: Post hoc analysis was performed of a multicenter trial on the efficacy of durvalumab in glioblastoma (n = 113). Radiomics tumor features on pretreatment and first on-treatment time point MR imaging were extracted. The random survival forest algorithm was applied to clinical and radiomics features from pretreatment and first on-treatment MR imaging from a subset of trial sites (n = 60-74) to train a model to predict long overall survival and progression-free survival and was tested externally on data from the remaining sites (n = 29-43). Model performance was assessed using the concordance index and dynamic area under the curve from different time points. RESULTS: The mean age was 55.2 (SD, 11.5) years, and 69% of patients were male. Pretreatment MR imaging features had a poor predictive value for overall survival and progression-free survival (concordance index = 0.472-0.524). First on-treatment MR imaging features had high predictive value for overall survival (concordance index = 0.692-0.750) and progression-free survival (concordance index = 0.680-0.715). CONCLUSIONS: A radiomics-based machine learning model from first on-treatment MR imaging predicts survival in patients with glioblastoma on programmed death-ligand 1 inhibition immunotherapy.

[Progress in pulmonary enteric adenocarcinoma].

Pulmonary enteric adenocarcinoma (PEAC), as a rare histologic subtype of primary lung adenocarcinoma, is defined as an adenocarcinoma in which the enteric component exceeds 50%. It is named after its shared morphological and immunohistochemical features with colorectal cancer. While with such similarity, the differential diagnosis of PEAC and lung metastatic colorectal cancer is a great challenge in the clinic. PEAC may originate from the intestinal metaplasia of respiratory basal cells stimulated by risk factors such as smoking. Current studies have found that KRAS is a relatively high-frequency mutation gene, and other driver gene mutations are rare. In terms of immunohistochemistry, in pulmonary enteric adenocarcinoma, the positive rate was 88.2% (149/169) for CK7, 78.1% (132/169) for CDX2, 48.2% (82/170) for CK20 and 38.8% (66/170) for TTF1. As for clinical features, the average age of onset for pulmonary enteric adenocarcinoma was 62 years, male patients accounted for 56.5% (35/62), smokers accounted for 78.8% (41/52), and 41.4% (24/58) of the primary lesion was located in the upper lobe of the right lung. In terms of treatment, conventional non-small cell lung cancer (NSCLC) regimens rather than colorectal cancer regimens are now recommended. There is still an urgent need for more basic and clinical research, in-depth exploration of its molecular feature and pathogenesis from the level of omics and other aspects, to help diagnosis and differential diagnosis, and find the optimal chemotherapy regimen, possibly effective targeted therapy and even immunotherapy.

Bone mineral density, osteopenia and osteoporosis among US adults with cancer.

BACKGROUND: Bone mineral deficits are one of the most common complications in cancer survivors. However, there are no studies evaluating bone mineral density (BMD) and the prevalence of osteopenia and osteoporosis among patients with different types of cancers. AIM: The objective was to assess BMD and evaluate the prevalence of osteopenia and osteoporosis among US adults with cancer. DESIGN: A cross-section propensity score matching study. METHODS: We extracted data from National Health and Nutrition Examination Survey database from 2005 to 2018. We compared BMD in participants with and without cancer which was further analyzed according to cancer type. We conducted logistic regression to evaluate adjusted odds ratios of osteopenia and osteoporosis and determine risk factors for their development. RESULTS: We found that BMD was significantly higher in participants without cancer than cancer patients. Furthermore, the median BMD of patients with breast cancer or skin cancer (including melanoma) was significantly lower than participants without cancer. People with breast, lung, genitourinary and skin cancers were more likely to incur osteopenia/osteoporosis than those without cancer. CONCLUSIONS: BMD differs depending upon type in survivors. Individuals with a history of cancer have a poor understanding of osteoporosis and its risk factors. Understanding risk factors in patients with cancers identified in our study may be helpful for preventing osteoporosis and fractures and the development of screening guidelines.

Investigating the mechanism of interactive regulation of B-cell lymphoma-2/Beclin 1 through electroacupuncture intervention during reperfusion in myocardial ischemia-reperfusion injury in a rat model.

To observe the regulation of B-cell lymphoma-2 (Bcl-2)/Beclin 1 interaction through electroacupuncture (EA) intervention during reperfusion and to investigate the EA mechanism of apoptosis-autophagy interactive regulation against myocardial ischemia-reperfusion injury (MIRI). A total of 48 adult Sprague Dawley (SD) rats were randomly divided into the sham-operated group (group Sham), the model group (group Model), the EA group (group EA), and the JNK inhibitor (SP600125) group (group JNK), with 12 rats in each group. Biospecimens were collected randomly from six rats in each group four hours after reperfusion. Evans Blue and triphenyl tetrazolium chloride double-staining were applied to observe each group's myocardial damage area and risk area. We collected 4 ml of blood by abdominal aortic method to detect serum troponin cTnI level by enzyme-linked immunosorbent assay (ELISA). For the remaining six in each group, a part of myocardial tissue below the ligation line was stored in 4% paraformaldehyde for immunohistochemistry and TUNEL staining; the other amount of myocardial tissue was detected by Western blotting to determine the expression levels of Bcl-2, Beclin1, and the phosphorylation levels of Thr69, Ser70, and Ser87 in Bcl-2. In results: electroacupuncture (EA) intervention during reperfusion significantly reduced the myocardial infarction area, cTnI level, and myocardial apoptosis, upregulated Bcl-2 expression, downregulated Beclin 1 expression and inhibited phosphorylation levels of Thr69, Ser70, and Ser87 in Bcl-2. We concluded that EA effectively inhibited apoptosis by upregulating Bcl-2 expression and inhibiting the phosphorylation of Thr69, Ser70, and Ser87 in Bcl-2. This reduced the separation of Bcl-2 and Beclin 1, restrains excessive autophagy, alleviates MIRI, and has a protective effect on myocardial tissue.

[Effect of circBIRC6 targeting miR-367-3p on cisplatin resistance of ovarian cancer cells].

Objective: To explore the effect of circBIRC6 on cisplatin resistance of ovarian cancer cells and the molecular mechanism. Methods: The ovarian cancer cell line SKOV3 and ovarian cancer cisplatin-resistant cell line SKOV3 / DDP were cultured in vitro, and treated with different concentrations of cisplatin. SKOV3 and SKOV3/DDP cells were transfected with si-NC, si-circBIRC6, si-circBIRC6+ anti-miR-NC, si-circBIRC6+ anti-miR-367-3p by liposome-mediated method, which were denoted as DDP+ si-NC group, DDP+ si-circBIRC6 group, DDP+ si-circBIRC6+ anti-miR-NC group and DDP+ si-circBIRC6+ anti-miR-367-3p group, respectively, and then were treated with 2 µg/ml cisplatin for 24 hours. The cell proliferation inhibition rate was detected by 3-(4, 5-dimethyl-2-thiazolyl)-2, 5-diphenyl-2H tetrazolium bromide method, and the half inhibitory concentration (IC(50)) value of cisplatin was calculated. Real-time fluorescent quantitative polymerase chain reaction was used to detect the transcriptional levels of circBIRC6 and miR-367-3p. Flow cytometry was used to detect the apoptotic rate. Dual luciferase report experiment verified the targeting relationship of circBIRC6 and miR-367-3p. Western blot was used to detect the expressions of Cyclin D1, Bcl-2, p21, Bax. Results: The expression levels of circBIRC6 in SKOV3 cells were 1.00±0.05, significantly lower than 3.04±0.24 in SKOV3/DDP cells (P<0.001). The expression levels of miR-367-3p in SKOV3 cells were 1.00±0.08, significantly higher than 0.54±0.05 in SKOV3/DDP cells (P<0.001). The cell proliferation inhibition rates of SKOV3 cells and SKOV3/DDP cells were (22.47±2.04)% and (8.84±0.71)%, the IC(50) values of SKOV3 cells and SKOV3/DDP cells were 6.65±0.94 and 28.18±4.91, respectively, with significant difference (P<0.05). The proliferation inhibition rate and apoptosis rate of SKOV3 cells in DDP+ si-NC group[(22.19±2.19)% and (10.98±1.12)%] were lower than those in DDP+ si-circBIRC6 group [(74.18±5.36)% and (32.91±3.19)%, all P<0.05]. The proliferation inhibition rate and apoptosis rate of SKOV3/DDP cells[(8.71±0.87)% and (7.39±0.63)%] were lower than those of DDP+ si-circBIRC6 group [(40.85±4.07)% and (25.31±2.53)%, all P<0.05]. The protein expression levels of Cyclin D1 and Bcl-2 in SKOV3 and SKOV3/DDP cells in DDP+ si-circBIRC6 group were lower than those in DDP+ si-NC group, and the protein expression levels of p21 and Bax were higher than those in DDP+ si-NC group (all P<0.05). The dual luciferase report experiment confirmed that circBIRC6 targeted miR-367-3p. Inhibition of miR-367-3p expression reduced the effect of circBIRC6 deletion on ovarian cancer cell proliferation, apoptosis and cisplatin resistance. Conclusion: Knockdown of circBIRC6 may inhibit the proliferation of ovarian cancer cisplatin-resistant cells and induce apoptosis by up-regulating the expression of miR-367-3p, therefore impair the cisplatin resistance of these cells.

Identification of peroxiredoxin 6 as a potential lung-adenocarcinoma biomarker for predicting chemotherapy response by proteomic analysis.

The prognosis of lung cancer remains poor due to the limited biomarker selection for treating patients with optimal chemotherapy. The aim of this study is to discover and identify new biomarkers with the value of predicting chemotherapy responses in a lung adenocarcinoma (AD) specimen. In this study, six pairs of pre-treatment fresh primary lung AD-cancer tumors with varied chemotherapy responses were used to discover new biomarkers by two-dimensional difference gel electrophoresis (2D DIGE). Among the matched protein spots, 19 were up-regulated and 18 were down-regulated in chemo-sensitive tumors versus chemo-resistant tissues. These differentially expressed proteins could be divided into five classes: redox regulation protein, the cytoskeletal protein, cell metabolism enzymes or proteins, apoptosis, signal transduction mediated molecules, and other functional proteins. Proteins of interest, including PRDX2, PRDX6, and Gelsolin, were differentially expressed in chemo-sensitive tumors versus chemo-resistant tissues and these observations were validated by immunohistochemistry in 92 formalin-fixed and paraffin-embedded (FFPE) specimens. Our results demonstrated that PRDX6 protein expression was closely related to tumor response (cc2 = 5.57, P < 0.05), whereas no relationship of PRDX2 and Gelsolin were obtained with tumor response (cc2 = 0.51 P > 0.05, cc2 = 0.41 P > 0.05). This tissue proteomics study provides evidence that PRDX6 may be regarded as a predictive biomarker for poor chemotherapy response, which can be helpful in guiding pretreatment protocols.

[Efficacy and safety of generic azacitidine in Chinese patients with higher-risk myelodysplastic syndromes: a multicenter, prospective, single-arm study].

Objective: To evaluate the efficacy, safety, and pharmacokinetics of the generic azacitidine in Chinese patients with higher-risk myelodysplastic syndromes(MDS). Methods: Between October 2013 and 2016, 72 patients were eligible for enrollment at 9 sites from China received generic subcutaneous azacitidine 75 mg·m(-2)·d(-1) for 7 days per 28-day cycle, for ≥6 cycles. Pharmacokinetic blood samples were collected on day 1 of a single-dose. Results: For each patient at cycle 6 or at the time of study discontinuation, whichever came first, the overall response rate, which included complete remission (CR)and partial remission(PR), was 6.9%(5/72), the rate of patients who had the best effect with CR or PR during the treatment was 12.5%(9/72). Patients who were dependent on red-blood-cell transfusions and platelet transfusions at baseline became transfusion independent were 46.3%(19/41)and 41.2% (7/17), respectively. The median time of treatment was 6 cycles, and the median OS was 16.1 months (95%CI 10.9-20.6 months). For 36 patients(50%)received treatment at ≥6 cycles, and the median OS was 22.3 months(95%CI 16.1- not evaluative). Most common grade Ⅲ-Ⅳ hematologic treatment-emergent adverse events were neutropenia(55%), leukopenia(47%), and thrombocytopenia(61%). Pharmacokinetic profiles were similar for generic and original azacitidine in Chinese patients. Conclusion: Generic azacitidine treatment was favorable and safe and can be used as a standard treatment for patients with higher-risk MDS.

[Aldo-keto reductase family 1 B10 participates in the regulation of hepatoma cell cycle through p27/p-Rb signaling pathway].

Objective: Aldo-keto reductase family 1 member B10 (AKR1B10) pathogenesis, early diagnosis and prognosis are closely related with hepatoma. Therefore, this study explores the effect and mechanism of AKR1B10 on cell cycle in hepatoma cells. Methods: HepG2 cells were infected with lentivirus LV-AKR1B10-shRNA or treated with epalrestat, an AKR1B10 inhibitor. The expression level of AKR1B10 was detected by Western blot assay and real-time fluorescence quantitative PCR (RT-qPCR). Decreased AKR1B10 activity was detected by reduced coenzyme II (NADPH) absorbance at 340 nm. The low expression of AKR1B10 and the effect of different concentrations of epalrestat on cell proliferation and cell cycle were detected by CCK-8 method and flow cytometry. The protein expression levels of p-rb, cyclin D1, E1, p27 in HepG2 cells were detected by Western blot. The mean of the two samples was tested using independent sample t-test. Results: AKR1B10 expression level in hepatoma cells was significantly increased compared to normal liver cells, and the relative expression level of AKR1B10 protein in HepG2 cells was 6.71 ± 1.11 (P = 0.012). Epalrestat was significantly inhibited with the enzymatic activity of AKR1B10 in a dose-dependent manner. AKR1B10 gene in HepG2 cells was effectively silenced. HepG2 cells treated with different concentrations of epalrestat (AKR1B10 inhibitor) for 24, 48 and 72 h had inhibited cell proliferation, promoted G0/G1 cell cycle arrest, reduced the expression of p-Rb, cyclin D1, and cyclin E1 and increased the expression of cyclin dependent kinase inhibitor p27 expression. Conclusion: AKR1B10 inhibitory expression and activity can promote G0/G1 cell cycle arrest in HepG2 cells through the p27 / p-Rb pathway.

[Thyroid disruptor p, p'-DDE inhibited the expression of LHX4 and DIS3L protein in Nthy-ori-3-1 cells].

Objective: To observe the changes of LHX4 and DIS3L mRNA and protein expression in Nthy-ori-3-1 cells after the treatment of thyroid disruptor p, p'-DDE. Methods: Nthy-ori-3-1 cells in logarithmic growth phase were treated with 0, 0.5, 1.0, 2.0 and 5.0 µg/ml p, p'-DDE solution. The growth state and morphology of the cells were observed by microscope. The mRNA levels of LHX4 and DIS3L were detected by real-time fluorescent quantitative PCR, and the protein expression levels of LHX4 and DIS3L were detected by Western blot. Results: when the concentrations of p, p'-DDE were 0, 0.5, 1.0 and 2.0 µg/ml, Nthy-ori-3-1 cells grew normally. There were 33 differential genes in 2.0 µg/ml group, among which 13 genes were down regulated and 20 genes were up-regulated. Compared with the control group, the protein expression levels of LHX4 and DIS3L in 1.0 and 2.0 µg/ml groups were significantly decreased (P<0.05) , and the relative expression levels of LHX4 and DIS3L protein mRNA in 1.0 µg/ml group were significantly decreased (P<0.05) . Conclusion: p, p'-DDE can affect the protein expression of LHX4 and dis3l in nthy-ori-3-1 cells.

Automatic Machine Learning to Differentiate Pediatric Posterior Fossa Tumors on Routine MR Imaging.

BACKGROUND AND PURPOSE: Differentiating the types of pediatric posterior fossa tumors on routine imaging may help in preoperative evaluation and guide surgical resection planning. However, qualitative radiologic MR imaging review has limited performance. This study aimed to compare different machine learning approaches to classify pediatric posterior fossa tumors on routine MR imaging. MATERIALS AND METHODS: This retrospective study included preoperative MR imaging of 288 patients with pediatric posterior fossa tumors, including medulloblastoma (n = 111), ependymoma (n = 70), and pilocytic astrocytoma (n = 107). Radiomics features were extracted from T2-weighted images, contrast-enhanced T1-weighted images, and ADC maps. Models generated by standard manual optimization by a machine learning expert were compared with automatic machine learning via the Tree-Based Pipeline Optimization Tool for performance evaluation. RESULTS: For 3-way classification, the radiomics model by automatic machine learning with the Tree-Based Pipeline Optimization Tool achieved a test micro-averaged area under the curve of 0.91 with an accuracy of 0.83, while the most optimized model based on the feature-selection method χ2 score and the Generalized Linear Model classifier achieved a test micro-averaged area under the curve of 0.92 with an accuracy of 0.74. Tree-Based Pipeline Optimization Tool models achieved significantly higher accuracy than average qualitative expert MR imaging review (0.83 versus 0.54, P < .001). For binary classification, Tree-Based Pipeline Optimization Tool models achieved an area under the curve of 0.94 with an accuracy of 0.85 for medulloblastoma versus nonmedulloblastoma, an area under the curve of 0.84 with an accuracy of 0.80 for ependymoma versus nonependymoma, and an area under the curve of 0.94 with an accuracy of 0.88 for pilocytic astrocytoma versus non-pilocytic astrocytoma. CONCLUSIONS: Automatic machine learning based on routine MR imaging classified pediatric posterior fossa tumors with high accuracy compared with manual expert pipeline optimization and qualitative expert MR imaging review.