logo logo
Deep Learning in Thyroid Ultrasonography to Predict Tumor Recurrence in Thyroid Cancers. Taehan Yongsang Uihakhoe chi Purpose:To evaluate a deep learning model to predict recurrence of thyroid tumor using preoperative ultrasonography (US). Materials and Methods:We included representative images from 229 US-based patients (male:female = 42:187; mean age, 49.6 years) who had been diagnosed with thyroid cancer on preoperative US and subsequently underwent thyroid surgery. After selecting each representative transverse or longitudinal US image, we created a data set from the resulting database of 898 images after augmentation. The Python 2.7.6 and Keras 2.1.5 framework for neural networks were used for deep learning with a convolutional neural network. We compared the clinical and histological features between patients with and without recurrence. The predictive performance of the deep learning model between groups was evaluated using receiver operating characteristic (ROC) analysis, and the area under the ROC curve served as a summary of the prognostic performance of the deep learning model to predict recurrent thyroid cancer. Results:Tumor recurrence was noted in 49 (21.4%) among the 229 patients. Tumor size and multifocality varied significantly between the groups with and without recurrence ( < 0.05). The overall mean area under the curve (AUC) value of the deep learning model for prediction of recurrent thyroid cancer was 0.9 ± 0.06. The mean AUC value was 0.87 ± 0.03 in macrocarcinoma and 0.79 ± 0.16 in microcarcinoma. Conclusion:A deep learning model for analysis of US images of thyroid cancer showed the possibility of predicting recurrence of thyroid cancer. 10.3348/jksr.2019.0147
Deep learning predicts cervical lymph node metastasis in clinically node-negative papillary thyroid carcinoma. Insights into imaging OBJECTIVES:Precise 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. METHODS:Preoperative 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. RESULTS:The 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). CONCLUSIONS:Deep 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 STATEMENT:Deep 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. 10.1186/s13244-023-01550-2
Prediction of lymph node metastasis in patients with papillary thyroid cancer based on radiomics analysis and intraoperative frozen section analysis: A retrospective study. Clinical otolaryngology : official journal of ENT-UK ; official journal of Netherlands Society for Oto-Rhino-Laryngology & Cervico-Facial Surgery INTRODUCTION:To evaluate the diagnostic efficiency among the clinical model, the radiomics model and the nomogram that combined radiomics features, frozen section (FS) analysis and clinical characteristics for the prediction of lymph node (LN) metastasis in patients with papillary thyroid cancer (PTC). METHODS:A total of 208 patients were randomly divided into two groups randomly with a proportion of 7:3 for the training groups (n = 146) and the validation groups (n = 62). The Least Absolute Shrinkage and Selection Operator (LASSO) regression was used for the selection of radiomics features extracted from ultrasound (US) images. Univariate and multivariate logistic analyses were used to select predictors associated with the status of LN. The clinical model, radiomics model and nomogram were subsequently established by logistic regression machine learning. The area under the curve (AUC), sensitivity and specificity were used to evaluate the diagnostic performance of the different models. The Delong test was used to compare the AUC of the three models. RESULTS:Multivariate analysis indicated that age, size group, Adler grade, ACR score and the psammoma body group were independent predictors of lymph node metastasis (LNM). The results showed that in both the training and validation groups, the nomogram showed better performance than the clinical model, albeit not statistically significant (p > .05), and significantly outperformed the radiomics model (p < .05). However, the nomogram exhibits a slight improvement in sensitivity that could reduce the incidence of false negatives. CONCLUSION:We propose that the nomogram holds substantial promise as an effective tool for predicting LNM in patients with PTC. 10.1111/coa.14162
Predicting central cervical lymph node metastasis in papillary thyroid carcinoma with Hashimoto's thyroiditis: a practical nomogram based on retrospective study. PeerJ Background:In papillary thyroid carcinoma (PTC) patients with Hashimoto's thyroiditis (HT), preoperative ultrasonography frequently reveals the presence of enlarged lymph nodes in the central neck region. These nodes pose a diagnostic challenge due to their potential resemblance to metastatic lymph nodes, thereby impacting the surgical decision-making process for clinicians in terms of determining the appropriate surgical extent. Methods:Logistic regression analysis was conducted to identify independent risk factors associated with central lymph node metastasis (CLNM) in PTC patients with HT. Then a prediction model was developed and visualized using a nomogram. The stability of the model was assessed using ten-fold cross-validation. The performance of the model was further evaluated through the use of ROC curve, calibration curve, and decision curve analysis. Results:A total of 376 HT PTC patients were included in this study, comprising 162 patients with CLNM and 214 patients without CLNM. The results of the multivariate logistic regression analysis revealed that age, Tg-Ab level, tumor size, punctate echogenic foci, and blood flow grade were identified as independent risk factors associated with the development of CLNM in HT PTC. The area under the curve (AUC) of this model was 0.76 (95% CI [0.71-0.80]). The sensitivity, specificity, accuracy, and positive predictive value of the model were determined to be 88%, 51%, 67%, and 57%, respectively. Conclusions:The proposed clinic-ultrasound-based nomogram in this study demonstrated a favorable performance in predicting CLNM in HT PTCs. This predictive tool has the potential to assist clinicians in making well-informed decisions regarding the appropriate extent of surgical intervention for patients. 10.7717/peerj.17108
A CT based radiomics analysis to predict the CN0 status of thyroid papillary carcinoma: a two- center study. Cancer imaging : the official publication of the International Cancer Imaging Society OBJECTIVES:To develop and validate radiomics model based on computed tomography (CT) for preoperative prediction of CN0 status in patients with papillary thyroid carcinoma (PTC). METHODS:A total of 548 pathologically confirmed LNs (243 non-metastatic and 305 metastatic) two distinct hospitals were retrospectively assessed. A total of 396 radiomics features were extracted from arterial-phase CT images, where the strongest features containing the most predictive potential were further selected using the least absolute shrinkage and selection operator (LASSO) regression method. Delong test was used to compare the AUC values of training set, test sets and cN0 group. RESULTS:The Rad-score showed good discriminating performance with Area Under the ROC Curve (AUC) of 0.917(95% CI, 0.884 to 0.950), 0.892 (95% CI, 0.833 to 0.950) and 0.921 (95% CI, 868 to 0.973) in the training, internal validation cohort and external validation cohort, respectively. The test group of CN0 with a AUC of 0.892 (95% CI, 0.805 to 0.979). The accuracy was 85.4% (sensitivity = 81.3%; specificity = 88.9%) in the training cohort, 82.9% (sensitivity = 79.0%; specificity = 88.7%) in the internal validation cohort, 85.4% (sensitivity = 89.7%; specificity = 83.8%) in the external validation cohort, 86.7% (sensitivity = 83.8%; specificity = 91.3%) in the CN0 test group.The calibration curve demonstrated a significant Rad-score (P-value in H-L test > 0.05). The decision curve analysis indicated that the rad-score was clinically useful. CONCLUSIONS:Radiomics has shown great diagnostic potential to preoperatively predict the status of cN0 in PTC. 10.1186/s40644-024-00690-y
Natural History and Prognostic Model of Untreated Papillary Thyroid Cancer: A SEER Database Analysis. Cancer control : journal of the Moffitt Cancer Center PURPOSE:This investigation leveraged the SEER database to delve into the progression patterns of PTC when left untreated. Furthermore, it aimed to devise and authenticate a nomogram for prognosis prediction for such patients. METHODS:We extracted data from the SEER database, focusing on PTC-diagnosed individuals from 2004-2020. To discern disease progression intervals, median survival times across stages were gauged, and the disease progression time was estimated by subtracting the median survival time of a more severe stage from its preceding stage. Prognostic determinants in the training set were pinpointed using both univariate and multivariate Cox regression. Using these determinants, a prognostic nomogram was crafted. RESULTS:In untreated PTC patients, those in stages I and II had a favorable prognosis, with 10-year overall survival rates of 86.34% and 66.03%, respectively. Patients in stages III and IV had a relatively poorer prognosis. The median survival time of stage III, stage IVA, stage IVB and stage IVC patients was 108months, 43 months, 20 months and 8 months, respectively. The deduced progression intervals from stages III-IVC were 65, 23, and 12 months. In the training set, age, tumor stage, gender, and marital status were identified as independent risk factors influencing the prognosis of untreated PTC, and a nomogram was constructed using these variables. CONCLUSION:In the absence of treatment intervention, early-stage PTC progressed slowly with an overall favorable prognosis. However, in mid to advanced-stage PTC, as tumor stage increased, disease progression accelerated, and prognosis gradually worsened. Age, tumor stage, marital status, and gender were independent risk factors influencing the prognosis of untreated PTC, and the nomogram based on these factors demonstrated good prognostic capability. 10.1177/10732748241253956
Predicting tall-cell subtype of papillary thyroid carcinomas independently with preoperative multimodal ultrasound. The British journal of radiology OBJECTIVES:This study aimed to explore the differences between tall-cell subtype of papillary thyroid carcinoma (TCPTC) and classical papillary thyroid carcinoma (cPTC) using multimodal ultrasound, and identify independent risk factors for TCPTC to compensate the deficiency of preoperative cytological and molecular diagnosis on PTC subtypes. METHODS:46 TCPTC patients and 92 cPTC patients were included. Each patient received grey-scale ultrasound, color Dopplor flow imaging (CDFI) and shear-wave elastography (SWE) preoperatively. Clinicopathologic information, grey-scale ultrasound features, CDFI features and SWE features of 98 lesion were compared using univariate analysis to find out predictors of TCPTC, based on which, a predictive model was built to differentiate TCPTC from cPTC and validated with 40 patients. RESULTS:Univariate and multivariate analysis identified that extrathyroidal extension (OR, 15.12; 95% CI, 2.26-115.44), aspect ratio (≥0.91) (OR, 29.34; 95% CI, 1.29-26.23), and maximum diameter ≥ 14.6 mm (OR, 20.79; 95% CI, 3.87-111.47) were the independent risk factors for TCPTC. Logistic regression equation: p = 1/1+ExpΣ[-5.099 + 3.004 × (if size ≥14.6 mm)+2.957 × (if aspect ratio≥0.91)+2.819 × (if extra-thyroidal extension)]. The prediction model had a good discrimination performance for TCPTC: the AUC, sensitivity and specificity were 0.928, 0.848 and 0.954 in cohort 1, and the corresponding values in cohort 2 were 0.943, 0.923 and 0.926. CONCLUSION:Ultrasound has potential for differential diagnosis of TCPTC from cPTC. A prediction model based on ultrasound characteristics (extrathyroidal extension, aspect ratio ≥0.91, and maximum diameter ≥14.6 mm) was useful to predict TCPTC. ADVANCES IN KNOWLEDGE:Multimodal ultrasound prediction of TCPTC were supplements to preoperative cytological diagnosis and molecular diagnosis on PTC subtypes. 10.1093/bjr/tqae103
Which Ultrasound Characteristics Predict Lymphatic Spread of Papillary Thyroid Cancer? The Journal of surgical research INTRODUCTION:The 2015 American Thyroid Association guidelines recommend lymph node mapping US in patients with definitive cytological evidence of thyroid cancer. Suspicious lymph node features on imaging including enlarged size (>1 cm in any dimension), architectural distortion, loss of fatty hilum, and microcalcifications often prompt evaluation with fine needle aspiration. There is no universally agreed upon model for determining which ultrasound characteristics most strongly correlate with metastatic disease. METHODS:A retrospective review of patients with confirmed papillary thyroid cancer (PTC) undergoing lymph node mapping ultrasound from 2013 to 2019 was performed. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value were calculated for each individual ultrasound characteristic as well as for characteristic combinations. RESULTS:Data from 119 lymph nodes were included. Malignant lymph nodes were more likely to be enlarged (71% versus 61%, P < 0.001) and to have each individual suspicious feature. Loss of fatty hilum had the highest sensitivity (89%) but was not specific (19%) for metastatic disease. Architectural distortion was found to have the highest specificity (87%). A combination of the four features was found to have higher specificity (97%) and PPV (88%) than any individual feature or combination of two/three features. CONCLUSIONS:A combination of four sonographic features correlates with metastatic PTC to lymph nodes and has the highest specificity and PPV for malignancy. A risk stratification model based on these features may lead to better classification of ultrasound findings in PTC patients with concern for nodal metastases. 10.1016/j.jss.2024.04.047
Machine learning algorithms for identifying contralateral central lymph node metastasis in unilateral cN0 papillary thyroid cancer. Frontiers in endocrinology Purpose:The incidence of thyroid cancer is growing fast and surgery is the most significant treatment of it. For patients with unilateral cN0 papillary thyroid cancer whether to dissect contralateral central lymph node is still under debating. Here, we aim to provide a machine learning based prediction model of contralateral central lymph node metastasis using demographic and clinical data. Methods:2225 patients with unilateral cN0 papillary thyroid cancer from Wuhan Union Hospital were retrospectively studied. Clinical and pathological features were compared between patients with contralateral central lymph node metastasis and without. Six machine learning models were constructed based on these patients and compared using accuracy, sensitivity, specificity, area under the receiver operating characteristic and decision curve analysis. The selected models were then verified using data from Differentiated Thyroid Cancer in China study. All statistical analysis and model construction were performed by R software. Results:Male, maximum diameter larger than 1cm, multifocality, ipsilateral central lymph node metastasis and younger than 50 years were independent risk factors of contralateral central lymph node metastasis. Random forest model performed better than others, and were verified in external validation cohort. A web calculator was constructed. Conclusions:Gender, maximum diameter, multifocality, ipsilateral central lymph node metastasis and age should be considered for contralateral central lymph node dissection. The web calculator based on random forest model may be helpful in clinical decision. 10.3389/fendo.2024.1385324