The Application Value of Deep Learning-Based Nomograms in Benign-Malignant Discrimination of TI-RADS Category 4 Thyroid Nodules

Ma Zhe500 enrolled

Overview

This retrospective study focuses on benign and malignant classification of thyroid nodules using deep learning techniques and evaluates the value of deep learning based nomograms in the classification of TI-RADS category 4 thyroid nodules to improve the accuracy of benign and malignant identification of TI-RADS category 4 thyroid nodules. Materials and methods: Patients who visited in The First Affiliated Hospital of Shandong First Medical University ＆ Shandong Provincial Qianfoshan Hospital were collected. Their general clinical features, information on preoperative ultrasound diagnosis, and postoperative pathologic data were reviewed.

Study Type

OBSERVATIONAL

Enrollment

500

Conditions

Thyroid Nodule

Eligibility

Sex: ALLMin age: 23 YearsMax age: 78 Years

Medical Language ↔ Plain English

Inclusion Criteria: 1. Ultrasound-confirmed diagnosis of thyroid nodules that are classified as TI-RADS category 4. 2. Availability of pathological results. Exclusion Criteria: 1. Lack of pathological diagnosis. 2. History of thyroid surgery or other treatments. 3. Poor quality of ultrasound images of thyroid nodules. 4. Incomplete clinical and imaging data of the patient.

Locations (1)

QianfoshanH

Jinan, Shandong, China

Outcomes

Primary Outcomes

deep learning prediction model（YOLOv3） and the model evaluation

Based on the characteristics of benign and malignant thyroid nodules, the dataset was divided into a training set and a test set using the cross-validation method, and the YOLOv3 model was trained using data from the training set, and the performance of the model was evaluated using data from the test set.The model is evaluated using a number of metrics such as: precision-recall curve, effective classification precision, confusion matrix and area under the curve.

Time frame: Immediately evaluated after the prediction model was built

nomogram prediction and assessment

Factoring clinical features, ultrasound grading and model predictions to map nomograms using R language.Evaluation of the nomogram using various metrics, including subject operating characteristic curves, calibration curves and decision curve analysis

Time frame: Immediately evaluated after the nomogram was built

Selection of clinical features and assessment

The researchers selected patients with TI-RADS category 4 thyroid nodules within 1 year to comprise the dataset. The researchers analyzed the clinical factors in the dataset and analyzed the significance of these clinical factors on the statistical results and clinical characteristics using the Wilcoxon two-sample rank sum test or chi-square test.

Time frame: After the dataset is collected and pathology results are obtained, the statistical results obtained are analyzed for clinical factors, averaging about 1 year.

Impact and assessment of ultrasound grading

The researchers selected patients with TI-RADS category 4 thyroid nodules within 1 year to comprise the dataset. The researchers analyzed the results of grading TI-RADS category 4 nodules in this dataset and determined the significance of ultrasound grading on the statistical results using the chi-square test.

Time frame: The graded results of the ultrasound examination were analyzed after the data set collection was completed, the ultrasound examination was completed and the final pathology results were obtained, on average about 1 year.