Machine learning and deep learning for classifying the justification of brain CT referrals.

OBJECTIVES: To train the machine and deep learning models to automate the justification analysis of radiology referrals in accordance with iGuide categorisation, and to determine if prediction models can generalise across multiple clinical sites and outperform human experts. METHODS: Adult brain computed tomography (CT) referrals from scans performed in three CT centres in Ireland in 2020 and 2021 were retrospectively collected. Two radiographers analysed the justification of 3000 randomly selected referrals using iGuide, with two consultant radiologists analysing the referrals with disagreement. Insufficient or duplicate referrals were discarded. The inter-rater agreement among radiographers and consultants was computed. A random split (4:1) was performed to apply machine learning (ML) and deep learning (DL) techniques to unstructured clinical indications to automate retrospective justification auditing with multi-class classification. The accuracy and macro-averaged F1 score of the best-performing classifier of each type on the training set were computed on the test set. RESULTS: 42 referrals were ignored. 1909 (64.5%) referrals were justified, 811 (27.4%) were potentially justified, and 238 (8.1%) were unjustified. The agreement between radiographers (κ = 0.268) was lower than radiologists (κ = 0.460). The best-performing ML model was the bag-of-words-based gradient-boosting classifier achieving a 94.4% accuracy and a macro F1 of 0.94. DL models were inferior, with bi-directional long short-term memory achieving 92.3% accuracy, a macro F1 of 0.92, and outperforming multilayer perceptrons. CONCLUSION: Interpreting unstructured clinical indications is challenging necessitating clinical decision support. ML and DL can generalise across multiple clinical sites, outperform human experts, and be used as an artificial intelligence-based iGuide interpreter when retrospectively vetting radiology referrals. CLINICAL RELEVANCE STATEMENT: Healthcare vendors and clinical sites should consider developing and utilising artificial intelligence-enabled systems for justifying medical exposures. This would enable better implementation of imaging referral guidelines in clinical practices and reduce population dose burden, CT waiting lists, and wasteful use of resources. KEY POINTS: Significant variations exist among human experts in interpreting unstructured clinical indications/patient presentations. Machine and deep learning can automate the justification analysis of radiology referrals according to iGuide categorisation. Machine and deep learning can improve retrospective and prospective justification auditing for better implementation of imaging referral guidelines.

Exploring the translational challenge for medical applications of ionising radiation and corresponding radiation protection research.

BACKGROUND: Medical applications of ionising radiation and associated radiation protection research often encounter long delays and inconsistent implementation when translated into clinical practice. A coordinated effort is needed to analyse the research needs for innovation transfer in radiation-based high-quality healthcare across Europe which can inform the development of an innovation transfer framework tailored for equitable implementation of radiation research at scale. METHODS: Between March and September 2021 a Delphi methodology was employed to gain consensus on key translational challenges from a range of professional stakeholders. A total of three Delphi rounds were conducted using a series of electronic surveys comprised of open-ended and closed-type questions. The surveys were disseminated via the EURAMED Rocc-n-Roll consortium network and prominent medical societies in the field. Approximately 350 professionals were invited to participate. Participants' level of agreement with each generated statement was captured using a 6-point Likert scale. Consensus was defined as median ≥ 4 with ≥ 60% of responses in the upper tertile of the scale. Additionally, the stability of responses across rounds was assessed. RESULTS: In the first Delphi round a multidisciplinary panel of 20 generated 127 unique statements. The second and third Delphi rounds recruited a broader sample of 130 individuals to rate the extent to which they agreed with each statement as a key translational challenge. A total of 60 consensus statements resulted from the iterative Delphi process of which 55 demonstrated good stability. Ten statements were identified as high priority challenges with ≥ 80% of statement ratings either 'Agree' or 'Strongly Agree'. CONCLUSION: A lack of interoperability between systems, insufficient resources, unsatisfactory education and training, and the need for greater public awareness surrounding the benefits, risks, and applications of ionising radiation were identified as principal translational challenges. These findings will help to inform a tailored innovation transfer framework for medical radiation research.

Potential for Dose Reduction in CT-Derived Left Ventricular Ejection Fraction: A Simulation Study.

BACKGROUND: Measuring left ventricular ejection fraction (LVEF) is important for detecting heart failure, e.g., in treatment with potentially cardiotoxic chemotherapy. MRI is considered the reference standard for LVEF, but availability may be limited and claustrophobia or metal implants still present challenges. CT has been shown to be accurate and would be advantageous, as LVEF could be measured in conjunction with routine chest-abdomen-pelvis oncology CT. However, the use of CT is not recommended due to the excessive radiation dose. This study aimed to explore the potential for dose reduction using simulation. Using an anthropomorphic heart phantom scanned at 13 dose levels, a noise simulation algorithm was developed to introduce controlled Poisson noise. Filtered backprojection parameters were iteratively tested to minimise differences in myocardium-to-ventricle contrast/noise ratio, as well as structural similarity index (SSIM) differences between real and simulated images at all dose levels. Fifty-one clinical CT coronary angiographies, scanned with full dose through end-systolic and -diastolic phases, were located retrospectively. Using the developed algorithm, noise was introduced corresponding to 25, 10, 5 and 2% of the original dose level. LVEF was measured using clinical software (Syngo.via VB50) with papillary muscles in and excluded from the LV volume. At each dose level, LVEF was compared to the 100% dose level, using Bland-Altman analysis. The effective dose was calculated from DLP using a conversion factor of 0.026 mSv/mGycm. RESULTS: In the clinical images, mean CTDIvol and DLP were 47.1 mGy and 771.9 mGycm, respectively (effective dose 20.0 mSv). Measurements with papillary muscles excluded did not exhibit statistically significant LVEF bias to full-dose images at 25, 10 and 5% simulated dose. At 2% dose, a significant bias of 4.4% was found. With papillary muscles included, small but significant biases were found at all simulated dose levels. CONCLUSION: Provided that measurements are performed with papillary muscles excluded from the LV volume, the dose can be reduced by a factor of 20 without significantly affecting LVEF measurements. This corresponds to an effective dose of 1 mSv. CT can potentially be used for LVEF measurement with minimal excessive radiation.

COMBINING HI-RESOLUTION SCAN MODE WITH DEEP LEARNING RECONSTRUCTION ALGORITHMS IN CARDIAC CT.

To investigate the impact of combining the high-resolution (Hi-res) scan mode with deep learning image reconstruction (DLIR) algorithm in CT. Two phantoms (Catphan600® and Lungman, small, medium, large size) were CT scanned using combinations of Hi-res/standard mode and high-definition (HD)/standard kernels. Images were reconstructed with ASiR-V and three levels of DLIR. Spatial resolution, noise and contrast-to-noise ratio (CNR) were assessed. The radiation dose was recorded. The spatial resolution increased using Hi-res & HD. Image noise in the Catphan600® (69%) and the Lungman (10-70%) significantly increased when Hi-res & HD was applied. DLIR reduced the mean noise (54%). The CNR was reduced (64%) for Hi-res & HD. The radiation dose increased for both small (+70%) and medium (+43%) Lungman phantoms but decreased slightly for the large ones (-3%) when Hi-res was applied. In conclusion, the Hi-res scan mode improved the spatial resolution. The HD kernel significantly increased the image noise. DLIR improved the image noise and CNR and did not affect the spatial resolution.

Justification of CT practices across Europe: results of a survey of national competent authorities and radiology societies.

OBJECTIVES: Published literature on justification of computed tomography (CT) examinations in Europe is sparse but demonstrates consistent sub-optimal application. As part of the EU initiated CT justification project, this work set out to capture CT justification practices across Europe. METHODS: An electronic questionnaire consisting of mostly closed multiple-choice questions was distributed to national competent authorities and to presidents of European radiology societies in EU member states as well as Iceland, Norway, Switzerland, and the UK (n = 31). RESULTS: Fifty-one results were received from 30 European countries. Just 47% (n = 24) stated that advance justification of individual CT examinations is performed by a medical practitioner. Radiologists alone mostly (n = 27, 53%) perform daily justification of CT referrals although this is a shared responsibility in many countries. Imaging referral guidelines are widely available although just 13% (n = 6) consider them in daily use. Four countries (Cyprus, Ireland, Sweden, UK) reported having them embedded within clinical decision support systems. Justification of new practices with CT is mostly regulated (77%) although three countries (Belgium, Iceland and Portugal) reported not having any national system in place for generic justification. Health screening with CT was reported by seven countries as part of approved screening programmes and by eight countries outside. When performed, CT justification audits were reported to improve CT justification rates. CONCLUSIONS: CT justification practices vary across Europe with less than 50% using advance justification and a minority having clinical decision support systems in place. CT for health screening purposes is not currently widely used in Europe.

Breast surface radiation dose during coronary CT angiography: reduction by breast displacement and lead shielding.

OBJECTIVE: The purpose of this study was to prospectively evaluate the effect of cranial breast displacement and lead shielding on in vivo breast surface radiation dose in women undergoing coronary CT angiography. SUBJECTS AND METHODS: Fifty-four women (mean age, 59.2 ± 9.8 years) prospectively underwent coronary 64-MDCT angiography for evaluation of chest pain. The patients were randomly assigned to a control group (n = 16), breast displacement group (n = 22), or breast displacement plus lead shielding group (n = 16). Thermoluminescent dosimeters (TLDs) were placed superficially on each breast quadrant and the areolar region of both breasts. Breast surface radiation doses, the degree of breast displacement, and coronary image quality were compared between groups. A phantom dose study was conducted to compare breast doses with z-axis positioning on the chest wall. RESULTS: A total of 1620 TLD dose measurements were recorded. Compared with control values, the mean breast surface dose was reduced 23% in the breast displacement group (24.3 vs 18.6 mGy, p = 0.015) and 36% in the displacement plus lead shielding group (24.3 vs 15.6 mGy, p = 0.0001). Surface dose reductions were greatest in the upper outer (displacement alone, 66%; displacement plus shielding, 63%), upper inner (65%, 58%), and areolar quadrants (44%, 53%). The smallest surface dose reductions were recorded for A-cup breasts: 7% for the displacement group and 3% for the displacement plus lead group (p = 0.741). Larger reductions in surface dose were recorded for B-cup (25% and 56%, p = 0.273), C-cup (38% and 60%, p = 0.001), and D-cup (31% and 25%, p = 0.095) sizes. Most of the patients (79%) had either good (< 50% of breast above scan range) or excellent (< 75% of breast above the scan range) breast displacement. No significant difference in coronary image quality was detected between groups. The phantom dose study showed that surface TLD measurements were underestimates of absorbed tissue dose by a mean of 9% and that a strong negative correlation exists between the amount of cranial displacement and breast dose. CONCLUSION: Use of breast displacement during coronary CTA substantially reduces the radiation dose to the breast surface.

Early experiences of radiographers in Ireland during the COVID-19 crisis.

BACKGROUND: Imaging is crucial for assessing the severity and progression of COVID-19. Radiographers are amongst the first-line health professionals that may be exposed to infected persons. This work describes the early experience of radiographers in Ireland to the impact of COVID-19 using two electronic surveys distributed 6 weeks apart. Results were analysed using descriptive statistics and thematic analysis. RESULTS: A total of 370 responded to the first survey and 276 the second, with all six Irish health regions represented. Three quarters of radiographers (77%) reported having adequate personal protective equipment (PPE) available to them. However, almost half of the radiographers were inadvertently exposed to COVID-19-positive patients without appropriate PPE, largely attributed to poor communication and testing. Anxiety levels while initially high, reduced substantially 6 weeks into the crisis period. However, obvious distress was noted amongst some respondents. Forty percent of radiographers reported burnout symptoms due to the COVID-19 crisis and 30% reported considering changing jobs or retiring since the COVID-19 outbreak. CONCLUSION: Clear communication regarding changing protocols and importantly patients' infectious status are essential to safeguard healthcare workers and to minimise unnecessary anxiety and distress. Attention is required to staff mental health including the identification of burnout symptoms to prevent long-term negative consequences of the pandemic on radiography services.

THE IMPACT OF OBESITY ON ABDOMINAL CT RADIATION DOSE AND IMAGE QUALITY.

The aim of this study was to evaluate how iterative reconstruction can compensate for the noise increase in low radiation dose abdominal computed tomography (CT) technique for large size patients and the general impact of obesity on abdominal organ doses and image quality in CT. An anthropomorphic phantom layered with either none or a single layer of 3-cm- thick circumferential animal fat packs to simulate obese patients was imaged using a 128MDCT scanner. Abdominal protocols (n = 12) were applied using automatic tube current modulation (ATCM) with various quality reference mAs (150, 200, 250 and 300). kVs of 100, 120 and 140 were used for each mAs selection. Metal oxide semiconductor field effect transistor dosimeters (MOSFET) measured internal organ dose. All images produced were reconstructed with filtered back projection (FBP) and sinogram affirmed iterative reconstruction (SAFIRE) (3, 4 and 5) and objective noise was measured within three regions of interest at the level of L4-L5. Organ doses varied from 0.12 to 41.9 mGy, the spleen received the highest doses for both phantom sizes. Compared to the phantom simulating average size, the obese phantom was associated with up to twofold increase in delivered mAs, dose length product (DLP) and computed tomography dose index (CTDIvol) for the matched mAs selection (p < 0.05). However, organ dose increased by 50% only. The use of 100 kV resulted in a 40% lower dose (p < 0.05) compared to 120 kV and the associated noise increase was improved by SAFIRE (5) use, which resulted in 60% noise reduction compared to FBP (p < 0.05). When combined with iterative reconstruction, low kV is feasible for obese patients to optimise radiation dose and maintain objective image quality.

Carotid atherosclerotic plaques standardised uptake values: software challenges and reproducibility.

BACKGROUND: Positron emission tomography-computed tomography (PET-CT) carotid standardised uptake values (SUV) of 18F-fluorodeoxyglucose (18FDG) have been proposed as an inflammatory biomarker for determining cerebrovascular diseases such as stroke. Consideration of varying methodological approaches and software packages is critical to the calculation of accurate SUVs in cross-sectional and longitudinal patient studies. The aim of this study was to investigate whether or not carotid atherosclerotic plaque SUVs are consistent and reproducible between software packages. 18FDG-PET SUVs of carotids were taken in 101 patients using two different software packages. Quality assurance checks were performed to standardise techniques before commencing the analysis where data from five to seven anatomical sites were measured. A total of ten regions of interest were drawn on each site analysed. Statistical analyses were then performed to compare SUV measurements from the two software packages and to explore reproducibility of measurements. Lastly, the time taken to complete each analysis was measured and compared. RESULTS: Statistically significant differences in SUV measurements, between the two software packages, ranging from 9 to 21.8% were found depending on ROI location. In 79% (n = 23) of the ROI locations, the differences between the SUV measurements from each software package were found to be statistically significant. The time taken to perform the analyses and export data from the software packages also varied considerably. CONCLUSIONS: This study highlights the importance of standardising all aspects of methodological approaches to ensure accuracy and reproducibility. Physicians must be aware that when a PET-CT data set is analysed, subsequent follow-ups must be verified, if possible, with the same software package or cross-calibration between packages should be performed.

Education, training, and professional issues of radiographers in six European countries: a comparative review.

Radiographers constitute an important part of a multidisciplinary radiation-based imaging and therapy chain. However, is there a common framework for assuring high education, training, and subsequent practice of profession among European countries? A study was conducted, based on a questionnaire that consisted of three parts, concerning education and training (Part A), national registry (Part B), and professional issues (Part C). Analysis of the collected data suggested that a common policy is generally followed in the countries investigated; however, differences were not negligible. A common framework of educational programmes among European countries could form the basis for overall standardisation at national and international level.

Establishment of diagnostic reference levels for CT trunk examinations in the western region of Saudi Arabia.

Diagnostic reference levels (DRLs) are an important optimisation tool, which aid in identifying abnormally high dose levels. These are currently not available in Saudi Arabia, and this research aims to remedy this. CT dose data (DLP and CTDIvol) were collected for a minimum number of 10 adult patients of average size (60-80 kg) presenting for a range of CT examinations from public hospitals in the western region of Saudi Arabia. These include routine chest, high-resolution chest (HRCT), pulmonary angiography (CTPA), abdomen and pelvis (AP) and the combined chest, abdomen and pelvis (CAP) CT examinations. Mean values for each site were calculated, and the 75th percentile of DLP and CTDIvol was used as a basis for DRLs. Data for 550 patients were collected from 14 hospitals over a 7-month period. The rounded third-quartile CTDIvol and DLP were 18 mGy and 630 mGy cm(-1) for chest CT, 20 mGy and 600mGy cm(-1) for HRCT, 18 mGy and 480 mGy cm(-1) for CTPA, 15 mGy and 800 mGy cm(-1) for AP, and 16 mGy and 1040 mGy cm(-1) for CAP, respectively. Regional DRLs have been proposed from this study. Dose variations across CT departments have identified an urgent need for optimisation to improve distribution of observed doses for CT examinations.