โก Quick Summary
This study presents a comprehensive ultrasonography dataset of thyroid nodules, comprising 8,508 images from 842 cases, annotated with pathological diagnoses. The dataset aims to enhance the development of deep learning algorithms for more accurate histological status inference from thyroid ultrasound images.
๐ Key Details
- ๐ Dataset: 8,508 ultrasound images from 842 cases
- ๐งฉ Features used: Pathological diagnosis annotations
- โ๏ธ Technology: Deep learning models for validation
- ๐ Objective: To improve diagnostic accuracy in thyroid nodule assessment
๐ Key Takeaways
- ๐ Ultrasonography of thyroid nodules is often time-consuming and inconsistent.
- ๐ก Fine needle aspiration biopsy (FNAB) is still required for definitive diagnosis despite TI-RADS staging.
- ๐ค Deep learning methods have been developed but often rely on TI-RADS reports as labels.
- ๐ The new dataset allows for direct inference of histological status from ultrasound images.
- ๐ฅ Potential applications in improving diagnostic workflows in clinical settings.
- ๐ Study conducted using data from two retrospective cohorts.
- ๐ PMID: 39580501
- ๐ Published in: Sci Data, 2024
๐ Background
The assessment of thyroid nodules through ultrasonography is a critical yet challenging aspect of endocrinology. Traditional methods often lead to variability in diagnosis and require additional invasive procedures like FNAB to confirm malignancy. The introduction of deep learning technologies in medical imaging holds promise for enhancing diagnostic accuracy and efficiency.
๐๏ธ Study
This study focused on creating a robust dataset of thyroid ultrasound images, annotated with pathological diagnoses. The dataset was compiled from two retrospective cohorts, providing a substantial resource for training deep learning models aimed at improving the diagnostic process for thyroid nodules.
๐ Results
The dataset comprises 8,508 ultrasound images from 842 cases, providing a rich foundation for developing and validating deep learning algorithms. The study also explored three different deep learning models, showcasing their potential in accurately inferring histological status from ultrasound images.
๐ Impact and Implications
The implications of this study are significant for the field of endocrinology and medical imaging. By leveraging deep learning algorithms trained on this comprehensive dataset, healthcare professionals could achieve more accurate and timely diagnoses of thyroid nodules, potentially reducing the need for invasive procedures and improving patient outcomes.
๐ฎ Conclusion
This study highlights the transformative potential of deep learning in the assessment of thyroid nodules through ultrasonography. With a well-annotated dataset, researchers and clinicians can work towards developing algorithms that enhance diagnostic accuracy, paving the way for improved patient care in endocrinology. Continued research in this area is essential for realizing the full benefits of AI in healthcare.
๐ฌ Your comments
What are your thoughts on the integration of deep learning in thyroid nodule assessment? We invite you to share your insights! ๐ฌ Leave your comments below or connect with us on social media:
An ultrasonography of thyroid nodules dataset with pathological diagnosis annotation for deep learning.
Abstract
Ultrasonography (US) of thyroid nodules is often time consuming and may be inconsistent between observers, with a low positivity rate for malignancy in biopsies. Even after determining the ultrasound Thyroid Imaging Reporting and Data System (TIRADS) stage, Fine needle aspiration biopsy (FNAB) is still required to obtain a definitive diagnosis. Although various deep learning methods were developed in medical field, they tend to be trained using TI-RADS reports as image labels. Here, we present a large US dataset with pathological diagnosis annotation for each case, designed for developing deep learning algorithms to directly infer histological status from thyroid ultrasound images. The dataset was collected from two retrospective cohorts, which consists of 8508 US images from 842 cases. Additionally, we explained three deep learning models used as validation examples using this dataset.
Author: [‘Hou X’, ‘Hua M’, ‘Zhang W’, ‘Ji J’, ‘Zhang X’, ‘Jiang H’, ‘Li M’, ‘Wu X’, ‘Zhao W’, ‘Sun S’, ‘Cao L’, ‘Wang L’]
Journal: Sci Data
Citation: Hou X, et al. An ultrasonography of thyroid nodules dataset with pathological diagnosis annotation for deep learning. An ultrasonography of thyroid nodules dataset with pathological diagnosis annotation for deep learning. 2024; 11:1272. doi: 10.1038/s41597-024-04156-5