Natural language processing in radiology: Clinical applications and future directions.

Natural language processing (NLP) is a wide range of techniques that allows computers to interact with human text. Applications of NLP in everyday life include language translation aids, chat bots, and text prediction. It has been increasingly utilized in the medical field with increased reliance on electronic health records. As findings in radiology are primarily communicated via text, the field is particularly suited to benefit from NLP based applications. Furthermore, rapidly increasing imaging volume will continue to increase burden on clinicians, emphasizing the need for improvements in workflow. In this article, we highlight the numerous non-clinical, provider focused, and patient focused applications of NLP in radiology. We also comment on challenges associated with development and incorporation of NLP based applications in radiology as well as potential future directions.

Impact of Radiologist-Driven Change-Order Requests on Outpatient CT and MRI Examinations.

PURPOSE: To assess impact of electronic medical record-embedded radiologist-driven change-order request on outpatient CT and MRI examinations. METHODS: Outpatient CT and MRI requests where an order change was requested by the protocoling radiologist in our tertiary care center, from April 11, 2017, to January 3, 2018, were analyzed. Percentage and categorization of requested order change, provider acceptance of requested change, patient and provider demographics, estimated radiation exposure reduction, and cost were analyzed. P < .05 was used for statistical significance. RESULTS: In 79,310 outpatient studies in which radiologists determined protocol, change-order requests were higher for MRI (5.2%, 1,283 of 24,553) compared with CT (2.9%, 1,585 of 54,757; P < .001). Provider approval of requested change was equivalent for CT (82%, 1,299 of 1,585) and MRI (82%, 1,052 of 1,283). Change requests driven by improper contrast media utilization were most common and different between CT (76%, 992 of 1,299) and MRI (65%, 688 of 1,052; P < .001). Changing without and with intravenous contrast orders to with contrast only was most common for CT (39%, 505 of 1,299) and with and without intravenous contrast to without contrast only was most common for MRI (26%, 274 of 1,052; P < .001). Of approved changes in CT, 51% (661 of 1,299) resulted in lower radiation exposure. Approved changes frequently resulted in less costly examinations (CT 67% [799 of 1,198], MRI 48% [411 of 863]). CONCLUSION: Outpatient CT and MRI orders are deemed incorrect in 2.9% to 5% of cases. Radiologist-driven change-order request for CT and MRI are well accepted by ordering providers and decrease radiation exposure associated with imaging.

Clinical Decision Support for Ordering CTA-PE Studies in the Emergency Department-A Pilot on Feasibility and Clinical Impact in a Tertiary Medical Center.

PURPOSE: To determine the feasibility and impact of Clinical Decision Support for imaging ordering. METHODS: A survey of 231 emergency providers identified Computed tomography angiography (CTA)-Pulmonary embolism (PE) as an overutilized study. We developed an algorithm that combined established risk scores to stratify patients for PE work-up (recommendations: CTA, D-dimer or no further testing); the algorithm was integrated into the Epic Radiology Information Ordering System. RESULTS: Among 872 studies requested, 479 (55%) received a recommendation to change their order: 6 (1.3%) were cancelled; 13 (2.7%) changed to a D-dimer, and 460 (96%) proceeded with CTA. Of the 853 studies conducted, 8.2% were positive for PE. The algorithm had good discriminatory power with positivity rates of 12.0% (CT), 10.0% (D-dimer), and 2.6% (no further testing). Compliance with the recommendation ranged from 12%-68% (mean 45%) with 10% correlation between compliance and positivity rates. CONCLUSION: While the CDS algorithm was accurate, it had only a minimal impact on ordering practices, in part due to heterogeneity in physician adherence.

Benefit of a Visual Aid in the Management of Moderate-Severity Contrast Media Reactions.

OBJECTIVE: The purpose of this study was to compare management of moderate-severity reactions with and without a visual aid or flowchart in contrast medium reaction simulations. SUBJECTS AND METHODS: All attending radiologists and trainees were requested to participate in a contrast medium reaction simulation program, including a moderate-severity reaction scenario, and were randomized to groups having or not having a visual aid. The time to administer intramuscular (IM) epinephrine via automated injector and errors in administration were recorded. After the simulation, all participants completed a survey assessing their comfort in treating reactions to contrast media with and without a visual aid. RESULTS: A total of 138 participants were divided into 21 sessions in the program, with 68 participants in the moderate-severity reaction scenario. Eleven groups were provided a visual aid; 10 groups were not. Errors in management occurred in 18.2% (2 of 11) of groups with visual aids versus 40% (4 of 10) in groups without (p = 0.35), with epinephrine self-administration reflecting the most common error. Excluding the groups with errors, the mean time to administration of IM epinephrine was 97 seconds with versus 152 seconds without the visual aid (p = 0.04). Of the 138 participants, 97.8% agreed that the poster would aid in medication administration, and 87% agreed that it would help decrease time to administer medications. CONCLUSION: A visual aid increased the subjective confidence of radiologists in the dose and route of medication administration in the contrast medium reaction simulation and led to faster administration of epinephrine. Self-administration IM epinephrine errors were common and seen in both groups.

JOURNAL CLUB: Benefit of Epinephrine Autoinjector for Treatment of Contrast Reactions: Comparison of Errors, Administration Times, and Provider Preferences.

OBJECTIVE: Given the rarity of contrast reactions in practice, most radiologists have little to no experience in their management, and errors are common. The purpose of this study was to compare treatment of a moderate-severity reaction with intramuscular epinephrine by either the traditional manual method of drawing up and delivering epinephrine with a needle and syringe or the use of an epinephrine autoinjector. SUBJECTS AND METHODS: All diagnostic radiologists at the study institution were requested to participate in an annual contrast reaction simulation program, which consisted of three simulation scenarios in a high-fidelity simulation laboratory. During the moderate-severity simulation scenario, the time to administer intramuscular epinephrine and any errors in administration were recorded. Groups were randomized to use an autoinjector device or manual delivery. All participants completed a survey assessing the experience with epinephrine and their comfort in treating contrast reactions using a traditional manual approach versus an epinephrine autoinjector. RESULTS: Among 189 participants in the contrast reaction simulation program, 76 participated in a moderate-severity reaction simulation two to five at a time in 25 sessions. Mean total time to administration was significantly longer for manual (108.8 seconds) than for autoinjector (38.7 seconds) delivery (p < 0.001). There were 11 errors in the manual group and one error in the autoinjector group (p = 0.005). Ninety-four percent of participants reported feeling very comfortable or comfortable with the autoinjector as opposed to 60% for manual delivery (p < 0.001). Overall, 96% of participants thought the autoinjector was easier to use. CONCLUSION: Use of an epinephrine autoinjector for treatment of contrast reactions was associated with a significantly greater degree of provider comfort, shorter time to administration, and fewer errors.

High-Fidelity Contrast Reaction Simulation Training: Performance Comparison of Faculty, Fellows, and Residents.

PURPOSE: Reactions to contrast material are uncommon in diagnostic radiology, and vary in clinical presentation from urticaria to life-threatening anaphylaxis. Prior studies have demonstrated a high error rate in contrast reaction management, with smaller studies using simulation demonstrating variable data on effectiveness. We sought to assess the effectiveness of high-fidelity simulation in teaching contrast reaction management for residents, fellows, and attendings. METHODS: A 20-question multiple-choice test assessing contrast reaction knowledge, with Likert-scale questions assessing subjective comfort levels of management of contrast reactions, was created. Three simulation scenarios that represented a moderate reaction, a severe reaction, and a contrast reaction mimic were completed in a one-hour period in a simulation laboratory. All participants completed a pretest and a posttest at one month. A six-month delayed posttest was given, but was optional for all participants. RESULTS: A total of 150 radiologists participated (residents = 52; fellows = 24; faculty = 74) in the pretest and posttest; and 105 participants completed the delayed posttest (residents = 31; fellows = 17; faculty = 57). A statistically significant increase was found in the one-month posttest (P < .00001) and the six-month posttest scores (P < .00001) and Likert scores (P < .001) assessing comfort level in managing all contrast reactions, compared with the pretest. Test scores and comfort level for moderate and severe reactions significantly decreased at six months, compared with the one-month posttest (P < .05). CONCLUSIONS: High-fidelity simulation is an effective learning tool, allowing practice of "high-acuity" situation management in a nonthreatening environment; the simulation training resulted in significant improvement in test scores, as well as an increase in subjective comfort in management of reactions, across all levels of training. A six-month refresher course is suggested, to maintain knowledge and comfort level in contrast reaction management.