Faculty Bibliography

BACKGROUND AND OBJECTIVE/UNASSIGNED:Intraoperative parathyroid gland recognition is a key step during thyroidectomy to decrease the risk of postoperative hypocalcemia and during parathyroidectomy to distinguish normal and abnormal glands. Current methods for intraoperative identification rely largely upon visual identification. Recent investigation of methods such as near-infrared (NIR) autofluorescence and indocyanine green (ICG) for enhanced recognition have demonstrated steep learning curves. Artificial intelligence (AI) augmentation of all methods of parathyroid gland identification may improve intraoperative recognition rates and ultimately decrease rates of postoperative hypoparathyroidism. This narrative review aims to summarize the status of intraoperative application of AI for parathyroid gland recognition. METHODS/UNASSIGNED:A systematic, comprehensive literature search was conducted using the search terms "artificial intelligence", "deep learning", "surgery", "parathyroid gland", and "parathyroid glands". Inclusion criteria included articles in English with the majority of the article devoted to intraoperative applications of AI on parathyroid gland recognition. Eleven studies were identified and included. KEY CONTENT AND FINDINGS/UNASSIGNED:Eight studies focused on utilizing AI intraoperatively to identify parathyroid glands from surrounding tissues. Three studies focused on using AI to predict abnormal from normal parathyroid glands. Five studies used NIR autofluorescence, two studies used visual recognition during open thyroidectomy, two studies used visual recognition during endoscopic thyroidectomy, one study used NIR autofluorescence with ICG angiography, and one study used coaxial dual-red-green-blue/near-infrared (dual-RGB/NIR) imaging system to identify parathyroid glands. Recall and precision scores for the models ranged from 50-95% and 72-94%, respectively. Four studies compared model performance with that of senior and junior surgeons and found that the models outperformed junior surgeons while performing comparably to senior surgeons. CONCLUSIONS/UNASSIGNED:AI augmentation of intraoperative parathyroid gland recognition demonstrates adequate accuracy results across a range of parathyroid gland recognition methods. Although these models are not currently available for widespread commercial use, the eventual integration into clinical practice may allow for enhanced intraoperative recognition of parathyroid glands, particularly in lower volume centers and for junior level surgeons.

INTRODUCTION/BACKGROUND:Papillary thyroid cancer (PTC) often follows an indolent course with a favorable prognosis. This has led to evolving guideline-based, low-intensity treatment options for low-risk patients. Recently, the purported benefit of total thyroidectomy (TT) over unilateral lobectomy for PTC with clinical lateral neck nodal metastasis (cN1b) has come into question. MATERIALS AND METHODS/METHODS:A retrospective analysis of the National Cancer Institute's Surveillance, Epidemiology, and End Results database was performed to study patients with PTC with ipsilateral cN1b disease from 1975 to 2020. Kaplan-Meier curves and log-rank tests were used to compare disease-specific survival (DSS) difference between lobectomy and TT at 10 y. Multivariable Cox proportional hazards analysis was performed to determine the independent association of lobectomy versus TT with DSS, correcting for age and lymph node ratio, defined as the ratio of pathologically positive lymph nodes to total number examined. RESULTS:Among 2943 patients (median [interquartile range] age, 45 [26] y), 42 underwent lobectomy and 2901 underwent TT. Unadjusted DSS at 10 y in the lobectomy and TT groups were 51.0% (95% confidence interval, 31.4%-82.8%) and 86.8% (95% confidence interval, 84.8%-88.9%), respectively. On multivariable analysis of all patients, older age (hazard ratio [HR], 1.08; P < 0.001) and male gender (HR, 1.74; P < 0.001) were associated with lower adjusted DSS, whereas treatment with TT (HR, 0.387; P = 0.005) and receipt of radioactive iodine (RAI) (HR, 0.604; P < 0.001) were associated with improved adjusted DSS. In addition, we observed that the magnitude of survival benefit conferred by RAI and TT were reduced with decreasing age (P < 0.001). CONCLUSIONS:This longitudinal cohort study suggests that, while TT is associated with a DSS benefit in most patients with PTC and ipsilateral cN1b disease, this association may not exist in a smaller cohort of younger patients. These findings raise the possibility that unilateral surgical clearance without RAI could offer adequate oncologic outcomes in selected younger individuals; however, further investigation is warranted to confirm these observations and inform clinical decision-making.

BACKGROUND:Four-dimensional computed tomography is routinely used to localize parathyroid disease, with consistently excellent parathyroid gland localization rates reported. This study evaluated whether pairing 4-dimensional computed tomography results with preoperative clinical variables can accurately predict single-gland disease in primary hyperparathyroidism. METHODS:Patients with primary hyperparathyroidism who underwent both 4-dimensional computed tomography imaging and parathyroidectomy between January 2019 and September 2021 at a large academic health system were included. Patient demographics, preoperative characteristics, and peri- and postoperative data were collected. The accuracy of 4-dimensional computed tomography in correctly identifying patients with single-gland disease with and without preoperative calcium and parathyroid hormone levels was calculated. Single-gland disease was defined by intraoperative parathyroid hormone decrease >50% and a hypercellular gland on pathology. RESULTS:One hundred seventy-five patients had 4-dimensional computed tomography results suggestive of single gland disease. One hundred fifty-two patients (87%) were predicted correctly to have single-gland disease. The predictive accuracy increased when stratifying by preoperative calcium (≥10.5 mg/dL, ≥11 mg/dL, and ≥12 mg/dL) and parathyroid hormone levels (≥65 pg/mL, ≥100 pg/mL, and ≥200 pg/dL). The accuracy further increased when stratifying by age (≤50 years). Accuracy for single gland disease was 100% when combined with any of the following: (1) calcium ≥12 mg/dL, (2) parathyroid hormone ≥200 pg/dL, or (3) calcium ≥11 mg/dL in patients ≤50 years. CONCLUSION/CONCLUSIONS:Four-dimensional computed tomography alone accurately predicted single gland disease in 87% of patients with primary hyperparathyroidism. When combined with preoperative calcium, parathyroid hormone and age thresholds, predictive accuracy for single-gland disease approached 100%. Given the high likelihood of single-gland disease in these scenarios, clinicians may consider offering focused unilateral parathyroidectomy without intraoperative parathyroid hormone monitoring in selected patients.

BACKGROUND:Racial and ethnic disparities in thyroid cancer care may be mitigated by improving enrollment of more diverse patient populations in clinical trials. We studied trial eligibility criteria and enrollment to assess barriers to equitable representation. METHODS:ClinicalTrials.gov was searched for studies on thyroid cancer treatment conducted between 1993 and 2023. The inclusion and exclusion criteria of each study were examined. For published studies, reported demographic information was collected. Observed enrollment by race was compared with the expected distribution as determined using data from the US Census and the National Cancer Institute's Surveillance, Epidemiology, and End Results (SEER) databases. Over- and under-representation was defined as the ratio of observed to expected (O/E) enrollment by the race and ethnicity group. RESULTS:Of 309 thyroid cancer-related trials, 23 (7.4%) used language as an exclusion criterion. Most were interventional (n = 239, 77.3%), university-initiated (194, 62.8%), and drug/device-focused (195, 63.1%). Of studies that excluded by language, 20 (87.0%) were university-initiated. Eighty-eight trials were subsequently published, with 16 (18.2%) reporting race and/or ethnicity distributions. When comparing O/E ratios, White American participants were over-represented (O/E ratio: 1.2, P < .0001). Under-represented groups included Asian/Native Hawaiian (O/E ratio: 0.6, P = .0085), Black (0.6, P = .014), Native American (0.2, P = .072), and Hispanic patients (0.2, P < .0001). CONCLUSION/CONCLUSIONS:Over the last 3 decades, 1 in 13 thyroid cancer-related clinical trials excluded patients based on language. In the fraction of published studies to report on racial and ethnic demographics, Asian/Native Hawaiian, Black, and Hispanic patients were under-represented. Improved reporting of demographics in published studies and elimination of exclusion criteria such as language that hinder enrollment of minority patients could improve equitable representation of patients in thyroid cancer clinical trials.

BACKGROUND:Patient electronic messaging has increased clinician workload contributing to burnout. Large language models can respond to these patient queries, but no studies exist on large language model responses in thyroid disease. METHODS:This cross-sectional study randomly selected 33 of 52 patient questions found on Reddit/askdocs. Questions were found through a "thyroid + cancer" or "thyroid + disease" search and had verified-physician responses. Additional responses were generated using ChatGPT-3.5 and GPT-4. Questions and responses were anonymized and graded for accuracy, quality, and empathy using a 4-point Likert scale by blinded providers, including 4 surgeons, 1 endocrinologist, and 2 physician assistants (n = 7). Results were analyzed using a single-factor analysis of variance. RESULTS:For accuracy, the results averaged 2.71/4 (standard deviation 1.04), 3.49/4 (0.391), and 3.66/4 (0.286) for physicians, GPT-3.5, and GPT-4, respectively (P < .01), where 4 = completely true information, 3 = greater than 50% true information, and 2 = less than 50% true information. For quality, the results were 2.37/4 (standard deviation 0.661), 2.98/4 (0.352), and 3.81/4 (0.36) for physicians, GPT-3.5, and GPT-4, respectively (P < .01), where 4 = provided information beyond what was asked, 3 = completely answers the question, and 2 = partially answers the question. For empathy, the mean scores were 2.37/4 (standard deviation 0.661), 2.80/4 (0.582), and 3.14/4 (0.578) for physicians, GPT-3.5, and GPT-4, respectively (P < .01), where 4 = anticipates and infers patient feelings from the expressed question, 3 = mirrors the patient's feelings, and 2 = contains no dismissive comments. Responses by GPT were ranked first 95% of the time. CONCLUSIONS:Large language model responses to patient queries about thyroid disease have the potential to be more accurate, complete, empathetic, and consistent than physician responses.

BACKGROUND:Radioactive iodine therapy (RAI) is a frequently chosen therapy for Graves' disease. The aim of this study was to determine whether RAI for Graves' disease increases the risk of thyroid malignancy. METHODS:A retrospective analysis was performed of all Graves' disease patients who underwent thyroidectomy at a single institution between 2013 and 2022. Comparative analyses were performed with cohorts based on RAI therapy as the primary grouping variable. RESULTS:413 patients were identified, of which 38 received RAI prior to surgery. RAI treated patients were more likely to undergo surgery for known malignancy or indeterminate nodules. RAI patients were also more likely to have malignancies larger than 1 ​cm. Among RAI treated patients, those who developed malignancy were older at the time of Graves' diagnosis and received early RAI therapy. CONCLUSIONS:Use of RAI for treatment of Graves' disease increases the progression of thyroid carcinoma, but not the prevalence. Older age and early RAI therapy may be risk factors for malignancy in RAI treated patients.

Thyroid carcinoma of follicular cell origin exists on a histopathologic and clinical spectrum. The authors focus on the category of tumors that fall between the very favorable well-differentiated thyroid carcinomas and the very unfavorable anaplastic thyroid carcinomas. These intermediately aggressive tumors include poorly differentiated thyroid carcinoma and the newly defined differentiated high-grade thyroid carcinoma. Both diagnoses require certain histopathologic requirements be met in order to accurately identify these tumors post-operatively. Management remains primarily surgical though adjunctive treatments such as molecular targeted therapies (eg, tyrosine kinase inhibitors) and differentiation therapy (to restore tumor response to radioactive iodine) are also becoming available.