Automated Detection of Pancreatic Cystic Lesions on CT Using Deep Learning.

Pancreatic cystic lesions (PCL) are a frequent and underreported incidental finding on CT scans and can transform into neoplasms with devastating consequences. We developed and evaluated an algorithm based on a two-step nnU-Net architecture for automated detection of PCL on CTs. A total of 543 cysts on 221 abdominal CTs were manually segmented in 3D by a radiology resident in consensus with a board-certified radiologist specialized in abdominal radiology. This information was used to train a two-step nnU-Net for detection with the performance assessed depending on lesions' volume and location in comparison to three human readers of varying experience. Mean sensitivity was 78.8 ± 0.1%. The sensitivity was highest for large lesions with 87.8% for cysts ≥220 mm3 and for lesions in the distal pancreas with up to 96.2%. The number of false-positive detections for cysts ≥220 mm3 was 0.1 per case. The algorithm's performance was comparable to human readers. To conclude, automated detection of PCL on CTs is feasible. The proposed model could serve radiologists as a second reading tool. All imaging data and code used in this study are freely available online.

The potential toxic impact of different gadolinium-based contrast agents combined with 7-T MRI on isolated human lymphocytes.

BACKGROUND: To investigate a potentially amplifying genotoxic or cytotoxic effect of different gadolinium-based contrast agents (GBCAs) in combination with ultra-high-field 7-T magnetic resonance imaging (MRI) exposure in separated human peripheral blood lymphocytes. METHODS: This in vitro study was approved by the local ethics committee and written informed consent was obtained from all participants. Isolated lymphocytes from twelve healthy donors were incubated with gadobutrol, gadoterate meglumine, gadodiamide, gadopentetate dimeglumine, or gadoxetate either alone or combined with 7-T MRI (1 h). Deoxyribonucleic acid (DNA) double-strand breaks were assessed 15 min after MRI exposure by automated γH2AX foci quantification. Cytotoxicity was determined at later endpoints by Annexin V/propidium iodide apoptosis assay (24 h) and [3H]-thymidine proliferation test (72 h). As a reference, lymphocytes from four different donors were exposed analogously to iodinated contrast agents (iomeprol, iopromide) in combination with computed tomography. RESULTS: Baseline γH2AX levels (0.08 ± 0.02 foci/cell) were not significantly (p between 0.135 and 1.000) enhanced after administration of GBCAs regardless of MRI exposure. In contrast to the two investigated macrocyclic GBCAs, lymphocytes exposed to the three linear GBCAs showed a dose-dependent increase in apoptosis (maximum 186% of unexposed control, p < 0.001) and reduced proliferation rate (minimum 0.7% of unexposed control, p < 0.001). However, additional 7-T MRI co-exposure did not alter GBCA-induced cytotoxicity. CONCLUSIONS: Exposure of lymphocytes to different GBCAs did not reveal significant induction of γH2AX foci, and enhanced cytotoxicity was only observed in lymphocytes treated with the linear GBCAs used in this study, independent of additional 7-T MRI co-exposure.

Transcatheter arterial embolization of acute bleeding as 24/7 service: predictors of outcome and mortality.

OBJECTIVE: To analyze times of occurrence and identify risk factors (RFs) for technical and clinical failure and mortality of transcatheter arterial embolization (TAE) of acute bleeding in a major hospital. METHODS: All TAEs performed at our hospital from 2006 to 2013 (n = 327) were retrospectively analyzed. RESULTS: TAEs were performed during regular weekday hours in 165 (50%) and during off-hours in 162 (50%) cases. With 40 regular and 128 off-hours/week, 3.25 times more TAEs were performed during regular hours. There was an even distribution across weekdays (Mon-Fri:16.9 ± 1.5%), while fewer TAEs were performed on weekends (Sat: 8.3%, Sun: 7.3%). Technical success of TAEs was 93.9% with a clinical success of 79.2% and a 30-day mortality of 18.4%. Shock was an RF for technical failure (p = 0.022). RFs for clinical failure were low hemoglobin (Hb) (p = 0.021) and transfusion of ≥6 units packed cells (p = 0.009). Independent RFs for mortality were clinical failure (p < 0.001), coagulopathy (p = 0.005), and shock (p < 0.001). CONCLUSION: Our results provide no evidence for a subjectively perceived increase in TAEs during off-hours but rather appear to show that most TAEs are performed during regular hours. Prompt TAE to control acute bleeding is crucial to prevent a drop in Hb with shock and the need for transfusion, which may promote coagulopathy and rebleeding, all of which are risk factors for a negative outcome. ADVANCES IN KNOWLEDGE: The presented analysis provides insights of occurrences and risk factors for success of transcatheter arterial embolization in acute bleeding in a large study population.

Assessment of Low-Grade Meniscal and Cartilage Damage of the Knee at 7 T: A Comparison to 3 T Imaging With Arthroscopic Correlation.

OBJECTIVES: The aim of this study was to compare the assessment of low-grade meniscal tears and cartilage damage in ultrahigh-field magnetic resonance imaging (MRI) at 7 T to routine clinical MRI at 3 T. MATERIALS AND METHODS: This study was approved by the local ethics committee, and written informed consent was obtained from each patient. Forty-one patients with suspected meniscal damage or mild osteoarthritis (Kellgren-Lawrence score, 0-2) received 7 T as well as routine clinical 3 T consecutively. The imaging protocol at both field strengths consisted of PD-weighted imaging with more than doubled resolution at 7 T. Images were read blinded regarding field strength and patient characteristics by 3 readers with different experience in musculoskeletal MRI (3 years, 6 years, and 10 years) according to a modified whole-organ MRI score of the knee in osteoarthritis and the Score of the International Cartilage Repair Society. Arthroscopic reports as a criterion standard were available for 12 patients. A multifactorial mixed model analysis was performed. RESULTS: The mean cumulated diagnostic score at 7 T was significantly closer to the criterion standard compared with 3 T in patients where criterion standard was available (P < 0.001). In all 41 patients, the damages were rated more severely at 7 T reflected by a mean higher cumulative score in cartilage (P < 0.001) and in the meniscus (P < 0.001). No difference in interreader variability between 3 T and 7 T was observed. Imaging acquisition time was nearly identical. CONCLUSIONS: Morphologic imaging of cartilage and meniscal damage of the knee in ultrahigh-field MRI at 7 T with PD-weighted TSE sequences seemed to have a significantly higher diagnostic accuracy than 3 T and can be performed with equal acquisition times while exploiting higher resolution of 7 T.

Feasibility study of MR-guided transgluteal targeted in-bore biopsy for suspicious lesions of the prostate at 3 Tesla using a freehand approach.

OBJECTIVE: The aim of our study was (1) to establish an in-bore targeted biopsy of suspicious prostate lesions, avoiding bowel penetration using a transgluteal approach and (2) to assess operator setup, patient comfort and safety aspects in the clinical setting for freehand real-time MR-guidance established for percutaneous procedures in an open MR-scanner. MATERIAL AND METHODS: 30 patients with suspect prostate lesions were biopsied in a cylindrical 3T-MRI system using a transgluteal approach in freehand technique. One to three biopsies were sampled using continuous dynamic imaging. Size, location and visibility of the lesion, intervention time, needle artefact size, interventional complications and histopathological diagnosis were recorded. RESULTS: All biopsies were technically successful. Nineteen patients showed evidence of prostate carcinoma. Cancer detection rate was 50 % in patients with previously negative TRUS-biopsy. The average intervention time was 26 min including a learning curve as the time was 13 min by the end of the study. No antibiotic prophylaxis was performed as none of the patients showed signs of infection. CONCLUSIONS: MR-guided targeted freehand biopsies of prostate lesions using a transgluteal approach are both technically feasible and time efficient in a standard closed-bore 3T-MR scanner as well as safe for the individual patient. KEY POINTS: • Open-bore freehand interventional principles were adapted to closed-bore systems. • Prostate MR-guided freehand biopsies were feasible in a clinical setting. • A transgluteal approach provides a short and simplified work flow. • An inoculation of the prostate with bowel flora is avoided. • The intervention time is comparable to the stereotactic approach.

Initial Assessment of the Efficacy of Irreversible Electroporation in the Focal Treatment of Localized Renal Cell Carcinoma With Delayed-interval Kidney Tumor Resection (Irreversible Electroporation of Kidney Tumors Before Partial Nephrectomy [IRENE] Trial-An Ablate-and-Resect Pilot Study).

OBJECTIVE: To assess the efficacy of irreversible electroporation (IRE) ablation of pT1a renal cell carcinoma (RCC) in the first prospective, monocentric phase 2a pilot ablate-and-resect study (Irreversible Electroporation of Kidney Tumors Before Partial Nephrectomy [IRENE] trial). It has been postulated that focal IRE can bring about complete ablation of soft-tissue tumors with protection of healthy peritumoral tissue and anatomic structures. PATIENTS AND METHODS: The first 7 study patients with biopsy-proven pT1a RCC (15-39 mm) underwent IRE. Percutaneous computed tomography-guided IRE was performed with electrocardiographic triggering under general anesthesia and deep muscle paralysis with 3-6 monopolar electrodes positioned within the renal tumor. Twenty-eight days later, the tumor region was completely resected to confirm tumor destruction pathologically. Individual results for these patients are displayed, described, and discussed. RESULTS: Technical feasibility was attained in all patients, but electrode placement and ablation were complex, with a mean overall procedure time of 129 minutes. There were no major complications. Partial kidney resection was performed in 5 patients, and radical nephrectomy was performed in 2 patients because of central tumor location and ablation areas. Resections revealed by tumor, node, and metastasis classification of the International Union for Cancer Control 2017 no residual tumor as complete ablation in 4 cases (ypT0V0N0Pn0R0) and microscopic residual tumor cells as incomplete ablation in the other 3 cases (ypT1aV0N0Pn0R1). CONCLUSION: Renal percutaneous IRE appears to be a safe treatment for pT1a RCC but requires substantial procedural effort. Resection specimens of the ablation zone revealed a high rate of microscopic incomplete ablation 4 weeks after IRE. According to these initial study results, curative, kidney-sparing ablation of T1a RCC appears possible but needs technical improvement to ensure complete ablation.

Cardiac Magnetic Resonance Imaging Using an Open 1.0T MR Platform: A Comparative Study with a 1.5T Tunnel System.

BACKGROUND: Cardiac magnetic resonance imaging (cMRI) has become the non-invasive reference standard for the evaluation of cardiac function and viability. The introduction of open, high-field, 1.0T (HFO) MR scanners offers advantages for examinations of obese, claustrophobic and paediatric patients.The aim of our study was to compare standard cMRI sequences from an HFO scanner and those from a cylindrical, 1.5T MR system. MATERIAL/METHOD: Fifteen volunteers underwent cMRI both in an open HFO and in a cylindrical MR system. The protocol consisted of cine and unenhanced tissue sequences. The signal-to-noise ratio (SNR) for each sequence and blood-myocardium contrast for the cine sequences were assessed. Image quality and artefacts were rated. The location and number of non-diagnostic segments was determined. Volunteers' tolerance to examinations in both scanners was investigated. RESULTS: SNR was significantly lower in the HFO scanner (all p<0.001). However, the contrast of the cine sequence was significantly higher in the HFO platform compared to the 1.5T MR scanner (0.685±0.41 vs. 0.611±0.54; p<0.001). Image quality was comparable for all sequences (all p>0.05). Overall, only few non-diagnostic myocardial segments were recorded: 6/960 (0.6%) by the HFO and 17/960 (1.8%) segments by the cylindrical system. The volunteers expressed a preference for the open MR system (p<0.01). CONCLUSIONS: Standard cardiac MRI sequences in an HFO platform offer a high image quality that is comparable to the quality of images acquired in a cylindrical 1.5T MR scanner. An open scanner design may potentially improve tolerance of cardiac MRI and therefore allow to examine an even broader patient spectrum.

Impact of in Vivo High-Field-Strength and Ultra-High-Field-Strength MR Imaging on DNA Double-Strand-Break Formation in Human Lymphocytes.

Purpose To determine the impact of different magnetic field strengths (1, 1.5, 3, and 7 T) and the effect of contrast agent on DNA double-strand-break (DSB) formation in patients undergoing magnetic resonance (MR) imaging. Materials and Methods This in vivo study was approved by the local ethics committee, and written informed consent was obtained from each patient. To analyze the level of DNA DSBs, peripheral blood mononuclear cells were isolated from blood samples drawn directly before, as well as 5 minutes and 30 minutes after MR imaging examination. After performing γH2AX immunofluorescence staining, DSBs were quantified with automated digital microscopy. MR group consisted of 43 patients (22 women, 21 men; mean age, 46.1 years; range, 20-77 years) and was further subdivided according to the applied field strength and administration of contrast agent. Additionally, 10 patients undergoing either unenhanced or contrast material-enhanced computed tomography (CT) served as positive control subjects. Statistical analysis was performed with Friedman test. Results Whereas DSBs in lymphocytes increased after CT exposure (before MR imaging: 0.14 foci per cell ± 0.05; 5 minutes after: 0.26 foci per cell ± 0.07; 30 minutes after: 0.24 foci per cell ± 0.07; P ≤ .05), no alterations were observed in patients examined with MR imaging (before MR imaging: 0.13 foci per cell ± 0.02; 5 minutes after: 0.12 foci per cell ± 0.02; 30 minutes after: 0.11 foci per cell ± 0.02; P > .05). Differentiated analysis of MR imaging subgroups again revealed no significant changes in γH2AX level. Conclusion Analysis of γH2AX foci showed no evidence of DSB induction after MR examination, independent of the applied field strength and administration of gadolinium-based contrast agent.

Haemorrhagic Complications and Symptomatic Venous Thromboembolism in Interventional Tumour Ablations: The Impact of Peri-interventional Thrombosis Prophylaxis.

AIM: The aim of this study was to assess the rates of haemorrhagic and thrombotic complications in patients undergoing interventional tumour ablation with and without peri-interventional low-molecular-weight heparin (LMWH) thrombosis prophylaxis. METHODS: Patients presented with primary and secondary neoplastic lesions in the liver, lung, kidney, lymph nodes and other locations. A total of 781 tumour ablations (radiofrequency ablation, n = 112; interstitial brachytherapy, n = 669) were performed in 446 patients over 22 months; 260 were conducted under peri-interventional thrombosis prophylaxis with LMWH (H-group;) and 521 without this (NH-group, in 143 of these, LMWH was given post-interventionally). RESULTS: Sixty-three bleeding events occurred. There were significantly more bleedings in the H-group than in the NH-group (all interventions, 11.66 and 6.26 %, p = 0.0127; liver ablations, 12.73 and 7.1 %, p = 0.0416). The rate of bleeding events Grade ≥ III in all procedures was greater by a factor of >2.6 in the H-group than in the NH-group (4.64 and 1.73 %, p = 0.0243). In liver tumour ablations, the corresponding factor was about 3.3 (5.23 and 1.54 %, p = 0.028). In uni- and multivariate analyses including covariates, the only factor constantly and significantly associated with the rate of haemorrhage events was peri-interventional LMWH prophylaxis. Only one symptomatic lung embolism occurred in the entire cohort (NH-group). The 30- and 90-day mortalities were significantly greater in the H-group than in the NH-group. CONCLUSIONS: Peri-interventional LMWH thrombosis prophylaxis should be considered with caution. The rate of clinically relevant thrombotic events was extremely low.

Feasibility study: 7 T MRI in giant cell arteritis.

OBJECTIVES: To assess the detectability of vasculitic changes of the superficial cranial arteries with contrast-enhanced 7 T MRI in three patients with GCA and intraindividual comparison with 3 T MRI. METHODS: Three patients (two female, one male) with suspected GCA underwent contrast-enhanced T1-weighted high-resolution MRI at 3 T and 7 T magnetic field strength. The clinical diagnosis, based on criteria of the American College of Rheumatology, was confirmed by biopsy of the superficial temporal artery after MRI. MR images were visually assessed for detection of vasculitic mural contrast enhancement and vessel wall thickening of the superficial cranial arteries. RESULTS: Both 3 T and 7 T MRI allows for visualisation of mural inflammatory changes and assessment of the vasculitic involvement pattern. Image quality of 7 T MRI was rated superior to image quality of 3 T MRI scans. CONCLUSIONS: 7 T MR imaging of vasculitic changes in patients with GCA is possible, and suggests superior image quality. Larger patient studies are necessary to quantify the diagnostic advantages of 7 T MRI.

DNA double-strand breaks and micronuclei in human blood lymphocytes after repeated whole body exposures to 7T Magnetic Resonance Imaging.

PURPOSE: To examine the extent of genetic damage, assessed from deoxyribonucleic acid (DNA) double-strand breaks (DSBs) and micronuclei (MN) in peripheral blood mononuclear cells obtained from individuals repeatedly exposed to 7T Magnetic Resonance Imaging (MRI). MATERIALS AND METHODS: The study protocol was approved by the local ethics committee. Informed consent was obtained from 22 healthy, non-smoking, non-alcoholic male individuals, who had never undergone radio-/chemo-therapy, scintigraphy, and had not undergone X-ray examination one year prior blood withdrawal. Eleven participants were repeatedly exposed to 7T and 3T MRI while working with/around scanners or frequently participating as 7T and lower field MRI research subjects (mean age 34±7years). The other half was never exposed to 7T or lower field MRI and served as controls (mean age 33±9years). The damage in lymphocytes was assessed using anti-γH2AX immunofluorescence staining of DNA DSBs and by quantification of MN. Isolated cells were further exposed in vitro to 7T MRI either alone or in the presence of the DNA damaging drug etoposide, to determine if there is any additional combined effect. The kinetics of DNA damage repair were examined. RESULTS: The mean base-level of γH2AX foci/cell and incidence of MN between repeatedly exposed and control group were not significantly different (P=0.618 and P=0.535, respectively). The additional in vitro exposure of cells to 7T MRI had no significant impact on MN frequencies and γH2AX foci at 1, 20 and 72h after exposure. CONCLUSION: Frequently repeated 7T MRI exposure did not result in a detectable increase in genotoxicity indices and alterations of DNA repair kinetics.

Analysis of DNA Double-Strand Breaks and Cytotoxicity after 7 Tesla Magnetic Resonance Imaging of Isolated Human Lymphocytes.

The global use of magnetic resonance imaging (MRI) is constantly growing and the field strengths increasing. Yet, only little data about harmful biological effects caused by MRI exposure are available and published research analyzing the impact of MRI on DNA integrity reported controversial results. This in vitro study aimed to investigate the genotoxic and cytotoxic potential of 7 T ultra-high-field MRI on isolated human peripheral blood mononuclear cells. Hence, unstimulated mononuclear blood cells were exposed to 7 T static magnetic field alone or in combination with maximum permissible imaging gradients and radiofrequency pulses as well as to ionizing radiation during computed tomography and γ-ray exposure. DNA double-strand breaks were quantified by flow cytometry and automated microscopy analysis of immunofluorescence stained γH2AX. Cytotoxicity was studied by CellTiter-Blue viability assay and [3H]-thymidine proliferation assay. Exposure of unstimulated mononuclear blood cells to 7 T static magnetic field alone or combined with varying gradient magnetic fields and pulsed radiofrequency fields did not induce DNA double-strand breaks, whereas irradiation with X- and γ-rays led to a dose-dependent induction of γH2AX foci. The viability assay revealed a time- and dose-dependent decrease in metabolic activity only among samples exposed to γ-radiation. Further, there was no evidence for altered proliferation response after cells were exposed to 7 T MRI or low doses of ionizing radiation (≤ 0.2 Gy). These findings confirm the acceptance of MRI as a safe non-invasive diagnostic imaging tool, but whether MRI can induce other types of DNA lesions or DNA double-strand breaks during altered conditions still needs to be investigated.

Sensory perceptions of individuals exposed to the static field of a 7T MRI: A controlled blinded study.

PURPOSE: To determine the subjective experience of subjects undergoing 7T magnetic resonance imaging (MRI) compared to a mock scanner with no magnetic field. METHODS AND MATERIALS: In all, 44 healthy subjects were exposed to both the B0 field of a 7T whole-body MRI and a realistic mock scanner with no magnetic field. Subjects were blinded to the actual field strength and no scanning was performed. After exposure, subjects rated their experience of potential sensory perceptions. RESULTS: The most frequently observed side effect was vertigo while entering the gantry, which was reported by 38.6% (n = 17). Other frequent side effects were the appearance of phosphenes (18.2%, n = 8), thermal heat sensation (15.9%), unsteady gait after exposure (13.6%, n = 6), and dizziness (13.6%). All side effects were reported significantly more often after 7T exposure. Nine subjects (20.5%) did not report any sensory perceptions at all, ie, neither in the 7T scanner nor in the mock scanner. CONCLUSION: Light, acute, and transient sensory perceptions can occur in subjects undergoing ultrahighfield MRI, of which vertigo seems to be the most frequently reported. Possible psychological effects might contribute to the emergence of such sensory perceptions, as some subjects also reported them to appear in a realistic mock scanner with no magnetic field.

Visualization of intrarenal vessels by 3.0-T MR angiography in comparison with digital subtraction angiography using renal specimens.

BACKGROUND: MRI at 3.0 T enables high-spatial resolution for renal MR angiography. OBJECTIVE: Evaluation of an arterial tree model in animal kidney specimens with comparison of digital subtraction angiography (DSA) and high-field MRI to find out the maximum spatial resolution of intrarenal vessels. It was considered that objective quantitation of angiogram quality could be achieved. MATERIALS AND METHODS: A total of 27 pig kidney specimens were examined by MR angiography (flash 3D) using a 3.0-T scanner (TRIO; Siemens, Erlangen, Germany) with an eight-channel head coil and a voxel size of 0.9x0.9x1.1 mm in the early arterial phase after implantation of a 4F catheter in the renal artery. DSA (Integris, Philips, Best, The Netherlands) was performed immediately after the MRI. With the help of semiautomated segmentation, all vessels were marked for comparison of the vessel trees. The Wilcoxon rank test was used for statistical evaluation of vessel numbers and branching depths. RESULTS: Objective comparison between DSA and MR angiography was achieved. High-field MR angiography had the ability to depict vessels up to the seventh branching on average. Significant differences in vessel delineation and counts were found from the fifth level of intrarenal branching onwards with DSA showing an advantage. CONCLUSION: High-field MRI has great potential in the detection of intrarenal arteries and is comparable to DSA in visualization of the central intrarenal vessel tree.