American College Of Radiology Guidelines Research Articles

Enhancing chatbot performance for imaging recommendations: Leveraging GPT-4 and context-awareness for trustworthy clinical guidance

PurposeTo investigate if GPT-4 improves the accuracy, consistency, and trustworthiness of a context-aware chatbot to provide personalized imaging recommendations from American College of Radiology (ACR) appropriateness criteria documents using semantic similarity processing: In addition, we sought to enable auditability of the output by revealing the information source the decision relies on. Material and MethodsWe refined an existing chatbot that incorporated specialized knowledge of the ACR guidelines by upgrading GPT-3.5-Turbo to its successor GPT-4 by OpenAI, using the latest version of LlamaIndex, and improving the prompting strategy. This chatbot was compared to the previous version, generic GPT-3.5-Turbo and GPT-4, and general radiologists regarding the performance in applying the ACR appropriateness guidelines. ResultsThe refined context-aware chatbot performed superior to the previous version using GPT-3.5-Turbo, generic chatbots GPT-3.5-Turbo and GPT-4, and general radiologists in providing “usually or may be appropriate” recommendations according to the ACR guidelines (all p < 0.001). It also outperformed GPT-3.5-Turbo and general radiologists in respect to “usually appropriate” recommendations (both p < 0.001). Moreover, the consistency in correct answers was higher with 78 % consistent correct “usually appropriate” answers and 94 % for “usually or may be appropriate” recommendations. In all cases, the same source documents were chosen, ensuring transparency. ConclusionOur study demonstrates the significance of context awareness in ensuring the use of appropriate knowledge and proposes a strategy to enhance trust in chatbot-based outputs to provide transparency. The improvements in accuracy, consistency, and source transparency address trust issues and enhance the clinical decision support process.Abbreviations: ACR, American College of Radiology; accGPT, appropriateness criteria context aware GPT; accGPT-4, appropriateness criteria context aware GPT using GPT-4; GPT, generative pre-trained transformer; LLM, Large Language Model.

Strategies for Identifying and Recruiting Women at High Risk for Breast Cancer for Research Outside of Clinical Settings: Observational Study.

Research is needed to understand and address barriers to risk management for women at high (≥20% lifetime) risk for breast cancer, but recruiting this population for research studies is challenging. This paper compares a variety of recruitment strategies used for a cross-sectional, observational study of high-risk women. Eligible participants were assigned female at birth, aged 25-85 years, English-speaking, living in the United States, and at high risk for breast cancer as defined by the American College of Radiology. Individuals were excluded if they had a personal history of breast cancer, prior bilateral mastectomy, medical contraindications for magnetic resonance imaging, or were not up-to-date on screening mammography per American College of Radiology guidelines. Participants were recruited from August 2020 to January 2021 using the following mechanisms: targeted Facebook advertisements, Twitter posts, ResearchMatch (a web-based research recruitment database), community partner promotions, paper flyers, and community outreach events. Interested individuals were directed to a secure website with eligibility screening questions. Participants self-reported method of recruitment during the eligibility screening. For each recruitment strategy, we calculated the rate of eligible respondents and completed surveys, costs per eligible participant, and participant demographics. We received 1566 unique responses to the eligibility screener. Participants most often reported recruitment via Facebook advertisements (724/1566, 46%) and ResearchMatch (646/1566, 41%). Community partner promotions resulted in the highest proportion of eligible respondents (24/46, 52%), while ResearchMatch had the lowest proportion of eligible respondents (73/646, 11%). Word of mouth was the most cost-effective recruitment strategy (US $4.66 per completed survey response) and paper flyers were the least cost-effective (US $1448.13 per completed survey response). The demographic characteristics of eligible respondents varied by recruitment strategy: Twitter posts and community outreach events resulted in the highest proportion of Hispanic or Latina women (1/4, 25% and 2/6, 33%, respectively), and community partner promotions resulted in the highest proportion of non-Hispanic Black women (4/24, 17%). Although recruitment strategies varied in their yield of study participants, results overall support the feasibility of identifying and recruiting women at high risk for breast cancer outside of clinical settings. Researchers must balance the associated costs and participant yield of various recruitment strategies in planning future studies focused on high-risk women.

Improving LDCT screening rates within a large multidisciplinary community based practice using integrated ACR accredited imaging centers.

e13582 Background: Lung cancer remains the leading cause of cancer-related death in the US, accounting for about 22% of such deaths. Lung cancer detected by low-dose computed tomography (LDCT) screening is discovered at an early stage 64-85% of the time. The benefits of LDCT screening have been demonstrated in numerous trials, but there has been difficulty in establishing programs that screen a high percentage of eligible patients. We present data from the first 18 months of our lung cancer screening program that is designed to provide the highest quality screening, follow-up and access to prevention, diagnosis, and treatment including clinical trials within their local communities. Methods: New York Cancer and Blood Specialists (NYCBS) is a community practice with over 50 locations across NY. While most lung cancer screening programs are hospital-based, NYCBS has a series of American College of Radiology (ACR) accredited imaging centers that perform these screenings and are closely affiliated with oncology, primary care, Ob-Gyn, nephrology, urology and other specialty medical and surgical practices. These centers are strategically located in areas that have high rates of tobacco use as suggested by heat map data from the NY City Community Health Survey . Upon initial patient presentation, all NYCBS and affiliated sites screen patients with tablet-based questionnaires that prompt additional questions during follow-up appointments. Patients with positive findings are discussed during a monthly, system-wide multidisciplinary lung nodule review conference. Follow-up interventions include referrals to pulmonology, interventional radiology, or further imaging with PET/CT. Per ACR guidelines, patients not requiring acute intervention continue with follow-up CT. Results: In 2023, 29,501 patients were screened with an in-office questionnaire. Of those patients, 3,606 were eligible for screening, 225 declined, and 896 (29%) high risk patients proceeded to undergo LDCT screening. In screened patients, 128 underwent follow-up CT, 36 patients had findings requiring immediate intervention, and 6 were diagnosed with lung cancer. 5 of these patients were diagnosed with stage I and 1 patient was diagnosed with stage IV lung cancer. Conclusions: LDCT screening is a proven method to discover early stage lung cancer and improve the survival rate of diagnosed patients but has not been well adopted in practice. With access to a diverse array of multidisciplinary practices across NY and integrated accredited imaging centers, we have increased the accessibility of lung cancer screening and improved transitions to follow-up care. Nationally, only 4.5% of patients at high risk are screened, however, we found that with our approach to patient care, we have achieved among the highest reported screening rates in eligible individuals.

Safety of Gadoxetate Disodium for Hepatobiliary MRI in Children and Adolescents.

Background Despite a proven role in the characterization of liver lesions, use of the gadolinium-based contrast agent (GBCA) gadoxetate disodium at MRI is limited in children due to a lack of comparative safety data. Purpose To evaluate the safety of the GBCA gadoxetate disodium (a linear ionic hepatobiliary contrast agent [HBA]) in children and adolescents, compared with extracellular contrast agents (ECA). Materials and Methods A retrospective analysis was conducted in children and adolescents aged 18 years or younger who underwent HBA-enhanced MRI at one of three tertiary hospitals from January 2010 to December 2022. The incidence of GBCA-associated acute adverse events was compared between MRI examinations with a HBA and those with ECA. Severity was categorized according to American College of Radiology guidelines (mild, moderate, or severe). (a) Propensity score matching using multivariable logistic regression models and (b) inverse probability of treatment weighting analysis based on nine covariates (age, sex, asthma, allergic rhinitis, chronic urticaria or atopy, food allergy, drug allergy, premedication, and history of GBCA-associated adverse events) were used for confounder adjustment. Results A total of 1629 MRI examinations (ECA, n = 1256; HBA, n = 373) in 1079 patients were included (mean age, 8.6 years ± 6.5; 566 girls). The per-examination incidence of GBCA-associated acute adverse events showed no evidence of a difference, with rates of 0.9% (11 of 1256 examinations) for ECA and 1.3% (five of 373 examinations) for HBA (odds ratio [OR], 1.55 [95% CI: 0.54, 4.46]; P = .42). Acute adverse events were all mild with ECA, whereas with HBA, they were mild for four patients and moderate for one patient. There was no evidence of a difference in the incidence of acute adverse events, even in propensity score matching (OR, 1.33 [95% CI: 0.30, 5.96]; P = .71) and inverse probability of treatment weighting analysis (OR, 0.84 [95% CI: 0.25, 2.86]; P = .78). Conclusion Gadoxetate disodium showed no difference in acute adverse events compared with ECA in children and adolescents, with further large-scale pediatric studies required to confirm its safety. © RSNA, 2024 Supplemental material is available for this article. See also the editorial by Otero in this issue.

Inappropriate wrist MRI: did guidelines have an impact?

To assess the national prevalence and cost of inappropriate MRI in patients with wrist pain prior to and following American College of Radiology (ACR) guidelinepublication. We used administrative claims from the IBM MarketScan Research Databases to evaluate the appropriateness of wrist MRI in a national cohort of patients with commercial insurance or Medicare Advantage. Adult patients with a diagnosis of wrist pain between 2016 and 2019 were included and followed for 1year. We made assessments of appropriateness based on ACR guidelines for specific wrist pain etiologies. We tabulated the total costs and out-of-pocket expenses associated with inappropriate MRI studies using weighted mean payments for facility and professional fees. We performed segmented logistic regression on interrupted time series data to identify predictors of receiving inappropriate imaging and the impact of guideline publication on MRI use. The study cohort consisted of 867,119 individuals. Of these, 40,164 individuals (4.6%) had MRI, of whom 52.6% received an inappropriate study. Inappropriate studies accounted for $44,493,234 in total payments and $8,307,540 in out-of-pocket expenses. The interrupted time series found an approximately 1% monthly decrease in the odds of receiving an inappropriate study after guidelinedissemination. MRI as a diagnostic tool for wrist pain is often inappropriate and expensive. Our findings support interventions to increase guideline adherence, such as integrated clinical decision support tools.

Generative pretrained transformer 4: an innovative approach to facilitate value-based healthcare

ObjectiveAppropriate medical imaging is important for value-based care. We aim to evaluate the performance of generative pretrained transformer 4 (GPT-4), an innovative natural language processing model, providing appropriate medical imaging automatically in different clinical scenarios. MethodsInstitutional Review Boards (IRB) approval was not required due to the use of nonidentifiable data. Instead, we used 112 questions from the American College of Radiology (ACR) Radiology-TEACHES Program as prompts, which is an open-sourced question and answer program to guide appropriate medical imaging. We included 69 free-text case vignettes and 43 simplified cases. For the performance evaluation of GPT-4 and GPT-3.5, we considered the recommendations of ACR guidelines as the gold standard, and then three radiologists analyzed the consistency of the responses from the GPT models with those of the ACR. We set a five-score criterion for the evaluation of the consistency. A paired t-test was applied to assess the statistical significance of the findings. ResultsFor the performance of the GPT models in free-text case vignettes, the accuracy of GPT-4 was 92.9%, whereas the accuracy of GPT-3.5 was just 78.3%. GPT-4 can provide more appropriate suggestions to reduce the overutilization of medical imaging than GPT-3.5 (t = 3.429, P = 0.001). For the performance of the GPT models in simplified scenarios, the accuracy of GPT-4 and GPT-3.5 was 66.5% and 60.0%, respectively. The differences were not statistically significant (t = 1.858, P = 0.070). GPT-4 was characterized by longer reaction times (27.1 s in average) and extensive responses (137.1 words on average) than GPT-3.5. ConclusionAs an advanced tool for improving value-based healthcare in clinics, GPT-4 may guide appropriate medical imaging accurately and efficiently.

A Context-based Chatbot Surpasses Trained Radiologists and Generic ChatGPT in Following the ACR Appropriateness Guidelines.

Background Radiological imaging guidelines are crucial for accurate diagnosis and optimal patient care as they result in standardized decisions and thus reduce inappropriate imaging studies. Purpose In the present study, we investigated the potential to support clinical decision-making using an interactive chatbot designed to provide personalized imaging recommendations from American College of Radiology (ACR) appropriateness criteria documents using semantic similarity processing. Methods We utilized 209 ACR appropriateness criteria documents as specialized knowledge base and employed LlamaIndex, a framework that allows to connect large language models with external data, and the ChatGPT 3.5-Turbo to create an appropriateness criteria contexted chatbot (accGPT). Fifty clinical case files were used to compare the accGPT's performance against general radiologists at varying experience levels and to generic ChatGPT 3.5 and 4.0. Results All chatbots reached at least human performance level. For the 50 case files, the accGPT performed best in providing correct recommendations that were "usually appropriate" according to the ACR criteria and also did provide the highest proportion of consistently correct answers in comparison with generic chatbots and radiologists. Further, the chatbots provided substantial time and cost savings, with an average decision time of 5 minutes and a cost of 0.19 € for all cases, compared to 50 minutes and 29.99 € for radiologists (both p < 0.01). Conclusion ChatGPT-based algorithms have the potential to substantially improve the decision-making for clinical imaging studies in accordance with ACR guidelines. Specifically, a context-based algorithm performed superior to its generic counterpart, demonstrating the value of tailoring AI solutions to specific healthcare applications.

Appropriateness of lumbar spine radiography and factors influencing imaging ordering patterns: paving the path toward value-driven health care.

The demand for diagnostic imaging continues to rise. Against the backdrop of rising health care costs and finite resources, this has prompted a paradigm shift toward value-driven patient care. Inappropriate imaging is a barrier toward achieving this goal, which runs counter to prevailing evidence-based guidelines and contributes to rising healthcare costs. Our objective was to evaluate the appropriateness of lumbar spine X-rays in a tertiary referral Emergency Department (ED) and assess whether physicians' specialization and years of experience influence appropriateness. A total of 1030 lumbar spine radiographs performed in the ED of an academic medical center over a consecutive 3-month period were reviewed retrospectively. Referral indications were reviewed for adherence to 2021 American College of Radiology appropriateness guidelines for lower back pain, and referral patterns were evaluated among physician groups based on specialists' training and years in practice. 63.8% of lumbar spine radiographs were appropriate, with trauma being the most common indication. 36.2% of orders were inappropriate, with lower back pain of <6 weeks duration being the most common indication. Significant differences in inappropriate orders were found (P < .001) across physician groups: qualified Emergency Medicine specialists (20.9% inappropriate orders), specialists in training (27.8%), and non-specialists with ≥3 (60.0%) and <3 (36.9%) years in practice, respectively. Approximately one-third of lumbar spine radiographs performed in the ED were inappropriately ordered by American College of Radiology guidelines; specialists training and years in practice affected referral patterns. Integrating evidence-based appropriateness guidelines into the physician order workflow and targeting older non-specialists may promote more judicious imaging and reduce health care costs.

The diagnostic efficacy and inappropriate biopsy rate of ACR TI-RADS and ATA guidelines for thyroid nodules in children and adolescents.

This study is aimed at evaluating the diagnostic efficacy and unnecessary fine-needle aspiration (FNA) rate of ultrasound-based risk stratification for thyroid nodules in the American College of Radiology (ACR) Thyroid Imaging Reporting and Data System (TI-RADS) and the American Thyroid Association (ATA) risk stratification systems. Children and adolescents with pathology confirmed thyroid nodules were retrospectively included in this study. A total of 217 thyroid nodules from multicenter of Union Medical College Hospital, China Japan Friendship Hospital and Civil Aviation Hospital were included, the diagnostic efficiency and unnecessary FNA rate were calculated according to ACR and ATA guidelines. Among all thyroid nodules, 139 nodules were malignant, and 78 nodules were benign. Choosing ATA high suspicion and ACR TI-RADS TR5 as benign and malignant cut-off points, the area under the curve and sensitivity of ATA were higher than ACR (AUC: 0.887 vs 0.840, p=0.0037; sensitivity 81.3% vs 71.0%, P <0.049;specificity 96.2% vs 97.4%, p=1.000;specificity both 85.9%); choosing high/intermediate suspicion in ATA and ACR TR4/5 as benign and malignant cut-off points, the two guidelines demonstrated similar diagnostic efficacy (AUC:0.890 vs 0.897, p=0.6038, sensitivity 92.1% vs 93.5%, P =0.817;specificity both 85.9%, p=1.000). The inappropriate FNA rate of ACR guideline was relatively lower (ATA 42.9% vs ACR 27.2%, P <0.001). If ACR TI-RADS TR5 nodules less than 1.0cm were included in the FNA indication, the unnecessary biopsy rate would be further reduced to 17.9%. This study indicated that both ATA and ACR TI-RADS risk stratification systems could provide a feasible differential diagnosis of benign and malignant thyroid nodules, while the ACR risk stratification system demonstrates a lower rate of inappropriate FNA rate. In addition, it was necessary to further study the minimum FNA threshold of thyroid nodules in Children and adolescents in order to reduce the missed biopsy rate of malignant nodules.

Multicentre Analysis of Cost, Uptake and Safety of Canadian Multidisciplinary Pancreatic Cyst Guidelines.

Pancreatic cystic lesions (PCLs) are common, with several guidelines providing surveillance recommendations. The Canadian Association of Radiologists published surveillance guidelines (CARGs) intended to provide simplified, cost-effective and safe recommendations. This study aimed to evaluate cost savings of CARGs compared to other North American guidelines including American Gastroenterology Association guidelines (AGAG) and American College of Radiology guidelines (ACRG), and to evaluate CARG safety and uptake. This is a multicentre retrospective study evaluating adults with PCL from a single health zone. MRIs completed from September 2018-2019, one year after local CARG guideline implementation, were reviewed to identify PCLs. All imaging following 3-4 years of CARG implementation was reviewed to evaluate true costs, missed malignancy and guideline uptake. Modelling, including MRI and consultation, predicted and compared costs associated with surveillance based on CARGs, AGAGs and ACRGs. 6698 abdominal MRIs were reviewed with 1001 (14.9%) identifying PCL. Application of CARGs over 3.1 years demonstrated a >70% cost reduction compared to other guidelines. Similarly, the modelled cost of surveillance for 10-years for each guideline was $516,183, $1,908,425 and $1,924,607 for CARGs, AGAGs and ACRGs respectively. Of patients suggested to not require further surveillance per CARGs, approximately 1% develop malignancy with fewer being candidates for surgical resection. Overall, 44.8% of initial PCL reports provided CARG recommendations while 54.3% of PCLs were followed as per CARGs. CARGs are safe and offer substantial cost and opportunity savings for PCL surveillance. These findings support Canada-wide implementation with close monitoring of consultation requirements and missed diagnoses.

Comparison of Various Ultrasound-Based Malignant Risk Stratification Systems on an Occasion for Assessing Thyroid Nodules in Hashimoto's Thyroiditis.

To compare the diagnostic performance and unnecessary ultrasound-guided fine-needle aspiration (US-FNA) biopsy rate of the 2015 American Thyroid Association (ATA), 2016 Korean Society of Thyroid Radiology (KSThR), and 2017 American College of Radiology (ACR) guidelines for patients with and without Hashimoto's thyroiditis (HT). This retrospective study included 716 nodules from 696 consecutive patients, which were classified using the categories defined by the three guidelines: ATA, KSThR, and ACR. The malignancy risk in each category was calculated and the diagnostic performance and unnecessary fine-needle aspiration (FNA) rates of the three guidelines were compared. In total, 426 malignant and 290 benign nodules were identified. Patients with malignant nodules had lower total thyroxine levels and higher thyroid-stimulating hormone, thyroid peroxidase antibody, and thyroglobulin antibody levels than those without malignant nodules (all P<0.01). The margin difference was significant in non-HT patients (P<0.01), but comparable in HT patients (P=0.55). The calculated malignancy risks of high and intermediate suspicion nodules in the ATA and KSThR guidelines and moderately suspicious nodules in the ACR guidelines were significantly lower in non-HT patients compared with HT patients (P<0.05). The ACR guidelines showed the lowest sensitivity, highest specificity, and lowest unnecessary FNA rates in patients with and without HT. Compared to non-HT patients, HT patients had significantly lower unnecessary FNA rates (P<0.01). HT was associated with a higher malignancy rate of thyroid nodules with intermediate suspicion according to the ATA, KSThR, and ACR guidelines. The three guidelines, especially ACR, were likely to be more effective and could allow a greater reduction in the percentage of benign nodules biopsied in patients with HT.

Bleeding Complications After Breast Core-needle Biopsy-An Approach to Managing Patients on Antithrombotic Therapy.

Image-guided core-needle breast and axillary biopsy (CNB) is the standard-of-care procedure for the diagnosis of breast cancer. Although the risks of CNB are low, the most common complications include bleeding and hematoma formation. Post-procedural bleeding is of particular concern in patients taking antithrombotic therapy, but there is currently no widely established standard protocol in the United States to guide antithrombotic therapy management. In the face of an increasing number of patients taking antithrombotic therapy and with the advent of novel classes of anticoagulants, the American College of Radiology guidelines recommend that radiologists consider cessation of antithrombotic therapy prior to CNB on a case-by-case basis. Lack of consensus results in disparate approaches to patients on antithrombotic therapy undergoing CNB. There is further heterogeneity in recommendations for cessation of antithrombotic therapy based on the modality used for image-guided biopsy, target location, number of simultaneous biopsies, and type of antithrombotic agent. A review of the available data demonstrates the safety of continuing antithrombotic therapy during CNB while highlighting additional procedural and target lesion factors that may increase the risk of bleeding. Risk stratification of patients undergoing breast interventional procedures is proposed to guide both pre-procedural decision-making and post-procedural management. Radiologists should be aware of antithrombotic agent pharmacokinetics and strategies to minimize post-procedural bleeding to safely manage patients.

Benign-appearing Incidental Adnexal Cysts at US, CT, and MRI: Putting the ACR, O-RADS, and SRU Guidelines All Together

Adnexal cysts are a common incidental finding at US, CT, and MRI but have historically caused a diagnostic dilemma for determining when to follow up and how to manage them. Characteristic imaging features of simple adnexal cysts include a simple fluid collection with smooth walls and no solid or vascular components. Day-to-day practice guidelines were recently updated to reflect the overwhelming evidence that incidental cystic adnexal masses are almost always benign. Three major consensus articles on adnexal cystic masses were published between 2019 and 2020: the Society of Radiologists in Ultrasound (SRU) consensus update on adnexal cysts, the Ovarian-Adnexal Reporting and Data System (O-RADS) US consensus guideline, and the American College of Radiology (ACR) white paper on the management for incidental adnexal findings at CT and MRI. All three standardize reporting terminology, are based on evidence-based data and institutional practice patterns, and apply to nonpregnant women of average risk for ovarian cancer. While there are small differences in follow-up recommendations based on size thresholds, the goal of each is the same-to limit unnecessary imaging follow-up and, by doing so, save the patient time, money, and anxiety. For the diagnostic radiologist to use these guidelines, it is essential that the entire mass is visualized well. Without adequate visualization, further characterization of the mass may be necessary. To put it all together, the SRU consensus guideline and ACR white paper are easily applied in day-to-day practice for masses that are O-RADS 2 and below. An invited commentary by Patel is available online.The online slide presentation from the RSNA Annual Meeting is available for this article.©RSNA, 2022.

Diagnostic performance of US-based FNAB criteria of the 2020 Chinese guideline for malignant thyroid nodules: comparison with the 2017 American College of Radiology guideline, the 2015 American Thyroid Association guideline, and the 2016 Korean Thyroid Association guideline.

To investigate the diagnostic performance of the ultrasonography-based fine-needle aspiration biopsy criteria of the Chinese Thyroid Imaging Reporting and Data System (C-TIRADS) for malignant nodules compared to 3 other guidelines. This study included 2,309 thyroid nodules in 1,697 patients with histopathological and cytopathological diagnoses of benign and malignant nodules from January 2018 to August 2020. The clinical and ultrasonographic features of the nodules were retrospectively reviewed and classified according to the Chinese guideline (C-TIRADS), the American College of Radiology guideline (ACR-TIRADS), the American Thyroid Association guideline (ATA guideline), and the Korean Thyroid Association guideline (K-TIRADS). The diagnostic performance of the guidelines and their unnecessary fine-needle aspiration biopsy rates were calculated using randomized, blinded trials. Of the 2,309 nodules, 1,418 (61.4%) were benign and 891 (38.6%) were malignant, with 884 (99.21%) papillary carcinomas. The accuracy of C-TIRADS was 84.71%, followed by the guidelines of ACR-TIRADS (82.11%), K-TIRADS (81.64%), and the ATA guideline (78.56%). Furthermore, the area under the receiver operating characteristic curve (AUC) was the highest for the C-TIRADS (0.905). Similar results were revealed for both the diagnostic performance and AUC of nodules smaller and larger than 10 mm. The ACR-TIRADS showed the lowest unnecessary biopsy rate (17.54%), followed by the C-TIRADS (22.61%), ATA guideline (27.90%), and the K-TIRADS (28.67%). The C-TIRADS demonstrated high diagnostic performance and a relatively low unnecessary biopsy rate in detecting thyroid cancer compared to the 3 other guidelines. However, further understanding of the ultrasonography-based fine-needle aspiration biopsy criteria of the C-TIRADS should be gained in the future.

Incidental Thyroid Nodules Found During Acute Stroke Angiography: Prevalence, Outcomes, and Suggested Management Guidelines.

In stroke patients, extensive interventions for incidental thyroid nodules can be burdensome and costly. It appears that the risk of malignancy has not been reported in angiographically detected nodules and outcomes have not yet been described in patients, receiving acute stroke work-up. Incidental thyroid nodules were found on neck computerized tomography angiography or magnetic resonance angiography performed during inpatient stroke workup (January 2017 to September 2019). These patient cases were reviewed based on sonography reports, diagnosis, and follow-up care. Of the 13 563 patients, 192 had incidental thyroid nodules (prevalence 1.4%). Twenty-six died from comorbidities and 22 received thyroid sonography. Twelve nodules from 10 patients had sonographic characteristics for biopsy: 10 benign, 1 indeterminate, and 1 papillary thyroid cancer (risk of malignancy: 8%). The cancer patient underwent hemithyroidectomy and is disease-free. Follow-up of the remaining patients showed no worsening or suspicious nodules. The American College of Radiology (ACR) guidelines would have prevented 8 unnecessary sonograms and 1 biopsy without missing malignancy. Although a small risk of malignancy was noted, 95% of patients undergoing additional diagnostic thyroid testing had clinically insignificant results. The ACR guidelines can prevent unnecessary interventions. Given the 14% mortality rate in the study cohort, it is proposed that a clinical evaluation of patients is important before undergoing further diagnostics, as comorbidities may be worse than a thyroid cancer.

Quality improvement time-saving intervention to increase use of a clinical decision support tool to reduce low-value diagnostic imaging in a safety net health system

ImportanceElectronic health record (EHR) clinical decision support (CDS) tools can provide evidence-based feedback at the point of care to reduce low-value imaging. Success of these tools has been limited partly...

Appropriateness of computed tomography and magnetic resonance imaging scans in a rural regional hospital in South Africa: A 6-year follow-up study.

Requests for computed tomography (CT) and magnetic resonance imaging (MRI) scans by doctors with different levels of experience have cost and risk implications globally. Evidence-based appropriateness criteria guide doctors to the suitable use of radiology imaging. There are few studies regarding appropriateness of CT requests in the South African (SA) context. Previous research in the Garden Route district of Western Cape Province, SA, evaluated the appropriateness of scans. To review the appropriateness of CT and MRI scans done in a 6-year follow-up study at one facility. This was a retrospective descriptive study. All CT and MRI scans performed during October 2018 at George Hospital were classified according to American College of Radiology guidelines as: usually appropriate (UA), might be appropriate (MBA), or not appropriate (NA). Stratified analysis allowed simple statistics and some comparison with the previous study. A total of 515 CT and MRI scans were included, of which 81.4% were UA, 7.4% MBA and 11.2% NA. Most scans were requested by medical officers (n=255), followed by consultants (n=126) and junior doctors (n=70). Medical officers made the majority of inappropriate requests. Second-year interns requested the lowest number of inappropriate scans, with registrars not requesting any inappropriate scans. Most of the inappropriate scans were requested after hours. Thirty-seven of the 123 (30.1%) after-hours scans were inappropriately requested compared with 21 of 392 (5.4%) scans during normal working hours, which were inappropriate. Although the majority of scans were being ordered appropriately, pre-authorisation by experienced physicians and incorporation of guidelines would make requests more complete and possibly more appropriate, especially after hours.

Combining radiomics with ultrasound-based risk stratification systems for thyroid nodules: an approach for improving performance

To develop a radiomics score using ultrasound images to predict thyroid malignancy and to investigate its potential as a complementary tool to improve the performance of risk stratification systems. We retrospectively included consecutive patients who underwent fine-needle aspiration (FNA) for thyroid nodules that were cytopathologically diagnosed as benign or malignant. Nodules were randomly assigned to a training and test set (8:2 ratio). A radiomics score was developed from the training set, and cutoff values based on the maximum Youden index (Rad_maxY) and for 5%, 10%, and 20% predicted malignancy risk (Rad_5%, Rad_10%, Rad_20%, respectively) were applied to the test set. The performances of the American College of Radiology (ACR) and the American Thyroid Association (ATA) guidelines were compared with the combined performances of the guidelines and radiomics score with interpretations from expert and nonexpert readers. A total of 1624 thyroid nodules from 1609 patients (mean age, 50.1 years [range, 18-90 years]) were included. The radiomics score yielded an AUC of 0.85 (95% CI: 0.83, 0.87) in the training set and 0.75 (95% CI: 0.69, 0.81) in the test set (Rad_maxY). When the radiomics score was combined with the ACR or ATA guidelines (Rad_5%), all readers showed increased specificity, accuracy, and PPV and decreased unnecessary FNA rates (all p < .05), with no difference in sensitivity (p > .05). Radiomics help predict thyroid malignancy and improve specificity, accuracy, PPV, and unnecessary FNA rate while maintaining the sensitivity of the ACR and ATA guidelines for both expert and nonexpert readers. • The radiomics score yielded an AUC of 0.85 and 0.75 in the training and test set, respectively. • For all readers, combining a 5% predicted malignancy risk cutoff for the radiomics score with the ACR and ATA guidelines significantly increased specificity, accuracy, and PPV and decreased unnecessary FNA rates, with no decrease in sensitivity. • Radiomics can help predict malignancy in thyroid nodules in combination with risk stratification systems, by improving specificity, accuracy, and PPV and unnecessary FNA rates while maintaining sensitivity for both expert and nonexpert readers.

Automatic Analysis of ACR Phantom Images in MRI.

Quality Assurance (QA) of Magnetic Resonance Imaging (MRI) system is an essential step to avoid problems in diagnosis when image quality is low. It is considered a patient safety issue. The accreditation program of the American College of Radiology (ACR) includes a standardized image quality measurement protocol. However, it has been shown that human testing by visual inspection is not objective and not reproducible. The overall goal of the present paper was to develop and implement a fully automated method for accurate image analysis to increase its objectivity. It can positively impact the QA process by decreasing the reaction time, improving repeatability, and by reducing operator dependency. The proposed QA procedures were applied to ten clinical MRI scanners. The performance of the automated procedure was assessed by comparing the test results with the decisions made by trained MRI technologists according to ACR guidelines. The p-value, correlation coefficient of the manual and automatic measurements were also computed using the Pearson test. Compared to the manual process, the use of the proposed approach can significantly reduce the time requirements while maintaining consistency with manual measurements and furthermore, decrease the subjectivity of the results. Accordingly, a strong correlation was found and the corresponding p-value was much lower than the significance level of 0.05 indicating a good agreement between the two measurements.

Discordance Between the American Thyroid Association and the American College of Radiology Guideline Systems for Thyroid Nodule Biopsy

BackgroundPreviously, fine-needle aspiration biopsy was recommended for any thyroid nodule >1.0 cm in size. In 2015, the American Thyroid Association (ATA) introduced a pattern-based approach for biopsy recommendations based on size and ultrasound (US) characteristics. In 2016, the American College of Radiology (ACR) published the Thyroid Imaging Reporting and Data System, using a point-based system that assesses risk of US characteristics. MethodsThis study aims to compare recommendations for thyroid nodule biopsy between the ATA and ACR systems and identify outcomes of nodules with discordant recommendations (DRs). US characteristics, fine-needle aspiration biopsy, and surgical pathology results were evaluated for all patients with >1.0 cm thyroid nodules treated at a single tertiary-care institution from 2010 to 2018. ResultsInclusion criteria were met by 1100 nodules from 687 patients; 42.8% (n = 471) had DR between the ATA and ACR guidelines. All (100%) DR nodules were not recommended for biopsy by ACR, though 53% were recommended to have follow-up. A majority (79%) of DR nodules were recommended for biopsy by ATA, with the remaining 21% recommended for follow-up. Among surgically excised DR nodules (n = 292), 10.3% (n = 30) nodules were found to be malignant, with the vast majority (90.3%) being well-differentiated carcinoma. Among malignant nodules, the ACR would not have recommended biopsy or follow-up for 26.7% (n = 8). ConclusionsThe ACR classification system is more restrictive compared with the ATA system for recommending thyroid nodule biopsy. This discrepancy could result in confusion for clinicians and delay in diagnosis or therapy for patients with thyroid cancer.