Search

Artificial intelligence (AI) systems play an increasingly important role in all parts of the imaging chain, from image creation to image interpretation to report generation. In order to responsibly manage radiology AI systems and make informed purchase decisions about them, radiologists must understand the underlying principles of AI. Our task force was formed by the Radiology Research Alliance (RRA) of the Association of University Radiologists to identify and summarize a curated list of current educational materials available for radiologists.

The emergence of the COVID-19 pandemic over a relatively brief interval illustrates the need for rapid data-driven approaches to facilitate clinical decision making. We examined a machine learning process to predict inpatient mortality among COVID-19 patients using clinical and chest radiographic data. Modeling was performed with a de-identified dataset of encounters prior to widespread vaccine availability. Non-imaging predictors included demographics, pre-admission clinical history, and past medical history variables. Imaging features were extracted from chest radiographs by applying a deep convolutional neural network with transfer learning. A multi-layer perceptron combining 64 deep learning features from chest radiographs with 98 patient clinical features was trained to predict mortality. The Local Interpretable Model-Agnostic Explanations (LIME) method was used to explain model predictions. Non-imaging data alone predicted mortality with an ROC-AUC of 0.87 ± 0.03 (mean ± SD), while the addition of imaging data improved prediction slightly (ROC-AUC: 0.91 ± 0.02). The application of LIME to the combined imaging and clinical model found HbA1c values to contribute the most to model prediction (17.1 ± 1.7%), while imaging contributed 8.8 ± 2.8%. Age, gender, and BMI contributed 8.7%, 8.2%, and 7.1%, respectively. Our findings demonstrate a viable explainable AI approach to quantify the contributions of imaging and clinical data to COVID mortality predictions.

Radiologists very frequently encounter incidental findings related to the thyroid gland. Given increases in imaging use over the past several decades, thyroid incidentalomas are increasingly encountered in clinical practice, and it is important for radiologists to be aware of recent developments with respect to workup and diagnosis of incidental thyroid abnormalities. Recent reporting and management guidelines, such as those from the American College of Radiology and American Thyroid Association, are reviewed along with applicable evidence in the literature. Trending topics, such as artificial intelligence approaches to guide thyroid incidentaloma workup, are also discussed.

Patients undergoing MRI may experience fear, claustrophobia, or other anxiety manifestations due to the typically lengthy, spatially constrictive, and noisy MRI acquisition process and in some cases are not able to tolerate completion of the study. This article discusses several patient-centered aspects of radiology practice that emphasize interpersonal interactions. Patient education and prescan communication represent 1 way to increase patients' awareness of what to expect during MRI and therefore mitigate anticipatory anxiety. Some patient interaction strategies to promote relaxation or calming effects are also discussed. Staff teamwork and staff training in communication and interpersonal skills are also described, along with literature evidence of effectiveness with respect to patient satisfaction and productivity endpoints. Attention to how radiologists, nurses, technologists, and other members of the radiology team interact with patients before or during the MRI scan could improve patients' motivation and ability to cooperate with the MRI scanning process as well as their subjective perceptions of the quality of their care. The topics discussed in this article are relevant not only to MRI operations but also to other clinical settings in which patient anxiety or motion represent impediments to optimal workflow.

Magnetic resonance elastography (MRE) is an imaging technique that can noninvasively assess the shear properties of the intervertebral disc (IVD). Unlike the standard gradient recalled echo (GRE) MRE technique, a spin-echo echo-planar imaging (SE-EPI) sequence has the potential to improve imaging efficiency and patient compliance.To validate the use of an SE-EPI sequence for MRE of the IVD compared against the standard GRE sequence.Cross-over.Twenty-eight healthy volunteers (15 males and 13 females, age range: 19-55).3 T; GRE, SE-EPI with breath holds (SE-EPI-BH) and SE-EPI with free breathing (SE-EPI-FB) MRE sequences.MRE-derived shear stiffnesses were calculated via principal frequency analysis. SE-EPI derived shear stiffness and octahedral shear strain signal-to-noise ratios (OSS-SNR) were compared against those derived using the GRE sequence. The reproducibility and repeatability of SE-EPI stiffness measurements were determined. Shear stiffness was evaluated in the nucleus pulposus (NP) and annulus fibrosus (AF) regions of the disc. Scan times between sequences were compared.Linear mixed models, Bland-Altman plots, and Lin's concordance correlation coefficients (CCCs) were used with P 0.05 considered statistically significant.Good correlation was observed between shear stiffnesses derived from the SE-EPI sequences with those derived from the GRE sequence with CCC values greater than 0.73 and 0.78 for the NP and AF regions, respectively. OSS-SNR was not significantly different between GRE and SE-EPI sequences (P 0.05). SE-EPI sequences generated highly reproducible and repeatable stiffness measurements with CCC values greater than 0.97 in the NP and AF regions and reduced scan time by at least 51% compared to GRE. SE-EPI-BH and SE-EPI-FB stiffness measurements were similar with CCC values greater than 0.98 for both regions.SE-EPI-based MRE-derived stiffnesses were highly reproducible and repeatable and correlated with current standard GRE MRE-derived stiffness estimates while reducing scan times.1 TECHNICAL EFFICACY STAGE: 1.

Evaluate trends and demographic predictors of imaging utilization at a university-affiliated health system.In this single-institution retrospective study, per capita estimates of imaging utilization among patients active in the health system were computed by cross-referencing all clinical encounters (2004-2016) for 1,628,980 unique patients with a listing of 6,157,303 diagnostic radiology encounters. Time trends in imaging utilization and effects of gender, race/ethnicity, and age were assessed, with subgroup analyses performed by imaging modality. Utilization was analyzed as both a continuous and binary outcome variable.Over 13 years, total diagnostic exams rose 6.8% a year (285,947-622,196 exams per annum), while the active population size grew 7.0% a year (244,238-543,290 active patients per annum). Per capita utilization peaked in 2007 at 1.33 studies/patient/year before dropping to 1.06 from 2011 to 2015. Latest per capita utilization was 0.22 for computed tomography, 0.10 for MR, 0.20 for US, 0.03 for NM, 0.51 for radiography, and 0.07 for mammography. Over the study period, ultrasound utilization doubled, whereas NM and radiography utilization decreased. computed tomography, MR, and mammography showed no significant net change. Univariate analysis of utilization as a continuous variable showed statistically significant effects of gender, race/ethnicity, and age (P0.0001), with utilization higher in males and Blacks and lower in Asian/Pacific Islanders and Hispanics. Utilization increased with age, except for a decline after age 75. Many of the effects of age, gender, and race/ethnicity were also found when analyzing the binarized utilization variable.Although absolute counts of imaging studies more than doubled, the net change in per capita utilization over the study period was minimal. Variations in utilization across age, gender, and race/ethnicity may reflect differential health needs and/or access disparities, warranting future studies.

Data-driven approaches have great potential to shape future practices in radiology. The most straightforward strategy to obtain clinically accurate models is to use large, well-curated and annotated datasets. However, patient privacy constraints, tedious annotation processes, and the limited availability of radiologists pose challenges to building such datasets. This review details model training strategies in scenarios with limited data, insufficiently labeled data, and/or limited expert resources. This review discusses strategies to enlarge the data sample, decrease the time burden of manual supervised labeling, adjust the neural network architecture to improve model performance, apply semisupervised approaches, and leverage efficiencies from pretrained models. Keywords: Computer-aided Detection/Diagnosis, Transfer Learning, Limited Annotated Data, Augmentation, Synthetic Data, Semisupervised Learning, Federated Learning, Few-Shot Learning, Class Imbalance

Early detection of brain metastases (BM) is one of the determining factors for the successful treatment of patients with cancer; however, the accurate detection of small BM lesions (< 15mm) remains a challenging task. We previously described a framework for the detection of small BM in single-sequence gadolinium-enhanced T1-weighted 3D MRI datasets. It combined classical image processing (IP) with a dedicated convolutional neural network, taking approximately 30 seconds to process each dataset due to computation-intensive IP stages. To overcome the speed limitation, this study aims to reformulate the framework via an augmented pair of CNNs (eliminating the IP) to reduce the processing times while preserving the BM detection performance. Our previous implementation of the BM detection algorithm utilized Laplacian of Gaussians (LoG) for the candidate selection portion of the solution. In this study, we introduce a novel BM candidate detection CNN (cdCNN) to replace this classical IP stage. The network is formulated to have (1) a similar receptive field as the LoG method, and (2) a bias for the detection of BM lesion loci. The proposed CNN is later augmented with a classification CNN to perform the BM detection task. The cdCNN achieved 97.4% BM detection sensitivity when producing 60K candidates per 3D MRI dataset, while the LoG achieved 96.5% detection sensitivity with 73K candidates. The augmented BM detection framework generated on average 9.20 false-positive BM detections per patient for 90% sensitivity, which is comparable with our previous results. However, it processes each 3D data in 1.9 seconds, presenting a 93.5% reduction in the computation time.

Claustrophobia, other anxiety reactions, excessive motion, and other unanticipated patient events in magnetic resonance imaging (MRI) not only delay or preclude diagnostic-quality imaging but can also negatively affect the patient experience. In addition, by impeding MRI workflow, they may affect the finances of an imaging practice. This review article offers an overview of the various types of patient-related unanticipated events that occur in MRI, along with estimates of their frequency of occurrence as documented in the available literature. In addition, the financial implications of these events are discussed from a microeconomic perspective, primarily from the point of view of a radiology practice or hospital, although associated limitations and other economic viewpoints are also included. Efforts to minimize these unanticipated patient events can potentially improve not only patient satisfaction and comfort but also an imaging practice's operational efficiency and diagnostic capabilities.

Patient comfort is an important factor of a successful magnetic resonance (MR) examination, and improvements in the patient's MR scanning experience can contribute to improved image quality, diagnostic accuracy, and efficiency in the radiology department, and therefore reduced cost. Magnet designs that are more open and accessible, reduced auditory noise of MR examinations, light and flexible radiofrequency (RF) coils, and faster motion-insensitive imaging techniques can all significantly improve the patient experience in MR imaging. In this work, we review the design, development, and implementation of these physics and engineering approaches to improve patient comfort.

Artificial intelligence, particularly deep learning, offers several possibilities to improve the quality or speed of image acquisition in magnetic resonance imaging (MRI). In this article, we briefly review basic machine learning concepts and discuss commonly used neural network architectures for image-to-image translation. Recent examples in the literature describing application of machine learning techniques to clinical MR image acquisition or postprocessing are discussed. Machine learning can contribute to better image quality by improving spatial resolution, reducing image noise, and removing undesired motion or other artifacts. As patients occasionally are unable to tolerate lengthy acquisition times or gadolinium agents, machine learning can potentially assist MRI workflow and patient comfort by facilitating faster acquisitions or reducing exogenous contrast dosage. Although artificial intelligence approaches often have limitations, such as problems with generalizability or explainability, there is potential for these techniques to improve diagnostic utility, throughput, and patient experience in clinical MRI practice.

For many patients, numerous unpleasant features of the magnetic resonance imaging (MRI) experience such as scan duration, auditory noise, spatial confinement, and motion restrictions can lead to premature termination or low diagnostic quality of imaging studies. This article discusses practical, patient-oriented considerations that are helpful for radiologists contemplating ways to improve the MRI experience for patients. Patient friendly scanner properties are discussed, with an emphasis on literature findings of effectiveness in mitigating patient claustrophobia, other anxiety, or motion and on reducing scan incompletion rates or need for sedation. As shorter scanning protocols designed to answer specific diagnostic questions may be more practical and tolerable to the patient than a full-length standard-of-care examination, a few select protocol adjustments potentially useful for specific clinical settings are discussed. In addition, adjunctive devices such as audiovisual or other sensory aides that can be useful distractive approaches to reduce patient discomfort are considered. These modifications to the MRI scanning process not only allow for a more pleasant experience for patients, but they may also increase patient compliance and decrease patient movement to allow more efficient acquisition of diagnostic-quality images.

Periprocedural anxiety is a major cause of morbidity, particularly for interventional radiology procedures that often depend on conscious sedation. Management of anxiety and pain during image-guided procedures has traditionally relied on pharmacologic agents such as benzodiazepines and opioids. Although generally safe, use of these medications risks adverse events, and newer noninvasive, nonpharmacologic techniques have evolved to address patient needs. In this review, we explore the roles of hypnosis, structured empathic attention, anodyne imagery, music, video glasses, and mobile applications in reducing procedural anxiety and pain with the goal of improving patient satisfaction, operational efficiency, and clinical outcomes.

Purpose: Our study investigates whether a machine-learning-based system can predict the rate of cognitive decline in mildly cognitively impaired patients by processing only the clinical and imaging data collected at the initial visit. Approach: We built a predictive model based on a supervised hybrid neural network utilizing a three-dimensional convolutional neural network to perform volume analysis of magnetic resonance imaging (MRI) and integration of nonimaging clinical data at the fully connected layer of the architecture. The experiments are conducted on the Alzheimer's Disease Neuroimaging Initiative dataset. Results: Experimental results confirm that there is a correlation between cognitive decline and the data obtained at the first visit. The system achieved an area under the receiver operator curve of 0.70 for cognitive decline class prediction. Conclusion: To our knowledge, this is the first study that predicts "slowly deteriorating/stable" or "rapidly deteriorating" classes by processing routinely collected baseline clinical and demographic data [baseline MRI, baseline mini-mental state examination (MMSE), scalar volumetric data, age, gender, education, ethnicity, and race]. The training data are built based on MMSE-rate values. Unlike the studies in the literature that focus on predicting mild cognitive impairment (MCI)-to-Alzheimer's disease conversion and disease classification, we approach the problem as an early prediction of cognitive decline rate in MCI patients.

We deem a computer to exhibit artificial intelligence (AI) when it performs a task that would normally require intelligent action by a human. Much of the recent excitement about AI in the medical literature has revolved around the ability of AI models to recognize anatomy and detect pathology on medical images, sometimes at the level of expert physicians. However, AI can also be used to solve a wide range of noninterpretive problems that are relevant to radiologists and their patients. This review summarizes some of the newer noninterpretive uses of AI in radiology.

Cerebrospinal fluid leaks pose a serious threat to patients as they represent an unchecked communication between the subarachnoid space and the extracranial environment. Accurate localization of the leakage site is essential for treatment planning. We describe the novel utilization of dual-energy computed tomography technology in cisternography in the evaluation of a patient with a cerebrospinal fluid leak.

Purpose To determine the repeatability of magnetic resonance (MR) elastography-derived shear stiffness measurements of the intervertebral disc (IVD) taken throughout the day and their relationship with IVD degeneration and subject age. Materials and Methods In a cross-sectional study, in vivo lumbar MR elastography was performed once in the morning and once in the afternoon in 47 subjects without current low back pain (IVDs = 230; age range, 20-71 years) after obtaining written consent under approval of the institutional review board. The Pfirrmann degeneration grade and MR elastography-derived shear stiffness of the nucleus pulposus and annulus fibrosus regions of all lumbar IVDs were assessed by means of principal frequency analysis. One-way analysis of variance, paired t tests, concordance and Bland-Altman tests, and Pearson correlations were used to evaluate degeneration, diurnal changes, repeatability, and age effects, respectively. Results There were no significant differences between morning and afternoon shear stiffness across all levels and there was very good technical repeatability between the morning and afternoon imaging results for both nucleus pulposus (R = 0.92) and annulus fibrosus (R = 0.83) regions. There was a significant increase in both nucleus pulposus and annulus fibrosus MR elastography-derived shear stiffness with increasing Pfirrmann degeneration grade (nucleus pulposus grade 1, 12.5 kPa ± 1.3; grade 5, 16.5 kPa ± 2.1; annulus fibrosus grade 1, 90.4 kPa ± 9.3; grade 5, 120.1 kPa ± 15.4), and there were weak correlations between shear stiffness and age across all levels (R ≤ 0.32). Conclusion Our results demonstrate that MR elastography-derived shear stiffness measurements are highly repeatable, weakly correlate with age, and increase with advancing IVD degeneration. These results suggest that MR elastography-derived shear stiffness may provide an objective biomarker of the IVD degeneration process.

ACR RADPEER® is the leading method of radiologic peer review in the United States. The program has evolved since its inception in 2002 and was most recently updated in 2016. In 2018, a survey was sent to RADPEER participants to gauge the current state of the program and explore opportunities for continued improvement. A total of 26 questions were included, and more than 300 practices responded. In this report, the ACR RADPEER Committee authors summarize the survey results and discuss opportunities for future iterations of the RADPEER program.

Given the rapid pace at which modern radiology is evolving and the associated paradigm shifts in health care delivery, it is critical that radiologists adapt and constantly update the skills and knowledge required to practice safe, patient-centered care. The Association of University Radiologists-Radiology Research Alliance Lifelong Learning Task Force convened to explore the current status and future directions of lifelong learning in radiology and summarized its finding in this article. We review the various learning platforms and resources available to radiologists in their self-motivated and self-directed pursuit of lifelong learning. We also discuss the challenges and perceived barriers to lifelong learning and strategies to mitigate those barriers and optimize learning outcomes.

The presence of a single nodal metastasis has significant prognostic and treatment implications for patients with head and neck cancer. This study aims to investigate whether dual-energy computed tomography (DECT)-derived iodine content and spectral attenuation curve analysis can improve detection of nodal metastasis in oropharyngeal carcinoma. Eight patients with newly diagnosed oropharyngeal squamous cell carcinoma and pathologically proven nodal metastatic disease (n = 13 metastatic nodes, n = 16 nonmetastatic nodes) who underwent contrast-enhanced DECT of the neck were retrospectively evaluated. DECT-derived iodine content (mg/mL) and monoenergetic attenuation values at 40 keV and 100 keV were obtained via circular regions of interest within metastatic and nonmetastatic cervical lymph nodes. Iodine content was significantly lower in metastatic nodes (0.96 ± 0.28 mg/mL) than in nonmetastatic nodes (1.65 ± 0.38 mg/mL, P = .002). Iodine spectral attenuation slope was significantly lower in metastatic nodes (1.33 ± 0.49 mg/mL) than in nonmetastatic nodes (1.91 ± 0.64 mg/mL, P = .015). A nodal iodine threshold of ≤1.3 mg/mL showed a sensitivity of 84.6% and a specificity of 75.0%, with an area under the curve of 0.839, P &lt, 0001. At a threshold value of ≤1.95 for nodal spectral attenuation slope, an optimized specificity of 92.3% and specificity of 50.0% was achieved, with an area under the curve of 0.68 (P = .049). DECT-derived quantitative iodine data and spectral attenuation curves may improve the diagnostic accuracy of computed tomography for nodal metastasis in patients with squamous cell carcinoma of the oropharynx.

Various imaging techniques used in stroke evaluation have complementary, evolving roles. This article reviews recent updates in emergent stroke neuroimaging. Multiple recent randomized trials showing benefit of endovascular therapy in strokes with confirmed large-vessel occlusion emphasize need for vessel imaging, typically CTA, in addition to unenhanced head CT. In addition, because recent trials published in 2018 showed efficacy of mechanical thrombectomy at up to 24 h from symptom onset in stroke patients with favorable ischemic core and/or penumbra assessments, imaging techniques to define viable and infarcted tissue, such as DWI or CT/MR perfusion imaging, have an increasing role in emergent stroke assessment. Other studies contributing to new AHA/ASA stroke guidelines are also discussed. Neuroimaging remains crucial to emergent stroke evaluation, with CTA, MRI, and perfusion studies playing an increasing role in identifying patients likely to benefit from treatment.

Purpose Examine the cost of MRI operations before and after implementation of interpersonal skills training to reduce unanticipated patient-related events in an academic medical center. Methods Teams at four MRI sites (two hospital-based, two freestanding) were trained in evidence-based communication skills in February to April 2015. Training was designed to enable staff members to help patients mobilize their innate coping skills in response to any distress they experienced during their MRI visit. Data were collected before training and afterward from January to June 2016. Staff reported the incidence of disruptive motion, sedation use, MRI delays, incomplete examinations, and no-shows. Cost and revenue associated with MRI operations and staff and physician costs were estimated using Medicare and private insurance rates and data from the US Bureau of Labor Statistics. Results The study included 12,930 outpatient MRI visits. From baseline to follow-up, average monthly patient volume increased from 1,105 to 1,463 at hospital MRI sites and from 245 to 313 at freestanding MRI sites. Patient factors necessitating sedation or interfering with image progression or quality decreased from 9.0% to 5.5% at hospital sites and from 3.1% to 1.2% at freestanding sites. These changes translated into a reduction in operational costs of $4,600 per 1,000 scheduled patients and an increase in profit of $8,370 per 1,000 scheduled patients in hospital MRI sites, and a corresponding increase in operational costs of $1,570 per 1,000 scheduled-patients and an increase in profit of $12,800 per 1,000 scheduled patients in freestanding MRI sites. Conclusions We found significant improvements in MRI operational efficiency after interpersonal skills team training, which were associated with reductions in costs and growth in revenue.

Thyroid cancer fulfills the criteria for overdiagnosis by having a reservoir of indolent cancers and practice patterns leading to the diagnosis of incidental cancers from the reservoir. The occurrence of overdiagnosis is also supported by population-based data showing an alarming rise in thyroid cancer incidence without change in mortality. Because one of the activities leading to overdiagnosis is the workup of incidental thyroid nodules detected on imaging, it is critical that radiologists understand the issue of overdiagnosis and their role in the problem and solution. This article addresses 1) essential thyroid cancer facts, 2) the evidence supporting overdiagnosis, 3) the role of radiology in overdiagnosis, 4) harms of overdiagnosis, and 5) steps radiologists can take to minimize the problem.

There are many disorders that can occur in the thyroid gland, ranging from benign to malignant entities. This article focuses on 5 common problems of the thyroid that require special consideration with regard to optimizing imaging strategies in a multidisciplinary and collaborative platform. These problems are the incidental thyroid nodule, preoperative evaluation of goiter, hyperthyroidism, invasive thyroid cancer, and recurrent thyroid cancer. For each problem essential facts, interesting updates, imaging approach, and management pearls are reviewed.

Purpose To assess how a patient's affect on presentation relates to the likelihood of adverse events during their subsequent interventional image-guided procedures. Materials and Methods A secondary analysis was performed of an existing dataset from a clinical trial with 230 patients who underwent percutaneous peripheral vascular and renal interventions and who had completed the positive affect (PA) negative affect (NA) schedule (PANAS) before their procedures. Summary PANAS scores were split over the median and used to classify the participants into those with high vs low PA and high vs low NA. Associations between affect and the absence or presence of adverse medical events were examined by two-sided Fisher exact tests. Results Patients with high baseline NA were significantly more likely to have adverse events during their procedures than those with low baseline NA (18% vs 8%; P = .030). High baseline PA was not associated with a significantly higher frequency of subsequent adverse events compared with low PA (15% vs 9%; P = .23). Patients with high NA requested and received significantly more sedative and opioid agents than those with low NA (2.0 vs 1.0 units requested [ P = .0009]; 3.0 vs 1.0 units received [ P = .0004]). PA levels did not affect medication use. Conclusions High NA, but not PA, was associated with an increased likelihood of adverse events. Improving patients' NA before procedures seems a more suitable target than attempting to boost PA to improve the procedural experience.

The ACR's RADPEER program is currently the leading method for peer review in the United States. To date, more than 18,000 radiologists and more than 1,100 groups participate in the program. The ABR accepted RADPEER as a practice quality improvement in 2009, which can be applied toward maintenance of certification; there are currently over 2,200 practice quality improvement participants. There have been ongoing deliberations regarding the utility of RADPEER, its goals, and its scoring system since the preceding 2009 white paper. This white paper reviews the history and evolution of RADPEER and eRADPEER, the 2016 ACR Peer Review Committee's discussions, the updated recommended scoring system and lexicon for RADPEER, and updates to eRADPEER including the study type, age, and discrepancy classifications. The central goal of RADPEER to aid in nonpunitive peer learning is discussed.

The aim of this study was to assess the effects of team training on operational efficiency during outpatient MRI.In this institutional review board-approved, HIPAA-compliant study, six MRI outpatient sites of a midwestern hospital system were randomized to serve as controls or have their teams trained in advanced communication skills. The fourth quarter of fiscal year 2015 was the trial baseline. The trial ended in the third quarter (Q3) of fiscal year 2016 (FY16). Equipment utilization (completed scans/available slots), hourly scan rates (total orders completed per machine per hour of operation), and no-show rates stratified by time were analyzed using the Cochran-Mantel-Haenszel method, with individual comparisons performed with Bonferroni correction.The study encompassed 27,425 MRI examinations. Overall volume peaked at baseline and then declined over the following quarters. Compared with baseline, untrained sites experienced significant drops in equipment utilization (P.01 for the first quarter of FY16 and P.0001 for the second quarter of FY16 and Q3 FY16), decreasing from 77% to 65% over the study period, corresponding to a decrease from 1.15 to 0.97 in hourly scan rates. For trained sites, these metrics showed no significant change, with maintenance of hourly scan rates of 1.23 and 1.27 and equipment utilization rates of 83% and 85% between baseline and Q3 FY16. No-show rates remained stable at trained sites but increased at untrained sites in the last two quarters (P.05). Nationally benchmarked patient satisfaction percentile ranking gradually increased at trained sites from 56th at baseline to 70th and successively decreased at untrained sites from 66th to 44th.MRI outpatient facilities trained in advanced communication techniques may have more favorable operational efficiency than untrained sites in a saturated market.

BACKGROUND AND PURPOSE: Thyroid nodules are common incidental findings on CT, but there are no clear guidelines regarding their further diagnostic work-up. This study compares the performance of 2 risk-categorization methods of selecting CT-detected incidental thyroid nodules for work-up. MATERIALS AND METHODS: The 2 categorization methods were method A, based on nodule size ≥10 mm, and method B, a 3-tiered system based on aggressive imaging features, patient age younger than 35 years or nodule size of ≥15 mm. In part 1, the 2 categorization methods were applied to thyroid cancers in the SEER data base of the National Cancer Institute to compare the cancer capture rates and survival. In part two, 755 CT neck scans at our institution were retrospectively reviewed for the presence of ITNs of ≥5 mm, and the same 2 categorization methods were applied to the CT cases to compare the number of patients who would theoretically meet the criteria for work-up. Comparisons of proportions of subjects captured under methods A and B were made by using the McNemar test. RESULTS: For 84,720 subjects in the SEER data base, methods A and B each captured 74% (62,708/84,720 and 62,586/84,720, respectively) of malignancies. SEER subjects who would not have met the criteria for further work-up by both methods had equally excellent 10-year cause-specific and relative survival of >99%. For part 2, the prevalence of ITNs of ≥5 mm at our institution was 133/755 (18%). The number of ITNs that would be recommended for work-up by method A was 57/133 (43%) compared with 31/133 (23%) for method B (P < .0005). CONCLUSIONS: Compared with using a 10-mm cutoff, the 3-tiered risk-stratification method identified fewer ITNs for work-up but captured the same proportion of cancers in a national data base and showed no difference in missing high-mortality cancers.