Hiroo Masuoka

Background: We previously demonstrated that the Ki‐67 labeling index (LI), a marker of cell proliferation, strongly predicts prognosis in patients with papillary thyroid carcinoma (PTC). We subsequently developed a prognostic scoring system incorporating Ki‐67 LI. Methods: We enrolled 8540 patients with PTC who underwent curative surgery (median follow‐up, 8.7 years). Multivariate Cox regression analysis identified prognostic factors, and regression coefficients from the final models were rounded to generate integer‐based scores. Results: To date, 368, 125, and 22 patients experienced local recurrence (LR), distant recurrence (DR), and carcinoma‐related death, respectively. Scores were assigned for age, tumor size, N factor, M factor, Ki‐67 LI, and extrathyroidal/extranodal invasion (Ex). The LR score comprised sex (0 or 1), age (0, 1, or 2), T (0 or 2), N (0, 2, or 4), Ki‐67 (0 or 4), and Ex (0, 1, or 2). Ten‐year LR‐free survival rates for low (≤ 2), intermediate (3–4), and high (≥ 5) scores were 98.1%, 95.2%, and 84.5%, respectively ( p < 0.001). The DR score comprised sex (0 or 1), age (0 or 3), T (0 or 3), N (0, 2, or 4), Ki‐67 (0 or 5), and Ex (0, 3, or 6). Ten‐year DR‐free survival rates for low (≤ 5), intermediate (6–9), and high (≥ 10) scores were 99.6%, 97.9%, and 85.8%, respectively ( p < 0.01). The cause‐specific survival (CSS) score for the entire series comprised sex (0 or 2), age (0 or 4), N (0, 1, or 2), M (0 or 4), Ki‐67 (0 or 3), and Ex (0, 1, or 2). Ten‐year CSS rates for low (≤ 3), intermediate (4–5), and high (≥ 6) scores were 100%, 99.8%, and 97.1%, respectively ( p < 0.01). In M‐negative patients, the CSS score comprised sex (0 or 1), age (0 or 3), N (0, 1, or 2), Ki‐67 (0 or 3), and Ex (0, 1, or 2). The 10‐year CSS rate for high‐score (≥ 6) patients was 96.9%, significantly lower than those for low‐ and intermediate‐score patients (100% and 99.8%, respectively; p < 0.01). Conclusions: MASSIN‐K, a prognostic scoring system incorporating Ki‐67 LI, showed good prognostic discrimination in patients with PTC. External validation is required as a further study.

Oncocytic thyroid carcinoma (OCA) is a rare thyroid malignancy that has only recently been recognized as a distinct entity from follicular thyroid carcinoma. Till date, studies regarding the prognosis of OCA remain inadequate. In this retrospective study, we investigated the prognosis of OCA among 137 Japanese patients who underwent surgery at our institution. Two patients had distant metastasis before surgery (M1), and two showed distant metastasis after surgery. The 10-year and 15-year distant recurrence-free survival rates were 100%, and 92.7%, respectively. No deaths occurred owing to thyroid carcinoma. Four patients with distant metastasis/recurrence were aged ≥60 years, and their tumors were >4 cm. Further, two patients with M1 had both vascular invasion (VI) and wide capsular invasion (CI). One patient had a poorly differentiated component (approximately 30%), and one showed high cell proliferating activity (Ki-67 labeling index [LI] ≥5%). Of the two patients who developed distant metastasis after surgery, one had VI and wide CI, and although another had minimal CI and was negative for VI, the Ki-67 LI of his tumor was >5%. These findings suggest that Japanese patients with OCA generally had excellent prognosis after surgery; however, older age, large tumor size, wide CI, VI positivity, high Ki-67 LI, and the presence of a poorly differentiated component may be indicators of aggressive behavior. Further multicenter studies with larger samples and longer follow-up durations are necessary to identify clinicopathological factors accurately predicting poor prognosis in patients with OCA.

Follicular thyroid carcinoma (FTC) is known to occasionally show distant recurrence (DR); however, no prognostic scoring system has been established. Furthermore, although prognostic factors have been identified, their weights as predictors of poor prognosis remain unclear. In this study, we aimed to establish a weighted scoring system for predicting DR in FTC using conventional prognostic factors and Ki-67 labeling index (LI). We analyzed data of 597 patients with FTC without distant metastasis at initial surgery. Tumor size >4 cm, male sex, age ≥60 years, Ki-67 LI >5%, and vascular invasion (VI) ≥4 (V2) significantly and VI 0-4 (V1) marginally affected DR. Wide capsular invasion was not significantly associated with DR. Stepwise model selection using the Akaike Information Criterion identified age ≥60 years, Ki-67 LI >5%, V1, V2, and tumor size >4 cm as suitable components of the score. According to the beta coefficient value, we assigned a score of 2 to age ≥60 years, Ki-67 LI >5%, and V2, and a score of 1 to tumor size >4 cm and V1, with the sum representing the final score. We categorized patients into three categories according to the DR rates: low-risk (score 0-2; n = 388, 65.0%), intermediate-risk (score 3-4; n = 156, 26.1%), and high-risk (score 5-7; n = 53, 8.9%). The 10-year DR-free survival rates declined from the low- to the high-risk categories (96.0% vs. 90.3% vs. 54.9%). These findings indicate that high-risk patients identified using this prognostic scoring model should be carefully monitored for DR following completion thyroidectomy.

Objective: Thyrotropin receptor antibodies (TRAbs) are central to Graves’ disease management, but their long-term kinetics after total thyroidectomy is poorly defined. We aimed to describe long-term TRAb trajectories and identify predictors of decline using clinically relevant thresholds. Methods: TRAb levels were measured serially in 1,516 patients after total thyroidectomy. Primary outcomes were time to TRAb < 10 and <2 IU/L, corresponding to a pragmatic fetal-risk threshold and assay negativity, respectively. Clinical factors (age, sex, body mass index, smoking, preoperative TRAb, and thyroid weight) were evaluated using Kaplan–Meier analyses and multivariable logistic regression. Results: TRAb levels declined rapidly, with median values falling below 10 IU/L within 2 years. Among patients with preoperative TRAb ≥ 10 IU/L, 84.8% reached <10 IU/L and 60.1% reached <2 IU/L within 5 years. In multivariable models, younger age, lower preoperative TRAb, and smaller thyroid weight independently predicted earlier decline, whereas sex, body mass index, and smoking did not. Conclusions: In this large post-operative cohort, younger age emerged as an independent predictor of accelerated TRAb decline after total thyroidectomy. Total thyroidectomy yields sustained TRAb decline and may promote immunological remission, particularly in younger patients, by removing the antigenic source before the establishment of long-lived plasma cells. These findings highlight age as a key determinant of TRAb kinetics and provide timelines for counselling and follow-up. Surgery may be especially valuable when a timely TRAb decline is required, such as in younger patients planning pregnancy or those with orbitopathy or antithyroid drug refractoriness.

Background: It has been 30 years since the initiation of active surveillance (AS) for adult patients with low-risk papillary thyroid microcarcinoma (PTMC). This study compared the long-term oncological outcomes of patients who underwent AS or immediate surgery (IS). Methods: This is a retrospective review of extended follow-up data from patients enrolled in a single-center, prospective observational study in Japan. In total, 5646 patients diagnosed with low-risk PTMC at Kuma Hospital between 1993 and 2019 were enrolled in this study. Of these, 3222 patients underwent AS (AS group), whereas 2424 underwent IS (IS group). The patients were followed up regularly, at least once per year. Descriptive outcome data were presented according to the treatment group. Results: In the AS group, 124 patients (3.8%) had tumor enlargement of ≥3 mm, and the 10- and 20-year enlargement rates were 4.7% and 6.6%, respectively. Novel lymph node metastases occurred in 27 patients (0.8%), and the 10- and 20-year nodal metastasis occurrence rates were 1.0% and 1.6%, respectively. In the IS group, 13 patients (0.5%) experienced lymph node recurrence postoperatively, and the 10- and 20-year nodal recurrence rates were 0.4% and 0.7%, respectively. Eighteen (1.4%) of the 1327 patients who underwent hemithyroidectomy experienced recurrence in the residual thyroid. The rate of lymph node metastasis was significantly higher in the AS group than in the IS group (1.1% vs. 0.4% and 1.7% vs. 0.7% at 10 and 20 years, respectively; p = 0.009), but the differences were small. However, the proportion of patients who underwent one or more and two or more surgeries was significantly higher in the IS group than in the AS group (100% vs. 12.3% and 1.07% vs. 0.09%, p < 0.01). Distant metastatic recurrence was observed in one patient after AS and conversion surgery and another after IS; however, they were alive (18.4 and 18.8 years after diagnosis, respectively). None of the patients in this study died of thyroid carcinoma. Conclusions: Long-term oncological outcomes of patients with PTMC generally did not differ clinically significantly between those undergoing AS and IS. AS is a viable initial management option for patients with low-risk PTMC.

Thyroid cancer often invades the recurrent laryngeal nerve (RLN), causing vocal cord paralysis. In such patients, the invaded portion of the RLN usually needs to be resected through curative surgery. We attempt to preserve the nerve by performing sharp dissection in such cases. During nerve dissection, an intraoperative nerve monitoring system helps identify the course of the RLN in the fibrous tissue around the tumor or even within the tumor, and also helps evaluate the nerve integrity. Because of extensive dissection, the preserved RLN may become much thinner than its original thickness. We refer to this procedure as “partial layer resection” of the RLN. In our cases, although the dissected RLNs became thinner, we found that vocal cord function recovered in most patients. If the RLN is fully involved by thyroid cancer or response of the vocal cord against electric stimulation to the RLN is lost, we resect the portion of the RLN together with the tumor and repair it using one of the reconstruction techniques. When a unilateral RLN is resected, the vocal cord on that side is paralyzed. Symptoms include hoarseness, mis-swallowing, and short phonation. RLN reconstruction using one of the reconstruction techniques leads to the recovery of phonatory and swallowing function, although the normal motion of the vocal cord on the side of the anastomosis is not restored. We used direct anastomosis, free nerve grafting, ansa cervicalis-RLN anastomosis, and vagus-RLN anastomosis to reconstruct the RLN. Thyroid cancer often invades the RLN near the Berry’s ligament. In such patients, surgeons might assume that reconstruction of the RLN may not be possible because the peripheral stump of the RLN cannot be observed. However, if we divide the inferior pharyngeal constrictor muscles along the lateral edge of the thyroid cartilage, the peripheral RLN can be identified, and nerve reconstruction can be performed. We refer to this procedure as “laryngeal approach”.In summary, of the patients with thyroid cancer who required resection of the RLN, RLN reconstruction led to the recovery of phonatory function. We suggest that all thyroid surgeons familiarize themselves with these reconstruction techniques.

Background The major sites of distant metastases of papillary thyroid carcinoma (PTC) are the lung and bone; metastasis to the liver is rare. Although the postoperative serum thyroglobulin (Tg) level after total thyroidectomy is a good prognostic indicator for PTC when anti-thyroglobulin antibody (TgAb) is negative, the presence of TgAb interferes with the Tg assay, making serum Tg levels unreliable. Here we report a case of liver metastasis of PTC that presented with elevated serum levels of carbohydrate antigen 19-9 (CA19-9), which is usually a serum marker of pancreatic and gastrointestinal neoplasias. Case presentation A 69-year-old man was diagnosed with PTC and underwent total thyroidectomy 16 years ago. The patient's serum Tg levels increased progressively during follow-up and his serum TgAb was negative. Positron emission tomography (PET) and computed tomography (CT) revealed metastases of the lung, cervical spine, mediastinum and liver. The liver lesion was a solitary tumor measuring 4.0 cm in the greatest dimension. His serum CA19-9 level was very high (326 U/mL), and intrahepatic cholangiocarcinoma was suspected from the results of various examinations including gastrointestinal endoscopic imaging and CT. Laparoscopic partial liver resection for segment 4 was performed. The histopathological diagnosis was a metastatic liver tumor from PTC. The immunohistological examination revealed that the liver tumor was positive for CA19-9 and Tg. The primary PTC, recovered from paraffin-embedded specimen, was also positive for CA19-9. After the surgery, his serum CA19-9 level as well as serum Tg level markedly decreased. Conclusions We presented the first reported case of liver metastasis of a PTC presenting with elevated serum levels of CA19-9 after total thyroidectomy. This case suggests that the serum CA19-9 levels may serve as a surrogate marker for PTC in place of the serum Tg level in patients with positive serum TgAb if the PTC and/or the metastatic lesions are positive for CA19-9 staining.

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.

Background: Acute suppurative thyroiditis through the congenital pyriform sinus fistula (PSF) often recurs if the fistula is not resected. Although endoscopic chemo-cauterization (ECC) to obliterate the orifice of the fistula is less invasive than open fistulectomy, it may require repeated treatments. We recently adopted an endoscopic diode laser-cauterization (ELC) system with the intention of improving treatment outcomes in PSF. Here, we describe ELC and compare the outcomes of these three modalities. Methods: We evaluated 83 patients with PSF who underwent treatment between 2007 and 2018 at Kuma Hospital, a tertiary thyroid treatment hospital. ECC and ELC were implemented in 2007 and 2015, respectively. Patients who were ineligible for the endoscopic procedures underwent open fistulectomy. Barium swallow studies and computed tomography scan under a trumpet maneuver were performed after treatment to evaluate obliteration or removal of the fistula. Results: In total, 70 of the 81 (86%) patients who underwent barium swallow studies after the first treatment achieved obliteration or removal of the fistula. The success rates for open fistulectomy, ECC, and ELC were 100% (9/9), 83% (49/59), and 100% (13/13), respectively. ECC and ELC had significantly shorter operative times and lower blood loss than open fistulectomy. Insufficient opening of the mouth was the major reason for converting endoscopic procedures to open fistulectomy. Conclusions: ELC may yield superior outcomes and is therefore the optimal treatment modality for PSF. However, it is still associated with certain limitations. Thus, treatment selection remains dependent on the shape and size of the PSF and the mouth opening of the individual patient. Graphical