Subjectively Reported Effects Experienced in an Actively Shielded 7T MRI: A Large‐Scale Study

Ultrahigh‐field (UHF) MRI advances towards clinical use. Patient compliance is generally high, but few large‐scale studies have investigated the effects experienced in 7T MRI systems, especially considering peripheral nerve stimulation (PNS) and caregiving.

field (causing, for example, projectile risk) [4][5][6] or translational forces and dizziness/vertigo originating from Lorenz forces due to ionic currents within the vestibular system, 7 the radiofrequency field (causing, for example, energy deposition, possible burns, or increase of temperature), and the gradient field (causing, for example, peripheral nerve stimulation [PNS] and acoustic noise). [4][5][6] Previous publications have shown that short-term effects-dizziness/vertigo, inconsistent movement, nausea, headache, and a metallic taste-may be experienced by subjects in the 7T environment, [8][9][10][11][12][13][14] but fewer sensations have been reported in the 1.5T environment, and with considerably (mean 84%) lower mean scores. 10 Evaluation often needs to rely on self-reporting of these effects, which further are often highly subjective, although in general are physiological and may, by some individuals, even be experienced in mock scanners with no static magnetic field, radiofrequency field, or gradient field. 15,16 Sensations might thus also be influenced by factors in the MRI environment that are not related to the MRI fields, such as experience of discomfort or anxiety.
The purpose of this large-scale study was to investigate the quantity of, the intensity of, and subjective experiences from the effects of 7T MRI, focusing on patient comfort and compliance.

Ethics
The study was approved by the appropriate Ethics Committee (entry nos. 2015/434 and 2016/126) and informed written consent was obtained from all subjects, as described by Hansson et al. 14

MRI System
Examinations were conducted in first-level controlled operating mode on an actively shielded 7T MRI scanner (Achieva; Philips, Best, the Netherlands) with the following specifications: gradient system with a combination of maximum amplitude 40 mT/m and maximum slew rate 200 mT/m/s, or maximum amplitude 60 mT/m and maximum slew rate 100 mT/m/s; tunnel diameter 58 cm; length of magnet 3.3 m; a maximum spatial field gradient (dB/dz) of the stray field of 7.86 T/m at 130 cm from isocenter. The 2Tx/32Rx Nova head coil (Nova Medical, Wilmington, MA) was used for the brain examinations; 28Rx Knee Coil QED (Quality Electrodynamics, Mayfield Village, OH) was used for the knee examinations;16Rx wrist array (RAPID MRI International, Columbus, OH) was used for the wrist examinations; 8Rx Breast array (RAPID MRI International, Columbus, OH) was used for the breast examinations; and 8Rx C-spine coil (Life Services, Minneapolis, MN) was used for the cspine examination. Scan protocols varied largely for different body parts and projects, including healthy research subjects and patients examined at the facility. Projects aimed at technical development, disease-oriented research (eg, systemic lupus erythematosus, epilepsy, brain tumors, dementia, psychiatric disorders, knee arthrosis), and neuroscience research (eg, cognition, motor skills, fear). The sequences used were both manufacturer-provided and programmed in-house. All scans were performed at first-level controlled operating mode and did not exceed the specific absorption rate (SAR) limit of whole-body 4 W/kg or head 3.2 W/kg. Used sequences ranged from standard morphological sequences to ultrashort echo time sequences and to functional, diffusion, perfusion, spectroscopy, flow, and chemical exchange saturation transfer MRI.

Study Subjects and Definition of Data Collected
Over a 3-year period, 1290 research and clinical 7T MRI examinations were performed and all subjects were invited to fill out the web-based questionnaire. Data on inclusion and exclusion of subjects and subject and examination demographics are given in Fig. 1. The participation rate was 74% (954 examinations, 801 individual subjects). The main reasons identified by staff for exclusion were unwillingness to participate due to time concerns, and earlier participation for subjects who underwent several 7T MRs (mainly healthy volunteers). A smaller portion of subjects could not participate due to their disease, due to being too tired, or because they did not feel capable of filling out the questionnaire under the prevailing circumstances (mainly patients). Of the questionnaires submitted, 3% had to be excluded due to being incomplete. For 627 examinations (66%) log files of the highest predicted PNS values could be retrieved, where 100% corresponds to 50% of subjects predicted to experience PNS, defined as anything between a mild tingling and painful contractions of muscles.
This study did not make use of examinations included in a previous study 14 and the questionnaire was updated with additional questions regarding experience and anxiety, and by exclusion of the visual analog scale (VAS) in favor of a sixpoint Likert scale and a seven-point adjectival scale. 17 Before the examination, all of the subjects underwent a strict MRI safety check, changed from their street clothes to hospital pants and gown, and were provided with hearing protection as in a previous study. 14 After the examination, a web-based questionnaire (REDCap; research electronic data capture; http://project-redcap.org) was used to collect data on demographics (gender (M/F), age (y), self-estimated sensitivity regarding motion sickness (kinetosis), on any short-term effects, and on body and room temperature experienced, scanner noise, the communication system, and willingness regarding any future 7T MRI examination. The operator entered the information on session parameters including the length of the examination (min), the part of the body that was examined, and the orientation of the body in the field (head-first or feet-first). Experience of the short-term effects dizziness, inconsistent movement, nausea, headache, and metallic taste were evaluated for four situations: lying on the table and moving into the scanner (in), being at the isocenter (inside), moving out of the scanner (out), and having gone outside the scanner room after the examination (outside), in accordance with a previous study. 14 These short-term effects, together with light flashes, self-estimated sensitivity regarding motion sickness, and PNS were evaluated regarding quantity and/or intensity using a six-point Likert scale: none, very little, little, moderate, much, and very much. 17 The term "inconsistent movement" refers to the experience of body movement in a direction other than the straight direction through the scanner tunnel, 14 and dizziness includes vertigo but also the feeling of presyncope, or disequilibrium and nonspecific feelings difficult for participants to describe or define as vertigo.
Experiencing of room and body temperature before, during, and after research examinations was measured using a seven-point adjectival scale 17 (uncomfortably cold, cold, slightly cold, comfortable, slightly warm, warm, and uncomfortably warm). Tolerance of maximum experienced scanner noise, functioning of the communication system, view of information from and contact with personnel, and willingness regarding future 7T MRI examinations as a research subject or as a clinical patient were measured with a six-point adjectival scale (strongly agree, agree, mildly agree, mildly disagree, disagree, and strongly disagree). At the end of the questionnaire, a free text option gave participants the opportunity to comment on information they had been lacking or suggestions for improvement.

Statistics
Descriptive statistics were used to present data. Nonparametric Spearman's rank correlation was used for analysis of ranked variables such as self-estimated sensitivity regarding motion sickness, highest recorded predicted PNS values, and quantity and intensity of different effects. Any P-value <0.05 was regarded as being statistically significant. The Mann-Whitney U-test was used to analyze differences in experience of effects between patients and healthy volunteers. A t-test was used to analyze the difference in anxiety levels between the first and second 7T examinations of the subjects who underwent more than one 7T examination.

Results
The proportion of subjects who experienced a short-term effect at any time (moving in, inside, moving out of the scanner, or being outside of the scanner after the examination) during individual examinations was 81% for dizziness, 68% for inconsistent movement, 63% for PNS, 40% for headache, 32% for nausea, 12% for metallic taste, and 8% for light flashes. The quantity and intensity of dizziness, inconsistent movement, nausea, headache, and metallic taste in relation to movement and position in and outside the scanner is summarized in Fig. 2. When comparing patients (n = 272) with healthy volunteers (n = 682), patients had significantly more often and more intense headache (P < 0.01; moving in, inside, moving out, or outside of the scanner) and metallic taste (P < 0.01; outside the scanner) but less intense dizziness (P = 0.01; outside) compared to healthy volunteers ( Table 1).
Analysis of the correlation between intensity of selfestimated sensitivity regarding motion sickness and the highest intensity of relevant short-term effects-during movement into the scanner or out of the scanner, or positioned in the scanner-with Spearman's rank correlation showed significant (P < 0.001) but very weak to weak correlations for nausea (ρ = 0.24), dizziness (ρ = 0.22), and inconsistent movement (ρ = 0.21).
Descriptive statistics for intensity of self-estimated sensitivity regarding motion sickness and intensity of relevant short-term effects are given in Table 2.
Subjects experienced light flashes in 80 examinations (8%) and rated the quantity as very little in 54 examinations, little in 17 examinations, moderate in six examinations, and much in one examination, with two examinations lacking data on quantity. Intensity was rated as very little in 51 examinations, little in 21 examinations, moderate in six examinations, and much in two examinations, with four examinations lacking data on intensity.
For 598 of the examinations (63%), the subjects reported that they had experienced PNS. The data on the quantity and intensity of PNS and how these were experienced are summarized in Fig. 3, differentiating data for examinations for which scanner log files were available (n = 627) from data for examinations for which such files were not available (n = 327)-for consideration of possible selection bias. Spearman's rank correlation showed a significant (P < 0.001) strong correlation between both the quantity (ρ = 0.87) and the intensity (ρ = 0.90) of PNS events associated with experiencing such PNS events. Furthermore, there was a significant (P < 0.001) and very strong correlation between the quantity and the intensity of the PNS events (ρ = 0.93), but there was a significant but only very weak correlation between highest predicted PNS value (ρ = 0.19) associated with the experience of PNS; there also was a significant (P < 0.001) but weak correlation between the highest predicted PNS value for each examination-for both quantity of PNS (ρ = 0.20) and intensity of PNS (ρ = 0.23). The relationships between the quantity and intensity of PNS experienced, the highest predicted PNS values, and how uncomfortable the PNS experienced was are illustrated in Fig. 4. Experience of PNS did not differ between patients and healthy volunteers (Table 1).
Scanner room temperature was generally experienced as being more comfortable before than during examinations, while the subject was inside the tunnel. The change in room temperature most commonly reported was a change towards warmer room temperature (338 subjects, as compared to 107 subjects who reported experiencing a decrease in room temperature). Table 3 is a summary of how subjects experienced scanner room temperature before, during, or after the examination and in which parts of the body a temperature change was felt during the examination. Experience of temperature did not differ between patients and healthy volunteers. The change in body temperature experienced was higher (warmer) for 374 subjects and lower for 119 subjects. Experiencing an increased body temperature was often associated with perception of a temperature increase in the face or head and upper extremities (hands and arms), whereas subjects who felt a decrease in body temperature mainly reported having cold feet. Table 4 is a summary of data on acceptability of maximum scanner noise levels, functioning of the communication system, view of information, and contact with personnel and shows that patients and healthy volunteers in 80% to 99% of the examinations agreed or strongly agreed on a positive perception of these aspects. Further, 57% of patients and 65% of healthy volunteers agreed or strongly agreed that the examination was comfortable and willingness to undergo future 7T examinations for research purposes and 82% was for clinical purposes 93%. Subjects who had experienced a previous MRI examination-which was true for 644 of all the examinations (77%)-rated the 7T MRI experience as being worse than previous MRI examinations (n = 174 examinations; 27%), as being the same as in previous MRI examinations (n = 323 examinations; 50%), and as being better than in previous MRI examinations (n = 147 examinations; 23%).
Anxiety before the examination was reported by 412 of the subjects (43%) (Fig. 5) with a mean anxiety level of 1.8 on a 6-grade Likert scale 1-6 for patients and a mean anxiety level of 1.6 for healthy volunteers. The difference in anxiety level reported by patients and healthy volunteers was significant, with a higher anxiety level for patients prior to the 7T MRI examination (Mann-Whitney U-test, P = 0.03), where the term "patients" covers those who underwent clinical scans and those who were included in a clinical, disease-specific research study. Anxiety levels were not significantly different between the 3 years of the study (Fig. 6). Subjects who had more than one 7T examination (n = 91; 10%) had the same anxiety score before the second examination as before the first examination in 67 cases (74%), a lower score in 17 cases (19%; 1 case being a patient), and a higher score in seven cases (7%). There was no significant difference in the anxiety levels reported before the first examination and the anxiety levels reported before the second examination (t-test, P = 0.3) in the 91 subjects who underwent more than one 7T examination. The statement that the total experience of the MRI examination was comfortable was strongly agreed or agreed with by 600 subjects (63%), whereas 246 (26%) neither agreed nor disagreed with this, and 108 (11%) disagreed or strongly disagreed (Table 3). Further, there was a significant difference in the comfort rate between patients and healthy volunteers (P = 0.007), with healthy volunteers experiencing the 7T examination as more comfortable compared to patients. In the optional free text at the end of the questionnaire, comments on potential improvements were suggested regarding 117 examinations; seven were positive comments, for example, that nothing required to be improved or that the hearing protection was very good. Improvements regarding information were suggested by 60 subjects, mainly regarding giving of information on when it was possible to adjust body position; remaining time; and warning prior to sequences with high acoustic noise levels or high risk of PNS. In 13 cases, the proposals for improvement concerned information specific to the study that the subjects were being scanned for. Thirty-five were complaints regarding poor sound quality or the volume of the communication system and/or music system. Fifteen were nursing care-related suggestions, mainly concerning better hearing protection.

Discussion
When UHF is used clinically, it is essential to minimize any undesired effects and optimize compliance. In addition to dizziness, and inconsistent movement, PNS is one of the most Very much 2 1 2 0 1 3 frequent undesired effects, despite vendor-implemented steps to limit its occurrence. However, most subjects who experience PNS report it as being "not uncomfortable at all" or "very little uncomfortable." This is in line with an overall high level of acceptance for UHF examinations both for patients and healthy volunteers. Regarding compliance and nursing care issues, the study revealed that data on the experience of short-term effects may differ between healthy volunteers and patients and that a large proportion of subjects-especially patients-stated that they had some degree of anxiety prior to the examination, and some had information-and communication-related complaints, leaving room for improvement in patient care (concerning handling and information) if we are to increase patient compliance in UHF examinations.
In comparison to an initial study, 14 most short-term effects that occurred during the movement of the subject into the scanner bore were less prevalent; dizziness had decreased    (7) 69 (24) 805 (222) 43 (16) 5 (2) 2 (0) During 4 (2) 42 (14) 67 (20) 488 (145) 245 (61) 92 (26) 16 ( from 84% to 68%; inconsistent movement from 70% to 54%; nausea from 52% to 23%; headache from 52% to 17%; and metallic taste from 43% to 5%. The numbers are still high in comparison to other publications. 12,13,18,19 Compared to the earlier publication, 14 the decrease in frequency of some of the short-term effects might partly be related to the change from using a VAS (with a slide bar) in the questionnaire to the use of six-or seven-step scales, avoiding positioning of the slide bar at very low numbers instead of an anticipated zero. Considering other studies performed on the use of passively shielded magnets, another explanation might be the adaptation theory-related to differences in the shortterm effects of passively and actively shielded 7T systemsbased on biological mechanisms, including adaptation to a continuous vestibular stimulation. The vicinity of the passively shielded system, and therefore the area in which the subjects are prepared on the table before entering the bore, has a higher stray field than the surroundings of the actively shielded system, so the subjects have more time to adapt to a higher field before going into the scanner. 14,20,21 To establish if differences in experience of effects do exist between actively vs. older passively shielded MRI scanners is mainly relevant from the perspective on how to translate established knowledge between such systems. Direct comparison is difficult, as large patient groups are required and scanners often differ regarding, not only in shield coils and resulting fringe fields, but also other aspects, as for example gradient coil design. We therefore advocate exploration of further large-scale populations from different types of actively shielded UHF systems.  The proportion of subjects who reported PNS was higher than in earlier publications 11,12,22 but was very similar to that in our previous study (63% and 67%, respectively), without any apparent effect of the change from VAS to step scales. Possible explanations for the high occurrence of PNS with 7T scanners might be the high level of dB/dt for protocols used in research studies, 14,23 a systematic difference to other systems in terms of geometry, or the current pattern of the gradient coil. The design (and especially the length) of the gradient coil has significance for PNS, as a longer gradient coil covers more of the body surface. Glover 24 carefully studied the causes and risks of PNS and concluded that, although the threshold limits for nerve stimulation are well known, there still are difficulties in applying them to a specific system and subject geometry. Our findings also indicate that the experience of PNS may be more dependent on the individual undergoing the scan than the predicted PNS value and the level of dB/dt to which the subject is exposed. Compared to the previous study, 14 the subjects were informed before the examination that PNS might occur, but they were not always prewarned about upcoming sequences when the system warning for high predicted PNS occurred. Subject feedback indicated a preference for being prepared prior to sequences with high predicted PNS. Of the subjects who experienced PNS, 1.5% rated the experience as "very uncomfortable" or "very much uncomfortable," but most of the subjects who experienced PNS rated the experience as "not uncomfortable at all" or "very little uncomfortable." The uncomfortable increase in body temperature experienced in a particular part of the body during the examination could be traced to a specific research protocol where several long functional (f)MRI echo-planar imaging (EPI) scans were included, and to proximity to hardware that might undergo heating during scanning, such as the inner wall of the bore. Variation in temperature in the scanner room (17 -22 C) has been an issue and, unfortunately, was beyond our control, although it would have been convenient to be able to adjust the scanner room temperature as appropriate. We tried to compensate for low temperature by offering the subject a blanket.
Concerning short-term effects when moving in the tunnel, nausea was not reported as often as dizziness and inconsistent movement, but there was a slightly higher correlation between nausea and self-estimated sensitivity regarding motion sickness. Although considered of relevance in other studies, 25 in the present study self-estimated sensitivity regarding motion sickness only showed a very weak to weak correlation with the occurrence of relevant short-term effects, and in our opinion premedication in a research or clinical setting is not indicated based on these correlations, but it might be considered for patients who are predisposed to diseaserelated nausea-especially in combination with a high selfestimated sensitivity regarding motion sickness.
As in previous studies, light flashes were reported very infrequently. 12,26 Phosphene is characterized by the experience of seeing light without actual light entering the eye. Magnetophosphene might occur when the retina is magnetically stimulated at a narrow frequency range of around 20 Hz. 16 The absolute majority of the study subjects were examined with the head coil. Investigation of differences in short-term effects depending on the part of the body examined would require a more diverse examination palette than represented in this study, which mainly consisted of brain examinations.
The general experience of having good contact with the personnel before the examination from the reporting in this study is in agreement with the reporting in the previous study 13 -99% and 98%, respectively, with 97% and 95%, respectively, during the examination. This might be a result of a focused effort to educate all personnel who come in contact with research subjects and patients. The noise generated primarily by the gradient system is a well-known issue in MRI. The degree of acceptance of noise levels has improved, from 74% in the previous study to 90% in the current study. This might be a result of improved skills in using hearing protection. None of the study subjects terminated the examination because of acoustic noise. However, improvements can still be made considering nursing efforts regarding hearing protection, as pointed out by study subjects in the free text comments, and also by vendors hopefully providing improved hearing protection or noise-cancelling headphones with a built-in communication system. This would not only allow improved communication and subject entertainment, but also improve prerequisites for many fMRI experiments.
That 644 subjects (67%) had experienced a previous MRI examination might introduce bias regarding the experiencing of 7T MRI. Considering this, however, we regard it notable that 412 of the subjects (43%) showed some level of anxiety before the examination, and that patients were significantly more anxious than volunteers. No significant difference in anxiety levels were observed between the first and second MRI examination in subjects undergoing two examinations in the study period, which speaks against a bias due to multiple examinations. Other studies confirm the high anxiety level prior to MRI examinations with 30-40%. 27,28 The main stressor in Lo Re et al 27 was the uncertainty of the diagnosis, therapy, and prognosis. These studies also stress the importance of professionalism of the radiological staff when they receive and inform the patient, and also during the examination with emotive involvement and targeted education. This has implications for both patient welfare and image quality. 27,28 Improved information, about the examination might reduce preexamination anxiety. To ease anxiety during the examination, some subjects made suggestions regarding the need for more information during the examination; for example, when they are allowed to move, the duration of the next scan, and notification before a high PNS risk sequence starts or acoustic noise levels. The majority of subjects experienced the examination as comfortable and considered the prospect of a further 7T MRI examination both as patient or research subject with a positive attitude. The findings in this study point towards a generally positive attitude towards 7T examinations and high patient comfort also seen by others, [9][10][11] but also show that there is absolutely room for improvement when 7T MRI now translates into clinical use.

Limitations
The lack of a control group or control situation with examinations performed at another field strength, a passively shielded scanner with otherwise comparable technical specifications, or a mock scanner is a limitation that we could not overcome considering the large number of subjects included in the study and the design of the study focusing on inclusion of all subjects examined at a certain scanner, independent of status as a healthy volunteer or patient, of indications of the study, or of examined body part. Considering the number of subjects, the variety of study protocols, and the fact that subjects were included after performance of a 7T MRI for other reasons than this study, we could not measure vital signs or neurocognitive functions.
It was only possible to retrieve 66% of the examination log files; however, there was no selection bias regarding experiencing of PNS between subjects with log files and those without log files available.
The temperature in the scanner room was not constant during the study, but varied between~17 C and 22 C. The staff compensated for the low temperature by offering the subject a blanket.
Differences in short-term effects depend on the part of the body examined, leading to differences in exposure to the static magnetic field, the gradient field, and the radiofrequency electromagnetic field. To evaluate such differences would require a more diverse examination palette than represented in this study, which mainly consisted of brain examinations.
Some of the aspects evaluated in this study might be influenced by personnel handling the subjects and thus the diversity of personnel: researchers, technicians, MRI physicists, and doctors. We tried to minimize this potential bias, as all personnel at the facility received training from only three technicians who work closely together and supervise or perform a majority of the scans. Examples of aspects that might be influenced by the diversity of personnel are application of hearing protection and communication skills, potentially influencing experience of noise, information, and communication.
Subjectively reported effects occur in actively shielded UHF MRI and include physiological responses and highly individual psychological issues. Although leaving room for improvement, only a few patients and healthy volunteers experienced these effects as being so uncomfortable that they would lead to aversion to future UHF examinations.

Conclusion
The effects experienced in actively shielded UHF MRI include physiological responses and also highly individual psychological issues. However, few subjects experienced effects, although frequent or intense, as being so uncomfortable that they would be reluctant to undergo possible UHF MRI examinations in the future. Considering the data, compliance and experience might be further improved by focusing on preexamination anxiety, communication, and supplying information before and during the examination in parallel to technical advances decreasing the physiological impact. statistical considerations. We thank Birger Balkenius for providing a four-dimensional viewpoint of the relationship between PNS values, quantity, strength, and experience.

Funding
This work was supported by Skåne University Hospital foundations and donations.