Background: The incidence and recurrence rate of papillary thyroid carcinoma (PTC) are high. Thus, it is critical to accurately identify patients at high risk of recurrence. Pyroptosis is a type of programmed cell death closely related to the progression and prognosis of cancer. However, the role of pyroptosis in PTC remains unclear. Methods: Transcriptome data for PTC patients were obtained from The Cancer Genome Atlas database. The expression level of pyroptosis-related genes (PRGs) in PTC and normal tissues was identified. Based on these differentially expressed genes, a risk score model of disease-free survival (DFS) was established using least absolute shrinkage and selection operator Cox regression. In-cluster and quantitative real-time PCR validations were carried out. A nomogram, in combination with clinical factors, was also established. In addition, its relationship with immune characteristics and tumor gene mutations is discussed. Results: A risk score model with four PRGs, including CASP6, CASP9, IL-18, and NOD1, was established. The samples were divided into high- and low-risk clusters, according to the risk score, revealing significant differences in DFS between the two clusters. A nomogram was established combining age, lymph node metastasis and extrathyroidal extension. The area under the curve (AUC) of predicting one-, five-, and 10-year DFS in PTC patients was 0.745, 0.801, and 0.803, respectively. The low-risk cluster showed higher levels of immune infiltration and immune checkpoint gene expression, while the high-risk cluster demonstrated a higher tumor mutation burden. Conclusion: A predictive DFS model was established, based on PRGs, which may aid in identifying patients at high risk of recurrence. The present study helps to better understand the role of pyroptosis in the progression and prognosis of PTC.

Objective: Accurately distinguishing between benign and malignant thyroid nodules (TNs) in the context of Hashimoto’s thyroiditis (HT) is challenging. This study aimed to explore the diagnostic efficacy of artificial intelligence models constructed using radiomics and deep learning (DL) features extracted from the ultrasound images of TNs in the setting of HT. The study also aimed to quantitatively compare the diagnostic performance of these models against that of fine-needle aspiration (FNA) cytology combined with BRAFV600E gene mutation testing so as to establish a superior diagnostic paradigm for TNs in HT. Methods: We analyzed the clinical data and preoperative ultrasound images of 1585 patients with HT admitted to eight hospitals in China between 1 January 2018, and 30 December 2023. Radiomics features were extracted from each manually annotated and standardized region of interest in the images. The DL features based on various convolutional neural network models were also extracted, including 11 pre-trained DL models based on real images and the image features extracted using a combined DL–radiomics (DLR) approach. Least absolute shrinkage and selection operator regression was used to select features with nonzero coefficients. Further, eight machine learning methods were employed to construct prediction models, namely the DLR models, incorporating both DL and radiomics features. The importance of features in contributing to the model was prioritized using the SHapley Additive exPlanations (SHAP) method for interpretation. The predictive performance of the models was evaluated using the area under the receiver operating characteristic curve (AUC), accuracy, sensitivity, and specificity. Two rounds of reader studies were conducted (first round: independent reading; second round: guided reading) to validate the clinical application value of the DLR model, with results compared with FNA biopsy outcomes. Results: A total of 1561 radiomics features and 256 DL features were extracted from the original images. The DLR model, leveraging the ResNet152 neural network, could effectively differentiate between benign and malignant nodules in the context of HT. The AUC, accuracy, sensitivity, and specificity of the logistic regression (LR) model, based on these features, were 0.917 [95% confidence interval (CI): 0.838–0.988], 85.7%, 81.8%, and 86.6%, respectively, in the validation cohort and 0.827(95% CI: 0.777–0.876), 93.9%, 80.9%, and 96.3%, respectively, in the external test cohort. The SHAP summary plot illustrated how feature values influenced their impact on the model, whereas the SHAP force plot showed the integrated impact of features on individual responses. Gradient-weighted class activation mapping (Grad-CAM)-generated heatmaps from DL models visually highlighted high-risk areas in HT. The DLR model outperformed five junior-level physicians in terms of diagnostic accuracy, sensitivity, and specificity in the validation cohort. The diagnostic performance of all clinicians was significantly enhanced with the assistance of the DLR model, with no statistical difference detected compared with the FNA biopsy results. Conclusions: The DLR model combined with the SHAP and Grad-CAM method can improve the diagnostic performance of radiologists in identifying benign and malignant TNs in the context of HT. The diagnostic efficacy of this visualization model is comparable to that of FNA cytology combined with gene mutation testing. The DLR model can enhance the diagnostic ability of radiologists in differentiating between benign and malignant TNs in the context of HT, thereby minimizing unnecessary biopsies. Additionally, it can aid clinicians in making personalized decisions regarding the necessity of biopsy or even surgery by providing intuitive visual explanations.

Objectives: To explore the preoperative high-risk clinical factors for contralateral medium-volume central lymph node metastasis (conMVCLNM) in unilateral papillary thyroid carcinoma (uPTC) and the indications for dissection of contralateral central lymph nodes (conCLN).Methods: Clinical and pathological data of 204 uPTC patients who underwent thyroid surgery at the Hangzhou First People’s Hospital from September 2010 to October 2022 were collected. Univariate and multivariate logistic regression analyses were conducted to determine the independent risk factors for contralateral central lymph node metastasis (conCLNM) and conMVCLNM in uPTC patients based on the preoperative clinical data. Predictive models for conCLNM and conMVCLNM were constructed using logistic regression analyses and validated using receiver operating characteristic (ROC) curves, concordance index (C-index), calibration curves, and decision curve analysis (DCA).Results: Univariate and multivariate logistic regression analyses showed that gender (P < 0.001), age (P < 0.001), tumor diameter (P < 0.001), and multifocality (P = 0.008) were independent risk factors for conCLNM in uPTC patients. Gender(P= 0.026), age (P = 0.010), platelet-to-lymphocyte ratio (PLR) (P =0.003), and tumor diameter (P = 0.036) were independent risk factors for conMVCLNM in uPTC patients. A predictive model was established to assess the risk of conCLNM and conMVCLNM, with ROC curve areas of 0.836 and 0.845, respectively. The C-index, the calibration curve, and DCA demonstrated that the model had good diagnostic value.Conclusion: Gender, age, tumor diameter, and multifocality are high-risk factors for conCLNM in uPTC patients. Gender, age, tumor diameter, and PLR are high-risk factors for conMVCLNM in uPTC patients, and preventive conCLN dissection should be performed.