Dataset Information

Predicting central cervical lymph node metastasis in papillary thyroid microcarcinoma using deep learning.

ABSTRACT:

Background

The aim of this study is to design a deep learning (DL) model to preoperatively predict the occurrence of central lymph node metastasis (CLNM) in patients with papillary thyroid microcarcinoma (PTMC).

Methods

This research collected preoperative ultrasound (US) images and clinical factors of 611 PTMC patients. The clinical factors were analyzed using multivariate regression. Then, a DL model based on US images and clinical factors was developed to preoperatively predict CLNM. The model's efficacy was evaluated using the receiver operating characteristic (ROC) curve, along with accuracy, sensitivity, specificity, and the F₁ score.

Results

The multivariate analysis indicated an independent correlation factors including age ≥55 (OR = 0.309, p < 0.001), tumor diameter (OR = 2.551, p = 0.010), macrocalcifications (OR = 1.832, p = 0.002), and capsular invasion (OR = 1.977, p = 0.005). The suggested DL model utilized US images achieved an average area under the curve (AUC) of 0.65, slightly outperforming the model that employed traditional clinical factors (AUC = 0.64). Nevertheless, the model that incorporated both of them did not enhance prediction accuracy (AUC = 0.63).

Conclusions

The suggested approach offers a reference for the treatment and supervision of PTMC. Among three models used in this study, the deep model relied generally more on image modalities than the data modality of clinic records when making the predictions.

SUBMITTER: Wang Y

PROVIDER: S-EPMC10984175 | biostudies-literature | 2024

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Predicting central cervical lymph node metastasis in papillary thyroid microcarcinoma using deep learning.

Wang Yu Y Tan Hai-Long HL Duan Sai-Li SL Li Ning N Ai Lei L Chang Shi S

PeerJ 20240329

<h4>Background</h4>The aim of this study is to design a deep learning (DL) model to preoperatively predict the occurrence of central lymph node metastasis (CLNM) in patients with papillary thyroid microcarcinoma (PTMC).<h4>Methods</h4>This research collected preoperative ultrasound (US) images and clinical factors of 611 PTMC patients. The clinical factors were analyzed using multivariate regression. Then, a DL model based on US images and clinical factors was developed to preoperatively predict ...[more]

PMID: 38563008

Similar Datasets

Project description:BackgroundPapillary thyroid carcinoma (PTC) is the most frequent malignant tumor in thyroid carcinoma. The aim of this study was to explore the risk factors associated with central lymph node metastasis in papillary thyroid microcarcinoma (PTMC) and establish a nomogram model that can assess the probability of central lymph node metastasis (CLNM).MethodsThe clinicopathological data of 377 patients with cN0 PTMC were collected and analyzed from The Second Affiliated Hospital of Fujian Medical University from July 1st, 2019 to December 30th, 2021. All patients were examined by underwent ultrasound (US), found without metastasis to central lymph nodes, and diagnosed with PTMC through pathologic examination. All patients received thyroid lobectomy or total thyroidectomy with therapeutic or prophylactic central lymph node dissection (CLND). R software (Version 4.1.0) was employed to conduct a series of statistical analyses and establish the nomogram.ResultsA total of 119 patients with PTMC had central lymph node metastases (31.56%). After that, age (P < 0.05), gender (P < 0.05), tumor size (P < 0.05), tumor multifocality (P < 0.05), and ultrasound imaging-suggested tumor boundaries (P < 0.05) were identified as the risk factors associated with CLNM. Subsequently, multivariate logistic regression analysis indicated that the area under the receiver operating characteristic (ROC) curve (AUC) of the training cohort was 0.703 and that of the validation cohort was 0.656, demonstrating that the prediction ability of this model is relatively good compared to existing models. The calibration curves indicated a good fit for the nomogram model. Finally, the decision curve analysis (DCA) showed that a probability threshold of 0.15-0.50 could benefit patients clinically. The probability threshold used in DCA captures the relative value the patient places on receiving treatment for the disease, if present, compared to the value of avoiding treatment if the disease is not present.ConclusionCLNM is associated with many risk factors, including age, gender, tumor size, tumor multifocality, and ultrasound imaging-suggested tumor boundaries. The nomogram established in our study has moderate predictive ability for CLNM and can be applied to the clinical management of patients with PTMC. Our findings will provide a better preoperative assessment and treatment strategies for patients with PTMC whether to undergo central lymph node dissection.

Project description:ObjectivesPrecise determination of cervical lymph node metastasis (CLNM) involvement in patients with early-stage thyroid cancer is fairly significant for identifying appropriate cervical treatment options. However, it is almost impossible to directly judge lymph node metastasis based on the imaging information of early-stage thyroid cancer patients with clinically negative lymph nodes.MethodsPreoperative US images (BMUS and CDFI) of 1031 clinically node negative PTC patients definitively diagnosed on pathology from two independent hospitals were divided into training set, validation set, internal test set, and external test set. An ensemble deep learning model based on ResNet-50 was built integrating clinical variables, BMUS, and CDFI images using a bagging classifier to predict metastasis of CLN. The final ensemble model performance was compared with expert interpretation.ResultsThe ensemble deep convolutional neural network (DCNN) achieved high performance in predicting CLNM in the test sets examined, with area under the curve values of 0.86 (95% CI 0.78-0.94) for the internal test set and 0.77 (95% CI 0.68-0.87) for the external test set. Compared to all radiologists averaged, the ensemble DCNN model also exhibited improved performance in making predictions. For the external validation set, accuracy was 0.72 versus 0.59 (p = 0.074), sensitivity was 0.75 versus 0.58 (p = 0.039), and specificity was 0.69 versus 0.60 (p = 0.078).ConclusionsDeep learning can non-invasive predict CLNM for clinically node-negative PTC using conventional US imaging of thyroid cancer nodules and clinical variables in a multi-institutional dataset with superior accuracy, sensitivity, and specificity comparable to experts.Critical relevance statementDeep learning efficiently predicts CLNM for clinically node-negative PTC based on US images and clinical variables in an advantageous manner.Key points• A deep learning-based ensemble algorithm for predicting CLNM in PTC was developed. • Ultrasound AI analysis combined with clinical data has advantages in predicting CLNM. • Compared to all experts averaged, the DCNN model achieved higher test performance.

Project description:BackgroundPapillary thyroid microcarcinoma (PTMC) is characterized by its favorable prognosis and potential for active surveillance (AS) as a management option. However, the presence of cervical lymph node (LN) metastasis, especially lateral LN metastasis, significantly impacts management and prognosis. Previous studies have focused on post-surgery risk factors for cervical LN metastasis. This study aims to identify predictors of lateral LN metastasis by analyzing pre-operative ultrasonographic findings alongside clinicopathological factors.MethodsA retrospective review of medical records was conducted for patients with PTMC who underwent surgery at Chonnam National University Hwasun Hospital between 2004 and 2013. This is a case-control study that compares patients with lateral LN metastasis (N1b) to age- and sex-matched patients without LN metastasis (N0). Subgroup analysis was performed to evaluate risk factors of skip metastasis.ResultsThe study included 90 patients with PTMC with lateral LN metastasis (N1b) and 268 age- and sex-matched patients without LN metastasis (N0). The mean age was 49.3 years, and female patients were dominant in both groups. Structural recurrences of 4.4% (4/90) were observed only in the N1b group. The N1b group exhibited a higher frequency of upper lobe tumor location compared to the N0 group (38.9% vs. 16.0%, p < 0.001). There was no significant difference in the locations with the presence of invasion to adjacent organs. A higher proportion of non-parallel shape was observed in the N1b group than the N0 group (80.0% vs. 66.0%, p = 0.013). There were no differences in echogenicity, sonographic feature, margin, and AP diameter of the thyroid gland between the two groups. In multivariate analysis, independent risk factors for lateral LN metastasis included extrathyroidal extension, multiplicity, upper lobe tumor location, and non-parallel shape. Skip metastasis in patients with PTMC was associated with upper lobe tumor location.ConclusionDetailed ultrasound examinations, evaluating tumor location, number, orientation, and the presence of ETE, are crucial in accurately predicting lateral LN metastasis especially when primary tumor was in the upper lobe to avoid missing skip metastasis. These evaluations can help guide the decision between AS and immediate surgery in patients with PTMC.

Dataset Information

Predicting central cervical lymph node metastasis in papillary thyroid microcarcinoma using deep learning.

Background

Methods

Results

Conclusions

Publications

Predicting central cervical lymph node metastasis in papillary thyroid microcarcinoma using deep learning.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets