Criteria for follow-up of thyroid nodules diagnosed as follicular neoplasm without molecular testing - The experience of a high-volume thyroid centre in Japan

Diagnostic Performance of Thyroid FNA: Experience of Belarusian National Referral Thyroid Cancer Center

BACKGROUND

A fine-needle aspiration (FNA) biopsy is a basic procedure to verify the morphology of thyroid nodules. However, significant variations in the diagnostic performance of FNA, as well as fluctuations in the rate of malignancy (ROM) in different categories of the Bethesda System for Reporting Thyroid Cytopathology, have been discovered in many studies. The development and employment of additional molecular tests in daily clinical practice require an understanding of the prevalence and structure of thyroid malignancy in each category of the Bethesda system.

METHODS

Current research is a continuous retrospective cohort study of the results of the examination and treatment of 1652 patients with thyroid tumors who have undergone primary surgery in 2021 at a national referral thyroid cancer center. FNA diagnostic performance was studied by comparing cytological diagnoses with histological outcomes. The authors evaluated ROM in each Bethesda category and identified the tumors representing a diagnostic pitfall.

RESULTS

The overall ROM was 72.4%. When considering ROM in each Bethesda category, it was 16.7% in Bethesda I, 10.1% in Bethesda II, 23.0% in Bethesda IV, 74.8% in Bethesda V, 99.4% in Bethesda VI. The false-positive rate was 65.0%, false negative-0.67%. Follicular and oncocytic adenomas, as well as the follicular variant of papillary thyroid cancer, were the most troublesome entities for patients' management.

CONCLUSION

ROM was comparable to the values implied by the Bethesda system. Follicular tumors were the main source of high false-positive rates. This necessitates the refinement of existing tests and the development of new diagnostic tests to overcome the abovementioned problems.

Evolutionary analysis of indeterminate cytology and risk of malignancy in a thyroid nodule unit

OBJECTIVE

The objective of this study was to analyze the evolution in the diagnosis and management of indeterminate thyroid nodules over three time periods.

METHODS

3020 patients with thyroid nodules underwent cytological evaluation during three periods (2006-2008, 2012-2014, 2017-2019). Distribution of diagnostic cytologies, risk of malignancy, diagnostic performance indices of FNA, and cytologic-histologic correlation in indeterminate cytologies were analyzed.

RESULTS

only 2.2% of cytology tests were insufficient for a diagnosis. 86.9% cytologies were benign, 1.7% malignant, and 11.4% indeterminate. Indeterminate cytology rates were 15.9% (2006-2008), 10.1% (2012-2014), and 10% (2017-2019). Surgery was performed in 13% of benign cytology, result-ing in malignant histology in 2.7%. All malignant and suspicious cytologies underwent surgery: malig-nancy confirmed in 98% and 77% of cases, respectively. All 'indeterminate with atypia' cytologies (2006-2008) and Bethesda IV (2012-2014; 2017-2019) un-derwent surgery, with malignancy confirmed in 19.6%, 43.8%, and 25.7%, respectively. In the 'inde-terminate without atypia' category (2006-2008) and Bethesda III (2012-2014; 2017-2019), diagnostic surgery was performed in 57.7%, 78.6%, and 59.4%, respectively, with malignancy confirmed in 3.3%, 20.5%, and 31.6%. The FNA sensitivity was 91.6% with a negative predictive value greater than 96% in all periods. The specificity exceeded 75% in the last two periods.

CONCLUSION

Bethesda system reduces indeterminate cytologies and improves the accuracy of FNA diagnosis. We reported a higher proportion of malignancy than expected in Bethesda III, underscoring the importance of having institution-specific data to guide decision-making. However, there is a need for risk stratification tools that allow for conservative management in low-risk cases.

Identification of invasive subpopulations using spatial transcriptome analysis in thyroid follicular tumors

BACKGROUND

Follicular tumors include follicular thyroid adenomas and carcinomas; however, it is difficult to distinguish between the two when the cytology or biopsy material is obtained from a portion of the tumor. The presence or absence of invasion in the resected material is used to differentiate between adenomas and carcinomas, which often results in the unnecessary removal of the adenomas. If nodules that may be follicular thyroid carcinomas are identified preoperatively, active surveillance of other nodules as adenomas is possible, which reduces the risk of surgical complications and the expenses incurred during medical treatment. Therefore, we aimed to identify biomarkers in the invasive subpopulation of follicular tumor cells.

METHODS

We performed a spatial transcriptome analysis of a case of follicular thyroid carcinoma and examined the dynamics of CD74 expression in 36 cases.

RESULTS

We identified a subpopulation in a region close to the invasive area, and this subpopulation expressed high levels of CD74. Immunohistochemically, CD74 was highly expressed in the invasive and peripheral areas of the tumor.

CONCLUSIONS

Although high CD74 expression has been reported in papillary and anaplastic thyroid carcinomas, it has not been analyzed in follicular thyroid carcinomas. Furthermore, the heterogeneity of CD74 expression in thyroid tumors has not yet been reported. The CD74-positive subpopulation identified in this study may be useful in predicting invasion of follicular thyroid carcinomas.

Application of deep learning as an ancillary diagnostic tool for thyroid FNA cytology

BACKGROUND

Several studies have used artificial intelligence (AI) to analyze cytology images, but AI has yet to be adopted in clinical practice. The objective of this study was to demonstrate the accuracy of AI-based image analysis for thyroid fine-needle aspiration cytology (FNAC) and to propose its application in clinical practice.

METHODS

In total, 148,395 microscopic images of FNAC were obtained from 393 thyroid nodules to train and validate the data, and EfficientNetV2-L was used as the image-classification model. The 35 nodules that were classified as atypia of undetermined significance (AUS) were predicted using AI training.

RESULTS

The precision-recall area under the curve (PR AUC) was >0.95, except for poorly differentiated thyroid carcinoma (PR AUC = 0.49) and medullary thyroid carcinoma (PR AUC = 0.91). Poorly differentiated thyroid carcinoma had the lowest recall (35.4%) and was difficult to distinguish from papillary thyroid carcinoma, medullary thyroid carcinoma, and follicular thyroid carcinoma. Follicular adenomas and follicular thyroid carcinomas were distinguished from each other by 86.7% and 93.9% recall, respectively. For two-dimensional mapping of the data using t-distributed stochastic neighbor embedding, the lymphomas, follicular adenomas, and anaplastic thyroid carcinomas were divided into three, two, and two groups, respectively. Analysis of the AUS nodules showed 94.7% sensitivity, 14.4% specificity, 56.3% positive predictive value, and 66.7% negative predictive value.

CONCLUSIONS

The authors developed an AI-based approach to analyze thyroid FNAC cases encountered in routine practice. This analysis could be useful for the clinical management of AUS and follicular neoplasm nodules (e.g., an online AI platform for thyroid cytology consultations).

The Significance of RAS-Like Mutations and MicroRNA Profiling in Predicting Malignancy in Thyroid Biopsy Specimens

In cytologically indeterminate thyroid nodules undergoing molecular testing, estimated risk of malignancy is variable. Identification of a non-cancer-specific mutation (RAS-like) confirms a neoplastic process but does not differentiate between benign, malignant, and low-risk neoplasms. This study aims to retrospectively evaluate institutional experience of Interpace (ThyGeNEXT^® and ThyraMIR^®; Pittsburgh, PA) testing and to determine the rate of malignancy in resected nodules, stratified by mutational analysis and microRNA profile. Of 1917 fine need aspirations, 140 (7.3%) underwent Interpace testing: 47 (33.6%) were molecular-not-benign (harbored mutation, fusion, and/or positive miRNA) and 93 (66.4%) were molecular-benign (no mutations or fusions and negative microRNA). Surgery was spared in 79.6% of molecular-benign and 61.4% of all tested patients. Fifty-four (38.6%) underwent resection. Seventeen (89.5%) of the resected molecular-benign were benign and 2 were malignant. Thirteen (37.1%) of the resected molecular-not-benign were benign, 7 (20%) were noninvasive follicular thyroid neoplasm with papillary-like nuclear features (NIFTP), and 15 (42.9%) were malignant (p < 0.05, negative predictive value (NPV) 89.4-95.6%, positive predictive value (PPV) 22.3-42.8%). Most molecular-not-benign (72.3%) had RAS-like mutation. Twenty-three were resected: 3 were malignant and 7 were NIFTP. Nodules with non-RAS-like mutations (BRAF V600E-like, others) were more likely to be malignant than RAS-like (H/N/KRAS, BRAF K601E) (p < 0.05, NPV 86.9-96.5%, PPV 100%). Most nodules had RAS-like mutations and most were benign or low-risk neoplasms (NIFTP). This study supports the role of histologic examination in the distinction of malignancy in RAS-like thyroid neoplasms and underscores the role of molecular testing in risk stratification, patient counseling, and operative management.