Purpose: Mass casualty incidents (MCIs) are increasing in frequency across Europe. Radiology plays a critical role in appropriate patient triage during an MCI. Major incident planning (MIP) can help optimize radiology’s response with the goal of improving patient outcomes. Historically, radiology is often underrepresented or excluded from MIP. Our goal was to identify key topics that should be addressed by radiology during MIP, thus promoting MCI awareness and education amongst radiologists.
Methods and Materials: A review of the literature was performed. The advanced search builder on PubMed was utilized. Search terms “radiology” and “mass casualty incident” were used in the primary search filter. External MCI protocols from level 1 trauma centres were obtained on request and reviewed.
Results: 27 relevant studies were identified on the primary search of PubMed and review of references. Trauma guidelines from the Royal College of Radiologists and NICE were also reviewed. Key topics include patient identification strategy, scan ordering, patient tracking, MCI scan protocols, result communication, PACS failure contingencies, scalability of response and individualized radiology staff action cards.
Conclusion: All radiology departments should partake in MIP and in MCI simulation to help maximize patient throughput and improve triage efficiency in the event of an MCI. We have provided an evidenced-based radiology MCI protocol template for use by radiologists during MIP.
Purpose: In trauma, every second counts: time = life. Whole-body computed tomography (WBCT) for trauma patients can achieve both significant time and survival benefits. Faster radiological diagnosis could be lifesaving, particularly for haemodynamically critical or MCI patients. Hence, high efficacy of the scanning workflow is paramount. However, detailed time claims of different scan setups are not known. The aim of our study was to investigate the impact of different WBCT procedures on trauma CT times.
Methods and Materials: Adult anthropomorphic phantoms were used to evaluate 6 different scan protocols in a trauma exercise, with 5 simulation tests for each protocol. Tested workflow settings were: single vs double scout acquisitions, acquisition direction, non-enhanced vs contrast-enhanced, and fixed-delay or bolus triggering contrast administration. Several start and end time points were measured for the complete trauma radiology period: scout acquisitions, CT acquisitions, and reconstruction times, respectively.
Results: Of the total trauma radiology time, up to 83% consisted of non-radiation time. Scout acquisition times claim between 7 and 24% of the examination time. Acquisition times of scouts were close (range 70-93%) to total CT acquisition times. The fastest protocol was a non-enhanced one-stop WBCT protocol, costing in mean only 79 seconds to completion (p<0.001). For contrast-enhanced scans, a fixed delay split bolus protocol could present all axial images 159 seconds after start of imaging.
Conclusion: Scout acquisition and non-acquisition times are significant time-limiting factors in WBCT. Optimising examination times of WBCT protocols can contribute to faster trauma radiology workflow.
Purpose: To evaluate dose savings and image quality after implementing adaptive statistical iterative reconstruction-V (ASiR-V) on a revised protocol for whole-body CT (WBCT) for trauma patients.
Methods and Materials: One hundred multi-trauma patients were scanned using a 256-section multidetector CT system (GE Healthcare Revolution system). They were randomized into two groups using two different scanning protocols. Group (A) (n=50, age 32.48±8.09) underwent conventional protocol including unenhanced scan for brain and cervical spines, then an contrast-enhanced arterial-phase of the thorax and abdomen followed by a venous and delayed scans of the abdomen and pelvis. Group (B) (n=50, age 35.94±13.57) underwent biphasic injection protocol including unenhanced scan for brain and cervical spines, followed by a one-step acquisition of the thorax, abdomen, and pelvis following a biphasic injection, the examination was ended by delayed phase for the abdomen and pelvis. All examination were done under 50 % ASiR-V. Image count, radiation dose, total acquisition time, mediastinal artifacts were compared between the groups. Two radiologists independently graded image quality from 1 to 5. In addition, contrast enhancement was measured in the pulmonary artery, aorta, inferior vena cava, portal vein, liver, spleen, and kidneys.
Results: The mean (±SD) dose length product for group (A) was 2202.3 ± 271.8 mGy*cm and markedly higher when compared to group (B) (p < 0.001) which was 1485.8 ± 489.2 mGy*cm. Protocol B gave a dose reduction of 32.5% and 7.7 % time reduction. The HU values of the aorta & liver were significantly higher in group (A) while kidneys values were higher in group (B). There was no significant difference between the image quality scores for both groups, however group (A) scored higher grades.
Conclusion: Implementing ASiR-V algorism into biphasic CT protocol markedly reduced radiation with maintenance of accuracy and image quality.
Purpose: To assess the diagnostic performance of CT regarding hollow viscus injury (HVI) from penetrating abdominal trauma.
Methods and Materials: Retrospective analysis of patients with penetrating abdominal trauma during 2013-2016. Data from Karolinska Trauma Registry, medical records, and from local PACS were reviewed. CT and surgical findings were compared.
Results: Of 636 patients with penetrating trauma, 177 had abdominal trauma (mean age 34 yrs, range 16-88 yrs; M 163, F 14), and 155 (85%, 155/177) were imaged with CT on arrival. 128 (83%) were stab wounds and 21 (14%) gunshot wounds. 6 patients had miscellaneous trauma mechanisms. Thereof 47 (47/155; 30%) had emergent surgery after CT imaging. Two patients were imaged using oral, rectal and i.v. contrast, 23 with rectal and i.v. contrast, and 22 with i.v. contrast only. Surgery revealed HVI in 26 patients. CT had an overall sensitivity of 69.2%, specificity 90.5%, PPV 90.0% and NPV 70.4%. CT with oral and/or rectal contrast (n = 25) had sensitivity of 66.7%, specificity 71.4%, PPV 85.7% and NPV 45.5%. CT with i.v. contrast only (n = 22) had sensitivity 75.0%, specificity 100%, PPV 100% and NPV 87.5%. Difference in sensitivity between the oral and/or rectal contrast group and i.v. contrast only was not statistically significant (p=1.00).
Conclusion: Stab wounds were the most common cause of penetrating abdominal trauma. The overall sensitivity and specificity of CT in detecting HVI were 69.2% and 90.5%, respectively. The use of oral and/or rectal contrast yielded similar sensitivity to the use of i.v. contrast only.
Purpose: Non-operative management (NOM) of blunt splenic and hepatic injuries has been getting increasingly common. Next to clinical parameters, CT-based grading systems are habitually used as screening tools in early management. Usually the AAST classification is used. Shortly, another scoring system, which contemplated contrast-media extravasation, was introduced for spleen injuries (CTSI). This study validates this system for spleen injuries and proposes/validates an adapted classification for liver injuries.
Methods and Materials: Retrospective analysis of patients with traumatic blunt liver/spleen lesions in the MUI from 2000 to 2016. CT imaging on admission was reevaluated by two radiologists, using the AAST and the CTSIs. Both classifications were examined regarding their capability to predict the necessity of operative treatment, the failing of NOM and the in-hospital-mortality.
Results: In total, 720 patients were analysed (median 32; 230 female), 276 with spleen, 364 with liver and 80 with combined injuries. The total mortality was 5.6%. 160 patients had severe (grade 4-5) lesions according to AAST, with a mortality of 6.0%/5.3% (severe/mild liver injuries, p=0.790) and 10.5%/3.2% (spleen injuries, p=0.008). When using CTSI, 33 patients with liver and 87 patients with spleen lesions had severe injuries (≥4a), associated with a higher in-hospital-mortality: 12.1%/4.9% for liver (p=0,095) and 10.3%/3.0% for spleen injuries (p=0.005). Both classifications showed a high-significant tendency (p < 0.001) of severe lesions more often requiring first-hand operative treatment and were equally good for predicting NOM failure.
Conclusion: The CTSI is able to predict in-hospital mortality and proved efficient as a management indication tool and as a predictor for NOM failure. Therefore, it should substitute the AAST classification as the gold standard.
Purpose: We compared patients with delayed events and immediate events who required splenic angioembolisation (SAE) for salvaging failures of non-operative management (NOM) after blunt splenic trauma.
Methods and Materials: From 2012 to 2017, 161 patients of blunt splenic trauma who were designated for NOM and treated with SAE at our institution were identified. Delayed event occurred at least 4 days after trauma, whereas immediate event occurred within 24 hours. We excluded 11 patients whose events fell between 1 and 4 days. The final inclusion was 150 patients (38 women, 112 men) with median of 33.0 (IQR 21, 47) years. Records were retrospectively reviewed for clinical and CT characteristics. Comparisons between delayed and immediate events as well as predictive factors for delayed events were computed.
Results: Delayed events occurred in 23 (15.3%) patients and were significantly associated with less vascular injuries on CT (3.5% vs. 96.5%), smaller CT score (3.3 vs. 5.2) and smaller hemoperitoneum score (2.9 vs. 4.3). Significant predictive factors for delayed events were CT score, hemoperitoneum score and initial platelet count. Among 23 delayed events, paired T test showed a worse CT score (3.2 vs 4.8; p=0.001) and lower haemoglobin (11.9 to 10.1 g/dL; p=0.004) on follow-up CT examinations and haemograms. Patients with immediate events had a significant longer length of stay after SAE than patients with delayed event, 12.7 days vs. 9.4 days.
Conclusion: Delayed events of blunt splenic trauma were associated with CT findings. The significant predictive factors were CT score, hemoperitoneum score and initial platelet count.
Purpose: In haemodynamically stable patients, the standard of care for blunt liver and spleen injuries is the non-operative management. In our Institution, MR was introduced for the follow-up of haemodynamically stable patients with blunt liver and spleen injuries. The aim is to describe the imaging protocol, the main findings and the potentialities in the use of MR in these patients.
Methods and Materials: From May 2018, patients with blunt liver and spleen injuries without MR contraindications, after admission CT, were followed up by MR. The imaging protocol was tailored to each patient. Post-contrast sequences were acquired especially in the early follow-up and in high grade injuries. In suspected biliary complications, MR cholangiographic sequences were added, and hepatobiliary-specific contrast agent (gadobenate dimeglumine) administration with delayed acquisitions was used.
Results: Seventeen patients with spleen (6 pts.) and liver (11 pts.) injuries of low (10 pts.) and high (7 pts.) grade, were followed-up by MR. The signal behaviour of the healing lesions was documented and described. In 4 patients, liver intraparenchymal collections were detected, MR allowed to characterize them and to examine the biliary ducts communication. In 1 case was observed an intraparenchymal pseudoaneurism of the spleen that was subsequently embolized.
Conclusion: MR may constitute a useful alternative to CT to follow-up patients with blunt liver and spleen injuries considering the panoramicity, the high resolution and the lack of ionizing radiation. In this preliminary experience, MR allowed to identify and monitoring, in a less invasive way, vascular and parenchymal liver and spleen injuries.
Purpose: To analysis risk factors to cause repeated imaging in revisiting emergency patients.
Methods and Materials: A retrospective review identified 12,357 patients returning to emergency room (ER) within 7 days after performed CT or MRI at the first visit and discharged from the ER between 2005 and 2013. Among them, 1250 patients underwent repeated image workup for the same or different part of the body. The patients were divided into two groups according to the revisiting time after discharge: early revisiting group (≤ 72 hours) and late revisiting group (> 72 hours). The factors to cause re-performing the imaging are classified as follows through medical record and image review: 1) radiological factor (RF) (misdiagnosis vs. inappropriate reporting turn-around time [R-TAT]); 2) non-radiology clinician’s factor (CF) (mistriage); and 3) patient’s factors (PF) (disease or symptom progression).
Results: The repeated imaging was performed in 10.12% of revisiting patients. The ratios of RF and CF in early revisiting group were significantly higher from those in late revisiting group (13.50% vs. 8.29% and 12.29% vs. 6.76%, respectively) (P < 0.001). In the radiological factors, there was significant difference between early revisiting group (inappropriate R-TAT 64.04%; and misdiagnosis 35.96%) and late revisiting groups (inappropriate R-TAT 30.61%; and misdiagnosis 69.38%) (P < 0.001).
Conclusion: RC and CF are the major causes of repeating imaging in early revisiting group than late revisiting group. In radiological factors, inappropriate R-TAT is the main cause in the early revisiting group and misdiagnosis in the late revisiting group.
Purpose: Whole-body computed tomography reveals beyond findings related to the suspected diagnosis often additional incidental findings. Aim of this investigation was the assessment of these findings in whole-body CT scans of patients admitted via the resuscitation room after suffering potential major trauma or life-threatening medical conditions.
Methods and Materials: Patients admitted via the resuscitation room were retrospectively included if they had received a whole-body computed tomography scan at admission. The final cohort consisted of 1165 patients (1038 trauma and 127 internal-neurological patients). Whole-body computed tomography reports screened for incidental findings. These findings were then classified as either clinically relevant or not.
Results: 465 incidental findings were reported in a total of 293 patients (25.1%). Relevant incidental findings could be detected in 5.8% of the study patients (68/1165). The group of internal-neurological patients was older than the trauma patients (61.6 years vs. 45.5 years). The rate of relevant incidental findings in the internal-neurological group was more than twice as high as in the trauma group (11.0% vs. 5.2%). Yet, in the relatively young trauma group one in 20 patients showed an incidental finding classified as clinically relevant as well.
Conclusion: Incidental findings are reported in ¼ of whole-body CT scans of patients admitted to the resuscitation room. About 6% of all patients had incidental findings being rated as clinically relevant. In the internal-neurological older group of patients the rate of incidental findings was doubled compared to the younger trauma group.
Purpose: To determine the discrepancy rate between the first and second readings. To estimate the clinical significance of errors. To identify latent factors associated with a higher incidence of errors. To discover weak points in the radiological work that can be solved by training or quality control measures.
Methods and Materials: Retrospective observational analysis of diagnostic discrepancies in emergency studies of adult patients from 37 health centres detected after a second reading at the company Iberorad SL from 2014 to 2016.
Results: We found a diagnostic discrepancy rate of 1.4% in emergency studies. The average rate of emergency studies with clinically significant discrepancies is 27.4% of total discrepancies. A statistically significant association was noticed with the diagnostic discrepancy, in decreasing order of magnitude, of the following: availability of previous imaging studies; average level of concordance between area of expertise of the radiologist and modality of the study; insufficient clinical information; non-use of speech recognition systems; low level of concordance between area of expertise of the radiologist and modality of the study; study of high complexity; body studies; use of a single monitor to visualise images; and MSK/spine studies.
Conclusion: The diagnostic discrepancy rate during the period analysed is similar to that reflected in the medical literature. Most discrepancies are not clinically significant. There are multiple factors associated with diagnostic error. The intensity in which each of these is associated with diagnostic error is different and variable determined by the rest of the characteristics present in the diagnostic process.
Purpose: To determine whether it would be hygienic to evaluate dogs and humans on the same MRI scanner.
Methods and Materials: We compared the bacterial load in colony-forming units (CFU) of human pathogenic microorganisms in specimens taken from 18 men and 30 dogs. In addition, we compared the extent of bacterial contamination of an MRI scanner shared by dogs and humans, with two other MRI scanners used exclusively by humans.
Results: Our study shows a significantly higher bacterial load in specimens taken from men’s beards compared with dogs’ furs (p = 0.036). All of the men (18/18) showed high microbial counts, whereas only 23/30 dogs had high microbial counts and 7 dogs moderate microbial counts. Furthermore, human pathogenic microorganisms were more frequently found in human beards (7/18) than in dog fur (4/30), although this difference did not reach statistical significance (p = 0.074). More microbes were found in human oral cavities than in dog oral cavities (p <0.001). After MRI of dogs, routine scanner disinfection was undertaken and the CFU found in specimens isolated from the MRI scanning table and receiver coils showed significantly lower bacteria count compared with “human” MRI scanners (p < 0.05).
Conclusion: Our study shows that bearded men harbour higher burden of microbes and more human pathogenic strains than dogs. As the MRI scanner used for both dogs and humans was routinely cleaned after animal scanning, there was substantially lower bacterial load compared with scanners used exclusively for humans.