We have a used Allegra you might be interested in.

The Ingenia has a worse gradient performance than the Achieva (the bore is wider). So ASL has a better SNR but DWI seems to be worse compared to the Achieva. Within the Philips range I would get an Achieva with digital upgrade. The competition for me would be a Siemens Trio.

The skyra and Verio should have the same problem as the Ingenia (wide bore) but no experience with that.

Siemens trio. Very popular, so you can get sequences straight from other research centers. You can also participate in many multi site studies, like fbirn. Its a great magnet in its own right too. Very flexible and capable, and extremely user friendly. A drunken monkey can run it well, so a grad student should be able to do almost as well.

I can only really speak about the systems I've used - which have all been Siemens. I currently use a Trio and Verio, both with the Tim system, and 32-channel head-coils. We get excellent fMRI signal-to-noise on both systems and have successfully done MRS and DTI on the trio. The Skyra looks like it should be a good option also. The key option is the 32-channel head-coil though, which really does make a big difference in signal-to-noise.

My sense is that Siemens systems are generally preferred over Philips for research purposes. I would probably avoid something more exotic like the Fonar systems, unless you have a particular need for an upright system.

I agree with Matthew. I've worked on a Siemens Allegra, Philips Achieva, Siemens Verio and now a Siemens Trio (an older platform than the Verio) personally, but our lab has experience with almost every platform. The consensus is that Siemens provides the best support for human imaging. The Siemens platform is more locked down than others, especially GE, but our experience with their support techs has been outstanding. The Siemens platform has GRAPPA and SENSE capabilities for parallel imaging and offers 32 channel head coils. You will see a marked improvement in SNR with the use of the 32-channel coil. (NOTE: we used a 32-channel coil with an Achieva system for DTI, fMRI and ASL with good results, as well). However, the Achieva had extremely high failure rates for its gradient amps (could have been a facilities problem though).

From what I am told, the SKYRA series from Siemens is really meant mainly for whole body scanning with its version of Tim integration. It would be a less attractive option than the Verio.

The Trio is no longer produced, but if you are into diffusion and fMRI I would get a Magnetom PRISMA with 64 Channel head coil 2Tx, and 80mT/m 200T/m/s gradients.

I would choose Siemens and more in details a Verio system more than a Skyra. We use Verio with great results, while I heard many people having problems with Skyra.

It would strongly depend on what you are intending to do with the system.

We have a Siemens Tim Trio and a Philips Ingenia (wide bore) 3T at our facility. Our research interests are in cardiovascular imaging and some body imaging, and our functional MR interests/needs are mostly clinical in nature.

I would say that the Philips Ingenia performs very well for cardiac imaging at 3.0T, and also in body imaging, particularly in two aspects.

First, the two independent RF transmit channels allows for reasonable B1 shimming. In our studies, we have found a substantial improvement in B1 homogeneity with RF shimming compared to without even in a small region such as the heart. This is essential, if you are planning to do some quantitative imaging.

The second advantage comes from the digitized receive coil chain, and its integration for whole body imaging. This is a slightly better iteration compared to the Tim trio - in our experience, to do whole body Dixon, or DWI. The posterior coil is integrated into the table top, and the work-flow is streamlined. Given the large number of coil elements that simultaneously receive signal, the choice of coil elements used in reconstruction, is automated, and the operator do not have to worry about it. It is a welcome change, if you have to deal with different patient sizes.

With respect to fMRI, we just received a 32 channel digital neuro head coil for the Ingenia and the images are quite good. We do not do a lot of fMRI.

For body/cardiac imaging, we have found Ingenia to be a very reliable system, with outstanding performance. We have not had any issues.

I'd go with Siemens Trio - or Verio. I agree with Michael Rubens above- it's the Honda Civic of scanners. Reliable, durable, easy to drive. That said, we have a Philips Achieva here at UW that is giving us very good SNR now that we have the new 32 channel head coil.

The new designed 3 T Siemens Prisma!

I agree with Torben above if the focus is on fMRI and DWI the gradient performance is essential and the upcoming Prisma indeed looks very interesting.

I am working with an Achieva and based in my experience with that scanner I would get an Ingenia. On Philips scanners you have access to a lot of parameters without actually having to do pulse programming.

There are now 7T Systems available. Best L.D.

7T systems are currently 4-6 times as expensive as 3T systems. The SNR improvement from 3T to 7T could in many cases be reached by scanning for longer time, remember 3T scanning time is 4-6 times cheaper. In many cases subject motion will be the limit for spatial resolution, and in that case I would spend my money on better motion tracking first. With 7T whole body gradients now being outperformed by 3T gradients, and T2 and T2* being smaller at 7T, it is not at all obvious to me that 7T is better for diffusion, and maybe not even for fMRI.

Wow... the Prisma looks bad-ass! Hadn't seen that before. Looks like it might be the best research system around at the moment.

Interesting thougsts, Torben. In principle, the higher field strength would allow for smaller voxels, which would result in reduced intravoxel dephasing, which is certainly a good thing in fMRI. However, if the gradients can't do single shot EPI at these high resolutions and at the reduced TE, the advantage is gone. I guess some of the gained SNR could be spent on higher SENSE factors. Does anyone here have experience with fMRI at 7T?

I can only speak about the field of neuroimaging. Both manufacturers offer very good imaging quality. If you need to cutomize your imaging sequences a lot e.g. for clinical studies or research, the philips operating system will probably perform better. On the other hand, if you are doing al lot of routine imaging, the siemens operating system may be more easy to handle for your technicians.

Once the protocol for a study is set up on our Philips Achieva, it becomes a routine imaging study and the technologists don't need to worry about the imaging parameters any more.

Personally, all brands have their strengths and weeknesses. At the end of the day I think it all depends on who is your closest service provider and how quickly they can be on site. Magnets need lots of upkeep. Your going to want access to them.

Specifically for brain and research, the new Prisma (Siemens) can compete with the Ingenia (Philips)

Elsewhere in the body and in a clinical setting, Ingenia (Philips) has clearly an outbreaking technology superior to the rest of options

I have been thinking about an MR with PET capabilities and I can't really come up with anything that makes such a machine a must have. I am sure this is mainly due to my ignorance. Does anyone know a good reason for buying a PET-MR?

We have an old non-Tim Trio that will soon need replacing. Personally I would like to know if the Toshiba 3T with Pianissimo gradients gives any DTI improvement, at present the vibration of our current system is limiting and I wondered whether the gradients enclosed in a vaccuum reduced the vibration associated with the sequence. The quiet scans would certainly benefit some of our fMRI research interests.

The last time (2 years ago) when I was shown EPI images from the 3T Toshiba, I was not impressed. There are no specs on the gradients apart from being silent. The Prisma has a new construction where the patient table is mechanically uncoupled from from the gradient coils, much like what you get with the anti-vibration kit for the trio. And for diffusion it should be a perfect match given the strong well cooled gradients.

COI disclosure: I have a research relationship with GE Healthcare.

From a practical standpoint, as others noted, much depends on what you wish to do on the scanner, service arrangements, and your relationship with the vendor. Capabilities of the scanner and specifications are obviously important too. I will echo that all the vendors have their strengths.

MR technology has competing demands. In the late 80's we had local head gradient coils that provided improved performance to enable fMRI to be done. See attached Eric Wong paper. Gradients within systems improved and obviated the need for the insert coil. In recent years patient comfort and accessibility stimulated development of wide bore scanners on the order of 70 cm. Gradient performance was reduced as a result limiting the number of DTI slices you can acquire in a given TR and the minimum slice thickness. The Prisma suggested by Dr. Lund is a return to the 55 cm bore, which was the diameter of my first 3T back in 2001, a GE 3T94 that ushered 3T into clinical practice. While the performance was admirable, 55 cm was responsible for turning many patients away because they simply didn't fit into the scanner. That being said, the benefits of better gradient performance in functional MR techniques is worthy of pursuit. Dennis Parker is actively developing insert gradients technology at the University of Utah that may return better performance to wider bore instruments. http://www.ucair.med.utah.edu/FacultyParker/ParkerResearch/Gradients.html

Dr. Parker's work is on Siemens and I am not aware of any commercial availability of his insert hardware.

We have a new neurosciences clinic under construction. DTI, fMRI, spectroscopy, and a complete array of whole body capabilities are going to be needed on this scanner. The GE MR750 provides a great balance of gradient performance, clinical innovation, and patient comfort. Our service relationship and clinical research partnership also factor in to my advocacy for this scanner.

I am not sure why Dr DeLano thinks that the Prisma is a 55cm bore. It says 60cm here: http://healthcare.siemens.com/magnetic-resonance-imaging/3t-mri-scanner/magnetom-prisma/technical-details

I know that the conectom scanner at MGH which is a Skyra with very aggressive gradients is below 60cm, but the Prisma is not as far as I can tell from the information available at this time!

Having said that the MR750 is a decent scanner when it comes to magnet and gradients, but there is room for improvement when it comes to coils and user interface.

For fMRI, DTI and MRS a large bore will always be a compromise. For patients 70cm is probably a wise choice, but for healthy volunteers I would definitely go for 60cm.

It is probably a competition between Siemen's PRISMA and a Philips Achieva with the d-stream (Ingenia) technology, with the Siemen's gradient performance a bit better and the Philips more flexible and better for patients and whole body imaging. Philips 32 ch head coil is better than the Siemens one but I haven't compared it with the new 64 ch one that Siemens have for the PRISMA. I gather that is a head-neck coil in any case so not sure whether the 64 channels are really 2 x 32 or not. The Ingenia is a wide bore system but the gradients are not as big as those of the Achieva, so to compare eggs with eggs you need to consider the Achieva with the digital upgrade vs the PRISMA. Both are pretty new so there isn't much comparative data available as yet.

Dr. Douraied Ben Salem, could you please explain better your answer?

Which is the problem for neuroradiology practice with the Skyra scanner?

This still doesn't answer why?

I would definitely go for a Siemens Prisma or Siemens Skyra! The coil concept of Siemens is superior to other vendors, especially if your "region of interest"/pathology expands over a large anatomic region. Coil-coverage will never be a problem with Siemens systems, and image quality is not compromized.

I have almost two years of experience with Skyra, scanning all parts human anatomy. It has been, and still is a pleasure!

Thank you everyone for the amazing feedback. This has been extremely helpful! Over the next few weeks I will collect sample data sets on the Ingenia, GE 750w, Prisma (and possibly Skyra). I will also be looking at hybrid PET/MR systems (GE, Siemens). Some of what I learn will be covered by NDAs, but I will share my general impressions here.

I will attempt to collect equivalent data for the following on each platform:

=neuro=

3D T1 weighted 1mm isotropic

3D T2 weighted 1mm iso

resting state BOLD

DTI, 30 (32?) directions

ASL

T2 Flair

SWI (if possible)

Phase Contrast

MRS (1cm iso voxel in the medial prefrontal lobe)

=cardio= (for a colleague)

stock Cardiac Sequences, SE w TR 2 heartbeats, (can do at 1 heart beat?)

CINE flash segmented k-space TR 7-8ms TE 3.5-4

true FISP TR 3.5 TE 1-1.5

breath hold IPAT 5 seconds

one short axis, one long axis

coronary artery image

FLASH

PSIR myocardial delayed contrast enhancement

velocity sensitive sequence

non-contrast renal artery, PCA

TOF of legs, no contrast

tagged MRI beating heart

it is unlikely that I will be able to collect all of this data in the time I have on each system, so I need to prioritize and eliminate some of the scans. The scans will be done on a volunteer. I will try to use the same volunteer at each site, but that currently seems unlikely. I will also collect some reference data using a phantom. Probably the ACR phantom.

Comments appreciated!

Well we have both a Trio and a Skyra (32ch head coil for both machines) and I would prefer the Skyra over the Trio any day as it provides much better images (higher SNR) and even offers new sequences not available with the Trio, e.g. ZoomIT.

So far we had no problems with the 32ch head coil, maybe the problems Dr. Douraied Ben Salem is talking about have been resolved in the last two years.

I would opt for PRISMA.

I have used TRIO TIM for many years including 32 ch head coil, and found it powerful and h flexible. Have not done cardiac but the liver work was good quality once we work around B1 problems.  The pelvic work prostate and cervix was a big surprise where we could use some extra SNR to achieve higher spatial resolution.

The PRISMA attracts me now because it ahs a health long magnet, we would likely refit the TRIO magnet.

Z direction FOV and Z direction homogeneity is far more important to y imaging throughput that a wider bore.

When Z is compromised throughput and flexibility falls,. We learned this with the  SONATA 15 years ago, I am sad to see the obsession with short Z magnets again limiting performance, particularly in hospitals where full spine and long bone exams are common.

I should say I don't have experience with TIMCT to try and address that issue.

80 mT/m, is also an attraction for HARDI tractography. I hope they offer a better on console tractography package soon, as DTI has declared its serious shortcomings for clinical and research applications

Though my previous post some time ago favored GE, after serious consideration I went with the Prisma.  Parallel transmit reduces B1 issues, better gradients reduce distortions on EPI and should give better echo spacing on everything, better coils facilitate better acceleration factors, and better software/multi-band and zoom it  for our DTI/DSI and fMRI work tipped the scales.  The scanner is impressively quiet.  We go live clinically soon--ran into a snag with a flood in the room of the new building it is going into.  The 64 channel head coil is 40 head/24 neck elements and there are 12 anterior and 12 posterior elements.   Haven't had the opportunity to use it and will repost after a bit of experience.

I received a lot of good feedback from this RG question. It led me down a long path which has ended in a very interesting way: by next year our institution will have recent 3T scanners from GE, Philips, and Siemens. We should be in a good position to do some direct comparisons.

Basically, before you purchase a system, if possible, test your protocols at a site that has the scanner and get sample data from the factory. That being said, here are musings based on my search for the right scanner.

Obviously it turned out to be much harder than I expected to directly compare scanner performance or to answer a naive question like "Which scanner is better?" After talking to a lot of people and playing with a lot of scanners, here is my subjective opinion and some observations:

For human imaging, GE, Philips, and Siemens all have great MRIs. But, I really like Siemens the best. However, an MRI can be like a musical instrument: you want a good one, but the person playing it is what really matters. (2) The facility around the MRI matters: setting up the facility to make it easy for investigators to do their work and facilitating development work can be more important than which scanner you get. I believe that ease of access and reducing administrative burden is important for good science. (3) Building a good relationship with the vendor is very important. Ideally, you want someone from the vendor on site as much as possible.

Here is how things turned out at our institution: I posted this question because we needed to replace an aging, head only, research dedicated MRI. We decided on a Siemens Prisma and it was installed Fall 2015. In a separate project made possible by a generous donor, we purchased a 3T GE PET/MR system which is based on the 750w platform. It was installed Fall 2015. The GE PET/MR is an amazing machine and a true work of art on the inside. Really mind blowing. Finally, we already had 2 Philips Ingenias on the clinical side, which we use do use for research. They work well for what we do.

I intend to update this thread occasionally as we have some experience with the new scanners. 

Thanks for your quick reply, I read the information to your posted question 6 years ago but I need an updated answer and with a nuance. Our new equipment will be especially dedicated to cardiac imaging but also we,ll perform studies of other anatomical areas. The priority is the cardiac imaging with clinical and research purposes