Search

The achievement of the pathologic complete response (pCR) has been considered a metric for the success of neoadjuvant chemotherapy (NAC) and a powerful surrogate indicator of the risk of recurrence and long-term survival. This study aimed to develop a multimodal deep learning model that combined clinical information and pretreatment MR images for predicting pCR to NAC in patients with breast cancer. The retrospective study cohort consisted of 536 patients with invasive breast cancer who underwent pre-operative NAC. We developed a deep learning model to fuse high-dimensional MR image features and the clinical information for the pretreatment prediction of pCR to NAC in breast cancer. The proposed deep learning model trained on all datasets as clinical information, T1-weighted subtraction images, and T2-weighted images shows better performance with area under the curve (AUC) of 0.888 as compared to the model using only clinical information (AUC = 0.827, P

BackgroundThe aim of this study was to develop a machine learning(ML) based model to accurately predict pathologic complete response(pCR) to neoadjuvant chemotherapy(NAC) using pretreatment clinical and pathological characteristics of electronic medical record(EMR) data in breast cancer(BC).Methods The EMR data from patients diagnosed with early and locally advanced BC and who received NAC followed by curative surgery were reviewed. A total of 16 clinical and pathological characteristics was selected to develop ML model. We practiced six ML models using default settings for multivariate analysis with extracted variables. ResultsIn total, 2,065 patients were included in this analysis. Overall, 30.6% (n=632) of patients achieved pCR. Among six ML models, the LightGBM had the highest area under the curve (AUC) for pCR prediction. After hyper-parameter tuning with Bayesian optimization, AUC was 0.810. Performance of pCR prediction models in different histology-based subtypes was compared. The AUC was highest in HR+HER2- subgroup and lowest in HR-/HER2- subgroup (HR+/HER2- 0.841, HR+/HER2+ 0.716, HR-/HER2 0.753, HR-/HER2- 0.653).ConclusionsA ML based pCR prediction model using pre-treatment clinical and pathological characteristics provided useful information to predict pCR during NAC. This prediction model would help to determine treatment strategy in patients with BC planned NAC.

We present a novel deep neural model for text detection in document images. For robust text detection in noisy scanned documents, the advantages of multi-task learning are adopted by adding an auxiliary task of text enhancement. Namely, our proposed model is designed to perform noise reduction and text region enhancement as well as text detection. Moreover, we enrich the training data for the model with synthesized document images that are fully labeled for text detection and enhancement, thus overcome the insufficiency of labeled document image data. For the effective exploitation of the synthetic and real data, the training process is separated in two phases. The first phase is training only synthetic data in a fully-supervised manner. Then real data with only detection labels are added in the second phase. The enhancement task for the real data is weakly-supervised with information from their detection labels. Our methods are demonstrated in a real document dataset with performances exceeding those of other text detection methods. Moreover, ablations are conducted and the results confirm the effectiveness of the synthetic data, auxiliary task, and weak-supervision. Whereas the existing text detection studies mostly focus on the text in scenes, our proposed method is optimized to the applications for the text in scanned documents.

Background Electrocardiographic (ECG) monitors have been widely used for diagnosing cardiac arrhythmias for decades. However, accurate analysis of ECG signals is difficult and time-consuming work because large amounts of beats need to be inspected. In order to enhance ECG beat classification, machine learning and deep learning methods have been studied. However, existing studies have limitations in model rigidity, model complexity, and inference speed. Objective To classify ECG beats effectively and efficiently, we propose a baseline model with recurrent neural networks (RNNs). Furthermore, we also propose a lightweight model with fused RNN for speeding up the prediction time on central processing units (CPUs). Methods We used 48 ECGs from the MIT-BIH (Massachusetts Institute of Technology-Beth Israel Hospital) Arrhythmia Database, and 76 ECGs were collected with S-Patch devices developed by Samsung SDS. We developed both baseline and lightweight models on the MXNet framework. We trained both models on graphics processing units and measured both models’ inference times on CPUs. Results Our models achieved overall beat classification accuracies of 99.72% for the baseline model with RNN and 99.80% for the lightweight model with fused RNN. Moreover, our lightweight model reduced the inference time on CPUs without any loss of accuracy. The inference time for the lightweight model for 24-hour ECGs was 3 minutes, which is 5 times faster than the baseline model. Conclusions Both our baseline and lightweight models achieved cardiologist-level accuracies. Furthermore, our lightweight model is competitive on CPU-based wearable hardware.

When a Liquified Natural Gas (LNG) receiving terminal is damaged by an earthquake, the socio-economic impact of it`s damage might be huge. Therefore, it is important to understand the seismic resistant capacity of such facilities, specially LNG storage tanks, in order for minimizing the earthquake hazard. Even though seismic design criteria are so strict that LNG storage tanks rarely suffer from earthquakes, we still need to know what to happen when a larger earthquake than a design earthquake occurs. A seismic fragility function is frequently used to express the seismic performance of structures. In this study, the seismic fragility functions of LNG storage tanks that are under operation in Korea are developed. A so-called `tuning-fork model` is used to evaluate the dynamic response of the LNG tank based on a series of time history analyses. The fragility functions are developed based on the performance limit states that are defined in HAZUS, a popular seismic performance evaluation program. Finally, the analytical fragility function is updated by combined by pre-defined empirical fragility functions. According to the fragility functions, the probabilities of having minor cracks and moderate cracks on the tank wall are 1.0% and 0.0%, respectively, for the design-level earthquake with the peak ground acceleration (PGA) 0.2 g. For PGA 0.4 g, the corresponding probabilities are 27.0% and 10.0%, respectively.