RESUMEN
OBJECTIVE: Deep learning (DL) has been applied in proofs of concept across biomedical imaging, including across modalities and medical specialties. Labeled data are critical to training and testing DL models, but human expert labelers are limited. In addition, DL traditionally requires copious training data, which is computationally expensive to process and iterate over. Consequently, it is useful to prioritize using those images that are most likely to improve a model's performance, a practice known as instance selection. The challenge is determining how best to prioritize. It is natural to prefer straightforward, robust, quantitative metrics as the basis for prioritization for instance selection. However, in current practice, such metrics are not tailored to, and almost never used for, image datasets. MATERIALS AND METHODS: To address this problem, we introduce ENRICH-Eliminate Noise and Redundancy for Imaging Challenges-a customizable method that prioritizes images based on how much diversity each image adds to the training set. RESULTS: First, we show that medical datasets are special in that in general each image adds less diversity than in nonmedical datasets. Next, we demonstrate that ENRICH achieves nearly maximal performance on classification and segmentation tasks on several medical image datasets using only a fraction of the available images and without up-front data labeling. ENRICH outperforms random image selection, the negative control. Finally, we show that ENRICH can also be used to identify errors and outliers in imaging datasets. CONCLUSIONS: ENRICH is a simple, computationally efficient method for prioritizing images for expert labeling and use in DL.
Asunto(s)
Diagnóstico por Imagen , Aprendizaje Automático , Humanos , Radiografía , Cuidados Paliativos , Procesamiento de Imagen Asistido por Computador/métodosRESUMEN
Congenital heart disease (CHD) is the most common birth defect. Fetal screening ultrasound provides five views of the heart that together can detect 90% of complex CHD, but in practice, sensitivity is as low as 30%. Here, using 107,823 images from 1,326 retrospective echocardiograms and screening ultrasounds from 18- to 24-week fetuses, we trained an ensemble of neural networks to identify recommended cardiac views and distinguish between normal hearts and complex CHD. We also used segmentation models to calculate standard fetal cardiothoracic measurements. In an internal test set of 4,108 fetal surveys (0.9% CHD, >4.4 million images), the model achieved an area under the curve (AUC) of 0.99, 95% sensitivity (95% confidence interval (CI), 84-99%), 96% specificity (95% CI, 95-97%) and 100% negative predictive value in distinguishing normal from abnormal hearts. Model sensitivity was comparable to that of clinicians and remained robust on outside-hospital and lower-quality images. The model's decisions were based on clinically relevant features. Cardiac measurements correlated with reported measures for normal and abnormal hearts. Applied to guideline-recommended imaging, ensemble learning models could significantly improve detection of fetal CHD, a critical and global diagnostic challenge.
Asunto(s)
Ecocardiografía Tridimensional/métodos , Cardiopatías Congénitas/diagnóstico , Cardiopatías Congénitas/patología , Diagnóstico Prenatal/métodos , Ultrasonografía Prenatal/métodos , Adulto , Biometría , Femenino , Feto/anomalías , Feto/diagnóstico por imagen , Corazón/diagnóstico por imagen , Humanos , Tamizaje Masivo/métodos , Miocardio/patología , Redes Neurales de la Computación , Embarazo , Segundo Trimestre del Embarazo , Sensibilidad y Especificidad , Tórax/diagnóstico por imagen , Adulto JovenRESUMEN
The evolution of communicative signals involves a major hurdle; signals need to effectively stimulate the sensory systems of their targets. Therefore, sensory specializations of target animals are important sources of selection on signal structure. Here we report the discovery of an animal signal that uses a previously unknown communicative modality, infrared radiation or "radiant heat," which capitalizes on the infrared sensory capabilities of the signal's target. California ground squirrels (Spermophilus beecheyi) add an infrared component to their snake-directed tail-flagging signals when confronting infrared-sensitive rattlesnakes (Crotalus oreganus), but tail flag without augmenting infrared emission when confronting infrared-insensitive gopher snakes (Pituophis melanoleucus). Experimental playbacks with a biorobotic squirrel model reveal this signal's communicative function. When the infrared component was added to the tail flagging display of the robotic models, rattlesnakes exhibited a greater shift from predatory to defensive behavior than during control trials in which tail flagging included no infrared component. These findings provide exceptionally strong support for the hypothesis that the sensory systems of signal targets should, in general, channel the evolution of signal structure. Furthermore, the discovery of previously undescribed signaling modalities such as infrared radiation should encourage us to overcome our own human-centered sensory biases and more fully examine the form and diversity of signals in the repertoires of many animal species.