Objectives
To assess undergraduate medical students’ attitudes towards artificial intelligence (AI) in radiology and medicine.
Materials and methods
A web-based questionnaire was designed using ...SurveyMonkey, and was sent out to students at three major medical schools. It consisted of various sections aiming to evaluate the students’ prior knowledge of AI in radiology and beyond, as well as their attitude towards AI in radiology specifically and in medicine in general. Respondents’ anonymity was ensured.
Results
A total of 263 students (166 female, 94 male, median age 23 years) responded to the questionnaire. Around 52% were aware of the ongoing discussion about AI in radiology and 68% stated that they were unaware of the technologies involved. Respondents agreed that AI could potentially detect pathologies in radiological examinations (83%) but felt that AI would not be able to establish a definite diagnosis (56%). The majority agreed that AI will revolutionise and improve radiology (77% and 86%), while disagreeing with statements that human radiologists will be replaced (83%). Over two-thirds agreed on the need for AI to be included in medical training (71%). In sub-group analyses male and tech-savvy respondents were more confident on the benefits of AI and less fearful of these technologies.
Conclusion
Contrary to anecdotes published in the media, undergraduate medical students do not worry that AI will replace human radiologists, and are aware of the potential applications and implications of AI on radiology and medicine. Radiology should take the lead in educating students about these emerging technologies.
Key Points
• Medical students are aware of the potential applications and implications of AI in radiology and medicine in general.
• Medical students do not worry that the human radiologist or physician will be replaced.
• Artificial intelligence should be included in medical training.
Cholinergic neurons of the medial forebrain are considered important contributors to brain plasticity and neuromodulation. A reduction of cholinergic innervation can lead to pathophysiological ...changes of neurotransmission and is observed in Alzheimer's disease. Here we report on six patients with mild to moderate Alzheimer's disease (AD) treated with bilateral low-frequency deep brain stimulation (DBS) of the nucleus basalis of Meynert (NBM). During a four-week double-blind sham-controlled phase and a subsequent 11-month follow-up open label period, clinical outcome was assessed by neuropsychological examination using the Alzheimer's Disease Assessment Scale-cognitive subscale as the primary outcome measure. Electroencephalography and (18)F-fluoro-desoxyglucose positron emission tomography were, besides others, secondary endpoints. On the basis of stable or improved primary outcome parameters twelve months after surgery, four of the six patients were considered responders. No severe or non-transitional side effects related to the stimulation were observed. Taking into account all limitations of a pilot study, we conclude that DBS of the NBM is both technically feasible and well tolerated.
Purpose
To investigate a benefit from virtual monoenergetic reconstructions (VMIs) for assessment of arterially hyper-enhancing liver lesions in phantom and patients and to compare hybrid-iterative ...and spectral image reconstructions of conventional images (CI-IR and CI-SR).
Methods
All imaging was performed on a SDCT (Philips Healthcare, Best, The Netherlands). Images of a non-anthropomorphic phantom with a lesion-mimicking insert (containing iodine in water solution) and arterial-phase images from contrast-enhanced patient examinations were evaluated. VMIs (40–200 keV, 10 keV increment), CI-IR, and CI-SR were reconstructed using different strengths of image denoising. ROIs were placed in lesions, liver/matrix, muscle; signal-to-noise, contrast-to-noise, and lesion-to-liver ratios (SNR, CNR, and LLR) were calculated. Qualitatively, 40, 70, and 110 keV and CI images were assessed by two radiologists on five-point Likert scales regarding overall image quality, lesion assessment, and noise.
Results
In phantoms, SNR was increased threefold by VMI
40keV
compared with CI-IR/SR (5.8 ± 1.1 vs. 18.8 ± 2.2,
p
≤ 0.001), while no difference was found between CI-IR and CI-SR (
p
= 1). Denoising was capable of noise reduction by 40%. In total, 20 patients exhibiting 51 liver lesions were assessed. Attenuation was the highest in VMI
40keV
, while image noise was comparable to CI-IR resulting in a threefold increase of CNR/LLR (CI-IR 1.3 ± 0.8/4.4 ± 2.0, VMI
40keV
: 3.8 ± 2.7/14.2 ± 7.5,
p
≤ 0.001). Subjective lesion delineation was the best in VMI
40keV
image (
p
≤ 0.01), which also provided the lowest perceptible noise and the best overall image quality.
Conclusions
VMIs improve assessment of arterially hyper-enhancing liver lesions since they increase lesion contrast while maintaining low image noise throughout the entire keV spectrum. These data suggest that to consider VMI screening after arterially hyper-enhancing liver lesions.
Objectives
To investigate whether the increased soft tissue contrast of virtual monoenergetic images (VMIs) obtained from a spectral detector computed tomography (SDCT) system improves washout ...assessment of arterially hyper-enhancing liver lesions.
Methods
Fifty-nine arterially hyper-enhancing lesions in 31 patients (age 65 ± 9 years, M/W 20/11) were included in this IRB-approved study. All patients underwent multi-phase SDCT for HCC screening. MRI, CEUS or biopsy within 3 months served as standard of reference to classify lesions as LiRADS 3 or 4/5. VMIs and conventional images (CIs) were reconstructed. Visual analysis was performed on 40, 60, and 80 kiloelectronvolt (keV) and CIs by 3 radiologists. Presence and visibility of washout were assessed; image quality and confidence of washout evaluation were evaluated on 5-point Likert scales. Signal-to-noise ratio (SNR), lesion-to-liver contrast-to-noise ratio (CNR) (|HU
lesion
–HU
liver
|/SD
liver
) and washout (|HU
lesion
–HU
liver
|) were calculated. Statistical assessment was performed using ANOVA and Wilcoxon test.
Results
On subjective lesion analysis, the highest level of diagnostic confidence and highest sensitivity for the detection of lesion washout were found for 40-keV VMIs (40 keV vs. CI, 81.3 vs. 71.3%). Image quality parameters were significantly better in low-kiloelectronvolt VMIs than in CIs (
p
< 0.05; e.g. SNR
liver
: 40 keV vs. CIs, 12.5 ± 4.1 vs. 5.6 ± 1.6). In LiRADS 4/5 lesions, CNR and quantitative washout values were significantly higher in 40-keV VMIs compared to CIs (
p
< 0.05; e.g. CNR and washout in 40 keV vs. CIs, 2.3 ± 1.6 vs. 0.8 ± 0.5 and 29.0 ± 19.1 vs. 12.9 ± 6.9 HU, respectively).
Conclusion
By increasing lesion contrast, low-kiloelectronvolt VMIs obtained from SDCT improve washout assessment of hyper-enhancing liver lesions with respect to washout visibility and diagnostic confidence.
Key Points
• Low-kiloelectronvolt virtual monoenergetic images from spectral detector CT facilitate washout assessment in arterially hyper-enhancing liver lesions.
• Image quality and quantitative washout parameters as well as subjective washout visibility and diagnostic confidence benefit from low-kiloelectronvolt virtual monoenergetic images.
Objective
To evaluate feasibility and diagnostic performance of multi-level calcium suppression in spectral detector computed tomography (SDCT) for assessment of bone metastasis.
Materials and ...methods
Retrospective IRB-approved study on 21 patients who underwent SDCT (120 kV, reference mAs 116) and MRI. Thoracic and lumbar vertebrae (
n
= 357) were included and categorized as normal (
n
= 133) or metastatic (
n
= 203) based on MRI (STIR, T1w, ±contrast). The multi-level virtual non-calcium (VNCa) algorithm computes dynamic soft tissue/calcium pairs allowing for computation of different suppression index levels to address inter-individual variance of prevalent calcium composition weights. We computed images with low, medium, and high calcium suppression indices and compared them with conventional images (VNCa_low/med/high and conventional images (CI)). For quantitative image analysis, regions of interest were placed in normal and metastatic bone. Two readers reviewed the datasets independently in multiple sessions. They determined the presence of vertebral metastases on a per vertebra basis using a binary scale. Statistic assessment was performed using ANOVA with Tukey HSD, Student’s
T
test, and ROC analysis.
Results
Attenuation of both normal and metastatic bone was lower in VNCa images than that in conventional images (e.g., CI/VNCa_low, − 46.3 to 238.8 HU/343.3–60.2 HU;
p
≤ 0.05). VNCa_low+med improved separation of normal and metastatic bone in ROC analysis (AUC, CI/VNCa_low/VNCa_med = 0.74/0.95/0.98;
p
≤ 0.05). In subjective analysis, both sensitivity and specificity were clearly improved in VNCa_low as compared with CI (0.85/0.84 versus 0.78/0.82). Readers showed a good inter-rater reliability (kappa = 0.65).
Conclusions
Multi-level VNCa reconstructed from SDCT improve quantitative separation of normal and metastatic bone and subjective determination of bone metastases when using low to intermediate calcium suppression indices.
Key Points
•
Spectral detector CT allows for multi-level calcium suppression in CT images and low and medium calcium suppression indices improved separation of normal and metastatic bone.
•
Thus, multi-level calcium suppression allows to optimize image contrast in regard to dedicated pathologies
.
•
Low-level virtual non-calcium images (index 25–50) improved diagnostic performance regarding detection of metastasis
.
Objectives
The blood of patients with anemia demonstrates distinctly lower attenuation in unenhanced CT images. However, the frequent usage of intravenous contrast hampers evaluation of anemia. ...Spectral detector computed tomography (SDCT) allows for reconstruction of virtual non-contrast images (VNC) from contrast-enhanced data (CE). The purpose of this study was to evaluate whether VNC allow for prediction of anemia.
Methods
Five hundred twenty-two patients with CE-SDCT of the chest and accessible serum hemoglobin (HbS) were retrospectively included. Patients were assigned to three groups (severe anemia, moderate/mild anemia, and healthy) based on recent lab tests (≤ 7 days) for HbS following gender and the WHO definition of anemia. CT attenuation was determined using two ROI in the left ventricular lumen and one ROI in the descending thoracic aorta. ROI were placed on CE and copied to VNC. ANOVA, linear regression, and receiver operating characteristics were used for statistic evaluation.
Results
Average HbS was 11.6 ± 2.4 g/dl. Attenuation on VNC showed significant differences between healthy patients, patients with mild/moderate anemia, and severely anemic patients (all
p
≤ 0.05). Applying cutoffs of 39.2/37.6 HU and 33.6/32.7 HU allowed to differentiate between healthy, mild/moderately, and severely anemic men/women (AUC 0.857/0.833 and 0.879/0.932). A linear relationship between HbS and attenuation on VNC was established (
r
2
= 0.54, HbS = − 0.875 + 0.329 × HU).
Conclusions
An approximation of HbS and presence of anemia can be conducted based on simple attenuation measurements in contrast-enhanced SDCT examinations enabled by VNC imaging.
Key Points
• While the attenuation of blood is a previously described biomarker for anemia in non-contrast images, virtual non-contrast images from spectral detector CT circumvent this limitation and allow for diagnosis of anemia in contrast-enhanced scans.
• Attenuation of blood in virtual non-contrast images derived from spectral detector CT shows a moderate correlation to serum hemoglobin levels.
• Presence of anemia be estimated in virtual non-contrast images using proposed cutoffs of 39.2 HU and 37.6 HU for men and women, respectively, to differentiate between healthy and anemic patients.
Image quality in head and neck imaging is often severely hampered by artifacts arising from dental implants. This study evaluates metal artifact (MA) reduction using virtual monoenergetic images ...(VMI) compared to conventional CT images (CI) from spectral-detector computed tomography (SDCT).
38 consecutive patients with dental implants were included in this retrospective study. All examinations were performed using a SDCT (IQon, Philips, Best, The Netherlands). Images were reconstructed as conventional images (CI) and as VMI in a range of 40–200 keV (10 keV increment). Quantitative image analysis was performed ROI-based by measurement of attenuation (HU) and standard deviation in most pronounced hypo- and hyperdense artifact, fat and soft tissue with presence of artifacts. Qualitatively, extent of artifact reduction, assessment of soft palate and cheeks were rated on 5-point Likert-scales by two radiologists. Statistical data evaluation included ANOVA and Wilcoxon-test with correction for multiple comparisons; interrater-agreement was determined by intraclass-correlation coefficient (ICC).
The hypo- and hyperattenuating artifacts showed an increase and decrease of HU-values in VMIhigh (CI/VMI200 keV: −218.7/−174.4 HU, p = 0.1; and 309.8/119.2, p ≤ 0.05, respectively). Artifacts in the fat, as depicted by image noise did also decrease in VMIhigh (CI/VMI200 keV: 23.9/16.4, p ≤ 0.05). Qualitatively, hyperdense artifacts were decreased significantly in VMI ≥100 keV (e.g. CI/VMI200 keV: 2(1–3)/3(1–5), p ≤ 0.05). Artifact reduction resulted in improved assessment of the soft palate and cheeks (e.g. CI/VMI200 keV: 2(1–4)/3(1–5) and 2(1–5)/3(1–5), p ≤ 0.05). Overall interrater agreement was good (ICC = 0.77).
Virtual monoenergetic images from SDCT reduce metal artifacts from dental implants and improve diagnostic assessment of surrounding soft tissue.
•VNC reconstructions of cCTA allow assessment of CACS without an additional unenhanced CT scan for calcium scoring.•This algorithm enables the application of VNC CACS in the clinical routine with a ...significant radiation dose reduction.•A correction factor is required to achieve a good agreement with the TNC standard technique.
To evaluate the clinical applicability of a prototype virtual non-contrast (VNC) reconstruction algorithm based on coronary CT angiography (cCTA) to assess calcified coronary plaques by calcium scoring (CACS).
Eighty consecutive patients suspected of coronary artery disease were retrospectively included. All patients underwent a cardiac CT using a dual-layer spectral-detector CT system. The standardized acquisition protocol included unenhanced CACS and cCTA. Datasets were acquired using 120 keV. VNC-reconstructions were calculated from the cCTA images at 2.5 mm (VNC group 1), 2.5 of 0.9 mm (group 2), and 0.9 mm (group 3) slice thickness. We compared the Agatston score and Coronary Artery Calcium Data and Reporting System (CAC-DRS) of all VNC reconstructions with the true non-contrast (TNC)-dataset as the gold standard.
In total, 73 patients were evaluated. Fifty patients (68.5 %) had a CACS > 0 based on TNC. We found a significant difference in the Agatston score comparing all VNC-reconstructions (1: 1.35, 2: 3.7, 3: 10.4) with the TNC dataset (3.8) (p < 0.001). Correlation analysis of the datasets showed an excellent correlation of the TNC results with the different VNC-reconstructions (r = 0.904–0.974, p < 0.001) with a slope of 1.89–2.53. Mean differences and limits of agreement by Bland–Altman analysis between TNC and group 1 were 83 and −196 to 362, respectively. By using the VNC-reconstructions, in group 1 23 patients (31.5 %), in group 2 10 (13.7 %), and in group 3 23 (31.5 %) were reclassified according to CAC-DRS compared to TNC. Classification according to CAC-DRS revealed a significant difference between TNC and group 1 (p = 0.024) and no significance compared to groups 2 and 3 (p = 0.670 and 0.273).
The investigated VNC reconstruction algorithm of routine cCTA allows the detection and evaluation of coronary calcium burden without the requirement for an additional acquisition of an unenhanced CT scan for CACS and, therefore, a reduction of radiation exposure.
Objectives
To systematically investigate the usability of virtual non-contrast reconstructions (VNC) derived from dual-layer CT (DLCT) for detection and size measurements of kidney stones with ...regards to different degrees of surrounding iodine-induced attenuation and radiation dose.
Methods
Ninety-two kidney stones of varying size (3–14 mm) and composition were placed in a phantom filled with different contrast media/water mixtures exhibiting specific iodine-induced attenuation (0–1500 HU). DLCT-scans were acquired using CTDI
vol
of 2 mGy and 10 mGy. Conventional images (CI) and VNC
0H-1500HU
were reconstructed. Reference stone size was determined using a digital caliper (Man-M). Visibility and stone size were assessed. Statistical analysis was performed using the McNemar test, Wilcoxon test, and the coefficient of determination.
Results
All stones were visible on CI
0HU
and VNC
200HU
. Starting at VNC
400 HU
, the detection rate decreased with increasing HU and was significantly lower as compared to CI
0HU
on VNC
≥ 600HU
(100.0 vs. 94.0%,
p
< 0.05). The overall detection rate was higher using 10 mGy as compared to 2 mGy protocol (87.9 vs. 81.8%;
p
< 0.001). Stone size was significantly overestimated on all VNC compared to Man-M (7.0 ± 3.5 vs. 6.6 ± 2.8 mm,
p
< 0.001). Again, the 10 mGy protocol tended to show a better correlation with Man-M as compared to 2 mGy protocol (
R
2
= 0.39–0.68 vs.
R
2
= 0.31–0.57).
Conclusions
Detection and size measurements of kidney stones surrounded by contrast media on VNC are feasible. The detection rate of kidney stones decreases with increasing iodine-induced attenuation and with decreasing radiation dose as well as stone size, while remaining comparable to CI
0HU
on VNC
≤ 400 HU
.
Key Points
• The detection rate of kidney stones on VNC depends on the surrounding iodine-induced attenuation, the used radiation dose, and the stone size.
• The detection rate of kidney stones on VNC decreases with greater iodine-induced attenuation and with lower radiation dose, particularly in small stones.
• The visibility of kidney stones on VNC
≤
400 HU
remains comparable to true-non-contrast scans even when using a low-dose technique.
To compare the quality of virtual mono-energetic (VMI) and polychromatic images reconstructed with hybrid iterative (PCIHIR) or model-based reconstruction (PCIMBR) derived from dual-layer spectral ...detector computed tomography (SDCT) in arterial phase images to visualise the aorta and abdominal main branches.
A retrospective review of 50 patients with abdominal arterial phase scans was undertaken. Attenuation, intraluminal noise, and signal-/contrast-to-noise ratio (S-/CNR) were assessed in the PCIHIR, PCIMBR and VMI40keV, VMI70keV, and VMI100keV images. Contrast, noise, and visualization of soft-plaque, and macro-/micro-calcifications were scored in a blinded reading by two radiologists.
VMI40keV yielded highest S-/CNR (p≤0.001). VMI70keV and PCIMBR showed comparable SNR (p≥0.999) and yielded higher SNR than PCIHIR. VMI70keV yielded higher CNR than PCIHIR (p<0.001) and PCIMBR (p<0.045). VMI100keV yielded lowest CNR (p≤0.001) and SNR (p≥0.104). In the subjective analysis, VMI40keV outperformed PCIMBR for contrast and noise, PCIMBR scored better than VMI70keV, and the latter scored better than PCIHIR for these categories (all p<0.001). PCIMBR was superior for depiction of soft-plaque and micro-calcifications (p<0.001). VMI100keV visualized micro-calcifications second best (p<0.001) and matched PCIMBR for the depiction of macro-calcifications (p>0.999), while VMI40keV scored second best for depiction of soft-plaque (p<0.020).
VMI40keV and VMI70keV yield better S-/CNR than PCIHIR and PCIMBR; however, PCIMBR visualized arteriosclerotic plaques best, followed by VMI40keV for depiction of soft-plaque and VMI100keV for macro- and micro-calcification. Based on the present findings, PCIMBR on conventional CT and VMI40keV supplemented by VMI100keV on SDCT are recommended for the diagnostic assessment of abdominal arteries.
•Model-based iterative reconstructions visualize arteriosclerosis best.•Virtual mono-energetic images at 40 keV show soft-plaque second best.•Virtual mono-energetic images at 100 keV show calcification second best.•VMI at 40 and 70 keV improve objective image quality over polychromatic images.