Novel machine learning method allerStat identifies statistically significant allergen-specific patterns in protein sequences.

Cutting-edge technologies such as genome editing and synthetic biology allow us to produce novel foods and functional proteins. However, their toxicity and allergenicity must be accurately evaluated. It is known that specific amino acid sequences in proteins make some proteins allergic, but many of these sequences remain uncharacterized. In this study, we introduce a data-driven approach and a machine-learning method to find undiscovered allergen-specific patterns (ASPs) among amino acid sequences. The proposed method enables an exhaustive search for amino acid subsequences whose frequencies are statistically significantly higher in allergenic proteins. As a proof-of-concept, we created a database containing 21,154 proteins of which the presence or absence of allergic reactions are already known and applied the proposed method to the database. The detected ASPs in this proof-of-concept study were consistent with known biological findings, and the allergenicity prediction performance using the detected ASPs was higher than extant approaches, indicating this method may be useful in evaluating the utility of synthetic foods and proteins.

Effects of rewarming therapies on outcomes in accidental hypothermia: A secondary analysis of a multicenter prospective study.

BACKGROUND: Rewarming therapies for accidental hypothermia (AH) include extracorporeal membrane oxygenation (ECMO) and non-ECMO related (conventional) therapies. However, there are limited data available to inform the selection of conventional rewarming therapy. The aim of the present study was to explore what patients' factors and which rewarming therapy predicted favorable prognosis. METHODS: This study is a secondary analysis of the Intensive Care with Extra Corporeal membrane oxygenation Rewarming in Accidentally Severe Hypothermia (ICE-CRASH) study, a multicenter prospective, observational study conducted in Japan. Enrolled in the ICE-CRASH study were patients aged ≥18 years with a core temperature of ≤32 °C who were transported to the emergency departments of 36 tertiary care hospitals in Japan between 1 December 2019 and 31 March 2022, among whom those who were rewarmed with conventional rewarming therapy were included in the present study. Logistic regression analysis was performed with 28-day survival as the objective variable; and seven factors including age, activities of daily living (ADL) independence, sequential organ failure assessment (SOFA) score, and each rewarming technique as explanatory variables. We performed linear regression analysis to identify whether each rewarming technique was associated with rewarming rate. RESULTS: Of the 499 patients enrolled in the ICE-CRASH study, 371 were eligible for this secondary analysis. The median age was 81 years, 50.9% were male, and the median initial body temperature was 28.8 °C. Age (odds ratio [OR]: 0.97, 95% confidence interval [CI]: 0.94-1.00) and SOFA score (OR: 0.73, 95% CI: 0.67-0.81) were associated with lower survival, whereas ADL independence (OR: 2.31, 95% CI: 1.15-4.63) was associated with higher survival. No conventional rewarming therapy was associated with 28-day survival. Hot bath was associated with a high rewarming rate (regression coefficient: 1.14, 95% CI: 0.75-1.53). CONCLUSION: No conventional rewarming therapy was associated with improved 28-day survival, which suggests that background factors such as age, ADL, and severity of condition contribute more to prognosis than does the selection of rewarming technique.

Generalized Low-Rank Update: Model Parameter Bounds for Low-Rank Training Data Modifications.

In this study, we have developed an incremental machine learning (ML) method that efficiently obtains the optimal model when a small number of instances or features are added or removed. This problem holds practical importance in model selection, such as cross-validation (CV) and feature selection. Among the class of ML methods known as linear estimators, there exists an efficient model update framework, the low-rank update, that can effectively handle changes in a small number of rows and columns within the data matrix. However, for ML methods beyond linear estimators, there is currently no comprehensive framework available to obtain knowledge about the updated solution within a specific computational complexity. In light of this, our study introduces a the generalized low-rank update (GLRU) method, which extends the low-rank update framework of linear estimators to ML methods formulated as a certain class of regularized empirical risk minimization, including commonly used methods such as support vector machines and logistic regression. The proposed GLRU method not only expands the range of its applicability but also provides information about the updated solutions with a computational complexity proportional to the number of data set changes. To demonstrate the effectiveness of the GLRU method, we conduct experiments showcasing its efficiency in performing cross-validation and feature selection compared to other baseline methods.

Early Initiation of Venovenous Extracorporeal Membrane Oxygenation for Critically Ill COVID-19 Patients.

The optimal timing for initiating extracorporeal membrane oxygenation (ECMO) after starting mechanical ventilation has yet to be clarified. We report herein the cases of two patients with coronavirus disease 2019 (COVID-19) acute respiratory distress syndrome (ARDS) who were successfully managed with an early ECMO induction strategy. Case 1 involved a 64-year-old man admitted in respiratory distress with polymerase chain reaction-confirmed COVID-19. On day 5 at hospital, he was intubated, but oxygenation remained unimproved despite mechanical ventilation treatment with high positive end-expiratory pressure (PEEP) (PaO2/FiO2 [P/F] ratio, 127; Respiratory ECMO Survival Prediction [RESP] score, 4). ECMO was initiated 4 hours after intubation, and stopped on day 16 at hospital. The patient was discharged from hospital on day 36. Case 2 involved a 49-year-old man who had been admitted 8 days prior. He was intubated on hospital on day 2. High PEEP mechanical ventilation did not improve oxygenation (P/F ratio, 93; RESP score, 7). ECMO was stopped on hospital on day 7 and he was discharged from hospital on day 21. The strategy of early initiation of ECMO in these two cases may have minimized the risk of ventilation-related lung injury and contributed to the achievement of favorable outcomes.

Rural-Urban Disparity in Emergency Care for Acute Myocardial Infarction in Japan.

BACKGROUND: There are few reports examining regional differences between rural prefectures and metropolitan areas in the management of acute myocardial infarction (AMI) in Japan.Methods and Results:In the Rural AMI registry, a prospective, multi-prefectural registry of AMI in 4 rural prefectures (Ishikawa, Aomori, Ehime and Mie), a total of 1,695 consecutive AMI patients were registered in 2013. Among them, 1,313 patients who underwent primary percutaneous coronary intervention (PPCI) within 24 h of onset were enrolled in this study (Rural group), and compared with the cohort data from the Tokyo CCU Network registry for AMI in the same period (Metropolitan group, 2,075 patients). The prevalence of direct ambulance transport to PCI-capable facilities in the Rural group was significantly lower than that in the Metropolitan group (43.8% vs. 60.3%, P<0.01), which resulted in a longer onset-to-balloon time (OTB: 225 vs. 210 min, P=0.02) and lower prevalence of PPCI in a timely fashion (OTB ≤2 h: 11.5% vs. 20.7%, P<0.01) in the Rural group. Multivariate analysis revealed that direct ambulance transport was the strongest predictor for PPCI in a timely fashion (odds ratio=4.13, P<0.001). CONCLUSIONS: AMI patients in rural areas were less likely to be transported directly to PCI-capable facilities, resulting in time delay to PPCI compared with those in metropolitan areas.

Inspection of Perirectal Lymph Nodes by One-Step Nucleic Acid Amplification Predicts Lateral Lymph Node Metastasis in Advanced Rectal Cancer.

BACKGROUND: Lateral lymph node dissection (LLND) is performed for advanced rectal cancers in Japan; however, it can cause sexual and urinary dysfunction. The incidence of lateral LN metastasis is estimated at 7-13.9%; therefore, excessive rectal surgery with LLND should be avoided, especially for prophylactic purposes. To identify the patients who require LLND, we examined metastases in perirectal LNs by using a one-step nucleic acid amplification (OSNA) assay to predict lateral LN metastases. METHODS: Twenty-five patients who underwent surgery with bilateral LN dissection due to T3-T4 rectal cancers were prospectively included in this study. Twenty-two patients (88.0%) received preoperative chemotherapy. Among 1052 LNs from 25 patients (median 40 per case), 135 perirectal LNs (median 6 per patient) were divided into three pieces and analyzed by OSNA, reverse transcriptase-polymerase chain reaction for carcinoembryonic antigen mRNA, and pathological examination after surgery. These results were compared with the pathological diagnosis of lateral LNs. RESULTS: Lateral LN metastases were present in 4 of 25 patients (16.0%). All of these patients were positive by OSNA for perirectal LN metastases. The OSNA assay had a sensitivity of 100%, specificity of 86%, positive predictive value of 57%, and negative predictive value (NPV) of 100% for predicting lateral LN metastases. CONCLUSIONS: The findings from this prospective study suggest that the OSNA assay of perirectal LNs may be useful for determining when LLND is necessary because of its high NPV, even in patients treated with preoperative chemotherapy.

Plasma soluble lectin-like oxidized low-density lipoprotein receptor-1 as a novel prognostic biomarker in patients with ST-segment elevation acute myocardial infarction.

BACKGROUND: Soluble lectin-like oxidized low-density lipoprotein receptor-1 (sLOX-1) level is a reliable prognostic biomarker in acute coronary syndrome. However, it is unclear whether its plasma level at acute phase is related to the long-term prognosis in patients with ST-segment elevation acute myocardial infarction (STEMI). METHODS AND RESULTS: We prospectively examined the relation between plasma sLOX-1 level on admission and prognosis in 153 consecutive STEMI patients admitted within 24 h of onset. Primary percutaneous coronary intervention was performed in 144 patients. The patients were divided into 2 groups by the median value (71 pg/ml) of plasma sLOX-1 level on admission [sLOX-1 level ≤71 pg/ml (n=77) and >71 pg/ml (n=76)], and were followed for median of 1,156 days. All-cause mortality and the combined endpoints of major adverse cardiovascular events (MACE) defined as cardiovascular mortality and recurrent MI were both significantly higher in patients with sLOX-1 values above median than in those below median (25.0% vs. 3.9%, P<0.001, and 19.4% vs. 6.5%, P=0.019 by log-rank test, respectively). Even after adjustment for confounders, a level of sLOX-1 above median was an independent predictor for all-cause mortality (hazard ratio (HR): 5.893; 95% confidence interval (CI): 1.665-20.854, P=0.006) and MACE (HR: 3.457; 95% CI: 1.164-10.270, P=0.030). CONCLUSIONS: Elevated plasma sLOX-1 level on admission independently predicts long-term all-cause mortality and MACE after STEMI.

Imidapril and enalapril similarly inhibit plasma matrix metalloproteinase activities and attenuate left ventricular remodeling in patients with acute myocardial infarction.

Matrix metalloproteinase (MMP) plays a critical role in the development of ventricular remodeling after acute myocardial infarction (AMI). Imidapril, an angiotensin-converting enzyme inhibitor, has been shown to inhibit MMP activity. We investigated whether imidapril inhibits plasma MMP activities and attenuates ventricular remodeling in patients with AMI in comparison with enalapril. We enrolled 70 patients with AMI. All patients underwent primary percutaneous coronary intervention and were randomly assigned either to imidapril (n = 35) or to enalapril (n = 35) treatment. Left ventriculography was performed in acute (day 14) and chronic (6 months) phases, and plasma MMP-2 and MMP-9 activities were measured by zymography. Any changes in left ventricular end-diastolic volume index and ejection fraction from acute to chronic phases did not differ between the 2 groups. The plasma MMP-2 and MMP-9 activities at day 14 were both significantly decreased compared with those at day 1 in both groups (all P < 0.05). At 6 months, MMP-9 activity still remained decreased in both groups (P < 0.05 vs. day 1). Overall, there were no differences between the 2 groups both in plasma MMP-2 and MMP-9 activities. These results demonstrate that imidapril exerts inhibitory effects on plasma MMP activities and attenuates left ventricular remodeling in patients with AMI similar to enalapril.

Multimodal Gated Mixture of Experts Using Whole Slide Image and Flow Cytometry for Multiple Instance Learning Classification of Lymphoma.

In this study, we present a deep-learning-based multimodal classification method for lymphoma diagnosis in digital pathology, which utilizes a whole slide image (WSI) as the primary image data and flow cytometry (FCM) data as auxiliary information. In pathological diagnosis of malignant lymphoma, FCM serves as valuable auxiliary information during the diagnosis process, offering useful insights into predicting the major class (superclass) of subtypes. By incorporating both images and FCM data into the classification process, we can develop a method that mimics the diagnostic process of pathologists, enhancing the explainability. In order to incorporate the hierarchical structure between superclasses and their subclasses, the proposed method utilizes a network structure that effectively combines the mixture of experts (MoE) and multiple instance learning (MIL) techniques, where MIL is widely recognized for its effectiveness in handling WSIs in digital pathology. The MoE network in the proposed method consists of a gating network for superclass classification and multiple expert networks for (sub)class classification, specialized for each superclass. To evaluate the effectiveness of our method, we conducted experiments involving a six-class classification task using 600 lymphoma cases. The proposed method achieved a classification accuracy of 72.3%, surpassing the 69.5% obtained through the straightforward combination of FCM and images, as well as the 70.2% achieved by the method using only images. Moreover, the combination of multiple weights in the MoE and MIL allows for the visualization of specific cellular and tumor regions, resulting in a highly explanatory model that cannot be attained with conventional methods. It is anticipated that by targeting a larger number of classes and increasing the number of expert networks, the proposed method could be effectively applied to the real problem of lymphoma diagnosis.

Severe thrombocytopenia caused by cytomegalovirus infection in an immunocompetent adult: A case report.

A 53-year-old otherwise healthy man was referred to our hospital with a fever of unknown origin, headache, and arthralgia. Four days earlier, he had a fever with chills. Treatment with antibiotics and acetaminophen proved ineffective, with the patient subsequently developing headache and joint pain. Blood analysis revealed elevated inflammatory markers, liver impairment, and severe thrombocytopenia (platelet count, 19,000/µL). Subsequent tests revealed elevated levels of anti-cytomegalovirus IgM and IgG. Based on these findings, the patient was diagnosed with severe thrombocytopenia associated with cytomegalovirus infection. Platelet counts increased spontaneously without antiviral therapy. Forty-five days after the initial visit, the symptoms improved, and blood tests revealed resolution of the inflammatory findings, with the platelet count recovering to 155,000/µL. Although the disease may resolve spontaneously, cytomegalovirus infection should be considered as a differential diagnosis in case of severe thrombocytopenia in immunocompetent adults.

Safe RuleFit: Learning Optimal Sparse Rule Model by Meta Safe Screening.

We consider the problem of learning a sparse rule model, a prediction model in the form of a sparse linear combination of rules, where a rule is an indicator function defined over a hyper-rectangle in the input space. Since the number of all possible such rules is extremely large, it has been computationally intractable to select the optimal set of active rules. In this paper, to solve this difficulty for learning the optimal sparse rule model, we propose Safe RuleFit (SRF). Our basic idea is to develop meta safe screening (mSS), which is a non-trivial extension of well-known safe screening (SS) techniques. While SS is used for screening out one feature, mSS can be used for screening out multiple features by exploiting the inclusion-relations of hyper-rectangles in the input space. SRF provides a general framework for fitting sparse rule models for regression and classification, and it can be extended to handle more general sparse regularizations such as group regularization. We demonstrate the advantages of SRF through intensive numerical experiments.

Efficient model selection for predictive pattern mining model by safe pattern pruning.

Predictive pattern mining is an approach used to construct prediction models when the input is represented by structured data, such as sets, graphs, and sequences. The main idea behind predictive pattern mining is to build a prediction model by considering unified inconsistent notation sub-structures, such as subsets, subgraphs, and subsequences (referred to as patterns), present in the structured data as features of the model. The primary challenge in predictive pattern mining lies in the exponential growth of the number of patterns with the complexity of the structured data. In this study, we propose the safe pattern pruning method to address the explosion of pattern numbers in predictive pattern mining. We also discuss how it can be effectively employed throughout the entire model building process in practical data analysis. To demonstrate the effectiveness of the proposed method, we conduct numerical experiments on regression and classification problems involving sets, graphs, and sequences.

Development of an Automatic Chronological Record Creation System Using Voice AI to Facilitate Information Aggregation and Sharing in the Event of a Disaster.

OBJECTIVE: In the event of a disaster, the chain of command and communication of each relevant agency is important. In this study, a chronological record creation system using voice AI (V-CRS) was developed, and an experiment was conducted to determine whether the obtained information could be quickly and easily summarized in chronological order. METHODS: After a lecture by Japanese Disaster Medical Assistant (DMAT) Team members and 8 medical clerks on how to use the developed tool, a comparison experiment was conducted between manual input and V-CRS utilization of the time to compile disaster information. RESULTS: Results proved that V-CRS can collect information gathered at headquarters more quickly than handwriting. It was also suggested that even medical clerks who have never been trained to record information during disasters could record information at the same speed as trained DMAT personnel. CONCLUSION: V-CRS can transcribe audio information even in situations where technical terms and physical units must be recorded, such as radiation disasters. It has been proven that anyone can quickly organize information using this method, to some extent.

Are Online Simulations for Radiation Emergency Medical Preparedness Less Effective in Teaching Than Face-to-Face Simulations?

OBJECTIVE: The study clarified differences in understanding and satisfaction between face-to-face and online training on radiation emergency medical preparedness (REMP) training. METHODS: The training was held at Hirosaki University between 2018 and 2022, with 46 face-to-face participants and 25 online participants. RESULTS: Face-to-face training was significantly more understandable than online for the use of the Geiger counter (P < 0.05), but the educational effect of virtual reality (VR) was not significantly different from the actual practice. For the team exercise of taking care of the victims, online resulted in a significantly higher understanding (P < 0.05). CONCLUSIONS: Interactive exercises can be done online with equipment sent to learners, and VR is also as effective. The use of videos was more effective for first-timers to learn the practical process from a bird's-eye view, especially for team-based medical procedures.

REGIV as a potential biomarker for peritoneal dissemination in gastric adenocarcinoma.

BACKGROUND: This study examined the clinical significance of regenerating islet-derived family member 4 (REGIV) in surgically resected gastric tumors. The potential of REGIV as a biomarker in gastric cancer was also assessed including its predictive value for prognosis and recurrence after surgery. METHODS: Immunohistochemistry was performed to assess the clinical significance of REGIV expression status in surgically resected specimens. The quantitative genetic diagnostic method, transcription-reverse transcription concerted reaction (TRC) that targeted REGIV mRNA was applied for prediction of peritoneal recurrence in gastric cancer. RESULTS: Positive immunostaining for REGIV was observed in 85 cases (52.5%), and correlated significantly with diffuse type histopathology (P = 0.001), advanced T stage (P = 0.022), and frequent peritoneal recurrence (P = 0.009). Multivariate analysis identified advanced T stage (P < 0.001) and REGIV expression (P = 0.034) as independent prognostic factors for peritoneal recurrence-free survival. Overexpression of REGIV protein was evident in the majority of peritoneal tumors (93.8%). REGIV mRNA assessed by TRC could be a predictive marker for peritoneal recurrence after curative operation. CONCLUSIONS: REGIV overexpression is common in primary gastric tumors and a potentially suitable marker of diffuse type histopathology and peritoneal dissemination. Overexpression of REGIV mRNA, assessed by the TRC method, is a potentially suitable marker of peritoneal recurrence after curative resection.

Japanese medical students' awareness of cardiopulmonary resuscitation in the context of the COVID-19 pandemic.

Aim: To evaluate Japanese medical students' awareness of newly recommended cardiopulmonary resuscitation (CPR) and airway management procedures in the context of the coronavirus disease (COVID-19) pandemic. Methods: An online survey was sent in December 2020 to all medical students at Hirosaki University in Japan. The survey included 15 questions and quizzes regarding prior experience of learning the new CPR guidelines in response to COVID-19, knowledge of conventional CPR, and COVID-19 context CPR and airway management procedures. Results: Of all medical students at the university, 457 (57.1%) responded to the survey. Among these, 22% reported that they were knowledgeable about CPR procedure in the COVID-19 pandemic setting. Prior knowledge of CPR in the context of COVID-19 was a significant positive predictor of quiz score regarding the CPR procedure (ß = 0.60, P < 0.01) and the airway management procedure (ß = 0.34, P = 0.02) in the context of the COVID-19 pandemic. Conclusions: Medical students with experience learning the new COVID-19 context CPR guidelines had sufficient knowledge of CPR and advanced airway management procedures in the setting of the COVID-19 pandemic. Implementation of a formal medical education curriculum based on the newly recommended CPR and advanced life support guidelines is needed to improve medical students' awareness and skills of CPR and airway management in the context of the COVID-19 pandemic.

Wood-Splitter-Related Upper-Limb Injuries: A Single-Centered Case-Series Study.

(1) Background: Injuries to the upper limbs during wood splitting can affect social and economic life. We aimed to describe the clinical information concerning these injuries in Japan. (2) Methods: We identified patients from our patient database from April 2015-November 2021 and extracted data from their medical records, which includes age, gender, occupation, month, time and location of the injury, diagnosis, duration of hospitalization, ICU admission, treatment interventions including surgery, outcome, and medical costs. (3) Result: Seventeen cases were identified. Most of the patients were male (n = 15), with median age being 68 years old. Regarding the patients' backgrounds, six were apple farmers and three were unemployed. Injuries to the index finger was most common (n = 9), followed by injuries to the thumb in five cases (n = 5). Most of the incidents occurred at home or on the patient's farm estate. No injuries were due to incidents at work. (4) Conclusion: The wood splitter-related injuries required long-term treatment and frequently damaged the thumb, a functionally important digit. All the injuries were sustained during non-occupational use of a wood splitter. Therefore, we suggest that safety training should be provided to prevent traumatic injuries when these products are being sold.