Background: Federated learning is a decentralized approach to machine learning; it is a training strategy that overcomes medical data privacy regulations and generalizes deep learning algorithms. Federated learning mitigates many systemic privacy risks by sharing only the model and parameters for training, without the need to export existing medical data sets. In this study, we performed ultrasound image analysis using federated learning to predict whether thyroid nodules were benign or malignant. Objective: The goal of this study was to evaluate whether the performance of federated learning was comparable with that of conventional deep learning. Methods: A total of 8457 (5375 malignant, 3082 benign) ultrasound images were collected from 6 institutions and used for federated learning and conventional deep learning. Five deep learning networks (VGG19, ResNet50, ResNext50, SE-ResNet50, and SE-ResNext50) were used. Using stratified random sampling, we selected 20% (1075 malignant, 616 benign) of the total images for internal validation. For external validation, we used 100 ultrasound images (50 malignant, 50 benign) from another institution. Results: For internal validation, the area under the receiver operating characteristic (AUROC) curve for federated learning was between 78.88% and 87.56%, and the AUROC for conventional deep learning was between 82.61% and 91.57%. For external validation, the AUROC for federated learning was between 75.20% and 86.72%, and the AUROC curve for conventional deep learning was between 73.04% and 91.04%. Conclusions: We demonstrated that the performance of federated learning using decentralized data was comparable to that of conventional deep learning using pooled data. Federated learning might be potentially useful for analyzing medical images while protecting patients’ personal information.

The presence of telomerase reverse transcriptase (TERT) promoter mutations have been associated with a poor prognosis in patients with papillary thyroid carcinomas (PTC). The frequency of TERT promoter mutations varies widely depending on the population and the nature of the study. Data were prospectively collected in 724 consecutive patients who underwent thyroidectomy for PTC from 2018 to 2019. Molecular testing for BRAF V600E and TERT promoter mutations was performed in all cases. TERT promoter alterations in two hotspots (C228T and C250T) and C216T were found in 16 (2.2%) and 4 (0.6%) of all PTCs, respectively. The hotspot mutations were significantly associated with older age at diagnosis, larger tumor size, extrathyroidal extension, higher pathologic T category, lateral lymph node metastasis, and higher American Thyroid Association recurrence risk. The patients with C216T variant were younger and had a lower American Thyroid Association recurrence risk than those with hotspot mutations. Concurrent BRAF V600E was found in 19 of 20 cases with TERT promoter mutations. Of 518 microcarcinomas measuring ≤1.0 cm in size, hotspot mutations and C216T variants were detected in five (1.0%) and three (0.6%) cases, respectively. Our study indicates a low frequency of TERT promoter mutations in Korean patients with PTC and supports previous findings that TERT promoter mutations are more common in older patients with unfavorable clinicopathologic features and BRAF V600E. TERT promoter mutations in patients with microcarcinoma are uncommon and may have a limited role in risk stratification. The C216T variant seems to have no clinicopathologic effect on PTC.

Triple-negative breast cancer (TNBC) is an aggressive subtype of breast cancer with poor prognosis. Current endocrine therapy or anti HER-2 therapy is not available for these patients. Chemotherapeutic treatment response varies among patients due to the disease heterogeneity. To overcome these challenges, theranostics for treating TNBC have been widely investigated. Anticancer material conjugated nanoparticles with target-binding ligand and tracer agents enable simultaneous drug delivery and visualization of the lesion with minimal off-target toxicity. In this review, we summarize recently FDA-approved targeted therapies for TNBC, such as poly-ADP-ribose polymerase (PARP) inhibitors, check point inhibitors, and antibody-drug conjugates. Particularly, novel theranostic approaches including lipid-based, polymer-based, and carbon-based nanocarriers are discussed, which can provide basic overview of nano-therapeutic modalities in TNBC diagnosis and treatment.

Thyroid nodules are primarily diagnosed using ultrasound imaging (USI), but its low specificity leads to unnecessary fine-needle aspiration biopsies (FNABs). In particular, USI’s limited ability to differentiate follicular neoplasms from benign nodules contributes to suboptimal biopsy decision-making. We propose a dual-modal imaging approach that combines multiparametric photoacoustic imaging (PAI) and USI to support smarter biopsy decisions. In 106 patients with 29 benign nodules, 45 papillary thyroid carcinomas, and 32 follicular neoplasms, three PAI-derived parameters—the photoacoustic spectral gradient, oxygen saturation, and skewness of the oxygen saturation distribution—were combined using a support vector machine. Following USI-based American Thyroid Association (ATA) guidelines, they were used to develop the ATA-Photoacoustic (ATAP) scoring system. The ATAP score achieved 97% sensitivity and 38% specificity in distinguishing nodules requiring FNAB. Our approach enabled better identification of benign nodules, reducing unnecessary FNAB in 11 of the 29 benign cases. This dual-modal strategy can assess thyroid nodules, effectively reducing unnecessary biopsies while maintaining high diagnostic accuracy.

Background: Many studies have shown the operative feasibility and safety of robotic thyroidectomy. However, there is still a concern on the operative invasiveness of robotic thyroidectomy owing to the wide flap dissection. The aim of this study was to introduce our novel robotic technique, which can minimize the dissection extent using the da Vinci SP robotic system. Methods: Ten patients underwent robotic transaxillary thyroidectomy using the da Vinci SP robotic system between December 2018 and April 2019 at Yonsei University Hospital (Seoul, Korea). All procedures were performed successfully using the single-port robotic system. Results: All 10 patients were women and diagnosed with papillary thyroid carcinoma. The mean operative time was 148.7 ± 26.8 minutes. The mean operative time using the gasless method was shorter than that using the gas insufflation method (130.5 ± 14.1 vs 176.0 ± 12.8 minutes). All patients were discharged on the third day after operation without any complications. Conclusions: Robotic transaxillary thyroidectomy using the da Vinci SP robotic system is technically feasible and safe with a short length of incision, shorter than what earlier reported. To our knowledge, this is the first study to use a single-port robotic system for thyroidectomy. To verify the superior operative outcomes, further clinical trials are necessary.

Male patients have a significantly higher prevalence of advanced-stage thyroid cancer. However, sex differences in the risk of differentiated thyroid carcinoma (DTC) recurrence have not been fully elucidated. Therefore, the present study aimed to investigate male sex as a prognostic factor for DTC. We assessed 5566 patients with DTC who underwent thyroid surgery between January 2009 and December 2015 at Seoul St. Mary’s Hospital (Seoul, Korea). Clinicopathological characteristics and long-term oncologic outcomes between female and male patients with DTC were compared using propensity score matching to reduce selection bias. The mean follow-up duration was 99.9 ± 18.7 months. The recurrence rate was significantly higher in male patients than female patients before matching (3.3% vs. 2.2%, p = 0.030), and there was no significant difference in recurrence rates between the matched groups after matching (3.0% vs. 2.5%, p = 0.591). Based on Kaplan–Meier analysis, the two groups did not significantly differ in disease-free survival after matching. Multivariate analysis revealed that male sex was not an independent prognostic factor of DTC recurrence. Male sex did not have a significant effect on DTC recurrence. Further studies with larger cohorts are required to validate the findings of this study.