Rationale: Deep learning is a powerful tool that may allow for improved outcome prediction. Objectives: To determine if deep learning, specifically convolutional neural network (CNN) analysis, could detect and stage chronic obstructive pulmonary disease (COPD) and predict acute respiratory disease (ARD) events and mortality in smokers. Methods: A CNN was trained using computed tomography scans from 7,983 COPDGene participants and evaluated using 1,000 nonoverlapping COPDGene participants and 1,672 ECLIPSE participants. Logistic regression (C statistic and the Hosmer-Lemeshow test) was used to assess COPD diagnosis and ARD prediction. Cox regression (C index and the Greenwood-Nam-D’Agnostino test) was used to assess mortality. Measurements and Main Results: In COPDGene, the C statistic for the detection of COPD was 0.856. A total of 51.1% of participants in COPDGene were accurately staged and 74.95% were within one stage. In ECLIPSE, 29.4% were accurately staged and 74.6% were within one stage. In COPDGene and ECLIPSE, the C statistics for ARD events were 0.64 and 0.55, respectively, and the Hosmer-Lemeshow P values were 0.502 and 0.380, respectively, suggesting no evidence of poor calibration. In COPDGene and ECLIPSE, CNN predicted mortality with fair discrimination (C indices, 0.72 and 0.60, respectively), and without evidence of poor calibration (Greenwood-Nam-D’Agnostino P values, 0.307 and 0.331, respectively). Conclusions: A deep-learning approach that uses only computed tomography imaging data can identify those smokers who have COPD and predict who are most likely to have ARD events and those with the highest mortality. At a population level CNN analysis may be a powerful tool for risk assessment.

While color video endoscopy has enabled wide-field examination of the gastrointestinal tract, it often misses or incorrectly classifies lesions. Many of these missed lesions exhibit characteristic three-dimensional surface topographies. An endoscopic system that adds topographical measurements to conventional color imagery could therefore increase lesion detection and improve classification accuracy. We introduce photometric stereo endoscopy (PSE), a technique which allows high spatial frequency components of surface topography to be acquired simultaneously with conventional two-dimensional color imagery. We implement this technique in an endoscopic form factor and demonstrate that it can acquire the topography of small features with complex geometries and heterogeneous optical properties. PSE imaging of ex vivo human gastrointestinal tissue shows that surface topography measurements enable differentiation of abnormal shapes from surrounding normal tissue. Together, these results confirm that the topographical measurements can be obtained with relatively simple hardware in an endoscopic form factor, and suggest the potential of PSE to improve lesion detection and classification in gastrointestinal imaging.

Introduction: The Agatston score is a well-established metric of cardiovascular disease related to clinical outcomes. It is computed from CT scans by a) measuring the volume and intensity of the atherosclerotic plaques and b) aggregating such information in an index. Objective: To generate a convolutional neural network that inputs a non-contrast chest CT scan and outputs the Agatston score associated with it directly, without a prior segmentation of Coronary Artery Calcifications (CAC). Materials and methods: We use a database of 5973 non-contrast non-ECG gated chest CT scans where the Agatston score has been manually computed. The heart of each scan is cropped automatically using an object detector. The database is split in 4973 cases for training and 1000 for testing. We train a 3D deep convolutional neural network to regress the Agatston score directly from the extracted hearts. Results: The proposed method yields a Pearson correlation coefficient of r = 0.93; p ≤ 0.0001 against manual reference standard in the 1000 test cases. It further stratifies correctly 72.6% of the cases with respect to standard risk groups. This compares to more complex state-of-the-art methods based on prior segmentations of the CACs, which achieve r = 0.94 in ECG-gated pulmonary CT. Conclusions: A convolutional neural network can regress the Agatston score from the image of the heart directly, without a prior segmentation of the CACs. This is a new and simpler paradigm in the Agatston score computation that yields similar results to the state-of-the-art literature.

Pulmonary Embolism (PE) is a respiratory disease caused by blood clots lodged in the pulmonary arteries, blocking perfusion, limiting blood oxygenation, and inducing a higher load on the right ventricle. Pulmonary embolism is diagnosed using contrast enhanced Computed Tomography Pulmonary Angiography (CTPA), resulting in a $3D$ image where the pulmonary arteries appear as bright structures, and emboli appear as filling defects, with these often being difficult to see, especially in the subsegmental case. In comparison to an expert panel, the average radiologist has a sensitivity of between 77% and $94\%$ . Computer Aided Detection (CAD) is regarded as a promising system to detect emboli, but current algorithms are hindered by a high false positive rate. In this paper, we propose a novel methodology for emboli detection. Instead of finding candidate points and characterizing them, we find emboli directly on the whole image slice. Detections across different slices are merged into a single detection volume that is post-processed to generate emboli detections. The system was evaluated on a public PE database of 80 scans. On 20 test scans, our system obtained a per-embolus sensitivity of 68% at a regime of one false positive per scan, improving on state-of-the-art methods. We therefore conclude that our multi-slice emboli segmentation CAD for PE method is a valuable alternative to the standard methods of candidate point selection and classification.

Biomarker estimation methods from medical images have traditionally followed a segment-and-measure strategy. Deep-learning regression networks have changed such a paradigm, enabling the direct estimation of biomarkers in databases where segmentation masks are not present. While such methods achieve high performance, they operate as a black-box. In this work, we present a novel deep learning network structure that, when trained with only the value of the biomarker, can perform biomarker regression and the generation of an accurate localization mask simultaneously, thus enabling a qualitative assessment of the image locus that relates to the quantitative result. We showcase the proposed method with three different network structures and compare their performance against direct regression networks in four different problems: pectoralis muscle area (PMA), subcutaneous fat area (SFA), liver mass area in single slice computed tomography (CT), and Agatston score estimated from non-contrast thoracic CT images (CAC). Our results show that the proposed method improves the performance with respect to direct biomarker regression methods (correlation coefficient of 0.978, 0.998, and 0.950 for the proposed method in comparison to 0.971, 0.982, and 0.936 for the reference regression methods on PMA, SFA and CAC respectively) while achieving good localization (DICE coefficients of 0.875, 0.914 for PMA and SFA respectively, p <; 0.05 for all pairs). We observe the same improvement in regression results comparing the proposed method with those obtained by quantify the outputs using an U-Net segmentation network (0.989 and 0.951 respectively). We, therefore, conclude that it is possible to obtain simultaneously good biomarker regression and localization when training biomarker regression networks using only the biomarker value.