Background/Objectives: Cystoscopy is necessary for diagnosing bladder cancer, but it has limitations in identifying ambiguous lesions, such as carcinoma in situ (CIS), which leads to a high recurrence rate of bladder cancer. With the significant advancements in deep learning in the medical field, several studies have explored its application in cystoscopy. This study aimed to utilize the VGG19 and Deeplab v3+ deep learning models to classify and segment cystoscope images, respectively. Methods: We classified cystoscope images obtained from 772 patients based on morphology (normal, papillary, flat, mixed) and biopsy results (normal, Ta, T1, T2, CIS, etc.). Experienced urologists annotated and labeled the lesion areas and image categories. The classification model for bladder cancer lesion, annotated with pathological results, was developed using VGG19 with an additional fully connected layer, utilizing sparse categorical cross-entropy as the loss function. The Deeplab v3+ model was used for segmenting each morphological type of bladder cancer in the cystoscope images, employing the dice coefficient loss function. The classification model was evaluated using validation accuracy and correlation with biopsy results, while the segmentation model was assessed using the Intersection over Union (IoU) combined with binary accuracy. Results: The dataset was split into training and validation sets with a 4:1 ratio. The VGG19 classification model achieved an accuracy score of 0.912. The Deeplab v3+ segmentation model achieved an IoU of 0.833 and a binary accuracy of 0.951. Visual analysis revealed a high similarity between the lesions identified by Deeplab v3+ and those labeled by experts. Conclusions: In this study, we applied two deep learning models using well-annotated datasets of cystoscopic images. Both VGG19 and Deeplab v3+ demonstrated high performance in classification and segmentation, respectively. These models can serve as valuable tools for bladder cancer research and may aid in the diagnosis of bladder cancer.

Respiratory diseases significantly affect respiratory function, making them a considerable contributor to global mortality. The respiratory muscles play an important role in disease prognosis; as such, quantitative analysis of the respiratory muscles is crucial to assess the status of the respiratory system and the quality of life in patients. In this study, we aimed to develop an automated approach for the segmentation and classification of three types of respiratory muscles from computed tomography (CT) images using artificial intelligence. With a dataset of approximately 600,000 thoracic CT images from 3,200 individuals, we trained the model using the Attention U-Net architecture, optimized for detailed and focused segmentation. Subsequently, we calculated the volumes and densities from the muscle masks segmented by our model and performed correlation analysis with pulmonary function test (PFT) parameters. The segmentation models for muscle tissue and respiratory muscles obtained dice scores of 0.9823 and 0.9688, respectively. The classification model, achieving a generalized dice score of 0.9900, also demonstrated high accuracy in classifying thoracic region muscle types, as evidenced by its F1 scores: 0.9793 for the pectoralis muscle, 0.9975 for the erector spinae muscle, and 0.9839 for the intercostal muscle. In the correlation analysis, the volume of the respiratory muscles showed a strong correlation with PFT parameters, suggesting that respiratory muscle volume may serve as a potential novel biomarker for respiratory function. Although muscle density showed a weaker correlation with the PFT parameters, it has a potential significance in medical research.

BACKGROUND Sepsis and mortality are the superior cause of death in intensive care unit patients. The ability to predict the onset of sepsis and mortality is of paramount importance for guiding clinical decision-making within hospitals. Many sepsis and mortality prediction studies have been developed to predict each task in advance. However, previous studies have concentrated on single tasks, namely, the prediction of sepsis or mortality. Furthermore, although it is essential to provide clinicians with real-time information that incorporates dynamic alterations in hospitalized patients, the emphasis is on enhancing the efficacy of long-term forecasting over a period exceeding one day. OBJECTIVE We propose a novel framework for multi-task short-term prediction that integrates gated recurrent units and bidirectional long- and short-term memory networks, optimized for the prediction of sepsis and mortality and used 6-hours of information. Therefore, the proposed framework could provide timely and accurate predictions of sepsis and mortality within 12-hour. Additionally, a subject-independent evaluation was performed to evaluate the generalization performance of the model. METHODS The performance of sepsis and mortality prediction was evaluated using the medical information mart for intensive care (MIMIC)-IV database. A comparison of single-task and multi-task existing methods revealed that the latter exhibited superior performance. To assess the generalizability of the proposed framework, the eICU Collaborative Research Database (eICU-CRD) was tested using a model trained on MIMIC-IV, which demonstrated enhanced performance compared with each method. RESULTS The proposed method was statistically superior to performance achieved by single- and multi-task predictions for MIMIC-IV. Next, the eICU-CRD was tested using a MIMIC-IV-trained model to check its generalization ability, and a statistically superior performance was achieved compared with the baselines. CONCLUSIONS Our novel framework for learning unique features provides stable predictive power in different environments within the intensive care unit. Therefore, the proposed multi-task, including the sepsis and mortality prediction system, allows clinicians to intervene early in patients through sepsis and mortality prediction information, and can be applied to clinical trials in digital health.

Continuous and noninvasive blood pressure (BP) monitoring is essential for the early detection and management of hypertension. Cuffless BP estimation using photoplethysmography (PPG) has emerged as a promising alternative to traditional methods. However, most calibration-based approaches underestimate the physiological link between PPG signals and BP values. In this study, we propose a lightweight and efficient framework using affine-transformed triple PPG signals and self-attention. The proposed framework predicts BP by leveraging both calibration BP values and multiple forms of PPG signals, including its first and second derivatives. The proposed model consists of a CNN-based feature extractor for encoding the input signals, an attention mechanism using a residual block with squeeze-and-excitation to enhance channel-wise representations, and a regression module for estimating systolic and diastolic BP. A key novelty lies in the use of an affine transformation that integrates calibration BP values into the signal features without relying on additional neural networks. The proposed framework is evaluated on two public datasets, VitalDB and BP-UCI, comprising over 4,900 independent cases. It satisfies both the Association for the Advancement of Medical Instrumentation standard and the British Hypertension Society Grade A criteria. Comparative and ablation studies demonstrate that our proposed model consistently outperforms state-of-the-art machine learning and deep learning models in both accuracy and clinical reliability. The results confirms that it provides a generalizable, low-cost solution for cuffless BP estimation, with strong potential for real-time deployment in wearable healthcare systems.