A Novel Ultrasound Probe for Thyroid Imaging and Machine Learning: A Pilot Study
试验速览
- 阶段
- 不适用
- 状态
- 招募中
- 发起方
- 入组人数
- 30
- 试验地点
- 1
- 主要终点
- The Ability to Visualization Thyroid Nodules and other key Anatomical Structures
研究概览
简要总结
The objective of the THYMAL 01 pilot study is to evaluate the Sound Blade Imaging System in subjects with benign thyroid nodules undergoing standard-of-care ultrasound imaging.
详细描述
Our purpose-built thyroid US platform, integrated with state-of-the-art machine-learning algorithms, will achieve accurate detection and delineation of thyroid nodules and neighboring critical structures, and will produce a trackable three-dimensional nodule model capable of guiding minimally invasive ablation while providing real-time volumetric quantification of treatment progress and residual disease.
STUDY DESIGN This study will consist of two parts. Part A - Prospective Thyroid Imaging Study We will conduct a prospective, paired imaging study (n=30) to determine whether the study US device acquires results/ability to detect thyroid nodules and key structures (common carotid artery, internal jugular vein, trachea, vagus nerve, recurrent laryngeal nerve, and parathyroid glands) that are comparable to the standard-of-care (SOC) US device.
Part B - Machine Learning Model Development US images and cine video acquired from the investigational device and radiologist annotations from Part A will be used to train and internally test an ML model for nodule and critical-structure segmentation. For external validation the ML model will be applied to anonymized thyroid US studies acquired on the standard-of-care (SOC) US and retrieved retrospectively from the provincial PACS (n≈30). Accuracy and precision will be assessed to establish technical validity and generalizability. ML outputs are for research only; no clinical decisions will be based on them.
研究设计
- 研究类型
- Interventional
- 分配方式
- Na
- 干预模型
- Single Group
- 主要目的
- Diagnostic
- 盲法
- None
入排标准
- 年龄范围
- 18 Years 至 —（Adult, Older Adult）
- 性别
- All
- 接受健康志愿者
- 否
入选标准
- •Participants must be ≥ 18 years of age
- •Able and willing to sign informed consent
- •Known thyroid nodule demonstrated on a prior thyroid US and is undergoing active surveillance of the nodule with serial USs
排除标准
- •Unable to lay flat for 15 minutes
- •Active neck wounds, dressings, or skin conditions that would interfere with neck US (e.g., preclude transducer placement)
- •Cervical spine or back disease that would prevent neck extension and would hinder the ability to obtain accurate thyroid US images
- •Previous thyroid surgeries, radiation of face and neck
- •Known inflammatory thyroid diseases or thyroiditis (e.g., Graves, Hashimoto)
研究组 & 干预措施
Thyroid Ultrasound
Patients who are scheduled for a SOC surveillance Thyroid Nodule Ultrasound will undergo a second ultrasound using the Sound Blade Ultrasound.
干预措施: Thyroid ultrasound (Diagnostic Test)
结局指标
主要结局
The Ability to Visualization Thyroid Nodules and other key Anatomical Structures
时间窗: Day 1
The primary endpoint is to evaluate the ability to visualize thyroid nodules and other key anatomical structures with the investigational device compared to the SOC US imaging device.
Evaluate the visibility/detection performance of a machine-learning model
时间窗: Day 1
A machine-learning model will be created to identify and segment thyroid nodules and key anatomical neck structures. The first outcome will evaluate performance of visibility and detection using standard classification measures
Evaluate the segmentation performance of a machine-learning model
时间窗: Day 1
A machine-learning model will be created to identify and segment thyroid nodules and key anatomical neck structures. The second outcome will evaluate performance on delineable cases.
次要结局
未报告次要终点