Computer-aided diagnosis of papillary thyroid carcinoma based on deep learning technology.

Background:Previous studies have demonstrated that ultrasonography is the recommended imaging modality for preoperative staging of papillary thyroid carcinomas (PTCs). However, only a few studies have kept watch on preoperative evaluation of capsular invasion (CI) or extracapsular extension (ECE) and cervical lymph node metastasis using preoperative ultrasonography. This study aimed to investigate the relationship between the CI or ECE and the cervical lymph node metastasis in PTCs using preoperative ultrasonography and postoperative pathology in Chinese patients.Methods:The data of preoperative ultrasonography and postoperative pathology of 166 patients who had a definitive diagnosis of PTCs from October 2011 to July 2014 at Xuanwu Hospital, Beijing were collected and reviewed. Preoperative ultrasonic parameters of thyroid nodules were compared with those of postoperative pathological diagnoses. All the patients were divided into bilateral PTCs group (n = 42, 78 nodules) and unilateral PTCs group (n = 124, 124 nodules), and the data of the nodule sizes, CI or ECE, and cervical lymph node metastasis by preoperative ultrasonography were compared between two groups.Results:A total of 202 nodules of 166 patients which were confirmed by preoperative ultrasonography and postoperative pathology were analyzed. Hypoechogenicity (n = 201, 99.5%) and irregular margins (n = 167, 82.7%) were the main ultrasonic characteristics of PTCs. A significant moderate agreement between preoperative ultrasonic examination and postoperative pathology for CI or ECE (κ = 0.622, P < 0.001) was observed. The diagnostic sensitivity was 92.0%, and specificity was 71.1%. In bilateral PTCs group, 81.0% had CI or ECE, and 61.9% had cervical lymph node metastasis. In unilateral PTCs group, 76.6% had CI or ECE, and 58.1% had cervical lymph node metastasis. There were no significant differences in the incidence of CI or ECE and cervical lymph node metastasis between two groups (all P > 0.05).Conclusions:Ultrasonography was proved to be a valuable method for preoperative diagnosis of PTCs. Hypoechogenicity and irregular margins were strongly associated with PTCs. CI or ECE in unilateral PTCs strongly implied the cervical lymph node metastasis. Therefore, the cervical lymph nodes should be carefully examined by ultrasonography in patients with PTCs.

Objective To investigate the relationship between serum microRNA (miRNA, miR)-34 level and clinical characteristics in patients with thyroid papillary carcinoma, and to analyze the value of serum miR-34 in the diagnosis and prognosis evaluation of thyroid papillary carcinoma. Methods The clinical data of 84 cases of papillary thyroid cancer, 45 cases of nodular goiter and 30 cases of normal thyroid in our department were analyzed. The serum levels of miR-34 in all subjects were detected by fluorescence quantitative PCR. The correlation between serum miR-34 level and clinical features of thyroid cancer patients was analyzed, and the diagnostic efficacy of serum miR-34 for thyroid cancer was also analyzed by receiver operating characteristic (ROC) curve method. The correlation between the expression of serum miR-34 and overall survival (OS) and progression-free survival (PFS) in patients with thyroid papillary cancer was analyzed, and the predictive value of serum miR-34 in prognosis of patients with thyroid papillary cancer was evaluated by Kaplan-Meier analysis. Results The relative expression level of serum miR-34 in patients with papillary thyroid cancer was significantly lower than that in patients with nodular goiter and normal persons (P 0.05). Spearman correlation coefficient test showed that the relative expression level of serum miR-34 was closely related to karnofsky performance status (KPS) score, World Health Organization (WHO) pathological grade and extraglandular infiltration (P<0.05). The area under the ROC curve (AUC) was 0.912 [95% confidence interval (CI): 0.831-0.973]; when the AUC <0.48 was used as the criterion for the diagnosis of papillary thyroid cancer, the sensitivity, specificity and Yoden index were 96.67%, 79.54% and 76.21%, respectively. The median PFS and median OS were 39.5 (95% CI: 37.916-41.655) months and 50.3 (95% CI: 43.143-56.651) months in the high-level group, 28.3 (95% CI: 24.146-32.618) months and 32.6 (95% CI: 28.781-37.765) months in the low-level group, respectively. PFS and OS in the low-level group were significantly longer than those in the high-level group (P<0.01). Conclusion Serum miR-34 level has good accuracy in the diagnosis of thyroid papillary carcinoma, and has a significant correlation with the prognosis of thyroid papillary carcinoma. Key words: Thyroid papillary carcinoma; MicroRNA-34; Diagnosis; Prognosis

Papillary thyroid carcinoma (PTC) is the most common endocrine cancer with a significantly increase of the incidence recently. Several cytokines, such as thyroid peroxidase (TPO), cluster of differentiation 56 (CD56), Galectin-3, mesothelial cell (MC), cytokeratin 19 (CK19) and BRAF (B-raf) were recommended to be tested by immunohistochemistry (IHC) for a definitive diagnosis, but were still limited in clinical use because of their relative lower sensitivity and specificity. MicroRNA (miRNA), as a new molecular biomarkers, however, has not been reported yet so far. To address this, hsa-miR-200a-5p, a miRNA, was selected and detected in PTC patients by in situ hybrization with benign thyroid tumor with papillary hyperplasia as a control, and the differential expression of hsa-miR-200a-5p between fresh PTC tissues and control was detected by qRT-PCR. Expressive levels of cytokines of TPO, CD56, Galectin-3, MC, CK19 and B-raf were also detected by immunohistochemistry. The correlation was analyzed by SPSS software using Spearman methods. As expected, the hsa-miR-200a-5p expressive level was significantly increased in PTC patients, compared to that of control, and was consistent with that of TPO, CD56, Galectin-3, MC, CK19 and B-raf. In addition, expression of hsa-miR-200a-5p showed negative correlation to that of TPO (rs = - 0.734; **: P < 0.01) and CD56 (rs = - 0.570; **: P < 0.01), but positive correlation to that of Galectin-3 (rs = 0.601; **: P < 0.01), MC (rs = 0.508; **: P < 0.01), CK19 (rs = 0.712; **: P < 0.01) and B-raf (rs = 0.378; **: P < 0.01). PTC and papillary benign thyroid papillary hyperplasia are difficult to distinguish in morphology, so requiring immunohistochemistry to further differentiate the diagnosis, however, for the existing clinical common diagnostic marker for immunohistochemistry, the sensitivity and accuracy are low, it is easy to miss diagnosis. Therefore, there is an urgent need for a rapid and sensitive molecular marker. So miR-200a-5p can be used to assist in the diagnosis of PTC at the molecular level, and as a biomarker, can be effectively used to distinguish between PTC and benign thyroid tumor with papillary hyperplasia.

Solitary bone and brain metastasis in a patient with papillary thyroid carcinoma mimicking cavernous angioma

Noninvasive follicular thyroid neoplasm with papillary-like nuclear features (NIFTP) is considered an indolent neoplasm of the thyroid. Currently, this entity presents a diagnostic challenge on preoperative fine-needle aspiration (FNA) cytology. Cases of NIFTP and noninvasive encapsulated follicular variant of papillary thyroid carcinoma from January 1994 to August 2018 were retrieved from our institution's pathology databases, and their clinical and cytopathologic features were reviewed. A total of 45 patients with NIFTP were identified. Thirty-nine of 45 patients had presurgical thyroid FNAs, 27 of which showed abnormal cytology (27/39). NIFTPs were most often classified in the indeterminate diagnostic categories on presurgical FNAs (78%), including suspicious for follicular neoplasm (8/27), atypia of undetermined significance (8/27) and suspicious for malignancy (5/27). Six patients had a cytologic diagnosis of papillary thyroid carcinoma (6/27) on thyroid FNAs. Fourteen out of 27 cases with abnormal FNAs had in-house cytology available for review. The most frequent cytologic features observed in NIFTPs included architectural abnormalities, such as crowded clusters (13/14), nuclear overlapping (13/14), predominance or presence of individual microfollicles (11/14), and abundant to scattered isolated cells (10/14). The papillary-like nuclear features observed included nuclear enlargement (14/14), eccentric nucleoli (13/14), nuclear membrane irregularity (12/14), pale chromatin (12/14), nuclear grooves (10/14), and elongated nuclear contour (8/14). The NIFTPs were most frequently categorized as atypia of undetermined significance and suspicious for follicular neoplasm on cytology. Recognition of the architectural and nuclear features of NIFTP can be helpful in distinguishing NIFTPs from other entities, such as papillary thyroid carcinoma and follicular neoplasms.

Background: Papillary thyroid carcinoma (PTC) is the commonest thyroid cancer. Cases in category-5a of Bethesda system (suspicious for papillary carcinoma) are treated by surgical lobectomy followedby total thyroidectomy if histopathology confirms papillary carcinoma. In order to reduce surgicalprocedures to one this was conducted.Objectives: evaluation of role of immunohistochemistry in pre-operative diagnosis of papillary thyroidcarcinoma on cell blocks.Materials and Method: Cell blocks were taken from cases labelled category-5a for histopathology andimmunohistochemistry using three markers (CK-19, Thyro-peroxidase, and BRAFv600E mutation).Results: were highly sensitive, and specific. The use of more than markers increases sensitivity of theprocedure.Conclusion: immunohistochemical stains on cell blocks is a reliable method for pre-operative diagnosisof papillary thyroid carcinoma.

Актуальність. Папілярну карциному (ПК) щитоподібної залози (ЩЗ), як і будь-яку іншу, важливо своєчасно ідентифікувати, оскільки від цього залежить алгоритм лікування пацієнта. Основним та ефективним методом її передопераційної діагностики є тонкоголкова аспіраційна пункційна біопсія (ТАПБ). Між тим, чітко виявити ПК на даному етапі вдається не завжди. Вирішенню цієї проблеми часто допомагають інтраопераційні експрес-гістологічні дослідження (ЕГД), однак не в усіх випадках вони ефективні. Тому актуальним залишається пошук додаткових методів виявлення ПК ЩЗ, а як одного з кандидатів можна розглядати інтраопераційну цитологію (ІЦ). Мета дослідження: оцінка результативності ІЦ як додаткового методу виявлення ПК ЩЗ. Матеріали та методи. ІЦ виконана в 213 випадках вузлових утворень ЩЗ, а її результати порівняні з ЕГД з урахуванням висновків ТАПБ, які були класифіковані відповідно до Бетездівської системи висновків тиреоїдної цитопатології. Результати. Серед проаналізованих випадків було 127 ПК, в яких чітко ідентифікувати карциному за допомогою ТАПБ не уявлялося можливим. Порівняння результатів ІЦ та ЕГД показало, що у випадках неінкапсульованих ПК ЩЗ результативність ЕГІ становила 100 %, а ІЦ їх лише дублювали. Однак при інкапсульованих ПК ефективність ІЦ та ЕГД була практично однаковою: 18 із 81 (22,2 %) та 20 із 81 (24,7 %) відповідно. При цьому виконання ІЦ дозволило виявити додатково 7 ПК: в категорії ТАПБ «підозра на карциному» — 5 із 34 (14,7 %), в категорії «фолікулярна неоплазія» — 1 із 30 (3,3 %), в категоріях «атипія/доброякісний процес» — 1 із 17 (5,9 %). Висновки. У випадках інкапсульованих утворень ЩЗ комплексне виконання ЕГД та ІЦ дозволяє підвищити ефективність інтраопераційної діагностики ПК ЩЗ.

TROP-2 immunohistochemistry: a highly accurate method in the differential diagnosis of papillary thyroid carcinoma

We undertook this study to estimate an accurate incidence and spread patterns of occult papillary thyroid carcinoma (PTC) in patients with a preoperative diagnosis of solitary PTC by using whole-specimen mapping of all specimens after a total thyroidectomy. Enrolled prospectively in this whole-thyroid mapping study are 82 consecutive patients who underwent a total thyroidectomy under a preoperative diagnosis of solitary PTC. All thyroidectomy specimens were serially sectioned in 2 mm thickness and whole-thyroid mapping was carried out for additional foci of occult PTC. The frequencies of occult lesions detected in the whole and contralateral lobe were determined, and clinicopathologic factors associated with multifocality were assessed. Whole-thyroid mapping revealed 66 occult PTC lesions missed by preoperative ultrasound in 37 (45.1%) of the 82 patients. The great majority (92.5%) of the occult PTC was smaller than 3 mm in size and 25 patients (30.5%) had contralateral lesions. We found that the male sex was an independent predictor of multifocality (odds ratio (OR), 3.00; 95% CI, 1.11-8.14), adjusting for preoperative findings. Analysis with pathologic parameters showed that the male sex (OR, 5.03; 95% CI, 1.68-15.08) and extrathyroidal extensions (OR, 3.03; 95% CI, 1.03-8.95) were associated with multifocal PTC. However, none of the clinicopathologic factors evaluated predicted contralateral PTC. Our study demonstrates the diagnostic limitations of ultrasound for the detection of multifocal PTC and the need to consider the possibility of occult lesions in the management of solitary PTC, especially in male patients.

Although infarction after fine-needle aspiration (FNA) is a rare occurrence, it is a known phenomenon that may lead to difficulties in interpretation for pathologists and in decision-making for head and neck surgeons. To characterize our experience with infarction in papillary thyroid carcinomas (PTCs) after FNA and review existing cases of infarcted PTCs in the literature to better understand this phenomenon. This was a retrospective case series and review of literature at a tertiary medical center (University of California, Los Angeles [UCLA], Medical Center). All patients who had a surgical pathologic diagnosis of infarcted PTC and who underwent FNA prior to surgery at UCLA from June 2006 to June 2012 were identified. There were 620 cases of PTC and 12 cases of infarcted PTC. Demographic data, FNA cytologic findings, and surgical pathologic data were gathered for each patient. A comprehensive literature search for infarcted PTC was performed. Twelve cases of infarcted PTC were found in a total of 620 cases of PTC (1.9%). The mean (SD) time interval between the last FNA and surgery was 52 (35) days (range, 13-133 days). All patients received a diagnosis of infarcted PTC after thyroidectomy was performed. Focal infarction was found in 4 patients (33%), and near-total infarction was found in 8 patients (67%). Five patients (47%) had the follicular variant of PTC, making it the most common subtype in our series. A thorough literature search yielded 11 articles reporting a total of 26 cases of infarcted PTC after FNA. To our knowledge, our case series on infarcted PTC is the largest reported series in the literature. Although infarction of PTC after FNA occurs infrequently, it may lead to difficulties in histologic diagnosis. Awareness of this phenomenon and its histologic associations, along with careful reevaluation of the FNA and surgical specimens, is important for appropriate diagnosis and subsequent treatment. At this point, infarction in PTC should not alarm a head and neck surgeon to change management, but future prospective studies with a large population of patients with infarcted PTCs are needed to establish the impact of infarction on differences in treatment outcomes for therapies that may be used in PTCs.

There are no known effective and reliable biomarkers to distinguish benign thyroid nodules from papillary thyroid carcinomas (PTC). Previous studies have indicated that serum microRNA (miRNA) profiles may be diagnostic and/or prognostic markers for numerous other cancers. We studied circulating miRNA profiles in patients with PTC or benign nodules and healthy controls to identify serum miRNA that may be useful as markers for PTC. Genome-wide serum miRNA expression profiles were determined using Solexa sequencing followed by extensive quantitative RT-PCR validation in 245 subjects (106 patients with PTC, 95 patients with benign nodules, and 44 healthy controls). A panel of miRNA was used to assess the expression of specific miRNA in the sera and thyroid tissues of patients with PTC or benign nodules. The expression of serum let-7e, miR-151-5p, and miR-222 was significantly increased in PTC cases relative to benign cases and healthy controls. Receiver operating characteristic curve analyses indicated that use of these three miRNA had a high diagnostic sensitivity and specificity for PTC. Serum let-7e, miR-151-5p, and miR-222 levels were found to be well correlated with certain clinicopathological variables, such as nodal status, tumor size, multifocal lesion status, and Tumor-Node-Metastasis stage. Expression of serum miR-151-5p and miR-222 in a subset of PTC patients decreased significantly after tumor excision. Increased expression of miR-151-5p and miR-222 was also found in the tissue of PTC patients. Our study demonstrates that serum miRNA profiles may be used as novel and minimally invasive diagnostic markers for PTC.

A story is told by Juan Rosai about an atypical melanocytic lesion sent to him in consultation from a pathologist friend in private practice. After a long study, Dr Rosai decided it was benign. His friend thanked him for his opinion, but told him that he was still going to diagnose it as malignant melanoma. He said that if he called it melanoma and it never came back, everyone would be happy that the patient was cured. If he called it benign and it came back, then he would end up in court. This is the dilemma that faces the practicing pathologist every day. The tendency to “overdiagnose” seems to be inevitable as long as there is a surfeit of aggressive, unforgiving malpractice attorneys as well as pathology experts and a limited risk to the “overdiagnosis.” This tendency to overdiagnose clearly is accentuated in private practice. Fueled by the perception that private practitioners are not as good as their academic counterparts, cases in growing numbers are being referred to the university setting for a second opinion, sometimes years after the case originally was signed out. If the diagnosis from this second opinion on a thyroid nodule is malignant when the original diagnosis was benign, the ramifications can be quite devastating. While there certainly are benefits to second opinions,1 and we are in no way arguing against them, the unintended result of this practice is to render the private pathologist more vulnerable, more prone to litigation, and more likely to overdiagnose. Defensive pathology is now so ingrained in the psyche of most private pathologists that it has become second nature. The purely follicular variant of papillary thyroid carcinoma (PTC) described by Lindsay in 19602 and later popularized by Chen and Rosai3 and Rosai et al4 is now well accepted. The nuclear features have become tantamount to papillary structures in the diagnosis of PTC. The nuclear features include the following: (1) overlapping ground-glass, optically clear, or “Orphan Annie” nuclei and (2) nuclear pseudoinclusions and grooves, which represent cytoplasmic invaginations into the nucleus. Despite the universal recognition of this variant, it is not well characterized. The conventional PTC is classically an invasive tumor even when the majority of it is composed of follicles. The purely follicular variant of PTC on the other hand is often well circumscribed and encapsulated.4 What clearly is lacking is a minimal histologic definition of the follicular variant of PTC. Is an occasional grooved nucleus with pale chromatin in a follicular neoplasm enough to warrant a diagnosis of malignancy? Should the nuclear changes be uniformly present throughout the nodule? Are the nuclear changes that we are seeing significant or are they the “pseudoclear nuclei” described by Hapke and Dehner in the 1970s?5 At present, there are only a few long-term follow-up studies (totaling only 107 patients, as nicely summarized by Chan6) of the encapsulated follicular variant of PTC. While it is true that only 1 of these 107 patients died of disease,7 more than 25% of these patients had regional lymph node metastases. In addition, a recent study by Baloch and LiVolsi8 described 5 cases in which the encapsulated follicular variant of PTC manifested as an occult primary with metastatic disease, or in which bone metastases appeared many years after resection, and invasion in the primary lesion was subtle enough to be overlooked. Another study documented rare clinical progression (lymph node metastases) in the macrofollicular variant of PTC despite the fact that the dominant cytology and architecture of the primary tumor very closely resembled goiter.9 In summary, although the follicular variant of PTC usually behaves in a very indolent fashion, it rarely can cause distal metastasis. Furthermore, the follicular variant

Potential diagnostic utility of CD56 and claudin-1 in papillary thyroid carcinoma and solitary follicular thyroid nodules

Background: Hemocytometer parameters can be important biomarkers for some types of cancers and diseases. There is a need to evaluate their biomarker potential in thyroid diseases. Objectives: The current study aimed at contributing to potential biomarker researches to diagnose papillary thyroid carcinoma (PTC) and benign multinodular goiter (BMG), evaluate the role of these markers to determine the PTC characteristics and contribute to clarification of controversial issues. Methods: The study was designed as a retrospective observational study. The study included 389 patients who underwent total thyroidectomy at private Sani Konukoglu hospital (Sanko University, School of Medicine, Gaziantep, Turkey) due to BMG or PTC diagnosis from November 2011 to May 2017. The study subjects were divided into 3 groups. Groups 1, 2, and 3 were subjects with BMG, PTC, and the control, respectively. The comparisons were made among the groups in terms of hemocytometer parameters. In the group with the thyroid papillary carcinoma diagnosis, the relationship between the hemocytometer parameters and the bilateral tumor presence, single-sided multicentrite, and tumor diameter (size) were investigated. Results: Red cell distribution width (RDW) levels were 15.50 ± 2.39, 15.68 ± 2.16, and 12.5 ± 1.51 in the BMG, PTC, and control groups, respectively. Mean platelet volumes (MPV) were 7.97 ± 1.19, 8.05 ± 1.20, and 7.23 ± 1.39 in the BMG, PTC, and control groups, respectively. MPV and RDW values were significantly lower in group 3 compared with the groups 1 and 2 (P < 0.0001). Plateletcrit (PCT) values were 0.22 ± 0.05, 0.23 ± 0.06, and 0.19 ± 0.05 in the BMG, PTC, and control groups, respectively. PCT was significantly lower in the group 3 compared with the groups 1 and 2 (P = 0.0001). In PTC group, no significant relationship was observed between any of the examined hematological parameters in terms of multicentric and bilateral tumor presence. Conclusions: RDW, MPV, haemoglobin content, and PCT were significantly higher in the BMG and PTC groups. This increase was not specific for either of the 2 groups. No significant difference was found between any of the hemocytometer parameters and the multicentricity and bilaterality of PTC. But, a possible correlation was observed between activated partial thromboplastin time (aPTT), hemoglobin content, and tumor size.

Objective To investigate the expression of miR-199b-5p in papillary thyroid carcinoma （ PTC） and its relationship with clinical features .Methods Total RNA was extracted from 36 cases of PTC and the adjacent normal thyroid tissues by reverse transcription quantitative real-time polymerase chain reaction （ qRT-PCR）method to detect the expression of miR-199b-5p, and to analyze its relationship with clinical features such as the capsule invasion and lymph node metastasis .Results miR-199b-5p expression in PTC was related to lymph node status（χ2 =9.20, P=0.01）, capsule invasion（U=36.00, P=0.047）, but had no correlation with other clinical characteristics such as age , sex, tumor size, the number of tumor foci （ U =151.00, 87.00, 64.00, 87.00 respectively, P〉0.05）.ROC curve analysis showed that the specificity and sensitivity of miR-199b-5p in diagnosis of PTC were 82.1% and 72.7% respectively.Conclusion The abnormal expression of miR-199b-5p may be related to the occurrence , development and invasion of PTC . Key words: Papillary thyroid carcinoma; microRNA; microRNA-199b-Sp