Using Natural Language Processing of Free-Text Radiology Reports to Identify Type 1 Modic Endplate Changes.

Electronic medical record (EMR) systems provide easy access to radiology reports and offer great potential to support quality improvement efforts and clinical research. Harnessing the full potential of the EMR requires scalable approaches such as natural language processing (NLP) to convert text into variables used for evaluation or analysis. Our goal was to determine the feasibility of using NLP to identify patients with Type 1 Modic endplate changes using clinical reports of magnetic resonance (MR) imaging examinations of the spine. Identifying patients with Type 1 Modic change who may be eligible for clinical trials is important as these findings may be important targets for intervention. Four annotators identified all reports that contained Type 1 Modic change, using N = 458 randomly selected lumbar spine MR reports. We then implemented a rule-based NLP algorithm in Java using regular expressions. The prevalence of Type 1 Modic change in the annotated dataset was 10%. Results were recall (sensitivity) 35/50 = 0.70 (95% confidence interval (C.I.) 0.52-0.82), specificity 404/408 = 0.99 (0.97-1.0), precision (positive predictive value) 35/39 = 0.90 (0.75-0.97), negative predictive value 404/419 = 0.96 (0.94-0.98), and F1-score 0.79 (0.43-1.0). Our evaluation shows the efficacy of rule-based NLP approach for identifying patients with Type 1 Modic change if the emphasis is on identifying only relevant cases with low concern regarding false negatives. As expected, our results show that specificity is higher than recall. This is due to the inherent difficulty of eliciting all possible keywords given the enormous variability of lumbar spine reporting, which decreases recall, while availability of good negation algorithms improves specificity.

Altered Mental Status in ICU Patients: Diagnostic Yield of Noncontrast Head CT for Abnormal and Communicable Findings.

OBJECTIVE: To evaluate the diagnostic yield of noncontrast head CT for acute communicable findings in ICU patients specifically scanned for altered mental status. DESIGN: Retrospective observational cohort study. SETTING: University Hospital Neuroscience, Medical, and Surgical ICUs. PATIENTS: ICU patients with new-onset altered mental status. INTERVENTION: Noncontrast head CT. MEASUREMENTS AND MAIN RESULTS: Reports on head CTs from two university hospitals performed for the sole indication of altered mental status in ICU patients between July 2011 and June 2013 were reviewed for 1) acute (new or worsening) hemorrhage, 2) mass effect/herniation, 3) infarction, and 4) hydrocephalus. Subgroup analyses of positive findings were performed by 1) ICU group type, 2) age, and 3) race. A total of 2,486 head CTs were performed in 1,357 patients whose age ranged from 14 to 116 years (median, 59; mean, 57.6 ± 16). Acute communicable findings in at least one of four categories were present in 22.8% (566/2,486) of examinations, with hydrocephalus being most common (11.5% [286/2,486]). The frequency of any acute communicable findings in neuroscience, medical, and surgical ICUs was 28.6% (471/1,648), 9.8% (43/440), and 13.1% (52/398), respectively. Neuroscience ICU head CTs had significantly higher rates of acute communicable findings in all categories, except for acute infarction, compared with the other two ICUs (p < 0.001). Acute hydrocephalus (13.6% vs 7.4%; p < 0.001) and mass effect (6.7% vs 4.3%; p = 0.01) were more common in patients less than 65 years. For other acute categories, no significant difference was noted by age. There was no significant difference in the likelihood of a positive examination by race. CONCLUSIONS: Almost one in four head CTs in a university ICU patient population performed for primary indication of altered mental status yields abnormal communicable findings. In this patient population, utilization management barriers to examination ordering should be minimized.

Multidetector CT of blunt cervical spine trauma in adults.

A number of new developments in cervical spine imaging have transpired since the introduction of 64-section computed tomographic (CT) scanners in 2004. An increasing body of evidence favors the use of multidetector CT as a stand-alone screening test for excluding cervical injuries in polytrauma patients with obtundation. A new grading scale that is based on CT and magnetic resonance (MR) imaging findings, the cervical spine Subaxial Injury Classification and Scoring (SLIC) system, is gaining acceptance among spine surgeons. Radiographic measurements described for the evaluation of craniocervical distraction injuries are now being reevaluated with the use of multidetector CT. Although most patients with blunt trauma are now treated nonsurgically, evolution in the understanding of spinal stability, as well as the development of new surgical techniques and hardware, has driven management strategies that are increasingly favorable toward surgical intervention. It is therefore essential that radiologists recognize findings that distinguish injuries with ligamentous instability or a high likelihood of nonfusion that require surgical stabilization from those that are classically stable and can be treated with a collar or halo vest alone. The purpose of this article is to review the spectrum of cervical spine injuries, from the craniocervical junction through the subaxial spine, and present the most widely used grading systems for each injury type.

64-MDCT angiography of blunt vascular injuries of the neck.

OBJECTIVE: CT angiography (CTA) using 64-MDCT enables timely evaluation of injuries associated with blunt neck trauma. The purpose of this article is to familiarize the reader with the most frequent CTA signs of blunt vascular injury. CONCLUSION: CTA is a valuable tool to detect blunt vascular injuries, especially using its multiplanar and 3D reconstruction capabilities.

Fostering a Healthy AI Ecosystem for Radiology: Conclusions of the 2018 RSNA Summit on AI in Radiology.

The 2018 RSNA Summit on AI in Radiology brought together a diverse group of stakeholders to identify and prioritize areas of need related to artificial intelligence in radiology. This article presents the proceedings of the summit with emphasis on RSNA's role in leading, organizing, and catalyzing change during this important time in radiology. © RSNA, 2019.

Radiographic Cobb Angle: A Feature of Congenital Lumbar Spine Stenosis.

PURPOSE: A low cost, reproducible radiographic method of diagnosing congenital lumbar spinal stenosis (CLSS) is lacking. We hypothesized that the Cobb angle for lumbar lordosis would be smaller in patients with CLSS, based on observations in our spine clinic patient population. Here, we compared lumbar lordosis Cobb angles with the radiographic ratio method in patients with normal spine imaging, degenerative spinal stenosis, and with CLSS. MATERIALS AND METHODS: Orthopedic surgeons categorized patients with low back pain as "Normal," "Degenerative spinal stenosis," and "CLSS" based on clinical presentation and findings on lumbar magnetic resonance imaging. We included 30 patients from each cohort who had undergone lateral lumbar spine radiographs and lumbar magnetic resonance imaging. For each lateral radiograph, 2 measurement methods were used (1) 4-line lumbosacral Cobb angle between L2-S1 and (2) the ratio of the anteroposterior vertebral body diameter and spinal canal anteroposterior diameter at the L3 level. We performed logistic regression analyses of CLSS prediction by Cobb angle vs the ratio method in all three cohorts. Covariates included age, gender, and body mass index. RESULTS: The radiographic Cobb angles were smaller in CLSS patients when compared to the degenerative disease and normal cohorts: a smaller radiographic Cobb angle showed higher odds ratio (OR) of predicting CLSS diagnosis compared to the radiographic ratio when compared with degenerative disease (OR = 0.28; 95% CI: 0.11-0.78, P = 0.01) and when compared with the normal cohort (OR = 0.46; 95% CI: 0.24-0.92, P = 0.03). Radiographic ratio measurements showed no difference between the three cohorts (P = 0.12). CLSS was associated with male gender (P = 0.04), younger age (P = 0.01), and higher body mass index (P = 0.01). CONCLUSION: The radiographic Cobb angle method for lumbar lordosis may be useful for raising the possibility of CLSS as the diagnosis.

Racial and Ethnic Differences in Diagnostic Imaging Utilization During Adult Emergency Department Visits in the United States, 2005 to 2014.

OBJECTIVE: To compare the use of medical imaging (x-ray [XR], CT, ultrasound, and MRI) in the emergency department (ED) for adult patients of different racial and ethnic groups in the United States from 2005 to 2014. METHODS: We performed a multilevel stratified regression analysis of the National Hospital Ambulatory Medical Care Survey ED Subfile, a nationally representative database of hospital-based ED visits. We examined race (white, black, Asian, other) and ethnicity (Hispanic versus non-Hispanic) as the primary exposures for the outcomes of ED medical imaging use (XR, CT, ultrasound, MRI, and any imaging). We controlled for other potential patient-level and facility-level determinants of ED imaging use. RESULTS: Approximately half (48.8%) of the 225,037 adult patient ED visits underwent imaging; 36.1% underwent XR, 16.4% CT, 4.1% ultrasound, and 0.8% MRI. White patients received imaging during 51.3% of their encounters, black patients received imaging during 43.6% of their encounters, Asians received imaging during 50.8% of their encounters, and other races received imaging during 46% of their encounters. As compared with white patients, black patients had decreased adjusted odds of receiving imaging in the ED (odds ratio [OR] = 0.86, 95% confidence interval [CI]: 0.84-0.89). Comparatively, black patients had a lower odds of CT scan (OR = 0.80, 95% CI: 0.77-0.83) or MRI (OR = 0.74, 95% CI: 0.65-0.85). Hispanic patients and Asian patients had a higher odds of receiving ultrasound (OR = 1.36, 95% CI: 1.27-1.44 and OR = 1.25, 95% CI: 1.10-1.42), respectively. IMPLICATIONS: We observed significant racial and ethnic differences in medical imaging use in the ED even after controlling for patient- and facility-level factors.

Advanced diagnostic imaging utilization during emergency department visits in the United States: A predictive modeling study for emergency department triage.

BACKGROUND: Emergency department (ED) crowding is associated with negative health outcomes, patient dissatisfaction, and longer length of stay (LOS). The addition of advanced diagnostic imaging (ADI), namely CT, ultrasound (U/S), and MRI to ED encounter work up is a predictor of longer length of stay. Earlier and improved prediction of patients' need for advanced imaging may improve overall ED efficiency. The aim of the study was to detect the association between ADI utilization and the structured and unstructured information immediately available during ED triage, and to develop and validate models to predict utilization of ADI during an ED encounter. METHODS: We used the United States National Hospital Ambulatory Medical Care Survey data from 2009 to 2014 to examine which sociodemographic and clinical factors immediately available at ED triage were associated with the utilization of CT, U/S, MRI, and multiple ADI during a patient's ED stay. We used natural language processing (NLP) topic modeling to incorporate free-text reason for visit data available at time of ED triage in addition to other structured patient data to predict the use of ADI using multivariable logistic regression models. RESULTS: Among the 139,150 adult ED visits from a national probability sample of hospitals across the U.S, 21.9% resulted in ADI use, including 16.8% who had a CT, 3.6% who had an ultrasound, 0.4% who had an MRI, and 1.2% of the population who had multiple types of ADI. The c-statistic of the predictive models was greater than or equal to 0.78 for all imaging outcomes, and the addition of text-based reason for visit information improved the accuracy of all predictive models. CONCLUSIONS: Patient information immediately available during ED triage can accurately predict the eventual use of advanced diagnostic imaging during an ED visit. Such models have the potential to be incorporated into the ED triage workflow in order to more rapidly identify patients who may require advanced imaging during their ED stay and assist with medical decision-making.

Predictive Value of First Posttreatment Imaging Using Standardized Reporting in Head and Neck Cancer.

OBJECTIVE: The Neck Imaging Reporting and Data System (NI-RADS) is a standardized numerical reporting template for surveillance of head and neck squamous cell carcinoma (HNSCC). Our aim was to analyze the accuracy of NI-RADS on the first posttreatment fluorodeoxyglucose positron emission tomography/contrast-enhanced computed tomography (PET/CECT). STUDY DESIGN: Retrospective cohort study. SETTING: Academic tertiary hospital. SUBJECT AND METHODS: Patients with HNSCC with a 12-week posttreatment PET/CECT interpreted using the NI-RADS template and 9 months of clinical and radiologic follow-up starting from treatment completion between June 2014 and July 2016 were included. Treatment failure was defined as positive tumor confirmed by biopsy or Response Evaluation Criteria in Solid Tumors criteria. Cox proportional hazards models were performed. RESULTS: This study comprised 199 patients followed for a median of 15.5 months after treatment completion (25% quartile, 11.8 months; 75% quartile, 20.2 months). The rates of treatment failure increased with each incremental increase in NI-RADS category from 1 to 3 (4.3%, 9.1%, and 42.1%, respectively). A Cox proportional hazards model demonstrated a strong association between NI-RADS categories and treatment failure at both primary and neck sites (hazard ratio [HR], 2.60 and 5.22, respectively; P < .001). In the smaller treatment subgroup analysis, increasing NI-RADS category at the primary site in surgically treated patients and treatment failure did not achieve statistically significant association (HR, 0.88; P = .82). CONCLUSION: Increasing NI-RADS category at the baseline posttreatment PET/CECT is strongly associated with increased risk of treatment failure in patients with HNSCC.

Challenges Related to Artificial Intelligence Research in Medical Imaging and the Importance of Image Analysis Competitions.

In recent years, there has been enormous interest in applying artificial intelligence (AI) to radiology. Although some of this interest may have been driven by exaggerated expectations that the technology can outperform radiologists in some tasks, there is a growing body of evidence that illustrates its limitations in medical imaging. The true potential of the technique probably lies somewhere in the middle, and AI will ultimately play a key role in medical imaging in the future. The limitless power of computers makes AI an ideal candidate to provide the standardization, consistency, and dependability needed to support radiologists in their mission to provide excellent patient care. However, important roadblocks currently limit the expansion of this field in medical imaging. This article reviews some of the challenges and potential solutions to advance the field forward, with focus on the experience gained by hosting image-based competitions.

Conventional and Advanced Imaging of Spine Oncologic Disease, Nonoperative Post-treatment Effects, and Unique Spinal Conditions.

In this review, we discuss the imaging features of diseases and conditions ranging from neoplastic to nonoperative post-treatment effects to unique conditions of the spine. Additionally, advanced imaging may increase diagnostic certainty in cases where conventional imaging characteristics of benign lesions and malignant pathology are variable.

Doubly-Robust Estimation of Effect of Imaging Resource Utilization on Discharge Decisions in Emergency Departments.

Cluster analysis provides a data-driven multidimensional approach for identifying distinct subgroups of patients in a cohort. Each of the clusters represents a particular health condition with specific clinical trajectory and medical needs. Patients visiting emergency rooms do not share the same health condition, therefore discriminating between groups may have implications for diagnostic testing and resource utilization. We carried out this retrospective cohort study on 13825 patients who visited the emergency rooms in three Emory hospitals presenting with head trauma and non-stroke-like non-specific neurologic symptoms from January 2010 to September 2015. We utilized k-means clustering to find five distinct subgroups. Then, we investigated if getting an emergency head CT scan could have a statistically significant effect on getting discharged from the hospital. Adjusted effect estimation method was applied on each cluster to estimate the association between receiving a diagnostic test (e.g., head CT scan) on the disposition status. Out of five patient subgroups in the cohort, the chance of getting discharged for two clusters were significantly affected by getting a head CT scan. They both include comparatively older, African American or black patients who arrived in the ER with EMS, the latter suggesting critical health conditions.

Trends in vertebral augmentation for spinal fractures in myeloma patients: a 2002-2012 population-based study using a large national cancer registry.

PURPOSE: To evaluate temporal trends and factors associated with vertebral augmentation use in myeloma patients with spinal fractures from 2002 to 2012. METHODS: This retrospective cohort study used the Surveillance, Epidemiology and End Results (SEER)-Medicare claims database for 2002 through 2012. We included patients age ≥66 years with myeloma and spinal fractures. First, we evaluated receipt of vertebral augmentation. Second, multivariate logistic regression was used to assess the impact of sociodemographic factors, treatment facility type, and underlying comorbidities on the odds of undergoing vertebral augmentation. RESULTS: Of 4725 myeloma patients with spinal fractures, 653 underwent vertebral augmentation. Procedures increased initially from <1.7% in 2002 to 21.0% (109/520) in 2007, 18.6% (81/435) in 2008, 21.4% (109/509) in 2009, and 17.5% (76/435) in 2011. Patients with a spinal fracture before myeloma diagnosis were twice as likely to undergo vertebral augmentation as patients with fracture after myeloma diagnosis (OR 2.06, 95% CI 1.55 to 2.75). Black patients were half as likely to undergo vertebral augmentation as white patients (OR 0.48, 95% CI 0.34 to 0.68). Patients with 3-5 comorbidities (OR 0.78, 95% CI 0.64 to 0.96) and ≥6 comorbidities (OR 0.69, 95% CI 0.54 to 0.87) were less likely than patients with 0-2 comorbidities to undergo vertebral augmentation. CONCLUSIONS: Vertebral augmentation for myeloma patients with spinal fractures peaked between 2007 and 2009 and then declined. Providers may have adopted vertebral augmentation in myeloma patients since its introduction, and potentially modified practice patterns following the publication of trials of vertebral augmentation in patients with osteoporotic spinal fractures.

How is the Doctor Feeling? ICU Provider Sentiment is Associated with Diagnostic Imaging Utilization.

The judgment of intensive care unit (ICU) providers is difficult to measure using conventional structured electronic medical record (EMR) data. However, provider sentiment may be a proxy for such judgment. Utilizing 10 years of EMR data, this study evaluates the association between provider sentiment and diagnostic imaging utilization. We extracted daily positive / negative sentiment scores of written provider notes, and used a Poisson regression to estimate sentiment association with the total number of daily imaging reports. After adjusting for confounding factors, we found that (1) negative sentiment was associated with increased imaging utilization $(p < 0.01)$, (2) sentiment's association was most pronounced at the beginning of the ICU stay $(p < 0.01)$, and (3) the presence of any form of sentiment increased diagnostic imaging utilization up to a critical threshold $(p < 0.01)$. Our results indicate that provider sentiment may clarify currently unexplained variance in resource utilization and clinical practice.

Translating New Imaging Technologies to Clinical Practice.

Radiology continues to benefit from constant innovation and technological advances. However, for promising new imaging technologies to reach widespread clinical practice, several milestones must be met. These include regulatory approval, early clinical evaluation, payer reimbursement, and broader marketplace adoption. Successful implementation of new imaging tests into clinical practice requires active stakeholder engagement and a focus on demonstrating clinical value during each phase of translation.

Spinal Diffusion Tensor Imaging in Evaluation of Preoperative and Postoperative Severity of Cervical Spondylotic Myelopathy: Systematic Review of Literature.

BACKGROUND: Diffusion tensor imaging (DTI) is increasingly investigated as a potential diagnostic and prognostic tool for symptomatic degenerative cervical pathology; however, it is yet to be validated for this purpose. OBJECTIVE: To investigate the association of preoperative DTI signal changes and postoperative outcomes in patients with cervical spondylotic myelopathy (CSM). METHODS: We performed a systematic literature review using PubMed for clinical studies using DTI in adults undergoing operative management for CSM. Data on preoperative clinical status, preoperative DTI metrics, and postoperative clinical outcomes were abstracted. Preoperative DTI parameters were correlated with preoperative severity and postoperative outcomes and pooled across studies. RESULTS: Nine studies met inclusion criteria for 238 patients who underwent operative management with mean follow-up time 310 days. Higher preoperative fractional anisotropy (FA) at the level of maximal compression correlates strongly with a higher preoperative modified Japanese Orthopaedic Association (mJOA) score (n = 192 patients, rho = 0.62, P < 0.001). Higher preoperative FA is associated with less postoperative mJOA change (n = 27, rho = -0.42, P = 0.02) but a greater recovery rate (n = 93, rho = 0.32, P < 0.001). Preoperative FA correlated with lower Neck Disability Index (n = 15, rho = -0.61, P = 0.04). Preoperative fiber tract ratio had a large positive correlation with a postoperative recovery rate (n = 20, rho = 0.61, P = 0.005). When reported, an apparent diffusion coefficient showed an inverse correlation compared with FA. CONCLUSION: DTI is associated with preoperative severity and postoperative outcomes in CSM patients, suggesting that DTI may become useful in identifying those most likely to benefit from operative intervention (Level 3 Evidence). Prospective trials with standardized DTI acquisition techniques and patient selection are required for higher-level evidence.

Imaging of Spinal Manifestations of Hematological Disorders.

Imaging manifestations of hematological diseases and their potential complications are broad, and there may be significant overlap in features of various disease processes. Knowledge of appropriate choice of imaging test, pertinent imaging patterns, and pathophysiology of disease can help the reader increase specificity in the diagnosis and treatment of the patient. Most importantly, we encourage readers of this review to engage their radiologists during the diagnostic, treatment, and management phases of care delivery.

Bibliometric Analysis of Manuscript Title Characteristics Associated With Higher Citation Numbers: A Comparison of Three Major Radiology Journals, AJNR, AJR, and Radiology.

Our purpose was to determine if associations exist between titles characteristics and citation numbers in Radiology, American Journal of Roentgenology (AJR), and American Journal of Neuroradiology (AJNR). This retrospective study is Institutional Review Board exempt. We searched Web of Science for all original research and review articles in Radiology, AJR, and AJNR between 2006 and 2012 and tabulated number of words in the title, presence of a colon symbol, and presence of an acronym. We used a Poisson regression model to evaluate the association between number of citations and title characteristics. We then used the Wald test to detect pairwise differences in the effect of title characteristics on number of citations among the 3 journals. Between 2006 and 2012, Radiology published 2662, AJR 3998, and AJNR 2581 original research and review articles. There was a citation number increase per title word increase of 1.6% for AJNR and 2.6% for AJR and decrease of 0.8% for Radiology. For all, P < 0.001. A title colon was associated with citation increases for AJNR (16%), Radiology (14%), and AJR (7.4%). Title acronym was associated with citation increases for AJNR (10%), Radiology (14%), and AJR (13.3%). All P < 0.001. AJR had greatest effect for number of words in title vs Radiology and AJNR (P < 0.001), AJNR for presence of colon vs Radiology (P < 0.001), and AJR for presence of acronym vs AJNR (P = 0.028). Title characteristics investigated here showed a strong association with higher citation numbers in Radiology, AJR, and AJNR.

Expectations of Medical Student Neuroradiology Education: A Survey of Practicing Neuroradiologists and Neurologists.

The purpose of this study is to evaluate which neuroradiological diseases neuroradiologists and neurologists believe medical students should be exposed to during their neuroradiology rotation. Members of the American Society of Neuroradiology (ASNR) and the American Academy of Neurology (AAN) were surveyed. Respondents were presented 32 diseases with neuroimaging findings and asked which ones medical students should be exposed to during a neuroradiology rotation. Using a 50% response threshold per disease entity, results were tabulated into 3 groups: diagnoses that (1) more than 50% of neuroradiologists and neurologists felt medical students should see radiologically by rotation completion, (2) less than 50% of respondents in both the groups felt were important, and (3) both the groups disagree are important. Both the groups thought medical students should be exposed to imaging of intraparenchymal hemorrhage (ASNR = 80.4% vs AAN = 84.3%; P = 0.346) and subarachnoid hemorrhage (ASNR = 74% vs AAN = 78%; P = 0.394). Both the groups (>50%) thought subdural hematoma, acute ischemic stroke, epidural hematoma, and spinal cord compression are important. Conditions such as spine fractures, nonacute stroke, arteriovenous malformation, and ear-nose-throat pathology showed varied results between both the groups. Varying degrees of similarity and differences exist between the expectations of neuroradiologists and neurologists regarding medical student neuroradiology education, presenting a positive opportunity for greater consensus, dialogue, and joint curriculum formation.