Chao Zhou's research while affiliated with The First People's Hospital of Changzhou and other places

Background: Dissection of the lymph nodes posterior to the right recurrent laryngeal nerve (LN-prRLNs) in papillary thyroid cancer (PTC) remains controversial. Objective: This study aimed to determine the capability of ultrasonography (US)-based radiomics for presurgical prediction of metastasis in LN-prRLNs in PTC. Methods: Patients were retrospectively enrolled and pathologically confirmed as LN-prRLN metastasis with PTC after surgery. Radiomic analysis based on preoperative US images with manual segmentation of targets was used to develop a radiomics model. US features described in ACR TI-RADS were collected to construct a clinical model. The Radiomics model, a combined model integrating radiomics and clinical model, was also developed for the presurgical prediction of metastasis in LN-prRLNs. Results: A total of 570 patients, including 488 patients with non-LN-prRLN metastasis and 82 with LN-prRLN metastasis, were assessed. The 15 topperforming features finally remained significant for constructing the radiomics model. The combined model showed that US measured tumor size (OR: 1.036, P = 0.044), US suspected lateral lymph node metastasis (OR: 2.247, P = 0.009), multifocality (OR: 1.920, P = 0.021), Delphian lymph node metastasis (DLNM) (OR: 2.300, P = 0.039), VIa compartment metastasis (OR: 5.357, P = 0.000), the radiomics score (OR: 1.003, P = 0.001) were significant risk factors for predicting LN-prRLN metastasis. The combined model achieved a higher AUC of 0.849 than that of the clinical model (AUC: 0.759) and radiomics model (AUC: 0.826). Conclusion: The US-based radiomics combined model can more effectively predict LN-prRLN metastasis in PTCs patients preoperatively. This approach had the potential to assist surgeons indecision-making regarding LN-prRLN dissection.

Background Delphian lymph node (DLN) has been considered to be a gate that predicts widespread lymph node involvement, higher recurrence and mortality rates of head and neck cancer. Objective This study aimed to establish a preoperative ultrasonography integrated machine learning prediction model to predict Delphian lymph node metastasis (DLNM) in patients with diagnosed papillary thyroid carcinoma (PTC). Methods Ultrasonographic and clinicopathologic variables of PTC patients from 2014 to 2021 were retrospectively analyzed. The risk factors associated with DLNM were identified and validated through a developed random forest (RF) algorithm model based on machine learning and a logistic regression (LR) model. Results A total of 316 patients with 402 thyroid lesions were enrolled for the training dataset and 280 patients with 341 lesions for the validation dataset, with 170 (28.52%) patients developed DLNM. The elastography score of ultrasonography, central lymph node metastasis, lateral lymph node metastasis, and serum calcitonin were predictive factors for DLNM in both models. The RF model has better predictive performance in the training dataset and validation dataset (AUC: 0.957 vs. 0.890) than that in the LR model (AUC: 0.908 vs. 0.833). Conclusion The preoperative ultrasonography integrated RF model constructed in this study could accurately predict DLNM in PTC patients, which may provide clinicians with more personalized clinical decision-making recommendations preoperatively. Machine learning technology has the potential to improve the development of DLNM prediction models in PTC patients.

Background This study aims to determine the prediction performance of a machine learning-based clinical model for cervical lymph node metastasis (CLNM) in micropapillary thyroid carcinoma (MPTC) with ultrasound (US). Methods Patients with MPTC who underwent total or hemithyroidectomy with unilateral or bilateral prophylactic central neck dissection were included (n = 692). Nodal status was pathologically determined. Clinical and US features and thyroid function markers were extracted to build a random forest model. A nomogram with the significant predictive risk factors from multivariable logistic regression analysis was built to visualize hazard rates. Finally, the predictive performances of the models were compared. Results Overall, 332 patients (47.98%) had CLNM. In multiple logistic regression, the strong predictive risk factors for CLNM were younger age, larger anteroposterior diameter, lower anteroposterior/transverse diameter (A/T) ratio, and higher thyroglobulin (TG) concentration (P < 0.05). The random forest and nomogram models showed good predictive performance with the area under the curves (AUCs) of 0.836 and 0.780, respectively, which were significantly higher than those without A/T ratio in the models (AUCs: 0.807 vs. 0.722, all P < 0.05). The AUC of the A/T ratio as a single feature for predicting CLNM was 0.744, while A/T ratio (≤ 0.828) combined with anteroposterior diameter (≥ 10 mm) yielded a higher AUC of 0.754 for predicting CLNM. Conclusions The machine learning-based clinical model with US had a good predictive performance for CLNM in MPTC patients. This clinical model may facilitate surgical decision-making for MPTC, especially regarding whether cervical lymph node dissection is warranted.