Deep learning assists in acute leukemia detection and cell classification via flow cytometry using the acute leukemia orientation tube.

This study aimed to evaluate the sensitivity of AI in screening acute leukemia and its capability to classify either physiological or pathological cells. Utilizing an acute leukemia orientation tube (ALOT), one of the protocols of Euroflow, flow cytometry efficiently identifies various forms of acute leukemia. However, the analysis of flow cytometry can be time-consuming work. This retrospective study included 241 patients who underwent flow cytometry examination using ALOT between 2017 and 2022. The collected flow cytometry data were used to train an artificial intelligence using deep learning. The trained AI demonstrated a 94.6% sensitivity in detecting acute myeloid leukemia (AML) patients and a 98.2% sensitivity for B-lymphoblastic leukemia (B-ALL) patients. The sensitivities of physiological cells were at least 80%, with variable performance for pathological cells. In conclusion, the AI, trained with ResNet-50 and EverFlow, shows promising results in identifying patients with AML and B-ALL, as well as classifying physiological cells.

Utility of polygenic scores across diverse diseases in a hospital cohort for predictive modeling.

Polygenic scores estimate genetic susceptibility to diseases. We systematically calculated polygenic scores across 457 phenotypes using genotyping array data from China Medical University Hospital. Logistic regression models assessed polygenic scores' ability to predict disease traits. The polygenic score model with the highest accuracy, based on maximal area under the receiver operating characteristic curve (AUC), is provided on the GeneAnaBase website of the hospital. Our findings indicate 49 phenotypes with AUC greater than 0.6, predominantly linked to endocrine and metabolic diseases. Notably, hyperplasia of the prostate exhibited the highest disease prediction ability (P value = 1.01 × 10-19, AUC = 0.874), highlighting the potential of these polygenic scores in preventive medicine and diagnosis. This study offers a comprehensive evaluation of polygenic scores performance across diverse human traits, identifying promising applications for precision medicine and personalized healthcare, thereby inspiring further research and development in this field.

Association of COVID-19 Infection with Subsequent Thyroid Dysfunction: An International Population-Based Propensity Score Matched Analysis.

Background: The COVID-19 pandemic's impact on thyroid function is a growing concern. Previous studies have produced inconclusive results, and there is a lack of comprehensive research into the long-term risks of thyroid dysfunction following COVID-19 infection. Methods: In this retrospective cohort study, we used data from the TriNetX international database, which includes electronic health records from a broad, diverse patient population. We compared patients with COVID-19 (cases) to those without (controls), matching for age, sex, race, and comorbidities using propensity score matching. The primary outcome was the diagnosis of thyroid dysfunction (thyrotoxicosis or hypothyroidism) within a 12-month period, analyzed using hazard ratios (HRs) and Kaplan-Meier curves, and stratified by age and sex. Results: Initially, the study included 1,379,311 COVID-19 patients and 6,896,814 non-COVID-19 patients from the TriNetX database. After matching, the cohorts were comparable in demographics and baseline characteristics. This study consistently demonstrated a significant increase in the risk of thyroid dysfunction, including thyrotoxicosis and hypothyroidism, among COVID-19 patients compared to non-COVID-19 patients. In the short term (3 months postexposure), the COVID-19 group exhibited a HR of 2.07 (95% confidence interval [CI] 2.01-2.12) for thyroid dysfunction, which included both thyrotoxicosis (HR 2.10, CI 1.92-2.29) and hypothyroidism (HR 2.08, CI 2.01-2.13). This heightened risk persisted over the long term (up to 12 months), with HRs indicating an ∼2.01-fold increased risk for overall thyroid dysfunction, a 1.8-fold increased risk for thyrotoxicosis, and a 2.04-fold increased risk for hypothyroidism. Subgroup analysis, stratified by age and sex, revealed a notably higher risk of thyroid dysfunction in patients aged 65 and above (HR 2.18, CI 2.11-2.25), compared to those in the under-65 age group (HR 1.97, CI 1.91-2.03). Both male and female patients were associated with an elevated risk, with females showing a slightly higher association with thyroid dysfunction (HR 2.12, CI 2.06-2.16) compared to males (HR 1.76, CI 1.69-1.82). Conclusions: COVID-19 infection was associated with an increased risk of thyroid dysfunction, including thyrotoxicosis and hypothyroidism, regardless of age or sex, during a 12-month follow-up period. Further research is required to validate these findings.

Pioglitazone, SGLT2 inhibitors and their combination for primary prevention of cardiovascular disease and heart failure in type 2 diabetes: Real-world evidence from a nationwide cohort database.

OBJECTIVE: To evaluate the effect of SGLT2is, pioglitazone, and their combination on the risk of major adverse cardiovascular events (MACE) and heart failure in type 2 diabetes mellitus (T2DM) patients without a history of cardiovascular disease. RESEARCH DESIGN AND METHODS: Using Taiwan National Health Insurance Research Database, we identified four groups based on medication use, including 1) both SGLT2is and pioglitazone, 2) SGLT2i, 3) pioglitazone and 4) non-study drugs (reference group). The four groups were matched by propensity score. The primary outcome was 3-point MACE, which included myocardial infarction, stroke, cardiovascular death, and the secondary outcome was incidence of heart failure. RESULTS: After propensity-matching, each group included 15,601 patients. Compared with the reference group, the pioglitazone/SGLT2i combination group had a significantly lower risk for MACE (aHR 0.76, 95 % CI 0.66-0.88) and heart failure (aHR 0.67, 95 % CI 0.55-0.82). Pioglitazone was associated with a lower risk of MACE (aHR 0.82, 95 % CI 0.71-0.94) and there was no difference in risk of heart failure compared with the reference group. The incidence of heart failure was significantly decreased in the SGLT2i group (aHR 0.7, 95 % CI 0.58-0.86). CONCLUSION: Combination therapy with pioglitazone and SGLT2is is an effective treatment in the primary prevention of MACE and heart failure in patients with type 2 diabetes.

Weakly supervised deep learning for prediction of treatment effectiveness on ovarian cancer from histopathology images.

Despite the progress made during the last two decades in the surgery and chemotherapy of ovarian cancer, more than 70 % of advanced patients are with recurrent cancer and decease. Surgical debulking of tumors following chemotherapy is the conventional treatment for advanced carcinoma, but patients with such treatment remain at great risk for recurrence and developing drug resistance, and only about 30 % of the women affected will be cured. Bevacizumab is a humanized monoclonal antibody, which blocks VEGF signaling in cancer, inhibits angiogenesis and causes tumor shrinkage, and has been recently approved by FDA as a monotherapy for advanced ovarian cancer in combination with chemotherapy. Considering the cost, potential toxicity, and finding that only a portion of patients will benefit from these drugs, the identification of new predictive method for the treatment of ovarian cancer remains an urgent unmet medical need. In this study, we develop weakly supervised deep learning approaches to accurately predict therapeutic effect for bevacizumab of ovarian cancer patients from histopathological hematoxylin and eosin stained whole slide images, without any pathologist-provided locally annotated regions. To the authors' best knowledge, this is the first model demonstrated to be effective for prediction of the therapeutic effect of patients with epithelial ovarian cancer to bevacizumab. Quantitative evaluation of a whole section dataset shows that the proposed method achieves high accuracy, 0.882 ± 0.06; precision, 0.921 ± 0.04, recall, 0.912 ± 0.03; F-measure, 0.917 ± 0.07 using 5-fold cross validation and outperforms two state-of-the art deep learning approaches Coudray et al. (2018), Campanella et al. (2019). For an independent TMA testing set, the three proposed methods obtain promising results with high recall (sensitivity) 0.946, 0.893 and 0.964, respectively. The results suggest that the proposed method could be useful for guiding treatment by assisting in filtering out patients without positive therapeutic response to suffer from further treatments while keeping patients with positive response in the treatment process. Furthermore, according to the statistical analysis of the Cox Proportional Hazards Model, patients who were predicted to be invalid by the proposed model had a very high risk of cancer recurrence (hazard ratio = 13.727) than patients predicted to be effective with statistical signifcance (p < 0.05).

ADAM Challenge: Detecting Age-Related Macular Degeneration From Fundus Images.

Age-related macular degeneration (AMD) is the leading cause of visual impairment among elderly in the world. Early detection of AMD is of great importance, as the vision loss caused by this disease is irreversible and permanent. Color fundus photography is the most cost-effective imaging modality to screen for retinal disorders. Cutting edge deep learning based algorithms have been recently developed for automatically detecting AMD from fundus images. However, there are still lack of a comprehensive annotated dataset and standard evaluation benchmarks. To deal with this issue, we set up the Automatic Detection challenge on Age-related Macular degeneration (ADAM), which was held as a satellite event of the ISBI 2020 conference. The ADAM challenge consisted of four tasks which cover the main aspects of detecting and characterizing AMD from fundus images, including detection of AMD, detection and segmentation of optic disc, localization of fovea, and detection and segmentation of lesions. As part of the ADAM challenge, we have released a comprehensive dataset of 1200 fundus images with AMD diagnostic labels, pixel-wise segmentation masks for both optic disc and AMD-related lesions (drusen, exudates, hemorrhages and scars, among others), as well as the coordinates corresponding to the location of the macular fovea. A uniform evaluation framework has been built to make a fair comparison of different models using this dataset. During the ADAM challenge, 610 results were submitted for online evaluation, with 11 teams finally participating in the onsite challenge. This paper introduces the challenge, the dataset and the evaluation methods, as well as summarizes the participating methods and analyzes their results for each task. In particular, we observed that the ensembling strategy and the incorporation of clinical domain knowledge were the key to improve the performance of the deep learning models.

The efficacy of pioglitazone for renal protection in diabetic kidney disease.

There is limited information on the efficacy of pioglitazone in diabetic kidney diseases (DKD). We evaluated whether pioglitazone exerts renal-protective effects in DKD patients. We designed a retrospective cohort study, which included 742 type 2 diabetes mellitus (T2DM) patients with DKD in Taiwan, with eGFR between 30 and 90 ml/min/1.73 m2 and UACR level 300-5000 mg/g. Patients not meeting the target range for HbA1c (above 7%) were given additional medication with pioglitazone (n = 111) or received standard care (non-pioglitazone group, n = 631). The primary endpoint was the occurrence of composite renal endpoints, which was defined as sustained eGFR<15 ml/min/1.73 m2 (confirmed by two measurements within 90 days); doubling of serum creatinine (compared to baseline); and the presence of hemodialysis or renal transplantation. The median follow-up duration was two years. At baseline, the mean HbA1C levels in the pioglitazone and non-pioglitazone groups were 8.8% and 8.1%, respectively; mean ages were 64.4 and 66.2 years old, respectively; diabetes durations were 14.3 and 12.3 years, respectively. Baseline eGFR showed no significant difference between the pioglitazone and non-pioglitazone groups (55.8 and 58.8 mL/min/1.73 m2, respectively). In terms of gender, 63% of patients were male in the pioglitazone group compared with 57% in the non-pioglitazone group. Pioglitazone use did not reduce the risk of composite renal endpoints in DKD patients (HR: 0.97, 95% CI = 0.53-1.77), including persistent eGFR<15 ml/min/1.73 m2 (HR = 1.07, 95% CI = 0.46-2.52), doubling of serum creatinine (HR = 0.97, 95% CI = 0.53-1.77), or ESRD (HR = 2.58, 95% CI = 0.29-23.04). The results were not changed after various adjustments. A non-significant albuminuria reduction was also noted after pioglitazone prescription in DKD patients. Further randomized controlled studies are needed to establish the effects of pioglitazone definitively.

Deep Learning for Automated Diabetic Retinopathy Screening Fused With Heterogeneous Data From EHRs Can Lead to Earlier Referral Decisions.

Purpose: Fundus images are typically used as the sole training input for automated diabetic retinopathy (DR) classification. In this study, we considered several well-known DR risk factors and attempted to improve the accuracy of DR screening. Metphods: Fusing nonimage data (e.g., age, gender, smoking status, International Classification of Disease code, and laboratory tests) with data from fundus images can enable an end-to-end deep learning architecture for DR screening. We propose a neural network that simultaneously trains heterogeneous data and increases the performance of DR classification in terms of sensitivity and specificity. In the current retrospective study, 13,410 fundus images and their corresponding nonimage data were collected from the Chung Shan Medical University Hospital in Taiwan. The images were classified as either nonreferable or referable for DR by a panel of ophthalmologists. Cross-validation was used for the training models and to evaluate the classification performance. Results: The proposed fusion model achieved 97.96% area under the curve with 96.84% sensitivity and 89.44% specificity for determining referable DR from multimodal data, and significantly outperformed the models that used image or nonimage information separately. Conclusions: The fusion model with heterogeneous data has the potential to improve referable DR screening performance for earlier referral decisions. Translational Relevance: Artificial intelligence fused with heterogeneous data from electronic health records could provide earlier referral decisions from DR screening.

Association of blood glucose and renal end points in advanced diabetic kidney disease.

AIMS: The association of blood glucose in advanced diabetic kidney disease (DKD) is unclear. This study investigated the association between blood glucose and renal endpoints in DKD patients. METHODS: This retrospective cohort study enrolled type 2 diabetic patients with advanced DKD with an estimated glomerular filtration rate (eGFR) between 30 and 90 ml/min/1.73 m2 and urine albumin-to-creatinine ratio (UACR) of 300-5000 mg/g. We classified patients into 2 groups according to their 1-year average HbA1c: <7% and >7%. We followed up the patients until the occurrence of primary renal endpoints. RESULTS: A total of 345 patients were included in the analysis for the period 2012-2018. Mean baseline eGFR was 58 ml/min/1.73 m2 and mean albuminuria levels were 1146 and 1313 mg/g, respectively. Median study duration was 3 years. The risk of primary renal endpoints was not decreased in patients with HbA1c less than 7% with an adjusted hazard ratio (aHR) of 0.62, 95% CI 0.26-1.45. The risks of persistent eGFR lower than 15 ml/min/1.73 m2 and doubling of serum creatinine level were similar between 2 group with aHR of 0.58 (95% CI 0.19-1.83) and 0.61 (95% CI 0.26-1.44), respectively. CONCLUSIONS: Intensive blood sugar control did not prevent renal failure in advanced DKD.

The Risk of Thyroid Cancer in Patients with Thyroid Nodule 3 Cm Or Larger.

OBJECTIVE: There are conflicting data on the risk of thyroid cancer in thyroid nodules 3 cm or larger, and few such studies on this issue have been conducted in Asia. This study aimed to examine the risk of thyroid cancer in patients with thyroid nodules 3 cm or larger. METHODS: This was a 7-year retrospective study conducted in a tertiary referral hospital in Taiwan. All patients with a thyroid nodule measuring ≥3 cm who underwent thyroid operation with or without fine-needle aspiration biopsy (FNAB) were included. The prevalence rate of thyroid cancer, as well as the sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and false-negative rate of FNAB for thyroid nodule ≥3 cm were also examined. RESULTS: A total of 132 patients were included in this study. Thyroid cancer was detected in 19 of 132 (14.4%) thyroid nodules measuring ≥3 cm. The performance of FNAB for detecting cancer in nodules 3 cm or larger without considering other ultrasonography parameters was relatively poor with a sensitivity of 50%, but the specificity (100%), PPV (100 %), and NPV (93.4 %) were excellent. CONCLUSION: The risk of thyroid cancer for thyroid nodules ≥3 cm in this study was low. The PPV and NPV of FNAB were high for the detection of cancer in large nodules. The decision to perform thyroidectomy should not be solely based on nodule size and should include other factors, such as ultrasound characteristics and surgical risk.

Dipeptidyl-Peptidase 4 Inhibitors did not Improve Renal Endpoints in Advanced Diabetic Kidney Disease.

OBJECTIVE: The efficacy of dipeptidyl-peptidase 4 inhibitors (DPP4is) in advanced diabetic kidney disease (DKD) is unknown. We investigated whether DPP4is confer renal protective benefits in DKD patients. METHODS: We conducted a retrospective cohort study between 2012 and 2018 in Taiwan. We only included type 2 diabetes patients with estimated glomerular filtration rate (eGFR) between 30 and 90 mL/min/1.73 m2 and urine albumin to creatinine ratio between 300 and 5,000 mg/g. Patients with DPP4i prescriptions were selected as cases, while non-DPP4i users served as controls. We followed these patients until the presence of composite primary renal endpoints, which was defined by the earliest occur-rence of clinical renal outcomes. RESULTS: A total of 522 patients were included in the analysis, comprising 273 patients with a DPP4i prescription who were selected as cases and 249 patients without DPP4i prescription who were assigned as controls. Median follow-up duration for DPP4i users and nonusers was 2.2 years and 3.4 years, respectively. At baseline, the mean glycated hemoglobin levels for DPP4i users and nonusers were 8.1% and 8.3%, respectively. Among patients with DPP4i prescriptions, there was no reduction in composite primary renal outcome, with a crude hazard ratio (HR) of 1.50 (95% confidence interval [CI], 0.95 to 2.36). Similar results were observed for the risk of persistent eGFR <15 mL/min/1.73 m2, with a HR of 1.68 (95% CI, 0.90 to 3.13), doubling of serum creatinine level, with a HR of 1.05 (95% CI, 0.15 to 7.45), and end-stage renal disease, with a HR of 0.87 (95% CI, 0.14 to 5.19). CONCLUSION: DPP4i prescription did not reduce the risk of composite renal endpoints in DKD patients. ABBREVIATIONS: BMI = body mass index; CI = confidence interval; CVOT = cardiovascular outcomes trial; DPP4i = dipeptidyl-peptidase 4 inhibitor; DKD = diabetic kidney disease; eGFR = estimated glomerular filtration rate; ESRD = end-stage renal disease; HbA1c = glycated hemoglobin; HR = hazard ratio; SGLT2i = sodium-glucose cotransporter 2 inhibitor; T2D = type 2 diabetes; UACR = urine albumin to creatinine ratio.

Early cardiovascular risk and all-cause mortality following an incident of severe hypoglycaemia: A population-based cohort study.

AIMS: Severe hypoglycaemia is associated with a high risk of cardiovascular events in patient with diabetes. The aim of this study was to clarify the temporal relationship between hypoglycaemia and cardiovascular events. MATERIALS AND METHODS: This observational cohort study was conducted using Taiwan's Longitudinal Cohort of Diabetes Patients Database, which included 360 000 patients with newly diagnosed diabetes during the period 1999 to 2001. Patients with the first severe hypoglycaemia after 2002 served as the study cohort. Each patient in the study cohort was matched with two control patients without severe hypoglycaemia, based on a propensity score. A joinpoint regression model was used to determine trends in all-cause mortality and incidence of cardiovascular disease (CVD) events in both cohorts. RESULTS: A total of 10 157 patients with severe hypoglycaemia and 20 314 matched controls were recruited. Patients with severe hypoglycaemia had a significantly higher risk of CVD (HR, 7.28; 95% CI, 5.19-10.20) and all-cause mortality (HR, 19.92; 95% CI, 13.42-29.56) during the first month compared with those without. In patients with severe hypoglycaemia, the incidence of CVDs dropped by 17.29% monthly during the first 4 months and slowly decreased (-0.67%) during subsequent months. All-cause mortality decreased by 16.55% and 3.24% monthly during months 0-6 and months 6-17, respectively. CONCLUSIONS: Severe hypoglycaemia is associated with a greater risk of cardiovascular events and death, especially during the first month following a hypoglycaemic episode. Patients prone to severe hypoglycaemia should be made aware of the elevated risk of subsequent cardiovascular events.

THE TRENDS OF HYPERTHYROIDISM TREATMENT IN TAIWAN: A NATIONWIDE POPULATION-BASED STUDY.

OBJECTIVE: Few studies on hyperthyroidism treatment have been reported in the past 3 decades. We used a nationwide population-based database to evaluate the current practices and management of hyperthyroidism in Taiwan. METHODS: This retrospective study included a random selection of 1 million people in Taiwan between 2004 and 2010. We identified patients with hyperthyroidism who received antithyroid drugs (ATD), radioactive iodine (RAI), or surgery. We calculated the proportions and treatment trends of those 3 treatment options annually. A Poisson regression model was used to determine whether trends changed. RESULTS: The prevalence of overt hyperthyroidism in Taiwan steadily increased from 2,666 (0.27%) in 2004 to 3,464 (0.37%) in 2010. The incidence of hyperthyroidism (per 1,000 persons) also increased from 0.97 in 2004 to 1.06 in 2010. The major proportion of hyperthyroidism in this study was Graves disease (95%), followed by toxic nodular goiter (2%), and other causes (3%). ATD is the most commonly used (96-97%) treatment for hyperthyroidism, followed by surgery (2-3%) and RAI (<1%). There was a significant decreasing trend for surgery, from 2.9% in 2004 to 2% in 2010, especially in female patients (3.3% in 2004 to 2.3% in 2010, P<.01) and patients younger than 40 (3.8% in 2004 to 2.9% in 2010, P<.01). Meanwhile, the proportions of ATD and RAI remained unchanged. The most common ATD prescription was methimazole (45-50%), followed by propylthiouracil (30-32%) and carbimazole (19-21%). CONCLUSION: Between 2004 and 2010, ATD was the treatment of choice in Taiwan, followed by surgery and RAI. ABBREVIATIONS: ATA = American Thyroid Association; ATD = antithyroid drug; LHID2005 = Longitudinal Health Insurance database 2005 dataset; NHI = National Health Insurance; RAI = radioactive iodine.

Progress of diabetes severity associated with severe hypoglycemia in Taiwan.

OBJECTIVES: The association between the progression of diabetes severity and risk of severe hypoglycemia is unknown. This study aimed to evaluate the association between the progression of diabetes severity and severe hypoglycemia in patients with diabetes. STUDY DESIGN: A 13-year population-based retrospective cohort study of patients with diabetes in Taiwan. METHODS: Diabetes progression was evaluated by the adapted Diabetes Complications Severity Index (aDCSI) score from index date to end of follow-up. The progression of diabetes severity was divided into 3 categories: slow, moderate, and rapid increase in aDSCI score. We further compared those 3 categories and evaluated the risk of first hospitalization due to severe hypoglycemia. RESULTS: A total of 330,831 patients with diabetes were recruited. The mean age of patients in this study was 56.8 years, and mean follow-up duration was 9.3 years. The mean initial aDCSI score was 0.7, whereas the mean aDCSI score at the event date or end date was 2.9. A rapid increase in aDCSI score was associated with higher risk of severe hypoglycemia compared with a slow increase (hazard ratio, 4.91; 95% CI, 4.65-5.18). The incidence densities of severe hypoglycemia (per 1000 person-years) for slow, moderate, and rapid increase in aDCSI score were 2.3, 2.5, and 11.4, respectively. CONCLUSIONS: This study demonstrated that rapid progression of diabetes complications was associated with higher risk of severe hypoglycemia. It is imperative that treating physicians identify patients with acute worsening of diabetes severity and provide proper hypoglycemia education and prevention care.

Low-Dose Aspirin Reduces Breast Cancer Risk in Women with Diabetes: A Nationwide Retrospective Cohort Study in Taiwan.

PURPOSE: Low-dose aspirin is commonly used for preventing cardiovascular disease in people with diabetes, but its association with cancer remains controversial. This study used a nationwide population-based reimbursement database to investigate the relationship between low-dose aspirin use and breast cancer incidence in women with diabetes. METHODS: This retrospective cohort study was conducted using data retrieved from the National Health Insurance Research Database in Taiwan from January 1, 1998 to December 31, 2011. Women diagnosed as having diabetes with low-dose aspirin use (75-165 mg daily) were identified as the study population, whereas those without low-dose aspirin use were selected as the comparison group. RESULTS: We analyzed 148,739 patients with diabetes. Their mean age (standard deviation) was 63.3 (12.8) years. A total of 27,378 patients were taking aspirin. Overall, the use of aspirin in patients with diabetes reduced the risk of breast cancer by 18% (hazard ratio [HR], 0.82; 95% confidence interval [CI], 0.71-0.94) after adjustment for potential confounders, namely age and comorbidities. Specifically, a cumulative dose of aspirin exceeding 88,900 mg was observed to reduce the risk of breast cancer by 47% (HR, 0.53, 95% CI, 0.43-0.67); however, low (<8,600 mg) and medium (8,600-88,900 mg) cumulative doses of aspirin did not reduce the risk of breast cancer. CONCLUSIONS: Our findings suggest that a cumulative aspirin dosage of more than 88,900 mg daily was associated with a reduced risk of breast cancer in women with diabetes. However, additional studies are necessary to confirm these findings.

Iodinated Contrast Media-Induced Thyroid Dysfunction in Euthyroid Nodular Goiter Patients.

BACKGROUND: The risks of thyroid dysfunction after iodinated contrast media exposure in patients with euthyroid nodular goiter are largely unknown. METHODS: This observational, retrospective cohort study included a random selection of one million people in Taiwan. All patients with iodinated contrast media exposure during this study period were selected. Patients with euthyroid nodular goiter were identified as cases, while patients without thyroid nodule were selected as controls. We followed these patients until the first event of thyroid dysfunction including hyperthyroidism or hypothyroidism after iodinated contrast media exposure. RESULTS: A total of 334 cases and 2672 matched controls were selected in this study. The mean age of cases and controls were 58.6 and 58.4 years old, and mean follow-up durations were 2.1 and 2 years respectively. After adjustment, patients with euthyroid nodular goiter had a higher risk of thyroid dysfunction (hazard ratio 5.43, [confidence interval (CI) 3.01-9.80]) compared with controls after iodinated contrast media exposure. In the subgroup analysis, the risks of hyperthyroidism and hypothyroidism in cases compared with controls were 5.77 [CI 2.64-12.62] and 4.95 [CI 2.15-11.40] respectively. Half of the euthyroid nodular goiter cases developed thyroid dysfunction within one year after iodinated contrast media exposure. Interestingly, all thyroid-related comorbidities and drug prescriptions did not increase the risk of thyroid dysfunction. CONCLUSIONS: Presence of euthyroid nodular goiter was associated with higher risk of thyroid dysfunction including hyperthyroidism and hypothyroidism after iodinated contrast media exposure.