Achieving a reliable and accurate biomedical image segmentation is a long-standing problem. In order to train or adapt the segmentation methods or measure their performance, reference segmentation masks are required. Usually gold-standard annotations, i.e. human-origin reference annotations, are used as reference although they are very hard to obtain. The increasing size of the acquired image data, large dimensionality such as 3D or 3D + time, limited human expert time, and annotator variability, typically result in sparsely annotated gold-standard datasets. Reliable silver-standard annotations, i.e. computer-origin reference annotations, are needed to provide dense segmentation annotations by fusing multiple computer-origin segmentation results. The produced dense silver-standard annotations can then be either used as reference annotations directly, or converted into gold-standard ones with much lighter manual curation, which saves experts' time significantly. We propose a novel full-resolution multi-rater fusion convolutional neural network (CNN) architecture for biomedical image segmentation masks, called DeepFuse, which lacks any down-sampling layers. Staying everywhere at the full resolution enables DeepFuse to fully benefit from the enormous feature extraction capabilities of CNNs. DeepFuse outperforms the popular and commonly used fusion methods, STAPLE, SIMPLE and other majority-voting-based approaches with statistical significance on a wide range of benchmark datasets as demonstrated on examples of a challenging task of 2D and 3D cell and cell nuclei instance segmentation for a wide range of microscopy modalities, magnifications, cell shapes and densities. A remarkable feature of the proposed method is that it can apply specialized post-processing to the segmentation masks of each rater separately and recover under-segmented object parts during the refinement phase even if the majority of inputs vote otherwise. Thus, DeepFuse takes a big step towards obtaining fast and reliable computer-origin segmentation annotations for biomedical images.

• MeSH
• Humans MeSH
• Neural Networks, Computer * MeSH
• Image Processing, Computer-Assisted * methods   MeSH
• Check Tag
• Humans MeSH
• Publication type
• Journal Article MeSH

INTRODUCTION: Pulsed Field Cryoablation (PFCA) is a dual-energy cardiac ablation modality consisting of short-duration ultra-low temperature cryoablation (ULTC) followed immediately by pulsed field ablation (PFA) delivered from the same catheter. It is hypothesized that PFCA may improve contact stability during PFA, while maintaining lesion depth and effectiveness of ULTC. METHODS: PARALELL is a first-in-human multicenter study evaluating safety and effectiveness of a novel PFCA catheter and system in patients with persistent atrial fibrillation (PsAF) using the combination of pulmonary vein (PVI) and posterior wall (PWI) isolation. RESULTS: Sixty-six patients were ablated at six sites. One groin hematoma and one intubation-related hospitalization were the only serious procedure- or device-related adverse events recorded in the study. Per protocol, acute effectiveness was evaluated in 46 patients, including 31 patients with post-hoc analysis of cryogenic energy per lesion. After an average of 21.1 ± 9.3 lesions per patient the rates of PVI and PWI were 95.7% (176/184) and 97.7% (42/43), respectively. The average cryogenic energy per patient was highly predictive of acute isolation success with ROC AUC = 0.944% and 100% rates of both PVI and PWI in 24 patients in the optimal energy cohort. Grade I microbubbles and faint muscle contractions were detected in 1.1% and 0.5% of ablations, respectively. CONCLUSION: This initial multi-center experience suggests that PFCA can be efficiently performed for PVI and PWI using a single versatile catheter system, with high acute success and good early safety profile. The evaluation of the chronic 12-month effectiveness of PFCA is ongoing.

• MeSH
• Action Potentials MeSH
• Time Factors MeSH
• Equipment Design MeSH
• Atrial Fibrillation * surgery   physiopathology   diagnosis   MeSH
• Cryosurgery * adverse effects   instrumentation   methods   MeSH
• Middle Aged MeSH
• Humans MeSH
• Risk Factors MeSH
• Aged MeSH
• Heart Rate MeSH
• Cardiac Catheters MeSH
• Pulmonary Veins * surgery   physiopathology   MeSH
• Treatment Outcome MeSH
• Check Tag
• Middle Aged MeSH
• Humans MeSH
• Male MeSH
• Aged MeSH
• Female MeSH
• Publication type
• Journal Article MeSH
• Clinical Trial MeSH
• Multicenter Study MeSH

INTRODUCTION: Up to 50% of diabetic patients with neuropathy suffer from chronic pain, namely painful diabetic neuropathy (PDN), an unmet medical need with significant impact on quality of life. Gabapentin is widely used for PDN, albeit with frequent dose-limiting effects. Trazodone, an antidepressant with multi-modal action, has shown promising results when given at low doses as an add-on to gabapentin. Upon previous clinical trials and experimental evidence, a fixed-dose combination (FDC) of both compounds, at low doses, was developed for neuropathic pain. METHODS: This was a phase II, randomized, double-blind, placebo and reference controlled, dose-finding, multicenter, international, prospective study. Male and female diabetic patients aged 18-75 years and affected by PDN were eligible for enrolment. Patients were randomized (1:1:1:1:2 ratio) to trazodone and gabapentin (Trazo/Gaba) 2.5/25 mg t.i.d. for 8 weeks, Trazo/Gaba 5/50 mg t.i.d. for 8 weeks, Trazo/Gaba 10/100 mg t.i.d. for 8 weeks, gabapentin (Gaba), or placebo (PLB). The aim of the study was to collect preliminary information on the effect of the 3 different FDCs of Trazo/Gaba on pain intensity based on the 11-point numeric rating score (NRS) after 8 weeks of treatment. The secondary objectives were the evaluation of the percentage of responders, neuropathic pain symptoms, anxiety, sleep, quality of life, safety, and tolerability. The primary efficacy endpoint was evaluated with last observation carried out forward (LOCF), using an analysis of covariance (ANCOVA), including treatment and centers as factors and baseline as covariate and applying linear contrast test, excluding the active treatment. Only if the linear contrast test was significant (p < 0.05), the step-down Dunnett test would be used to determine the minimum effective dose significantly different from PLB. If linearity was not verified, an adjusted ANCOVA model and comparisons with Dunnett test were performed. Before the application of the ANCOVA model, the non-significance of interaction treatment per baseline was verified. RESULTS: A total of 240 patients were included in the modified intention-to-treat (m-ITT) population: 39 in Trazo/Gaba 2.5/25 mg, 38 in Trazo/Gaba 5/50 mg, 37 in Trazo/Gaba 10/100 mg, 83 in PLB, and 43 in Gaba. After 8 weeks of treatment, changes of the average daily pain score based on the 11-point NRS from baseline were - 2.52 ± 2.31 in Trazo/Gaba 2.5/25 mg group, - 2.24 ± 1.96 in Trazo/Gaba 5/50 mg group, - 2.46 ± 2.12 in Trazo/Gaba 10/100 mg group, - 1.92 ± 2.21 in Gaba group, and - 2.02 ± 1.95 in the PLB group. The linear contrast test did not result in significant differences (p > 0.05) among treatment groups. Consequently, the minimum effective dose against PLB was not determined. The multiple comparison with Dunnett adjustment did not show any statistically significant differences vs. PLB after 8 weeks of treatment: Trazo/Gaba 2.5/25 mg (95% confidence interval (CI) - 1.2739, 0.2026; p = 0.1539); Trazo/Gaba 5/50 mg (95% CI - 0.9401, 0.5390; p = 0.5931); Trazo/Gaba 10/100 mg (95% CI - 1.0342, 0.4582; p = 0.4471). However, patients receiving the lowest dose of Trazo/Gaba 2.5/25 mg showed a statistically significant difference to PLB after 6 weeks of treatment (95% CI - 1.6648, - 0.2126; p = 0.0116). Positive results were also found for responder patients, other items related to the pain, anxiety, depression, sleep, and quality of life, consistently in favor to the lowest Trazo/Gaba FDC. Two serious adverse events (SAEs) occurred but were judged unrelated to the study treatment. Treatment-emergent adverse events (TEAEs) were mainly mild-to-moderate in intensity and involved primarily nervous system, gastrointestinal disorders, and investigations. CONCLUSIONS: The primary end point of the study was the change from baseline of the average daily pain score based on the 11-point NRS after 8 weeks of treatment. While the primary endpoint was not reached, patients treated with Trazo/Gaba 2.5/25 mg t.i.d. showed statistically significant improvement of pain and other scores after 6 weeks and reported consistent better results in comparison to PLB on primary and secondary endpoints for the overall study duration. According to these results, the lowest dose of Trazo/Gaba FDC may be the best candidate for further clinical development to confirm the potential benefits of the FDC drug for this condition. CLINICAL TRIAL REGISTRATION: NCT03749642.

• Publication type
• Journal Article MeSH

Machine learning has been proven to provide good performances on stress detection tasks using multi-modal sensor data from a smartwatch. Generally, machine learning techniques need a sufficient amount of data to train a robust model. Thus, we need to collect data from several users and send them to a central server to feed the algorithm. However, the uploaded data may contain sensitive information that can jeopardize the user's privacy. Federated learning can tackle this challenge by enabling the model to be trained using data from all users without the user's data leaving the user's device. In this study, we implement federated learning-based stress detection and provide a comparative analysis between individual, centralized, and federated learning. The experiment was conducted on WESAD dataset by using Logistic Regression as the classifier. The experiment results show that in terms of accuracy, federated learning cannot reach the performance level of both individual and centralized learning. The individual learning strategy performs best with an average accuracy of 0.9998 and an average F1-measure of 0.9996.

• Publication type
• Journal Article MeSH

Gliomas are the most common central nervous tumors in children and adolescents. However, spinal cord low-grade gliomas (sLGGs) are rare, with scarce information on tumor genomics and epigenomics. To define the molecular landscape of sLGGs, we integrated clinical data, histology, and multi-level genetic and epigenetic analyses on a consecutive cohort of 26 pediatric patients. Driver molecular alteration was found in 92% of patients (24/26). A novel variant of KIAA1549:BRAF fusion (ex10:ex9) was identified using RNA-seq in four cases. Importantly, only one-third of oncogenic drivers could be revealed using standard diagnostic methods, and two-thirds of pediatric patients with sLGGs required extensive molecular examination. The majority (23/24) of detected alterations were potentially druggable targets. Four patients in our cohort received targeted therapy with MEK or NTRK inhibitors. Three of those exhibited clinical improvement (two with trametinib, one with larotrectinib), and two patients achieved partial response. Methylation profiling was implemented to further refine the diagnosis and revealed intertumoral heterogeneity in sLGGs. Although 55% of tumors clustered with pilocytic astrocytoma, other rare entities were identified in this patient population. In particular, diffuse leptomeningeal glioneuronal tumors (n = 3) and high-grade astrocytoma with piloid features (n = 1) and pleomorphic xanthoastrocytoma (n = 1) were present. A proportion of tumors (14%) had no match with the current version of the classifier. Complex molecular genetic sLGGs characterization was invaluable to refine diagnosis, which has proven to be essential in such a rare tumor entity. Moreover, identifying a high proportion of drugable targets in sLGGs opened an opportunity for new treatment modalities.

• MeSH
• Astrocytoma * genetics   MeSH
• Child MeSH
• Genomics MeSH
• Glioma * genetics   pathology   MeSH
• Humans MeSH
• Mitogen-Activated Protein Kinase Kinases MeSH
• Adolescent MeSH
• Spinal Cord Neoplasms * genetics   MeSH
• Brain Neoplasms * genetics   MeSH
• Proto-Oncogene Proteins B-raf genetics   MeSH
• Check Tag
• Child MeSH
• Humans MeSH
• Adolescent MeSH
• Publication type
• Journal Article MeSH

BACKGROUND: Integration of multi-omics data can provide a more complex view of the biological system consisting of different interconnected molecular components, the crucial aspect for developing novel personalised therapeutic strategies for complex diseases. Various tools have been developed to integrate multi-omics data. However, an efficient multi-omics framework for regulatory network inference at the genome level that incorporates prior knowledge is still to emerge. RESULTS: We present IntOMICS, an efficient integrative framework based on Bayesian networks. IntOMICS systematically analyses gene expression, DNA methylation, copy number variation and biological prior knowledge to infer regulatory networks. IntOMICS complements the missing biological prior knowledge by so-called empirical biological knowledge, estimated from the available experimental data. Regulatory networks derived from IntOMICS provide deeper insights into the complex flow of genetic information on top of the increasing accuracy trend compared to a published algorithm designed exclusively for gene expression data. The ability to capture relevant crosstalks between multi-omics modalities is verified using known associations in microsatellite stable/instable colon cancer samples. Additionally, IntOMICS performance is compared with two algorithms for multi-omics regulatory network inference that can also incorporate prior knowledge in the inference framework. IntOMICS is also applied to detect potential predictive biomarkers in microsatellite stable stage III colon cancer samples. CONCLUSIONS: We provide IntOMICS, a framework for multi-omics data integration using a novel approach to biological knowledge discovery. IntOMICS is a powerful resource for exploratory systems biology and can provide valuable insights into the complex mechanisms of biological processes that have a vital role in personalised medicine.

• MeSH
• Algorithms MeSH
• Bayes Theorem MeSH
• Gene Regulatory Networks MeSH
• Humans MeSH
• Colonic Neoplasms * MeSH
• Systems Biology methods   MeSH
• DNA Copy Number Variations * MeSH
• Check Tag
• Humans MeSH
• Publication type
• Journal Article MeSH

INTRODUCTION: This study presents a novel molecular parameter potentially co-defining tumor biology-the total tumor suppressor gene (TSG) count at chromosomal loci harboring genes rearranged in fusion-defined tumors. It belongs to the family of molecular parameters created using a black-box approach. METHOD: It is based on a public curated Texas TSG database. Its data are regrouped based on individual genes loci using another public database (Genecards). The total TSG count for NTRK (NTRK1; OMIM: 191315; NTRK2; OMIM: 600456; NTRK3; OMIM: 191316), NRG1 (OMIM: 142445), and RET (OMIM: 164761) rearranged tumors in patients treated with a theranostic approach is calculated using the results of recently published studies. RESULTS: Altogether 138 loci containing at least three TSGs are identified. These include 21 "extremely hot" spots, with 10 to 28 TSGs mapping to a given locus. However, the study falls short of finding a correlation between tumor regression or patient survival and the TSG count owing to a low number of cases meeting the study criteria. CONCLUSION: The total TSG count alone cannot predict the biology of translocation-defined tumors. The addition of other parameters, including microsatellite instability (MSI), tumor mutation burden (TMB), homologous recombination repair deficiency (HRD), and copy number heterogeneity (CNH), might be helpful. Thus a multi-modal data integration is advocated. We believe that large scale studies should evaluate the significance and value of the total TSG count.

• MeSH
• Gene Fusion MeSH
• Genomics MeSH
• Humans MeSH
• Mutation MeSH
• Biomarkers, Tumor genetics   MeSH
• Neoplasms * genetics   pathology   MeSH
• Translocation, Genetic MeSH
• Check Tag
• Humans MeSH
• Publication type
• Journal Article MeSH

AIM: To describe clinical characteristics, procedural details, specific challenges, and outcomes in patients with HeartMate3TM (HM3), a left ventricular assist device system with a magnetically levitated pump, undergoing ventricular tachycardia ablation (VTA). METHODS AND RESULTS: Data were collected from patients with an HM3 system who underwent VTA in seven tertiary centres. Data included baseline patient characteristics, procedural data, mortality, and arrhythmia-free survival. The study cohort included 19 patients with cardiomyopathy presenting with ventricular tachycardia (VT) (53% with VT storm). Ventricular tachycardias were induced in 89% of patients and a total of 41 VTs were observed. Severe electromagnetic interference was present on the surface electrocardiogram. Hence, VT localization required analysis of intra-cardiac signals or the use of filter in the 40-20 Hz range. The large house pump HM3 design obscured the cannula inflow and therefore multi imaging modalities were necessary to avoid catheter entrapment in the cannula. A total of 32 VTs were mapped and were successfully ablated (31% to the anterior wall, 38% to the septum and only 9% to the inflow cannula region). Non-inducibility of any VT was reached in 11 patients (58%). Over a follow-up of 429 (interquartile range 101-692) days, 5 (26%) patients underwent a redo VT ablation due to recurrent VTA and 2 (11%) patients died. CONCLUSIONS: Ventricular tachycardia ablation in patients with HM3 is feasible and safe when done in the appropriate setup. Long-term arrhythmia-free survival is acceptable but not well predicted by non-inducibility at the end of the procedure.

• MeSH
• Cardiomyopathies * etiology   MeSH
• Catheter Ablation * adverse effects   methods   MeSH
• Tachycardia, Ventricular * diagnosis   surgery   MeSH
• Humans MeSH
• Heart-Assist Devices * MeSH
• Recurrence MeSH
• Treatment Outcome MeSH
• Check Tag
• Humans MeSH
• Publication type
• Journal Article MeSH

BACKGROUND: Radical nephroureterectomy (RNU) with the concomitant excision of the distal ureter and bladder cuff is the current standard of care for the treatment of muscle invasive and/or high-risk upper tract urothelial carcinoma (UTUC). In small uncontrolled studies, laparoscopic RNU has been suggested to be associated with better perioperative outcomes compared to open RNU. The aim of our study was to compare the perioperative oncological and functional outcomes of open RNU versus laparoscopic RNU after adjusting for preoperative baseline patient-related characteristics. METHODS: We evaluated a multi-institutional retrospective database composed by 1512 patients diagnosed with UTUC and treated with open or laparoscopic RNU between 1990 and 2016. Perioperative outcomes included operative time, blood loss, and length of hospital stay, as well as postoperative complications, readmission, reoperation, and mortality rates at 30 and 90 days from surgery. A 1:1 propensity score matching estimated using logistic regression with the teffects psmatch function of STATA 13® (caliper 0.2, no replacement; StataCorp LLC; College Station, TX, USA) was performed using preoperative parameters such as: age, gender, Body Mass Index (BMI), and American Society of Anesthesiologists (ASA) Score. RESULTS: Overall, 1007 (66.6%) patients were treated with open and 505 (33.4%) with laparoscopic RNU. Open RNU resulted into shorter median operative time (180 vs. 230 min, P<0.001) and longer median hospital stay (10 vs. 7 days, P<0.001) in comparison to laparoscopic RNU. No statistically significant difference was identified for the other variables of interest (all P>0.05). At multivariable linear regression after propensity score matching adjusted for lymph node dissection and year of surgery, laparoscopic RNU resulted in longer operative time (coefficient 43.6, 95% CI 27.9-59.3, P<0.001) and shorter hospital stay (coefficient -1.27, 95% CI -2.1 to -0.3, P=0.01) compared to open RNU, but the risk of other perioperative complications remained similar between the two treatments. CONCLUSIONS: Laparoscopic RNU is associated with shorter hospital stay, but longer operative time in comparison to open RNU. Otherwise, there were no differences in other perioperative outcomes between these surgical modalities even after propensity score matching. The choice to offer laparoscopic or open RNU in the treatment of UTUC should not be based on concerns of different safety outcomes.

• MeSH
• Carcinoma, Transitional Cell * MeSH
• Laparoscopy * adverse effects   methods   MeSH
• Humans MeSH
• Urinary Bladder Neoplasms * surgery   MeSH
• Ureteral Neoplasms * MeSH
• Nephroureterectomy methods   MeSH
• Retrospective Studies MeSH
• Propensity Score MeSH
• Check Tag
• Humans MeSH
• Publication type
• Journal Article MeSH