Abstract:Abstract Accurate identification and localization of multiple abnormalities are crucial steps in the interpretation of chest X-rays (CXRs); however, the lack of a large CXR dataset with bounding boxes severely constrains accurate localization research based on deep learning. We created a large CXR dataset named CXR-AL14, containing 165,988 CXRs and 253,844 bounding boxes. On the basis of this dataset, a deep-learning-based framework was developed to identify and localize 14 common abnormalities and calculate the cardiothoracic ratio (CTR) simultaneously. The mean average precision values obtained by the model for 14 abnormalities reached 0.572-0.631 with an intersection-over-union threshold of 0.5, and the intraclass correlation coefficient of the CTR algorithm exceeded 0.95 on the held-out, multicentre and prospective test datasets. This framework shows an excellent performance, good generalization ability and strong clinical applicability, which is superior to senior radiologists and suitable for routine clinical settings.

What problem does this paper attempt to address?

The main problem that this paper attempts to solve is to improve the ability to accurately identify and localize multiple abnormalities in chest X - rays (CXR). Specifically: 1. **Create a large - scale dataset**: Researchers created a large - scale dataset named CXR - AL14, which contains 165,988 chest X - rays and 253,844 bounding boxes. These bounding boxes label 14 common chest abnormalities. This is currently the chest X - ray dataset in the world with the most real - world bounding boxes. 2. **Develop a deep - learning framework**: Based on the CXR - AL14 dataset, researchers developed a deep - learning framework that can simultaneously identify and localize 14 chest abnormalities and calculate the cardiothoracic ratio (CTR) for evaluating the degree of cardiac enlargement. 3. **Improve diagnostic efficiency**: The performance of this framework on multiple test datasets is better than that of senior radiologists, showing excellent performance, good generalization ability and strong clinical applicability, which helps to reduce the workload of radiologists and improve diagnostic efficiency. ### Main contributions - **Dataset**: The creation of the CXR - AL14 dataset provides a large amount of data support with real - world bounding boxes for the application of deep learning in chest X - rays. - **Model performance**: The developed deep - learning framework performs well in identifying and localizing 14 chest abnormalities. The mean average precision (mAP) reaches 0.572 - 0.631 (IoU threshold is 0.5), and the intra - class correlation coefficient (ICC) of the CTR algorithm exceeds 0.95. - **Clinical application**: This framework shows good generalization performance and clinical applicability in multi - center and prospective validations, and is suitable for routine clinical use. ### Problems solved - **Lack of large - scale datasets with bounding boxes**: Most of the existing public chest X - ray datasets only have class labels and lack accurate localization information, which limits the performance of deep - learning models. - **Improve diagnostic accuracy**: By automatically identifying and localizing multiple abnormalities, the rates of missed diagnosis and misdiagnosis are reduced, and the diagnostic efficiency of radiologists is improved. - **Calculate the cardiothoracic ratio**: A fast and accurate CTR calculation method is provided to assist in evaluating the degree of cardiac enlargement. In conclusion, this research significantly improves the ability to identify and localize multiple abnormalities in chest X - rays by creating a large - scale chest X - ray dataset with bounding boxes and developing an efficient deep - learning framework, and has important clinical application value.

A deep-learning-based framework for identifying and localizing multiple abnormalities and assessing cardiomegaly in chest X-ray

Abnormalities detection on chest radiograph with bounding box-based lungs extraction and object detection algorithm

Development of a multipotent diagnostic tool for chest X-rays by multi-object detection method

Size-adaptive Mediastinal Multilesion Detection in Chest CT Images Via Deep Learning and a Benchmark Dataset

Automatic Localization and Identification of Thoracic Diseases from Chest X-rays with Deep Learning.

Can Deep Learning Reliably Recognize Abnormality Patterns on Chest X-rays? A Multi-Reader Study Examining One Month of AI Implementation in Everyday Radiology Clinical Practice

Abnormality Detection and Localization in Chest X-Rays using Deep Convolutional Neural Networks

Deep learning to detect left ventricular structural abnormalities in chest X-rays

ChestX-ray8: Hospital-scale Chest X-ray Database and Benchmarks on Weakly-Supervised Classification and Localization of Common Thorax Diseases

Identifying cardiomegaly in chest x-rays using dual attention network

Multi-task Learning for Chest X-ray Abnormality Classification on Noisy Labels

Chest X-Ray Image Analysis With Combining 2D and 1D Convolutional Neural Network Based Classifier for Rapid Cardiomegaly Screening

Evaluation of the feasibility of explainable computer-aided detection of cardiomegaly on chest radiographs using deep learning

Deep-Learning-Based Diagnosis of Bedside Chest X-ray in Intensive Care and Emergency Medicine

Machine-Learning-Based Multiple Abnormality Prediction with Large-Scale Chest Computed Tomography Volumes

DeepCXray: Automatically Diagnosing Diseases on Chest X-Rays Using Deep Neural Networks

Development and validation of a deep learning algorithm detecting 10 common abnormalities on chest radiographs

Low-Shot Multi-label Incremental Learning for Thoracic Diseases Diagnosis

Development and External Validation of an Artificial Intelligence-Based Method for Scalable Chest Radiograph Diagnosis: A Multi-Country Cross-Sectional Study

Identifying Cardiomegaly in ChestX-ray8 Using Transfer Learning

Development and validation of an abnormality-derived deep-learning diagnostic system for major respiratory diseases