Deep Convolutional Neural Network With a Multi-Scale Attention Feature Fusion Module for Segmentation of Multimodal Brain Tumor

End-to-End Multi-task Learning Architecture for Brain Tumor Analysis with Uncertainty Estimation in MRI Images

Brain tumors are a threat to life for every other human being, be it adults or children. Gliomas are one of the deadliest brain tumors with an extremely difficult diagnosis. The reason is their complex and heterogenous structure which gives rise to subjective as well as objective errors. Their manual segmentation is a laborious task due to their complex structure and irregular appearance. To cater to all these issues, a lot of research has been done and is going on to develop AI-based solutions that can help doctors and radiologists in the effective diagnosis of gliomas with the least subjective and objective errors, but an end-to-end system is still missing. An all-in-one framework has been proposed in this research. The developed end-to-end multi-task learning (MTL) architecture with a feature attention module can classify, segment, and predict the overall survival of gliomas by leveraging task relationships between similar tasks. Uncertainty estimation has also been incorporated into the framework to enhance the confidence level of healthcare practitioners. Extensive experimentation was performed by using combinations of MRI sequences. Brain tumor segmentation (BraTS) challenge datasets of 2019 and 2020 were used for experimental purposes. Results of the best model with four sequences show 95.1% accuracy for classification, 86.3% dice score for segmentation, and a mean absolute error (MAE) of 456.59 for survival prediction on the test data. It is evident from the results that deep learning-based MTL models have the potential to automate the whole brain tumor analysis process and give efficient results with least inference time without human intervention. Uncertainty quantification confirms the idea that more data can improve the generalization ability and in turn can produce more accurate results with less uncertainty. The proposed model has the potential to be utilized in a clinical setup for the initial screening of glioma patients.

Incidence of Dementia Before Age 65 Years Among World Trade Center Attack Responders

Importance

Reports suggest that the individuals who served in rescue operations following the terrorist attacks on the World Trade Center (WTC) have poorer brain health than expected.

Objective

To assess the incidence of dementia before age 65 years in a prospective study of WTC responders and to compare incidence among responders with severe exposures to debris vs responders not exposed to building debris or who wore personalized protective equipment (PPE).

Design, Setting, and Participants

This prospective cohort study was conducted from November 1, 2014, to January 1, 2023, in an academic medical monitoring program available to verified WTC responders residing on Long Island, New York. Responders 60 years of age or younger without dementia at the time of their first cognitive assessment were followed up every 18 months, on average, for up to 5 years.

Exposures

Exposure severity was based on responses to a detailed questionnaire of WTC exposures and exposure-related activities that included exposures to fine particulate dust and potentially neurotoxic debris, duration of work, and the use of PPE. Exposure level was divided into 5 categories ranging from low to severe.

Main Outcomes and Measures

Incidence of all-cause dementia before age 65 years was the primary outcome. Dementia was diagnosed following standard guidelines relying on repeated measures of cognition.

Results

Of 9891 responders, 5010 were eligible for inclusion in this study of cognitive function (median [IQR] age, 53 [48-57] years; 4573 [91.3%] male). There were 228 cases of dementia identified during 15 913.1 person-years of follow-up. Increasing WTC exposure severity was associated with incremental increases in the incidence rate of dementia per 1000 person-years (low, 2.95 [95% CI, 1.07-11.18]; mild, 12.16 [95% CI, 10.09-14.79]; moderate, 16.53 [95% CI, 13.30-20.81]; high, 30.09 [95% CI, 21.35-43.79]; and severe, 42.37 [95% CI, 24.86-78.24]). Adjusting for social, demographic, and relevant medical factors, each unit increase in exposure severity was associated with increased incidence of dementia (adjusted hazard ratio, 1.42 [95% CI, 1.18-1.71]; P < .001; mean risk difference, 9.74 [95% CI, 2.94-32.32] per 1000 person-years; P < .001).

Conclusions and Relevance

In this cohort study of WTC responders who survived these unique exposures and participated in a longitudinal follow-up study of cognition from 2014 through 2022, when compared with responders with the lowest exposure levels or responders who used PPE, more severe exposure to dust or debris was significantly associated with a higher risk of dementia before 65 years of age. This study suggests that the reliable use of PPE might help prevent the onset of dementia before age 65 years among individuals exposed to an uncontrolled building collapse. Future research is warranted to determine cerebral biomarkers for individuals with exposure-associated dementia.

Unsupervised Black-Box Model Domain Adaptation for Brain Tumor Segmentation

Unsupervised domain adaptation (UDA) is an emerging technique that enables the transfer of domain knowledge learned from a labeled source domain to unlabeled target domains, providing a way of coping with the difficulty of labeling in new domains. The majority of prior work has relied on both source and target domain data for adaptation. However, because of privacy concerns about potential leaks in sensitive information contained in patient data, it is often challenging to share the data and labels in the source domain and trained model parameters in cross-center collaborations. To address this issue, we propose a practical framework for UDA with a black-box segmentation model trained in the source domain only, without relying on source data or a white-box source model in which the network parameters are accessible. In particular, we propose a knowledge distillation scheme to gradually learn target-specific representations. Additionally, we regularize the confidence of the labels in the target domain via unsupervised entropy minimization, leading to performance gain over UDA without entropy minimization. We extensively validated our framework on a few datasets and deep learning backbones, demonstrating the potential for our framework to be applied in challenging yet realistic clinical settings.