Evolutionary Model for Brain Cancer-Grading and Classification

IEEE Access, Journal Year: 2023, Volume and Issue: 11, P. 126182 - 126194

Published: Jan. 1, 2023

Brain cancer is a bad disease and affects millions of people in worldwide. Approximately 70% patients diagnosed with this do not survive. The Machine learning promising recent development area. However, very limited research performed direction. Therefore, research, we propose an evolutionary lightweight model aimed at detecting brain classification, starting from the analysis magnetic resonance images. proposed named ensemble combines (weighted average multiple XGBoost decision trees) modified version Multimodal Lightweight XGBoost. Herein, provide prediction explain ability by considering preprocessing Magnetic Resonance Imaging (MRI) data feature extraction (Intensity, texture, shape). process involves various step - first, prepare data, extract important features, finally, merge together using special kind classification called classification. We evaluate our BraTS 2020 dataset. dataset consists 285 MRI scans gliomas. simulation results showed that achieved 93.0% accuracy, 0.94 precision, 0.93 recall, F1 score, area under Receiver Operating Characteristic Curve (AUC-ROC) value 0.984. efficient demonstrate effectiveness for tumor grading four grades. show potential approach as valuable tool early diagnosis effective treatment planning tumors. Finally, holds promise aiding treatment.

Language: Английский

---

Multimodal deep learning for predicting in-hospital mortality in heart failure patients using longitudinal chest X-rays and electronic health records

The International Journal of Cardiovascular Imaging, Journal Year: 2025, Volume and Issue: 41(3), P. 427 - 440

Published: Jan. 9, 2025

Language: Английский

---

Risk factors for the in‑hospital and 1‑year mortality of elderly patients hospitalized due to COVID‑19‑related pneumonia

Experimental and Therapeutic Medicine, Journal Year: 2023, Volume and Issue: 27(1)

Published: Nov. 20, 2023

Coronavirus disease 2019 (COVID-19) is characterized by poor outcomes and a high mortality rate, particularly among elderly patients. Since the beginning of pandemic, an older age has been recognized as critical risk factor for severity, with increasing rates in each decade life. This phenomenon may be consequence previous health status, higher prevalence pre-existing comorbidities degree frailty. The majority studies on factors patients refer to first waves pandemic predictors in-hospital these aim present study was provide detailed description clinical characteristics management cohort (≥65 years age) who were hospitalized COVID-19-related pneumonia all phases presenting their outcomes, investigating out-of-hospital over period 1 year this vulnerable population. A total 1,124 (603 males, 53.7%) mean 78.51±7.42 median Charlson comorbidity index (CCI) 5 included study. Of patients, 104 (9.3%) during original strain Wuhan, 385 (34.3%) Alpha variant, 221 (19.7%) Delta 414 (36.8%) Omicron variant. Overall, rate 33.4% (375 patients), 1-year 44.7% (502 patients). had not vaccinated or completed full vaccination against severe acute respiratory syndrome coronavirus-2 (843 75%), given infection. Age, immature granulocytes, lactate dehydrogenase (LDH) levels, ferritin chest X-ray score, well absence vaccination, cough fatigue, statistically significantly independently associated mortality, while age, LDH alanine aminotransferase CCI, history dementia mortality. On whole, demonstrates that both due are high.

Language: Английский

---

Predicting omicron pneumonia severity and outcome: a single-center study in Hangzhou, China

Frontiers in Medicine, Journal Year: 2023, Volume and Issue: 10

Published: May 26, 2023

Background In December 2022, there was a large Omicron epidemic in Hangzhou, China. Many people were diagnosed with pneumonia variable symptom severity and outcome. Computed tomography (CT) imaging has been proven to be an important tool for COVID-19 screening quantification. We hypothesized that CT-based machine learning algorithms can predict disease outcome pneumonia, we compared its performance the index (PSI)-related clinical biological features. Methods Our study included 238 patients variant who have admitted our hospital China from 15 2022 16 January 2023 (the first wave after dynamic zero-COVID strategy stopped). All had positive real-time polymerase chain reaction (PCR) or lateral flow antigen test SARS-CoV-2 vaccination no previous infections. recorded patient baseline information pertaining demographics, comorbid conditions, vital signs, available laboratory data. CT images processed commercial artificial intelligence (AI) algorithm obtain volume percentage of consolidation infiltration related pneumonia. The support vector (SVM) model used Results receiver operating characteristic (ROC) area under curve (AUC) classifier using PSI-related features 0.85 (accuracy = 87.40%, p < 0.001) predicting while only 0.70 76.47%, 0.014). If combined, AUC not increased, showing 0.84 84.03%, 0.001). Trained on prediction, reached 85.29%, 0.001), which higher than (AUC 0.67, accuracy 75.21%, integrated showed slightly 0.86 86.13%, Oxygen saturation, IL-6, great importance both Conclusion provided comprehensive analysis comparison between chest assessment prediction predictive accurately predicts infection. found biomarkers. This approach potential provide frontline physicians objective manage more effectively time-sensitive, stressful, potentially resource-constrained environments.

Language: Английский

---

Enhancing pneumonia detection with masked neural networks: a deep learning approach

Neural Computing and Applications, Journal Year: 2024, Volume and Issue: 36(29), P. 18433 - 18444

Published: July 28, 2024

Language: Английский

---