Development and external validation of deep learning clinical prediction models using variable-length time series data.

OBJECTIVES: To compare and externally validate popular deep learning model architectures and data transformation methods for variable-length time series data in 3 clinical tasks (clinical deterioration, severe acute kidney injury [AKI], and suspected infection). MATERIALS AND METHODS: This multicenter retrospective study included admissions at 2 medical centers that spanned 2007-2022. Distinct datasets were created for each clinical task, with 1 site used for training and the other for testing. Three feature engineering methods (normalization, standardization, and piece-wise linear encoding with decision trees [PLE-DTs]) and 3 architectures (long short-term memory/gated recurrent unit [LSTM/GRU], temporal convolutional network, and time-distributed wrapper with convolutional neural network [TDW-CNN]) were compared in each clinical task. Model discrimination was evaluated using the area under the precision-recall curve (AUPRC) and the area under the receiver operating characteristic curve (AUROC). RESULTS: The study comprised 373 825 admissions for training and 256 128 admissions for testing. LSTM/GRU models tied with TDW-CNN models with both obtaining the highest mean AUPRC in 2 tasks, and LSTM/GRU had the highest mean AUROC across all tasks (deterioration: 0.81, AKI: 0.92, infection: 0.87). PLE-DT with LSTM/GRU achieved the highest AUPRC in all tasks. DISCUSSION: When externally validated in 3 clinical tasks, the LSTM/GRU model architecture with PLE-DT transformed data demonstrated the highest AUPRC in all tasks. Multiple models achieved similar performance when evaluated using AUROC. CONCLUSION: The LSTM architecture performs as well or better than some newer architectures, and PLE-DT may enhance the AUPRC in variable-length time series data for predicting clinical outcomes during external validation.

Identifying infected patients using semi-supervised and transfer learning.

OBJECTIVES: Early identification of infection improves outcomes, but developing models for early identification requires determining infection status with manual chart review, limiting sample size. Therefore, we aimed to compare semi-supervised and transfer learning algorithms with algorithms based solely on manual chart review for identifying infection in hospitalized patients. MATERIALS AND METHODS: This multicenter retrospective study of admissions to 6 hospitals included "gold-standard" labels of infection from manual chart review and "silver-standard" labels from nonchart-reviewed patients using the Sepsis-3 infection criteria based on antibiotic and culture orders. "Gold-standard" labeled admissions were randomly allocated to training (70%) and testing (30%) datasets. Using patient characteristics, vital signs, and laboratory data from the first 24 hours of admission, we derived deep learning and non-deep learning models using transfer learning and semi-supervised methods. Performance was compared in the gold-standard test set using discrimination and calibration metrics. RESULTS: The study comprised 432 965 admissions, of which 2724 underwent chart review. In the test set, deep learning and non-deep learning approaches had similar discrimination (area under the receiver operating characteristic curve of 0.82). Semi-supervised and transfer learning approaches did not improve discrimination over models fit using only silver- or gold-standard data. Transfer learning had the best calibration (unreliability index P value: .997, Brier score: 0.173), followed by self-learning gradient boosted machine (P value: .67, Brier score: 0.170). DISCUSSION: Deep learning and non-deep learning models performed similarly for identifying infection, as did models developed using Sepsis-3 and manual chart review labels. CONCLUSION: In a multicenter study of almost 3000 chart-reviewed patients, semi-supervised and transfer learning models showed similar performance for model discrimination as baseline XGBoost, while transfer learning improved calibration.

MCIndoor20000: A fully-labeled image dataset to advance indoor objects detection.

A fully-labeled image dataset provides a unique resource for reproducible research inquiries and data analyses in several computational fields, such as computer vision, machine learning and deep learning machine intelligence. With the present contribution, a large-scale fully-labeled image dataset is provided, and made publicly and freely available to the research community. The current dataset entitled MCIndoor20000 includes more than 20,000 digital images from three different indoor object categories, including doors, stairs, and hospital signs. To make a comprehensive dataset addressing current challenges that exist in indoor objects modeling, we cover a multiple set of variations in images, such as rotation, intra-class variation plus various noise models. The current dataset is freely and publicly available at https://github.com/bircatmcri/MCIndoor20000.

Multi-Modal Medical Image Registration with Full or Partial Data: A Manifold Learning Approach.

Multi-modal image registration is the primary step in integrating information stored in two or more images, which are captured using multiple imaging modalities. In addition to intensity variations and structural differences between images, they may have partial or full overlap, which adds an extra hurdle to the success of registration process. In this contribution, we propose a multi-modal to mono-modal transformation method that facilitates direct application of well-founded mono-modal registration methods in order to obtain accurate alignment of multi-modal images in both cases, with complete (full) and incomplete (partial) overlap. The proposed transformation facilitates recovering strong scales, rotations, and translations. We explain the method thoroughly and discuss the choice of parameters. For evaluation purposes, the effectiveness of the proposed method is examined and compared with widely used information theory-based techniques using simulated and clinical human brain images with full data. Using RIRE dataset, mean absolute error of 1.37, 1.00, and 1.41 mm are obtained for registering CT images with PD-, T1-, and T2-MRIs, respectively. In the end, we empirically investigate the efficacy of the proposed transformation in registering multi-modal partially overlapped images.