Label-Free Rapid Intelligent Diagnosis of Thyroid Cancer

10020 Background: Childhood cancer survivors are at increased risk of developing a subsequent thyroid cancer, particularly following radiotherapy. In the general population, thyroid cancer has a very low mortality rate. Mortality after a diagnosis of subsequent thyroid cancer in survivors is unknown. Methods: We calculated the standardized mortality ratio (SMR) following the development of subsequent thyroid cancer in a cohort of 24,683 5-year survivors of childhood cancer diagnosed between 1970 and 1999 using the age-sex-calendar-year-specific general population all-cause mortality rates from the CDC as the reference rates. We estimated all-cause mortality post the diagnosis of thyroid cancer (time-dependent covariate), adjusting for development of other subsequent malignant neoplasms (SMN) and chronic health conditions (CHC), using a piecewise exponential model. Thyroid cancer-specific mortality among survivors was compared to SEER cases with thyroid cancer, adjusting for age, sex, race and calendar-year. SEER data was also used to compare thyroid cancer characteristics in childhood cancer survivors with thyroid cancer patients without a history of childhood cancer. Results: Among 397 survivors with subsequent thyroid cancer, 63% were female, 83% had received radiotherapy for treatment of their primary childhood cancer with fields that included the thyroid gland, and 92% had at least one severe or life-threatening chronic condition. Thyroid tumor size was significantly smaller in survivors, with 33% of cases in survivors and 24% in SEER being less than 1 cm (p < 0.001). There were 82 deaths with 7 deaths due to thyroid cancer. Within the cohort of survivors of childhood cancer, the rate of all-cause mortality did not increase with a diagnosis of thyroid cancer, adjusting for development of other SMNs and CHCs (RR = 1.0, p = 0.96), but it was 7 times higher than that of the general population (SMR = 6.9, 95% CI 5.5-8.5). Compared to adults diagnosed with thyroid cancer in the general population, survivors with subsequent thyroid cancer did not have an increased risk of thyroid cancer-specific death (RR = 0.9, 95% CI 0.4-1.9). Mortality risk was higher among those with older age at subsequent thyroid cancer diagnosis, male sex, Black and Hispanic race and ethnicity and tumor size > 1 cm. Conclusions: The rate of all-cause mortality does not increase with a diagnosis of subsequent thyroid cancer in childhood cancer survivors. This finding suggests that thyroid cancer screening in this population should be based on reducing morbidity since it likely will not provide survival benefit. Enhanced attention to CHC management may be critical for long-term survival.

Aim . To establish the diagnostic value of liquid and traditional cytological diagnosis with immunocytochemical (ICC) detection proteins (galectin-3, nucleophosmin, Ki-67) in the preoperative diagnosis of well-differentiated thyroid cancer and analysis of the effect of the autoimmune process in the thyroid gland in the diagnostic accuracy. Materials and methods . The traditional liquid-based cytology was performed on 107 samples and immunocytochemistry performed on 56 samples with histological findings: colloid goiter in 24 cases, 4 cases of Hashimoto’s thyroiditis, 4 cases of follicular adenoma, 22 cases of papillary cancer. Results . The results of a comparison of cytological and histological findings. The results of studies of diagnostic accuracy expression of galectin-3, nucleophosmin, Ki-67 for the diagnosis of thyroid cancer. A formula was developed to help determine the presence of malignancy in thyroid with high accuracy. Conclusion . The diagnostic accuracy of the method of liquid-based cytology is higher than the traditional method of cytology. ICC expression of Ki-67 method has 81.8% of sensitivity and 100% of specificity for the preoperative diagnosis of thyroid cancer. Conjoint definition HS Ki-67 and liquid-based cytology increases the sensitivity and specificity of the diagnosis of well-differentiated thyroid cancer preoperative to 100%. There no detected relations between the expression of galectin-3, NFM, Ki-67 and the presence an autoimmune process in the thyroid.

The aim of this study was to explore the osteopontin expression and microvascular density in thyroid cancer, compare computed tomography (CT) and ultrasound in diagnosis of thyroid cancer and investigate the correlations of CT Features and thyroid cancer. A total of 80 patients with thyroid masses admitted to our hospital from April 2017 to August 2019 were selected, of which there were 40 with benign tumor and 40 with malignant tumor. All patients with thyroid cancer confirmed by pathological tissue biopsy were examined by ultrasound (ultrasound group) and CT (CT group). The expression of osteopontin was detected by PCR while microvascular density was tested by immunohistochemistry. Then univariate analysis and multivariate logistic regression analysis of risk factors were carried out for CT imaging diagnosis of thyroid cancer. The levels of osteopontin and microvascular density in malignant group were significantly higher than those in benign group. The incidence rates of unclear boundary and peripheral lymph node enlargement in CT group were remarkably higher than those in ultrasound group. The diagnostic rate of masses ≥1 cm in diameter was notably higher than that of masses <1 cm in diameter in thyroid cancer patients in CT group and ultrasound group (P<0.05). In addition, the diagnostic rates of follicular carcinoma and papillary carcinoma were higher, whereas those of medullary carcinoma and undifferentiated carcinoma were lower in CT group and ultrasound group. There was no significant difference in the accuracy of thyroid cancer diagnosis between CT group and ultrasound group. Moreover, diameter ≥1 cm, irregular shape, unclear boundary, calcified foci, uneven density/echo and peripheral lymph node enlargement were related risk factors for the CT imaging diagnosis of thyroid cancer, in which irregular shape, unclear boundary, calcified foci and uneven density/echo were independent risk factors for the CT imaging diagnosis of thyroid cancer. The levels of osteopontin and microvascular density were increased in thyroid cancer. CT examination may be of higher diagnostic value in diagnosis of thyroid cancer compared with ultrasound. Irregular shape, unclear boundary, calcified foci, and uneven density/echo were independent risk factors for the CT imaging diagnosis of thyroid cancer.

The aim of this study was to explore the osteopontin expression and microvascular density in thyroid cancer, compare computed tomography (CT) and ultrasound in diagnosis of thyroid cancer and investigate the correlations of CT Features and thyroid cancer. A total of 80 patients with thyroid masses admitted to our hospital from April 2017 to August 2019 were selected, of which there were 40 with benign tumor and 40 with malignant tumor. All patients with thyroid cancer confirmed by pathological tissue biopsy were examined by ultrasound (ultrasound group) and CT (CT group). The expression of osteopontin was detected by PCR while microvascular density was tested by immunohistochemistry. Then univariate analysis and multivariate logistic regression analysis of risk factors were carried out for CT imaging diagnosis of thyroid cancer. The levels of osteopontin and microvascular density in malignant group were significantly higher than those in benign group. The incidence rates of unclear boundary and peripheral lymph node enlargement in CT group were remarkably higher than those in ultrasound group. The diagnostic rate of masses ≥1 cm in diameter was notably higher than that of masses <1 cm in diameter in thyroid cancer patients in CT group and ultrasound group (P<0.05). In addition, the diagnostic rates of follicular carcinoma and papillary carcinoma were higher, whereas those of medullary carcinoma and undifferentiated carcinoma were lower in CT group and ultrasound group. There was no significant difference in the accuracy of thyroid cancer diagnosis between CT group and ultrasound group. Moreover, diameter ≥1 cm, irregular shape, unclear boundary, calcified foci, uneven density/echo and peripheral lymph node enlargement were related risk factors for the CT imaging diagnosis of thyroid cancer, in which irregular shape, unclear boundary, calcified foci and uneven density/echo were independent risk factors for the CT imaging diagnosis of thyroid cancer. The levels of osteopontin and microvascular density were increased in thyroid cancer. CT examination may be of higher diagnostic value in diagnosis of thyroid cancer compared with ultrasound. Irregular shape, unclear boundary, calcified foci and uneven density/echo were independent risk factors for the CT imaging diagnosis of thyroid cancer.

Epidemiology of Thyroid Cancer.

228 Background: In a multicenter cohort of over 1,200 patients, we have shown that thyroid cancer survivors report a decrease in quality of life (QoL) that is significant and equal to other common cancers. While these findings are striking, there have been no studies assessing physician perceptions regarding how a diagnosis of thyroid cancer affects QoL. We hypothesize that there is a large discrepancy between what physicians perceive the diagnosis and treatment of thyroid cancer has on QoL and what patients report. Methods: Physicians were recruited from two national organizations. A 37-question survey was administered evaluating demographic and treatment-related information as well as questions specific to perceptions of patient satisfaction with various aspects of treatment, complications and overall effects on QoL. Responses were categorized into physical, psychological, social and outcomes. Regression analysis was performed to determine trends in physician perceptions. Results: 105 physicians completed the survey. The majority of participants were endocrine surgeons (n=59, 56.2%), followed by endocrinologists (n= 31, 29.5%), general surgeons (n=8, 7.6%), otolaryngologists (n=2, 1.9%), and medical oncologists (n=1, 1%). The mean number of years in practice was 18.78 years (n=80). There was no difference in response by physician specialty. There was no difference in physician perception on psychological and social effects of thyroid cancer compared to those reported by patients (p=0.643). However, regarding physical concerns, physicians report significantly lower rates of physical concerns than reported by patients with regards to fatigue (p&lt;0.01), weight gain (p&lt;0.01), voice changes (p=0.0427), and heat/cold intolerance (p&lt;0.01). Conclusions: There is a discrepancy between physician perceptions on the effect of thyroid cancer on QoL in thyroid cancer survivors and what these survivors report. As such, there needs to be greater awareness about the effects of thyroid cancer on QoL. In order to better assess and monitor QoL in thyroid cancer survivors, a validated thyroid cancer-specific survivorship care plan incorporating these findings should be created to better assess and monitor patients with a diagnosis of thyroid cancer.

Background/Objectives: Thyroid cancer is the most common endocrine malignancy and accounts for 1% of cancers. In recent years, there has been much interest in the feasibility of using miRNAs or miRNA panels as biomarkers for the diagnosis of thyroid cancer. miRNAs are noncoding RNAs with 21–23 nucleotides that are highly conserved during evolution. They have been proposed as regulators of gene expression, apoptosis, cancer, and cell growth and differentiation. Methods: The Directory of Open Access Journals (DOAJ), Google Scholar, PubMed (NLM), LISTA (EBSCO), and Web of Science were searched. Results: The serum level of miRNAs (miRNA-375, 34a, 145b, 221, 222, 155, Let-7, 181b) can be used as molecular markers for the diagnosis and prognosis of thyroid cancer in the serum samples of patients with thyroid glands. Conclusions: Given that most common methods for the screening of thyroid cancer cannot detect the disease in its early stages, identifying miRNAs that are released in the bloodstream during the gradual progression of the disease is considered a key method in the early diagnosis of thyroid cancers.

In this review we have brought forward various nuclear imaging modalities used in the diagnosis, staging, and management of thyroid cancer. Thyroid cancer is the most common endocrine malignancy, accounting for approximately 3% of all new cancer diagnoses. Nuclear imaging plays an important role in the evaluation of thyroid cancer, and the use of radioiodine imaging, FDG imaging, and somatostatin receptor imaging are all valuable tools in the management of this disease. Radioiodine imaging involves the use of Iodine-123 [I-123] or Iodine-131 [I-131] to evaluate thyroid function and detect thyroid cancer. I-123 is a gamma-emitting isotope that is used in thyroid imaging to evaluate thyroid function and detect thyroid nodules. I-131 is a beta-emitting isotope that is used for the treatment of thyroid cancer. Radioiodine imaging is used to detect the presence of thyroid nodules and evaluate thyroid function. FDG imaging is a PET imaging modality that is used to evaluate the metabolic activity of thyroid cancer cells. FDG is a glucose analogue that is taken up by cells that are metabolically active, such as cancer cells. FDG PET/CT can detect primary thyroid cancer and metastatic disease, including lymph nodes and distant metastases. FDG PET/CT is also used to monitor treatment response and detect the recurrence of thyroid cancer. Somatostatin receptor imaging involves the use of radiolabeled somatostatin analogues to detect neuroendocrine tumors, including thyroid cancer. Radiolabeled somatostatin analogues, such as Indium-111 octreotide or Gallium-68 DOTATATE, are administered to the patient, and a gamma camera is used to detect areas of uptake. Somatostatin receptor imaging is highly sensitive and specific for the detection of metastatic thyroid cancer. A comprehensive search of relevant literature was done using online databases of PubMed, Embase, and Cochrane Library using the keywords "thyroid cancer," "nuclear imaging," "radioiodine imaging," "FDG PET/CT," and "somatostatin receptor imaging" to identify relevant studies to be included in this review. Nuclear imaging plays an important role in the diagnosis, staging, and management of thyroid cancer. The use of radioiodine imaging, thyroglobulin imaging, FDG imaging, and somatostatin receptor imaging are all valuable tools in the evaluation of thyroid cancer. With further research and development, nuclear imaging techniques have the potential to improve the diagnosis and management of thyroid cancer and other endocrine malignancies.

Intracranial metastases from thyroid cancer are rare. Although the prognosis of thyroid cancer patients is generally favorable, the prognosis of patients with intracranial metastases from thyroid cancer has been considered unfavorable owing to lower survival rates among such patients compared to those without intracranial involvement. Many questions about their management remain unclear. The aim of the present study was to analyze the characteristics, treatment modalities, and outcomes of patients with brain metastases from thyroid cancer. Among 4320 patients with thyroid cancer recorded in our institutional database over a 30-year period, the data of 20 patients with brain metastasis were retrospectively collected and analyzed. The clinical characteristics, histological type of primary cancer and metastatic brain tumor, additional previous distant metastasis, treatment modalities, locations and characteristics on radiologic findings, time interval between the first diagnosis of primary thyroid cancer and brain metastasis, and survival were analyzed. Among our patient cohort, the mean age at initial diagnosis was 59.3 ± 14.1 years, and at the manifestation of diagnosis of cerebral metastasis, the mean age was found to be 64.8 ± 14.9 years. The histological types of primary thyroid cancer were identified as papillary in ten patients, follicular in seven, and poorly differentiated carcinoma in three. The average interval between the diagnosis of thyroid cancer and brain metastasis was 63.4 ± 58.4 months (range: 0-180 months). Ten patients were identified as having a single intracranial lesion, and ten patients were found to have multiple lesions. Surgical resection was primarily performed in fifteen patients, and whole-brain radiotherapy, radiotherapy, or tyrosine kinase inhibitors were applied in the remaining five patients. The overall median survival time was 15 months after the diagnosis of BMs from TC (range: 1-252 months). Patients with thyroid cancer can develop brain metastasis even many years after the diagnosis of the primary tumor. The results of our study demonstrate increased overall survival in patients younger than 60 years of age at the time of diagnosis of brain metastasis. There was no difference in survival between patients with brain metastasis from papillary carcinoma and those with follicular thyroid carcinoma.

Background:: Ultrasound test is one of the routine tests for the diagnosis of thyroid cancer. The diagnosis accuracy depends largely on the correct interpretation of ultrasound images of thyroid nodules. However, human eye-based image recognition is usually subjective and sometimes error-prone especially for less experienced doctors, which presents a need for computeraided diagnostic systems. Objective: : To our best knowledge, there is no well-maintained ultrasound image database for the Chinese population. In addition, though there are several computational methods for image-based thyroid cancer detection, a comparison among them is missing. Finally, the effects of features like the choice of distance measures have not been assessed. The study aims to give the improvement of these limitations and proposes a highly accurate image-based thyroid cancer diagnosis system, which can better assist doctors in the diagnosis of thyroid cancer. Methods:: We first establish a novel thyroid nodule ultrasound image database consisting of 508 images collected from the Third Hospital of Hebei Medical University in China. The clinical information for the patients is also collected from the hospital, where 415 patients are diagnosed to be benign and 93 are malignant by doctors following a standard diagnosis procedure. We develop and apply five machine learning methods to the dataset including deep neural network, support vector machine, the center clustering method, k-nearest neighbor, and logistic regression. Results:: Experimental results show that deep neural network outperforms other diagnosis methods with an average cross-validation accuracy of 0.87 in 10 runs. Meanwhile, we also explore the performance of four image distance measures including the Euclidean distance, the Manhattan distance, the Chebyshev distance, and the Minkowski distance, among which the Chebyshev distance is the best. The resource can be directly used to aid doctors in thyroid cancer diagnosis and treatment. Conclusions: : The paper establishes a novel thyroid nodule ultrasound image database and develops a high accurate image-based thyroid cancer diagnosis system which can better assist doctors in the diagnosis of thyroid cancer.

We are concerned that the momentum gained toward preventing the overdiagnosis of thyroid cancer will decrease because of the encouragement of studies concerning environmental risk factors, as discussed in the article by Bernier et al.1 It is important to note that in an editorial, Chen and Davies2 mentioned that there is a need to examine whether the increase in pediatric thyroid cancer is attributable to overdiagnosis, even though there might be new genetic, environmental, or dietary causes of thyroid cancer. Furthermore, Goldenberg3 indicated that overdiagnosis absolves researchers of the duty to seek causes of the real increase in thyroid cancer. Overdiagnosis is a serious problem that should be considered during the process of diagnosing cancer. Every clinician and scientist should be aware that the diagnosis of indolent cancer in a patient's lifetime (overdiagnosis) leads to physical and psychological burdens, as well as individual and social economic costs during the process of diagnostic examinations and subsequent potential overtreatment. Living as a patient with cancer over a lifetime involves not only anxiety and unhappiness, but also amplified public risk perception of the disease. In addition, there are social disadvantages and discrimination faced when obtaining employment, insurance coverage, and a marriage partner.4, 5 The harm of overdiagnosis in young people is more serious because of their longer lifespan,5 and the experience of overdiagnosis can negatively affect decision making in their lives. These disadvantages must be avoided if cancer is overdiagnosed. Even in the case of a real increase in the incidence of pediatric thyroid cancer without overdiagnosis, extremely early diagnosis also is considered to be harmful. This is because of the good prognosis of patients diagnosed with pediatric thyroid cancer and the possibility of growth arrest in thyroid cancer in young individuals.6 We have no evidence that the extremely early diagnosis of thyroid cancer decreases mortality. Therefore, an extremely early diagnosis could have harms similar to those of overdiagnosis (physical, psychological, economic, and social issues). Extremely early diagnosis makes patients experience such problems over a longer lifetime. However, medical professionals and the public tend to regard early diagnosis as a good thing. There is a misinterpretation that an early diagnosis would be more beneficial than harmful for any type of cancer. Therefore, we need to consider the balance between benefits and harms in the early diagnosis of pediatric thyroid cancer. Bernier et al1 described the importance of investigating environmental, dietary, and genetic factors as contributors to the rising incidence of pediatric thyroid cancer. This would require large-scale epidemiological studies because of the rarity of these cancers. In Fukushima, an epidemiological survey demonstrated that using highly sensitive ultrasonography for thyroid cancer screening led to overdiagnosis and/or extremely early diagnosis in the young population.7 Surveys should be implemented with special notes made of the conditions for which benefits overweigh harms by avoiding overdiagnosis and extremely early diagnosis, and also by effective communication with patients and their guardians. We strongly hope that the article by Bernier et al1 will be useful for the prevention of overdiagnosis and the promotion of primary prevention without the extremely early diagnosis of pediatric thyroid cancer. No specific funding was disclosed. The authors made no disclosures.

Thyroid diseases are divided into papillary carcinoma and nodular diseases, which are very harmful to the human body. Ultrasound is a common diagnostic method for thyroid diseases. In the process of diagnosis, doctors need to observe the characteristics of ultrasound images, combined with professional knowledge and clinical experience, to give the disease situation of patients. However, different doctors have different clinical experience and professional backgrounds, and the diagnosis results lack objectivity and consistency, so an intelligent diagnosis technology for thyroid diseases based on the ultrasound image is needed in clinic, which can give objective and reliable diagnosis opinions on thyroid diseases by extracting the texture, shape, and other information of the image and assist doctors in clinical diagnosis. This paper mainly studies the intelligent ultrasonic diagnosis of papillary thyroid cancer based on machine learning, compares the ultrasonic characteristics of PTMC diagnosed by using the new ultrasound technology (CEUS and UE), and summarizes the differential diagnosis effect and clinical application value of the two technology methods for PTMC. In this paper, machine learning, diffuse thyroid image features, and RBM learning methods are used to study the ultrasonic intelligent diagnosis of papillary thyroid cancer based on machine learning. At the same time, the new contrast-enhanced ultrasound (CEUS) technology and ultrasound elastography (UE) technology are used to obtain the experimental phenomena in the experiment of ultrasonic intelligent diagnosis of papillary thyroid cancer. The results showed that 90% of the cases were diagnosed by contrast-enhanced ultrasound and confirmed by postoperative pathology. CEUS and UE have reliable practical value in the diagnosis of PTMC, and the combined application of CEUS and UE can improve the sensitivity and accuracy of PTMC diagnosis.

<h3>Objectives</h3> Overdiagnosis, diagnosis of a cancer that would not ultimately cause symptoms or death, has become relatively common in a variety of cancers, including thyroid cancer (ThCa). The trend of ThCa incidence, which has increased substantially without a corresponding change in mortality, typifies overdiagnosis. An analysis of Surveillance, Epidemiology, and End Results (SEER) program data has suggested that overdiagnosis may be responsible for up to 60% of diagnosed papillary ThCa cases, the most common histotype of ThCa. Our main objective is to assess whether thyroid abnormalities discovered incidentally on imaging studies carried out in the National Lung Screening Trial (NLST) are associated with overdiagnosis. <h3>Method</h3> The National Lung Screening Trial (NLST), a randomized trial of almost 54 000 current and former smokers that compared low-dose computed tomography (LDCT) to chest radiography (CXR) for the early detection of lung cancer, is a resource to investigate the potential impact of incidental overdiagnosis of ThCa even when screening for an unrelated cancer in anatomic proximity to the thyroid gland. Previously, Pinsky and colleagues used NLST data to examine renal-related abnormalities outside the intended lung field and found that renal tumors could be incidentally detected by LDCT. A similar phenomenon can be investigated in ThCa, examining the upper portion of the LDCT field. <h3>Results</h3> Preliminary analyses show more diagnosed thyroid cancers in the LDCT as compared to CXR arm during the screening phase (first 3 years) of the trial (23 versus 11) and overall (35 versus 25). We will examine the relationship of reported thyroid abnormalities on LDCT and subsequent diagnoses of thyroid cancer. Specifically, we will assess the proportion of thyroid cancers diagnosed within one year of an LDCT screen that had a reported thyroid abnormality, as well as the overall rate of thyroid abnormalities seen on LDCT screens. We will also examine survival of thyroid cancers by trial arm and mode of diagnosis (observed on LDCT scan or not). <h3>Conclusions</h3> Preliminary data suggest that more thyroid cancers were diagnosed in the LDCT than in the CXR arm. This suggests a possible association of overdiagnosis of ThCa with the more sensitive screening technology (LDCT).

Rationale and Objectives: Accurately establishing the diagnosis and staging of cervical and thyroid cancers is essential in medical practice in determining tumor extension and dissemination and involves the most accurate and effective therapeutic approach. For accurate diagnosis and staging of cervical and thyroid cancers, we aim to create a diagnostic method, optimized by the algorithms of artificial intelligence and validated by achieving accurate and favorable results by conducting a clinical trial, during which we will use the diagnostic method optimized by artificial intelligence (AI) algorithms, to avoid errors, to increase the understanding on interpretation computer tomography (CT) scan, magnetic resonance imaging (MRI) of the doctor and improve therapeutic planning. Materials and Methods: The optimization of the computer assisted diagnosis (CAD) method will consist in the development and formation of artificial intelligence models, using algorithms and tools used in segmental volumetric constructions to generate 3D images from MRI/CT. We propose a comparative study of current developments in “DICOM” image processing by volume rendering technique, the use of the transfer function for opacity and color, shades of gray from “DICOM” images projected in a three-dimensional space. We also use artificial intelligence (AI), through the technique of Generative Adversarial Networks (GAN), which has proven to be effective in representing complex data distributions, as we do in this study. Validation of the diagnostic method, optimized by algorithm of artificial intelligence will consist of achieving accurate results on diagnosis and staging of cervical and thyroid cancers by conducting a randomized, controlled clinical trial, for a period of 17 months. Results: We will validate the CAD method through a clinical study and, secondly, we use various network topologies specified above, which have produced promising results in the tasks of image model recognition and by using this mixture. By using this method in medical practice, we aim to avoid errors, provide precision in diagnosing, staging and establishing the therapeutic plan in cancers of the cervix and thyroid using AI. Conclusion: The use of the CAD method can increase the quality of life by avoiding intra and postoperative complications in surgery, intraoperative orientation and the precise determination of radiation doses and irradiation zone in radiotherapy.

Background Incidence rate of thyroid cancer is steadily increasing due to overdiagnosis and overtreatment. Thyroid ultrasound is commonly used to diagnose thyroid cancer. The aim of this study is to examine the accuracy of using deep convolutional neural network (DCNN) models to improve diagnosis of thyroid cancer by analyzing sonographic imaging data from clinical thyroid ultrasound. Methods A total of 131,731 sonographic images from 17,627 thyroid cancer patients and 180,668 sonographic images from 25,325 controls used as training set were obtained from Tianjin Cancer Hospital. Images from anatomical sites that did not have cancer according to location sign on the image were not included. All thyroid cancer patients and 13·2% of controls (51,255 images) were confirmed by pathological reports. DCNN is a specific type of neural network optimized for image recognition. We trained two DCNN models on the training set and subsequently evaluated the performance on one independent internal (Tianjin, 1,118 individuals) and two external (Jilin,154 individuals; Weihai, 1,420 individuals) validation sets. Individuals in the validation sets all have pathological examinations. We compared the specificity/sensitivity of DCNN models with the performance of six thyroid ultrasound radiologists on these three validation sets. Findings DCNN model achieved high performance in identifying thyroid cancer patients versus six experience radiologists: for Tianjin validation set, sensitivity was 92·2% versus 96·9% (95% CI 89·7% - 94·3% vs. 93·9% - 98·6%; p = 0·003), and specificity was 85·6% versus 59·4% (95% CI 82·4% - 88·4% vs. 53% - 65·6%; p &amp;lt; 0·0001); for Jilin validation set, sensitivity was 84·3% versus 92·9% (95% CI 73·6% - 91·9% vs. 84·1% - 97·6%; p = 0·05), and specificity was 86·9% versus 57·1% (95% CI 77·8% - 93·3% vs. 45·9% - 67·9%; p &amp;lt; 0·0001); for Weihai validation set, sensitivity was 84·5% versus 89% (95% CI 81·2% - 87·4% vs. 81·9% - 94%; p = 0·2), and specificity was 87·5% versus 68·6% (95% CI 85·1% - 89·6% vs. 60·7% - 75·8%; p &amp;lt; 0·0001). Interpretation DCNN models exhibited high accuracy, sensitivity, and specificity in identifying thyroid cancer patients at levels comparable to or higher than six experienced radiologists. Conferred by the high specificity of DCNN models, the rate of overdiagnosis and overtreatment of patients with thyroid cancer is expected to decrease. This supports future application of the deep learning models to clinical practice for thyroid cancer diagnosis. However, further validation of these DCNN models in prospective clinical trials is warranted. Funding The Program for Changjiang Scholars and Innovative Research Team in University in China (IRT_14R40), National Natural Science Foundation of China (31801117). Citation Format: Xiangchun Li, Sheng Zhang, Qiang Zhang, Xi Wei, Yi Pan, Jing Zhao, Xiaojie Xin, Xiaoqing Wang, Fan Yang, Jianxin Li, Meng Yang, Qinghua Wang, Xiangqian Zheng, Yanhui Zhao, Lun Zhang, Xudong Wang, Zhimin Zheng, Christopher T. Whitlow, Metin N. Gurcan, Boris Pasche, Ming Gao, Wei Zhang, Kexin Chen. Diagnosis of thyroid cancer using deep convolutional neural network models applied to sonographic images from clinical ultrasound exams [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 1394.