Summary Background The clinical outcome of extranodal natural killer T-cell lymphoma (ENKTL) has improved substantially as a result of new treatment strategies with non-anthracycline-based ...chemotherapies and upfront use of concurrent chemoradiotherapy or radiotherapy. A new prognostic model based on the outcomes obtained with these contemporary treatments was warranted. Methods We did a retrospective study of patients with newly diagnosed ENKTL without any previous treatment history for the disease who were given non-anthracycline-based chemotherapies with or without upfront concurrent chemoradiotherapy or radiotherapy with curative intent. A prognostic model to predict overall survival and progression-free survival on the basis of pretreatment clinical and laboratory characteristics was developed by filling a multivariable model on the basis of the dataset with complete data for the selected risk factors for an unbiased prediction model. The final model was applied to the patients who had complete data for the selected risk factors. We did a validation analysis of the prognostic model in an independent cohort. Findings We did multivariate analyses of 527 patients who were included from 38 hospitals in 11 countries in the training cohort. Analyses showed that age greater than 60 years, stage III or IV disease, distant lymph-node involvement, and non-nasal type disease were significantly associated with overall survival and progression-free survival. We used these data as the basis for the prognostic index of natural killer lymphoma (PINK), in which patients are stratified into low-risk (no risk factors), intermediate-risk (one risk factor), or high-risk (two or more risk factors) groups, which were associated with 3-year overall survival of 81% (95% CI 75–86), 62% (55–70), and 25% (20–34), respectively. In the 328 patients with data for Epstein-Barr virus DNA, a detectable viral DNA titre was an independent prognostic factor for overall survival. When these data were added to PINK as the basis for another prognostic index (PINK-E)—which had similar low-risk (zero or one risk factor), intermediate-risk (two risk factors), and high-risk (three or more risk factors) categories—significant associations with overall survival were noted (81% 95% CI 75–87%, 55% (44–66), and 28% (18–40%), respectively). These results were validated and confirmed in an independent cohort, although the PINK-E model was only significantly associated with the high-risk group compared with the low-risk group. Interpretation PINK and PINK-E are new prognostic models that can be used to develop risk-adapted treatment approaches for patients with ENKTL being treated in the contemporary era of non-anthracycline-based therapy. Funding Samsung Biomedical Research Institute.
The clinical outcome of advanced-stage Extranodal NK/T cell lymphoma (ENKTL) patients using conventional chemotherapy is extremely poor. The aim of this study was to investigate the outcomes of ...advanced-stage ENKTL patients treated with non-anthracycline-based chemotherapy followed by upfront autologous stem cell transplant (ASCT). From 8 institutions, 27 patients were recruited from February 2016 to May 2019. Patients were treated with 4 cycles of VIDL induction chemotherapy. Patients who achieved complete response (CR) or partial response (PR) underwent upfront ASCT. This study is registered with clinicaltrial.gov, # NCT02544425. Twenty patients (74.1%) completed 4 cycles of VIDL induction. The overall response rate of VIDL was 74.1%, including 17 (63.0%) with CR and 3 (11.1%) with PR. Primary toxicity of the induction regimen was grade 3 or 4 neutropenia, and no treatment-related mortality was reported. Seventeen patients proceeded with upfront ASCT, and 9 patients relapsed after ASCT, among whom, 4 was central nervous system (CNS) relapse. The median duration of response was 15.2 months (95% CI, 6.3-24.1 months). This study suggested that VIDL induction chemotherapy followed by upfront ASCT is feasible and effective for the treatment of advanced-stage ENKTL. However, CNS relapse prevention is needed in the treatment of advanced-stage ENKTL.
To apply emotion recognition and classification technology to the field of human-robot interaction, it is necessary to implement fast data processing and model weight reduction. This paper proposes a ...new photoplethysmogram (PPG) and galvanic skin response (GSR) signals-based labeling method using Asian multimodal data, a real-time emotion classification method, a 1d convolutional neural network autoencoder model, and a lightweight model obtained using knowledge distillation. In addition, the model performance was verified using the public DEAP dataset and the Asian multi-modal dataset 'MERTI-Apps'. For emotion classification, bio-signal data were window-sliced in 1-pulse units, and the label was reset to reflect the characteristics of the PPG and GSR signals. Simple data pre-processing, such as the prevention of loss and waveform duplication, was performed without using handcrafted features. The experiment showed that the accuracy of the proposed model using MERTI-Apps was 79.18% and 74.84% in the case of arousal and valence, respectively, for 3-class criteria, and the accuracy of the proposed model using DEAP was 81.33% and 80.25% in the case of arousal and valence, respectively, for 2-class criteria. The accuracy of the lightweight model was 77.87% and 73.49% in the case of arousal and valence, respectively, for 3-class criteria and its calculation time was reduced by more than 80% compared to the proposed 1d convolutional autoencoder model. We also confirmed that the proposed model improved computational time and accuracy compared to previous studies using MERTI-Apps and the lightweight model used in limited hardware environments enabled fast computation and real-time emotion classification.
Federated learning (FL) is a distributed machine learning paradigm in which multiple clients collaboratively learn a global model without sharing local training data. This study presents an adaptive ...model selection algorithm to enhance existing aggregation for FL. The proposed method uses sparsely learned local models and adaptively selects the best models based on the entropy-inspired information using the neural nodes of fully connected layers. This enables substantial nodes to be included in federated averaging (FedAvg), unlike standard FedAvg. In the case of a client model (at an embedded edge cluster), we exploited sparse learning to absorb more features associated with new classes by retraining weak neural nodes. For global model averaging, we used the proposed model selection technique to enhance the performance of traditional FedAvg by up to 20% in terms of global accuracy in the same number of rounds, which, in turn, has increased the convergence in federated rounds. For local sparse learning models, we simulated clients at edge clusters using the Docker container on embedded boards (NVIDIA Jetson Series). To ensure the efficacy of our proposed technique, we rigorously tested our algorithm on various datasets (CIFAR 10, MNIST, OCS lab(driving), UCI-HAR) using different models for image classification (EfficientNet, MobileNetv2, ResNet18) and time-series classification (FCN-LSTM, DepthConv-LSTM), and compared our algorithm with several state-of-the-art methods, such as FedAvg and FedFS. The proposed method outperformed the others in terms of generalizability and assists in converging earlier than existing algorithms in federated rounds.
Recently, the endothelial Activation and Stress Index (EASIX) score has been reported to predict overall survival (OS) after allogeneic stem cell transplantation. This study evaluated the prognostic ...role of EASIX score in patients with newly diagnosed multiple myeloma (MM).
This retrospective study analyzed the records of 1177 patients with newly diagnosed MM between February 2003 and December 2017 from three institutions in the Republic of Korea. Serum lactate dehydrogenase (LDH), creatinine, and platelet count at diagnosis were measured in all included patients. EASIX scores were calculated using the formula-LDH (U/L) × Creatinine (mg/dL) / platelet count (10
/L) and were evaluated based on log2 transformed values.
The median age of patients was 63 years (range, 22-92), and 495 patients (42.1%) underwent autologous stem cell transplantation (ASCT). The median log2 EASIX score at diagnosis was 1.1 (IQR 0.3-2.3). Using maximally selected log-rank statistics, the optimal EASIX cutoff value for OS was 1.87 on the log2 scale (95% CI 0.562-0.619, p < 0.001). After median follow-up for 50.0 months (range, 0.3-184.1), the median OS was 58.2 months (95% CI 53.644-62.674). Overall, 372 patients (31.6%) showed high EASIX scores at diagnosis, and had significantly inferior OS compared to those with low EASIX (log2 EASIX ≤1.87) (39.1 months vs. 67.2 months, p < 0.001). In multivariate Cox analysis, high EASIX was significantly associated with poor OS (HR 1.444, 95% CI 1.170-1.780, p = 0.001). In the subgroup analysis of patients who underwent ASCT, patients with high EASIX showed significantly inferior OS compared to those with low EASIX (52.8 months vs. 87.0 months, p < 0.001). In addition, in each group of ISS I, II, and III, high EASIX was associated with significantly inferior OS (ISS 1, 45.2 months vs. 76.0 months, p = 0.001; ISS 2, 42.3 months vs. 66.5 months, p = 0.002; ISS 3, 36.8 months vs. 55.1 months, p = 0.001).
EASIX score at diagnosis is a simple and strong predictor for OS in patients with newly diagnosed MM.
Celotno besedilo
Dostopno za:
DOBA, IZUM, KILJ, NUK, PILJ, PNG, SAZU, SIK, UILJ, UKNU, UL, UM, UPUK
This study aimed to evaluate the prognostic significance of the revised European LeukemiaNet (ELN)-2022 risk stratification model for 123 elderly acute myeloid leukemia (AML) patients treated with ...decitabine chemotherapy.
Based on the ELN-2022 risk stratification, 15 (12.2%), 51 (41.5%), and 57 (46.3%) patients were classified as having favorable, intermediate, and high-risk AML, respectively. In comparison with the ELN-2017 risk stratification, the ELN-2022 risk stratification re-assigned 26 (21.1%) and three (2.4%) patients to the adverse and favorable risk groups, respectively. Survival analysis revealed distinctive overall survival (OS) outcomes among the ELN-2022 risk groups (6-month OS rate: 73.3%, 52.9%, and 47.7% for favorable, intermediate, and adverse risk, respectively;
= 0.101), with a parallel trend observed in the event-free survival (EFS) (6-month EFS rate: 73.3%, 52.9%, and 45.6% for favorable, intermediate, and adverse risk, respectively;
= 0.049). Notably, both OS and EFS in the favorable risk group were significantly superior in comparison to that of the adverse risk group (OS:
= 0.040, EFS:
= 0.030). Although the ELN-2022 C-index (0.559) was greater than the ELN-2017 C-index (0.539), the result was not statistically significant (
= 0.059). Based on the event net reclassification index, we consistently observed significant improvements in the ELN-2022 risk stratification for overall survival (0.21 at 6 months).
In conclusion, the revised ELN-2022 risk stratification model may have improved the risk classification of elderly AML patients treated with hypomethylating agents compared to the ELN-2017 risk stratification model.
Celotno besedilo
Dostopno za:
DOBA, IZUM, KILJ, NUK, PILJ, PNG, SAZU, UILJ, UKNU, UL, UM, UPUK
Potential synergism between Bruton's tyrosine kinase (BTK) inhibitor and lenalidomide in treating aggressive B-cell lymphoma has been suggested. Here, the authors report a single-arm phase II ...clinical trial of combination of acalabrutinib, lenalidomide and rituximab (R2A) in patients with aggressive relapsed/refractory aggressive (R/R) B-cell non-Hodgkin lymphoma (NHL). The primary endpoint of this study is objective response rate (ORR), and the secondary endpoints are complete remission (CR) rate, duration of response (DoR), progression-free survival (PFS) and overall survival (OS). A total of 66 patients are enrolled mostly with diffuse large B-cell lymphoma. The ORR is 54.5% and CR rate is 31.8% meeting the primary end point. The median DoR is 12.9 months, and 1-year PFS and OS rate is 33.1% and 67.5% respectively. Adverse events (AE) are manageable with the most frequent AE being neutropenia (31.8%). Patients with MYD88 mutations, subtypes known for NF-κB activation, and high BTK expression by immunohistochemistry respond well. Overall, these results show a significant efficacy of the R2A regimen in patients with aggressive R/R B-cell NHL, with exploratory biomarkers suggesting potential associations with response. (ClinicalTrials.gov 51 identifier: NCT04094142).
Emotion recognition is a very important technique for ultimate interactions between human beings and artificial intelligence systems. For effective emotion recognition in a continuous-time domain, ...this article presents a multimodal fusion network which integrates video modality and electroencephalogram (EEG) modality networks. To calculate the attention weights of facial video features and the corresponding EEG features in fusion, a multimodal attention network, that is utilizing bilinear pooling based on low-rank decomposition, is proposed. Finally, continuous domain valence values are computed by using two modality network outputs and attention weights. Experimental results show that the proposed fusion network provides an improved performance of about 6.9% over the video modality network for the MAHNOB human computer interface (MAHNOB-HCI) dataset. Also, we achieved the performance improvement even for our proprietary dataset.
Key-value stores based on a log-structured merge (LSM) tree have emerged in big data systems because of their scalability and reliability. An LSM-tree offers a multilevel data structure with a simple ...interface. However, it performs file rewrites at the disk level, which causes write amplification. This study is concerned with this problem in relation to an embedded board environment, which can be used in edge computing. Addressing the major problems associated with an LSM-tree, we propose a new key-value store named CaseDB, which aggressively separates keys and bloom filters on the non-volatile memory express (NVMe) drive and stores the values on the SSD. Our solution reduces the I/O cost and enhances the overall performance in a cost-efficient manner. CaseDB employs a memory component, CBuffer, to avoid small write operations, and a delayed value compaction technique that guarantees the sorted order for both keys and values. CaseDB also utilizes deduction-based data deduplication to prevent space amplification in the values layer. The experiments show that CaseDB outperforms LevelDB and WiscKey 5.7 and 1.8 times, respectively, with respect to data writes, and additionally improves the read performance by 1.5 times. CaseDB also avoids the space amplification of WiscKey.