Fine-grained attention & knowledge-based collaborative network for diabetic retinopathy grading.

Accurate diabetic retinopathy (DR) grading is crucial for making the proper treatment plan to reduce the damage caused by vision loss. This task is challenging due to the fact that the DR related lesions are often small and subtle in visual differences and intra-class variations. Moreover, relationships between the lesions and the DR levels are complicated. Although many deep learning (DL) DR grading systems have been developed with some success, there are still rooms for grading accuracy improvement. A common issue is that not much medical knowledge was used in these DL DR grading systems. As a result, the grading results are not properly interpreted by ophthalmologists, thus hinder the potential for practical applications. This paper proposes a novel fine-grained attention & knowledge-based collaborative network (FA+KC-Net) to address this concern. The fine-grained attention network dynamically divides the extracted feature maps into smaller patches and effectively captures small image features that are meaningful in the sense of its training from large amount of retinopathy fundus images. The knowledge-based collaborative network extracts a-priori medical knowledge features, i.e., lesions such as the microaneurysms (MAs), soft exudates (SEs), hard exudates (EXs), and hemorrhages (HEs). Finally, decision rules are developed to fuse the DR grading results from the fine-grained network and the knowledge-based collaborative network to make the final grading. Extensive experiments are carried out on four widely-used datasets, the DDR, Messidor, APTOS, and EyePACS to evaluate the efficacy of our method and compare with other state-of-the-art (SOTA) DL models. Simulation results show that proposed FA+KC-Net is accurate and stable, achieves the best performances on the DDR, Messidor, and APTOS datasets.

Attention-guided multi-scale context aggregation network for multi-modal brain glioma segmentation.

BACKGROUND: Accurate segmentation of brain glioma is a critical prerequisite for clinical diagnosis, surgical planning and treatment evaluation. In current clinical workflow, physicians typically perform delineation of brain tumor subregions slice-by-slice, which is more susceptible to variabilities in raters and also time-consuming. Besides, even though convolutional neural networks (CNNs) are driving progress, the performance of standard models still have some room for further improvement. PURPOSE: To deal with these issues, this paper proposes an attention-guided multi-scale context aggregation network (AMCA-Net) for the accurate segmentation of brain glioma in the magnetic resonance imaging (MRI) images with multi-modalities. METHODS: AMCA-Net extracts the multi-scale features from the MRI images and fuses the extracted discriminative features via a self-attention mechanism for brain glioma segmentation. The extraction is performed via a series of down-sampling, convolution layers, and the global context information guidance (GCIG) modules are developed to fuse the features extracted for contextual features. At the end of the down-sampling, a multi-scale fusion (MSF) module is designed to exploit and combine all the extracted multi-scale features. Each of the GCIG and MSF modules contain a channel attention (CA) module that can adaptively calibrate feature responses and emphasize the most relevant features. Finally, multiple predictions with different resolutions are fused through different weightings given by a multi-resolution adaptation (MRA) module instead of the use of averaging or max-pooling to improve the final segmentation results. RESULTS: Datasets used in this paper are publicly accessible, that is, the Multimodal Brain Tumor Segmentation Challenges 2018 (BraTS2018) and 2019 (BraTS2019). BraTS2018 contains 285 patient cases and BraTS2019 contains 335 cases. Simulations show that the AMCA-Net has better or comparable performance against that of the other state-of-the-art models. In terms of the Dice score and Hausdorff 95 for the BraTS2018 dataset, 90.4% and 10.2 mm for the whole tumor region (WT), 83.9% and 7.4 mm for the tumor core region (TC), 80.2% and 4.3 mm for the enhancing tumor region (ET), whereas the Dice score and Hausdorff 95 for the BraTS2019 dataset, 91.0% and 10.7 mm for the WT, 84.2% and 8.4 mm for the TC, 80.1% and 4.8 mm for the ET. CONCLUSIONS: The proposed AMCA-Net performs comparably well in comparison to several state-of-the-art neural net models in identifying the areas involving the peritumoral edema, enhancing tumor, and necrotic and non-enhancing tumor core of brain glioma, which has great potential for clinical practice. In future research, we will further explore the feasibility of applying AMCA-Net to other similar segmentation tasks.

Delineation of clinical target volume and organs at risk in cervical cancer radiotherapy by deep learning networks.

PURPOSE: Delineation of the clinical target volume (CTV) and organs-at-risk (OARs) is important in cervical cancer radiotherapy. But it is generally labor-intensive, time-consuming, and subjective. This paper proposes a parallel-path attention fusion network (PPAF-net) to overcome these disadvantages in the delineation task. METHODS: The PPAF-net utilizes both the texture and structure information of CTV and OARs by employing a U-Net network to capture the high-level texture information, and an up-sampling and down-sampling (USDS) network to capture the low-level structure information to accentuate the boundaries of CTV and OARs. Multi-level features extracted from both networks are then fused together through an attention module to generate the delineation result. RESULTS: The dataset contains 276 computed tomography (CT) scans of patients with cervical cancer of staging IB-IIA. The images are provided by the West China Hospital of Sichuan University. Simulation results demonstrate that PPAF-net performs favorably on the delineation of the CTV and OARs (e.g., rectum, bladder and etc.) and achieves the state-of-the-art delineation accuracy, respectively, for the CTV and OARs. In terms of the Dice Similarity Coefficient (DSC) and the Hausdorff Distance (HD), 88.61% and 2.25 cm for the CTV, 92.27% and 0.73 cm for the rectum, 96.74% and 0.68 cm for the bladder, 96.38% and 0.65 cm for the left kidney, 96.79% and 0.63 cm for the right kidney, 93.42% and 0.52 cm for the left femoral head, 93.69% and 0.51 cm for the right femoral head, 87.53% and 1.07 cm for the small intestine, and 91.50% and 0.84 cm for the spinal cord. CONCLUSIONS: The proposed automatic delineation network PPAF-net performs well on CTV and OARs segmentation tasks, which has great potential for reducing the burden of radiation oncologists and increasing the accuracy of delineation. In future, radiation oncologists from the West China Hospital of Sichuan University will further evaluate the results of network delineation, making this method helpful in clinical practice.

A data mining framework for time series estimation.

Time series estimation techniques are usually employed in biomedical research to derive variables less accessible from a set of related and more accessible variables. These techniques are traditionally built from systems modeling approaches including simulation, blind decovolution, and state estimation. In this work, we define target time series (TTS) and its related time series (RTS) as the output and input of a time series estimation process, respectively. We then propose a novel data mining framework for time series estimation when TTS and RTS represent different sets of observed variables from the same dynamic system. This is made possible by mining a database of instances of TTS, its simultaneously recorded RTS, and the input/output dynamic models between them. The key mining strategy is to formulate a mapping function for each TTS-RTS pair in the database that translates a feature vector extracted from RTS to the dissimilarity between true TTS and its estimate from the dynamic model associated with the same TTS-RTS pair. At run time, a feature vector is extracted from an inquiry RTS and supplied to the mapping function associated with each TTS-RTS pair to calculate a dissimilarity measure. An optimal TTS-RTS pair is then selected by analyzing these dissimilarity measures. The associated input/output model of the selected TTS-RTS pair is then used to simulate the TTS given the inquiry RTS as an input. An exemplary implementation was built to address a biomedical problem of noninvasive intracranial pressure assessment. The performance of the proposed method was superior to that of a simple training-free approach of finding the optimal TTS-RTS pair by a conventional similarity-based search on RTS features.

A comparison of vital signs charted by nurses with automated acquired values using waveform quality indices.

OBJECTIVE: (1) To investigate if there exist any discrepancies between the values of vital signs charted by nurses and those recorded by bedside monitors for a group of patients admitted for neurocritical care. (2) To investigate possible interpretations of discrepancies by exploring information in the alarm messages and the raw waveform data from monitors. METHODS: Each charted vital sign value was paired with a corresponding value from data collected by an archival program of bedside monitors such that the automatically archived data preceded the charted data and had minimal time lag to the charted value. Next, the absolute differences between the paired values were taken as the discrepancy between charted and automatically-collected data. Archived alarm messages were searched for technical alarms of sensor/lead failure types. Additionally, 7-min waveform data around the place of large discrepancy were analyzed using signal abnormality indices (SAI) for quantifying the quality of recorded signals. RESULTS: About 31,145 pairs of systolic blood pressure (BP-S) and 67,097 pairs of SpO(2) were investigated. Seven and a half percent of systolic blood pressure pairs had a discrepancy greater than 20 mmHg and less than one percent of the SpO2 pairs had a discrepancy greater than 10. We could not find any technical alarms from the monitors that could explain the large difference. However, SAI calculated for the waveforms associated with this group of cases was significantly larger than the SAI values calculated for the control waveform data of the same patients with small discrepancies. CONCLUSION: Charted vital signs reflect in large the raw data as reported by bedside monitors. Poor signal quality could partially explain the existence of cases of large discrepancies.