The impacts of MRI innovation on patient management and clinical outcomes are difficult to assess. Studies require a broad range of patients with several clinical conditions, and long and complete patient follow-up. The outcomes need to be clinically relevant and, ideally, the impact on diagnosis, patient management, or patient outcomes needs to be assessed. Studies with hundreds of patients per study arm can be required to measure the small incremental differences between similar technologies such as different magnet strengths in MRI. MRI has applications in many clinical areas, and each can have unique considerations.
Advice about the selection of MRI type was sought from Dr. Ian Smith, the Director General, Institute for Biodiagnostics, National Research Council Canada. “The first question I ask people looking for advice is, ‘What will you do with the MRI?’ If it is routine scanning of brains and joints, a 1.5 is fine. If it is sophisticated measurements such as functional MRI or diffusion tensor measurements, the 3 T is essential.” (Dr. Ian Smith, National Research Council Canada, Winnipeg, personal communication, 2011 Jan)
6.1. Summary of Results
No identified studies examined whether the use of 3.0 T MRI scanners would result in a change in patient or health outcomes, or a change in clinical management, compared with 1.5 T MRI scanners.
All of the 25 included studies reported on clinical test parameters. The six clinical areas were neurology (mainly MS), cerebrovascular conditions, renal artery stenosis, CAD, musculoskeletal disorders, and oncology (breast cancer, liver cancer, prostate cancer, endometrial cancer, and cervical cancer). The authors most commonly reported that 3.0 T MRI was equivalent to 1.5 T MRI for various diagnostic and technical outcomes. In a few cases, 1.5 T MRI scanners were found to be better than 3.0 T MRI scanners; for example, tumour delineation of the prostate. And, in some other instances, 3.0 T MRI scanners outperformed 1.5 T MRI scanners. For example, advantage for 3.0 T MRI was seen in:
lesion detection in MS
identification of single or multi-vessel disease in patients with CAD
identification of disc shape and position for TMJ
nerve visibility for brachial plexus
visibility of anatomic structures in the wrist
identification of fibrocartilage lesions
diagnostic accuracy for hepatic metastases
sensitivity for detecting hepatic metastases.
All the identified studies were observational and thus did not stringently control for potential biases that may result in a higher chance of differences being falsely detected or actual differences not being detected. Appropriate to these study goals, a small number of patients (maximum 65) prospectively received repeat testing with 1.5 T MRI and 3.0 T MRI within a short time frame. Two or more interpreters (usually radiologists), generally blinded to patient details and magnet size, assessed the images using standardized quantitative measurements and qualitative questionnaires. In some cases, the findings were recorded independently and then compared. In other cases, the findings were agreed to by consensus.
Safety information collected from reviews, not individual studies, indicated the greater magnetic effect of 3.0 T MRI scanners may make them unsuitable for patients with specific implanted devices; to date, more than 1,000 devices and other objects that have not yet been deemed to be safe with the use of the 3.0 T MRI. Increased heat and increased noise with 3.0 T MRI may also be of concern.
One relevant study by Ohba et al.38 was identified through the alert process. The study was a non-randomized, prospective comparative study and the results did not affect the conclusions of the systematic review. Ohba et al. provided evidence that 3.0 T MRI and 1.5 T MRI were similar in identifying 58 malignant pulmonary nodules in 76 patients when using diffusion-weighted imaging. The authors also noted that further software developments for 3.0 T MRI would reduce lung artifacts, thus improving the correlation with apparent diffusion coefficient values and the F-fluorodeoxyglucose uptake on the positron-emission tomography.
6.2. Limitations of Assessment
6.2.1. Clinical benefits, limitations, and safety
The main limitation for the systematic review was the lack of evidence linking the clinical test findings from using different MRI technologies to an impact on clinically meaningful outcomes; that is, diagnosis, patient management, and clinical outcomes. Although some studies reported that 3.0 T MRI was superior in technical outcomes such as image detail, it was unclear that this would make a difference to patients and what the magnitude of the difference would be. Studies also tended to be small, generally with 20 patients to 30 patients enrolled. Several articles acknowledged this limitation and suggested that studies would need to be larger, enrol a broader spectrum of patients, and include more extensive patient follow-up to draw clinically valid conclusions.
The included literature for the systematic review was limited to those studies meeting selection criteria. Therefore, a number of indications for MRI were excluded; for example, brain tumours, epilepsy, breast imaging, and knee and shoulder pathology. Similarly, all included studies involved adult populations and pediatric populations were not studied.
Although the funding source was sought for each included study, 22 of the 25 (88%) included studies did not report funding or conflicts of interest. Of the remaining three studies, one, each, was funded by the German Research Foundation, Dutch Foundation for MS Research, and Pfizer. Regarding industry affiliation, studies reported one author employed by GE, two by Philips, and one by Pfizer.
An issue in the interpretation of the results of these studies is the increasing sophistication and changing performance of MRI devices. Although only recent studies were included (published in 2005 or later), some studies were performed as early as 2003 when 3.0 T MRI was in the early stages of introduction. Current 1.5 T MRI and 3.0 T MRI machines would perform differently from those that were used in the studies, suggesting that the findings from the earlier studies would not be reproducible today.
The MRI literature is limited in part due to federal regulations that only require device manufacturers to provide proof of safety and technical performance consistency according to specifications (scientific evidence of clinical utility or patient benefit before licensing is unnecessary). This does not provide an impetus for manufacturers to conduct studies that explore the impact of device technology on clinical outcomes.
6.2.2. Service delivery, personnel, and structural differences
Short time lines for report completion, time of year (December and January), and extensiveness of the survey requests limited the information received from OEMs. The information that is needed to adequately assess service delivery, personnel, and structural differences is often unpublished, inaccessible, and anecdotal. Similarly, data on utilization were limited to information that was collected in January 2009 (2010 data have been collected but are not yet released by CIHI). More than the other sections of this report, the information on service delivery and personnel are not immediately transferable to other jurisdictions.