Use of natural language processing to extract and classify papillary thyroid cancer features from surgical pathology reports.

BACKGROUND: We aim to use Natural Language Processing (NLP) to automate the extraction and classification of thyroid cancer risk factors from pathology reports. METHODS: We analyzed 1,410 surgical pathology reports from adult papillary thyroid cancer patients from 2010 to 2019. Structured and non-structured reports were used to create a consensus-based ground truth dictionary and categorized them into modified recurrence risk levels. Non-structured reports were narrative, while structured reports followed standardized formats. We developed ThyroPath, a rule-based NLP pipeline, to extract and classify thyroid cancer features into risk categories. Training involved 225 reports (150 structured, 75 unstructured), with testing on 170 reports (120 structured, 50 unstructured) for evaluation. The pipeline's performance was assessed using both strict and lenient criteria for accuracy, precision, recall, and F1-score, a metric that combines precision and recall evaluation. RESULTS: In extraction tasks, ThyroPath achieved overall strict F-1 scores of 93% for structured reports and 90% for unstructured reports, covering 18 thyroid cancer pathology features. In classification tasks, ThyroPath-extracted information demonstrated an overall accuracy of 93% in categorizing reports based on their corresponding guideline-based risk of recurrence: 76.9% for high-risk, 86.8% for intermediate risk, and 100% for both low and very low-risk cases. However, ThyroPath achieved 100% accuracy across all risk categories with human extracted pathology information. CONCLUSIONS: ThyroPath shows promise in automating the extraction and risk recurrence classification of thyroid pathology reports at large scale. It offers a solution to laborious manual reviews and advancing virtual registries. However, it requires further validation before implementation.

Interventions to improve symptomatology in patients with hypothyroidism and persistent symptoms: A systematic review.

BACKGROUND: Levothyroxine (LT4) monotherapy is the standard treatment for hypothyroidism; however, 10-15% of patients have persistent hypothyroid symptoms despite normalizing thyroid hormone levels with LT4. This study aims to summarize the best available evidence on interventions to improve symptomatology in patients with hypothyroidism and persistent symptoms. METHODS: A systematic search was conducted in March 2022 for randomized controlled trials and observational studies on interventions for adult patients with persistent hypothyroid symptoms despite biochemical euthyroidism on thyroid hormone replacement. RESULTS: A total of 277 articles were reviewed and seven fulfilled the inclusion criteria. 455 participants were included. Most intervention participants were female (78.6%) with a mean age of 47.5 (±2.8) years. Five clinical trials evaluating ginger (vs. starch), L-carnitine (vs. placebo), combination LT4 and liothyronine (LT3) (vs. LT4 or placebo), and surgery for patients with serum antithyroid peroxidase (TPO Ab) titers greater than 1000 IU/ml (vs. LT4) found inconsistent improvement in hypothyroidism related symptoms and general health. The two clinical trials with the largest improvement in fatigue scores were the use of ginger and surgery. One observational study comparing thyroidectomy vs observation found no significant difference on general health. Lastly, another observational study evaluating combination LT4/LT3 (vs. LT4 monotherapy) found improvement in fatigue and quality of life. There were 31 (12%) adverse events in the intervention group and 18 (10.8%) in the comparator group. CONCLUSIONS: There is no high-quality evidence supporting any intervention for persistent symptoms in hypothyroidism. Available evidence, limited by the risk of bias, inconsistency, and heterogeneity, suggests that some persistent symptoms, particularly fatigue, could improve with ginger and thyroidectomy.

Development and pilot testing of a conversation aid to support the evaluation of patients with thyroid nodules.

OBJECTIVE: To support patient-centred care and the collaboration of patients and clinicians, we developed and pilot tested a conversation aid for patients with thyroid nodules. DESIGN, PATIENT AND MEASUREMENTS: We developed a web-based Thyroid NOdule Conversation aid (TNOC) following a human-centred design. A proof of concept observational pre-post study was conducted (TNOC vs. usual care [UC]) to assess the impact of TNOC on the quality of conversations. Data sources included recordings of clinical visits, post-encounter surveys and review of electronic health records. Summary statistics and group comparisons are reported. RESULTS: Sixty-five patients were analysed (32 in the UC and 33 in the TNOC cohort). Most patients were women (89%) with a median age of 57 years and were incidentally found to have a thyroid nodule (62%). Most thyroid nodules were at low risk for thyroid cancer (71%) and the median size was 1.4 cm. At baseline, the groups were similar except for higher numeracy in the TNOC cohort. The use of TNOC was associated with increased involvement of patients in the decision-making process, clinician satisfaction and discussion of relevant topics for decision making. In addition, decreased decisional conflict and fewer thyroid biopsies as the next management step were noted in the TNOC cohort. No differences in terms of knowledge transfer, length of consultation, thyroid cancer risk perception or concern for thyroid cancer diagnosis were found. CONCLUSION: In this pilot observational study, using TNOC in clinical practice was feasible and seemed to help the collaboration of patients and clinicians.

Pituitary adenomas in the setting of multiple endocrine neoplasia type 1: a single-institution experience.

OBJECTIVE: Multiple endocrine neoplasia type 1 (MEN1) is a rare, autosomal-dominant tumor disorder characterized by the development of pituitary tumors and other endocrine neoplasms. Diagnosis is made clinically based on the development of 2 or more canonical lesions (parathyroid gland, anterior pituitary, and enteropancreatic tumors) or in family members of a patient with a clinical diagnosis of MEN1 and the occurrence of one of the MEN1-associated tumors. The goal of this study was to characterize pituitary tumors arising in the setting of MEN1 at a single institution. The probability of tumor progression and the likelihood of surgical intervention in patients with asymptomatic nonfunctional pituitary adenomas were also analyzed. METHODS: A retrospective review of a prospectively maintained institutional database was performed for patients with MEN1 diagnosed from 1970 to 2017. Data included patient demographics, tumor characteristics, treatment strategies, and outcomes. RESULTS: A review of the database identified 268 patients diagnosed with MEN1, of whom 158 (59%) were female. Among the 268 patients, 139 (51.8%) had pituitary adenomas. There was a higher prevalence in women than in men (65% vs 35%, p < 0.005). Functional adenomas (57%) were more common. Prolactin-secreting adenomas were the most common functional tumors. Macroadenomas were seen in 27% of patients and were more likely to be symptomatic and locally aggressive (p < 0.001). Forty-nine patients (35%) underwent transsphenoidal resection at some point during their disease course. In 52 patients who were initially observed with MEN1 asymptomatic nonfunctional adenomas, only 5 (10%) progressed to need surgery. In MEN1 patients, an initial parathyroid lesion is most likely followed in order by pituitary, pancreatic, adrenal, and, finally, rare carcinoid tumors. CONCLUSIONS: Asymptomatic nonfunctional pituitary adenomas in patients with MEN1 may be followed safely with MRI. In this series, parathyroid tumors developed at the lowest median age of all cardinal tumors, and development of additional cardinal MEN1 lesions followed a predictable pattern. This pattern of disease progression could have significant implications for disease surveillance in clinical practice and may help to target clinical resources to the lesions most likely to develop next. This may aid with early detection and treatment and warrants further study.

Ethanol Ablation for the Treatment of Cystic and Predominantly Cystic Thyroid Nodules.

OBJECTIVE: To determine the efficacy and safety of percutaneous ethanol injection (PEI) for the treatment of symptomatic cystic thyroid nodules. PATIENTS AND METHODS: Retrospective analysis of patients with benign cystic thyroid nodules treated with PEI from February 1, 2000, through October 31, 2016. The main outcomes were efficacy, defined as symptom relief or reduction in nodule volume of 50% or more, and safety, defined as no or minor adverse events. RESULTS: Twenty patients had PEI. Mean age at the time of PEI was 50 years, and 13 (65%) were women; all patients were euthyroid. Twelve patients (60%) had complex cystic thyroid nodules (>50% cystic component), with the rest being purely cystic. The median largest diameter of the thyroid cyst was 4.5 cm (interquartile range [IQR], 3.2-5.3 cm; range, 2.3-8.0 cm); the median volume pre-PEI was 19.6 mL (IQR, 10.4-48.5 mL; range, 2.8-118.1 mL). The median amount of cystic fluid drained before PEI was 13.5 mL (IQR, 6.8-32.3 mL), and the median amount of ethanol administered was 3 mL (IQR, 2-5 mL; range, 0.5-20 mL). After median follow-up of 2 years, 17 of 19 patients (89%) were asymptomatic. Of 10 patients with available imaging on follow-up, 7 (70%) had a 50% or greater reduction in nodule volume (median volume decrease, 75.64% [IQR, 41.40%-91.99%]). Adverse effects occurred in 4 patients (20%) and were mild and temporary (slight pain, vagal reaction, and bleeding into the cyst). CONCLUSION: Percutaneous ethanol injection seems to be a safe and effective alternative to surgical resection for patients with purely or predominantly cystic thyroid nodules and compressive symptoms who decline surgery or are not good surgical candidates.

Outcomes of Parathyroidectomy in Patients with Primary Hyperparathyroidism: A Systematic Review and Meta-analysis.

BACKGROUND: Parathyroidectomy is a definitive treatment for primary hyperparathyroidism. Patients contemplating this intervention will benefit from knowledge regarding the expected outcomes and potential risks of the currently available surgical options. PURPOSE: To appraise and summarize the available evidence regarding benefits and harms of minimally invasive parathyroidectomy (MIP) and bilateral neck exploration (BNE). DATA SOURCES: A comprehensive search of multiple databases (MEDLINE, EMBASE, and Scopus) from each database's inception to September 2014 was performed. STUDY SELECTION: Eligible studies evaluated patients with primary hyperparathyroidism undergoing MIP or BNE. DATA EXTRACTION: Reviewers working independently and in duplicate extracted data and assessed the risk of bias. DATA SYNTHESIS: We identified 82 observational studies and 6 randomized trials at moderate risk of bias. Most of them reported outcomes after MIP (n = 71). Using random-effects models to pool results across studies, the cure rate was 98 % (95 % CI 97-98 %, I (2) = 10 %) with BNE and 97 % (95 % CI 96-98 %, I (2) = 86 %) with MIP. Hypocalcemia occurred in 14 % (95 % CI 10-17 % I (2) = 93 %) of the BNE cases and in 2.3 % (95 % CI 1.6-3.1 %, I (2) = 87 %) with MIP (P < 0.001). There was a statistically significant lower risk of laryngeal nerve injury with MIP (0.3 %) than with BNE (0.9 %), but similar risk of infection (0.5 vs. 0.5 %) and mortality (0.1 vs. 0.5 %). LIMITATIONS: The available evidence, mostly observational, is at moderate risk of bias, and limited by indirect comparisons and inconsistency for some outcomes (cure rate, hypocalcemia). CONCLUSION: MIP and BNE are both effective surgical techniques for the treatment of primary hyperparathyroidism. The safety profile of MIP appears superior to BNE (lower rate of hypocalcemia and recurrent laryngeal nerve injury).

PREDICTORS OF BIOCHEMICAL REMISSION AND RECURRENCE AFTER SURGICAL AND RADIATION TREATMENTS OF CUSHING DISEASE: A SYSTEMATIC REVIEW AND META-ANALYSIS.

OBJECTIVE: We conducted a systematic review and meta-analysis to synthesize the evidence about predictors that may affect biochemical remission and recurrence after transsphenoidal surgery (TSS), radiosurgery (RS), and radiotherapy (RT) in Cushing disease. METHODS: We searched multiple databases through December 2014 including original controlled and uncontrolled studies that enrolled patients with Cushing disease who received TSS (first-line), RS, or RT. We extracted data independently, in duplicates. Outcomes of interest were biochemical remission and recurrence. A meta-analysis was conducted using the random-effects model to estimate event rates with 95% confidence intervals (CIs). RESULTS: First-line TSS was associated with high remission (76% [95% CI, 72 to 79%]) and low recurrence rates (10% [95% CI, 6 to 16%]). Remission after TSS was higher in patients with microadenomas or positive-adrenocorticotropic hormone tumor histology. RT was associated with a high remission rate (RS, 68% [95% CI, 61 to 77%]; RT, 66% [95% CI, 58 to 75%]) but also with a high recurrence rate (RS, 32% [95% CI, 16 to 60%]; RT, 26% [95% CI, 14 to 48%]). Remission after RS was higher at short-term follow-up (≤2 years) and with high-dose radiation, while recurrence was higher in women and with lower-dose radiation. Remission was after RT in adults who received TSS prior to RT, and with lower radiation doses. There was heterogeneity (nonstandardization) in the criteria and cutoff points used to define biochemical remission and recurrence. CONCLUSION: First-line TSS is associated with high remission and low recurrence, while RS and RT are associated with reasonable remission rates but important recurrence rates. The current evidence warrants low confidence due to the noncomparative nature of the studies, high heterogeneity, and imprecision.