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 to automate the extraction and classification of thyroid cancer risk factors from pathology reports. METHODS: We analyzed 1410 surgical pathology reports from adult papillary thyroid cancer patients from 2010 to 2019. Structured and nonstructured reports were used to create a consensus-based ground truth dictionary and categorized them into modified recurrence risk levels. Nonstructured reports were narrative, while structured reports followed standardized formats. We developed ThyroPath, a rule-based Natural Language Processing 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.

A Systematic Review of Natural Language Processing Methods and Applications in Thyroidology.

This study aimed to review the application of natural language processing (NLP) in thyroid-related conditions and to summarize current challenges and potential future directions. We performed a systematic search of databases for studies describing NLP applications in thyroid conditions published in English between January 1, 2012 and November 4, 2022. In addition, we used a snowballing technique to identify studies missed in the initial search or published after our search timeline until April 1, 2023. For included studies, we extracted the NLP method (eg, rule-based, machine learning, deep learning, or hybrid), NLP application (eg, identification, classification, and automation), thyroid condition (eg, thyroid cancer, thyroid nodule, and functional or autoimmune disease), data source (eg, electronic health records, health forums, medical literature databases, or genomic databases), performance metrics, and stages of development. We identified 24 eligible NLP studies focusing on thyroid-related conditions. Deep learning-based methods were the most common (38%), followed by rule-based (21%), and traditional machine learning (21%) methods. Thyroid nodules (54%) and thyroid cancer (29%) were the primary conditions under investigation. Electronic health records were the dominant data source (17/24, 71%), with imaging reports being the most frequently used (15/17, 88%). There is increasing interest in NLP applications for thyroid-related studies, mostly addressing thyroid nodules and using deep learning-based methodologies with limited external validation. However, none of the reviewed NLP applications have reached clinical practice. Several limitations, including inconsistent clinical documentation and model portability, need to be addressed to promote the evaluation and implementation of NLP applications to support patient care in thyroidology.

Inclusion of Rurality and Social Determinants of Health in Documents for the Primary Prevention of Type 2 Diabetes: A Systematic Review.

Purpose: The type 2 diabetes (T2D) burden is disproportionately concentrated in low- and middle-income economies, particularly among rural populations. The purpose of the systematic review was to evaluate the inclusion of rurality and social determinants of health (SDOH) in documents for T2D primary prevention. Methods: This systematic review is reported following Preferred Reporting Items for Systematic Reviews and Meta-Analysis guidelines. We searched 19 databases, from 2017-2023, for documents on rurality and T2D primary prevention. Furthermore, we searched online for documents from the 216 World Bank economies, categorized by high, upper-middle, lower-middle, and low income status. We extracted data on rurality and the ten World Health Organization SDOH. Two authors independently screened documents and extracted data. Findings: Based on 3318 documents (19 databases and online search), we selected 15 documents for data extraction. The 15 documents applied to 32 economies; 12 of 15 documents were from nongovernment sources, none was from low-income economies, and 10 of 15 documents did not define or describe rurality. Among the SDOH, income and social protection (SDOH 1) and social inclusion and nondiscrimination (SDOH 8) were mentioned in documents for 25 of 29 high-income economies, while food insecurity (SDOH 5) and housing, basic amenities, and the environment (SDOH 6) were mentioned in documents for 1 of 2 lower-middle-income economies. For U.S. documents, none of the authors was from institutions in noncore (most rural) counties. Conclusions: Overall, documents on T2D primary prevention had sparse inclusion of rurality and SDOH, with additional disparity based on economic status. Inclusion of rurality and/or SDOH may improve T2D primary prevention in rural populations.

Thyroid Ultrasound Appropriateness Identification Through Natural Language Processing of Electronic Health Records.

Objective: To address thyroid cancer overdiagnosis, we aim to develop a natural language processing (NLP) algorithm to determine the appropriateness of thyroid ultrasounds (TUS). Patients and Methods: Between 2017 and 2021, we identified 18,000 TUS patients at Mayo Clinic and selected 628 for chart review to create a ground truth dataset based on consensus. We developed a rule-based NLP pipeline to identify TUS as appropriate TUS (aTUS) or inappropriate TUS (iTUS) using patients' clinical notes and additional meta information. In addition, we designed an abbreviated NLP pipeline (aNLP) solely focusing on labels from TUS order requisitions to facilitate deployment at other health care systems. Our dataset was split into a training set of 468 (75%) and a test set of 160 (25%), using the former for rule development and the latter for performance evaluation. Results: There were 449 (95.9%) patients identified as aTUS and 19 (4.06%) as iTUS in the training set; there are 155 (96.88%) patients identified as aTUS and 5 (3.12%) were iTUS in the test set. In the training set, the pipeline achieved a sensitivity of 0.99, specificity of 0.95, and positive predictive value of 1.0 for detecting aTUS. The testing cohort revealed a sensitivity of 0.96, specificity of 0.80, and positive predictive value of 0.99. Similar performance metrics were observed in the aNLP pipeline. Conclusion: The NLP models can accurately identify the appropriateness of a thyroid ultrasound from clinical documentation and order requisition information, a critical initial step toward evaluating the drivers and outcomes of TUS use and subsequent thyroid cancer overdiagnosis.

Physical Examination of the Thyroid: Accuracy in Detecting Thyroid Nodules and Frequency of Additional Findings.

OBJECTIVE: Thyroid palpation is a common clinical practice to detect thyroid abnormalities. However, its accuracy and potential for additional findings remain unclear. This study aimed to assess the diagnostic accuracy of physical exams in detecting thyroid nodules. METHODS: A retrospective observational study was conducted on a random sample of adult patients who underwent their first-time thyroid ultrasound between January 2015 and September 2017, following a documented thyroid physical exam. The study assessed the performance of thyroid palpation in detecting 1 or multiple thyroid nodules, as well as the proportion of additional findings on ultrasounds due to false positive thyroid palpation. RESULTS: We included 327 patients, mostly female (65.1%), white (84.1%), and treated in a primary care setting (54.4%) with a mean age of 50.8 years (SD 16.9). For solitary thyroid nodules, the physical exam had a sensitivity of 20.3%, specificity of 79.1%, an accuracy of 68.5%, negative predictive value of 81.8%, and positive predictive value of 17.6%. For detecting a multinodular goiter, physical exams demonstrated a sensitivity of 10.8%, specificity of 96.5%, accuracy of 55.4%, negative predictive value of 53.9, and positive predictive value of 73.9%. Among 154 cases with palpable nodules, 60% had additional nodules found in subsequent thyroid ultrasound. CONCLUSION: Thyroid physical exam has limited diagnostic performance and leads to additional findings when followed by a thyroid ultrasound. Future efforts should be directed at improving the accuracy of thyroid physical exams or re-evaluating its routine use.

Prognostic and Diagnostic Utility of Serum Biomarkers in Pediatric Traumatic Brain Injury.

Traumatic brain injury (TBI) remains a major cause of morbidity and death among the pediatric population. Timely diagnosis, however, remains a complex task because of the lack of standardized methods that permit its accurate identification. The aim of this study was to determine whether serum levels of brain injury biomarkers can be used as a diagnostic and prognostic tool in this pathology. This prospective, observational study collected and analyzed the serum concentration of neuronal injury biomarkers at enrollment, 24h and 48h post-injury, in 34 children ages 0-18 with pTBI and 19 healthy controls (HC). Biomarkers included glial fibrillary acidic protein (GFAP), neurofilament protein L (NfL), ubiquitin-C-terminal hydrolase (UCH-L1), S-100B, tau and tau phosphorylated at threonine 181 (p-tau181). Subjects were stratified by admission Glasgow Coma Scale score into two categories: a combined mild/moderate (GCS 9-15) and severe (GCS 3-8). Glasgow Outcome Scale-Extended (GOS-E) Peds was dichotomized into favorable (≤4) and unfavorable (≥5) and outcomes. Data were analyzed utilizing Prism 9 and R statistical software. The findings were as follows: 15 patients were stratified as severe TBI and 19 as mild/moderate per GCS. All biomarkers measured at enrollment were elevated compared with HC. Serum levels for all biomarkers were significantly higher in the severe TBI group compared with HC at 0, 24, and 48h. The GFAP, tau S100B, and p-tau181 had the ability to differentiate TBI severity in the mild/moderate group when measured at 0h post-injury. Tau serum levels were increased in the mild/moderate group at 24h. In addition, NfL and p-tau181 showed increased serum levels at 48h in the aforementioned GCS category. Individual biomarker performance on predicting unfavorable outcomes was measured at 0, 24, and 48h across different GOS-E Peds time points, which was significant for p-tau181 at 0h at all time points, UCH-L1 at 0h at 6-9 months and 12 months, GFAP at 48h at 12 months, NfL at 0h at 12 months, tau at 0h at 12 months and S100B at 0h at 12 months. We concluded that TBI leads to increased serum neuronal injury biomarkers during the first 0-48h post-injury. A biomarker panel measuring these proteins could aid in the early diagnosis of mild to moderate pTBI and may predict neurological outcomes across the injury spectrum.

Drivers of Thyroid Ultrasound Use: A Retrospective Observational Study.

OBJECTIVE: Excessive use of thyroid ultrasound (TUS) contributes to the overdiagnosis of thyroid nodules and thyroid cancer. In this study, we evaluated drivers of and clinical trajectories following TUS orders. METHODS: We conducted a retrospective review of 500 adult patients who underwent an initial TUS between 2015 and 2017 at Mayo Clinic in Rochester, MN. A framework was employed to classify the indication for TUS, and it was characterized as inappropriate when ordered without a guideline-based indication. Medical records were reviewed for up to 12 months following the TUS, and clinical outcomes were evaluated. RESULTS: The mean age mean age (SD) was 53.6 years (16.6), 63.8% female, and 86.6% white. TUS orders were triggered by incidental findings on unrelated imaging (31.6%), thyroid symptoms (20.4%), thyroid abnormalities on routine physical examination (17.2%), and thyroid dysfunction workup (11.8%). In females and males, the most common reason were incidental findings on imaging (female, 91/319, 28.5% and male, 67/181, 37.0%). In primary care practice, TUS orders were mostly triggered by symptoms (71/218, 32.5%), while thyroid dysfunction workup was the primary reason in endocrinology (28/100, 28.0%). We classified 11.2% (56/500) TUS orders as likely to have been ordered inappropriately based on current guidelines. Finally, 119 patients (119/500, 23.8%) had a thyroid biopsy with 11.8% had thyroid cancer (14/119. 11.8%). CONCLUSIONS: Incidental findings on imaging, symptoms, and routine physical exam findings in asymptomatic patients were the most prevalent drivers of TUS. Furthermore, 1 in 10 TUS were likely inappropriately ordered based on current practice guidelines.

Artificial Intelligence in Thyroidology: A Narrative Review of the Current Applications, Associated Challenges, and Future Directions.

Background: The use of artificial intelligence (AI) in health care has grown exponentially with the promise of facilitating biomedical research and enhancing diagnosis, treatment, monitoring, disease prevention, and health care delivery. We aim to examine the current state, limitations, and future directions of AI in thyroidology. Summary: AI has been explored in thyroidology since the 1990s, and currently, there is an increasing interest in applying AI to improve the care of patients with thyroid nodules (TNODs), thyroid cancer, and functional or autoimmune thyroid disease. These applications aim to automate processes, improve the accuracy and consistency of diagnosis, personalize treatment, decrease the burden for health care professionals, improve access to specialized care in areas lacking expertise, deepen the understanding of subtle pathophysiologic patterns, and accelerate the learning curve of less experienced clinicians. There are promising results for many of these applications. Yet, most are in the validation or early clinical evaluation stages. Only a few are currently adopted for risk stratification of TNODs by ultrasound and determination of the malignant nature of indeterminate TNODs by molecular testing. Challenges of the currently available AI applications include the lack of prospective and multicenter validations and utility studies, small and low diversity of training data sets, differences in data sources, lack of explainability, unclear clinical impact, inadequate stakeholder engagement, and inability to use outside of the research setting, which might limit the value of their future adoption. Conclusions: AI has the potential to improve many aspects of thyroidology; however, addressing the limitations affecting the suitability of AI interventions in thyroidology is a prerequisite to ensure that AI provides added value for patients with thyroid disease.