Analysis of the lower limb muscles on MRI

Below are listed some tips on the tricker aspects of analysing the lower-limb muscles on magnetic resonance imaging (MRI). It essentially details the approach we have used in prior studies of lower-limb muscle atrophy in prolonged bed-rest.

I have focussed on the main difficult things in the MR image analysis. If you are starting to learn how to do MR-image analysis, use of the following text book:

Harold Ellis (1994) Human cross-sectional anatomy: pocket atlas of body sections and CT images. CRC Press (13 July 1994). 180 Pages. ISBN-10 0750620285, ISBN-13 9780750620284

..will be of invaluable help in guiding your analysis. There does appear to be a new version of this book out which I haven't looked at yet: Bari M Logan, Adrian Dixon, Harold Ellis (2009): Human Sectional Anatomy: Pocket Atlas of Body Sections, CT and MRI Images, Third Edition. Hodder Arnold (25 September 2009) - 288 Seiten. ISBN-10 034098516X, ISBN-13 9780340985168.

You can download a PDF-version of the following information here: Download PDF

Magnetic resonance imaging analysis of the lower limb muscles

Hamstrings

There are four different members of hamstrings which can be readily separated on MRI: semimembranosus (SEMI_M), semitendinosus (SEMI_T), biceps femoris long head (BF_LH), biceps femoris short head (BF_SH). The images here help to describe where the muscles typically start and their anatomical relationship to one another.

Below: Semitendinosus (SEMI_T) typically starts at about the level of the lesser trochanter

Below: Biceps femoris long head (BF_LH) comes in as we move more distally. Sometimes you need to be careful that you do not mistake gluteus maximus (G_MAX) with the start of BF_LH. On the right side of the image the border between these two muscles is indicated. Measuring BF_LH before you measure G_MAX will help to avoid any problems.

Below: Semimembranosus (SEMI_M) starts as a very small triangle of muscle next to semitendinosus. Note that on the right side of the image, the SEMI_M is not quite yet visible.

Below: the start of biceps femoris short head (BF_SH). This muscle starts as a small “sliver” next to the vasti (V). It is important to measure this muscle before the vasti because you will otherwise sometimes by mistake include the very proximal portion of BF_SH with the vasti.

Below: the border between BF_SH and the vasti has been indicated with arrows (zoom of the previous image).

Below: the members of hamstrings distally. The demarcations between the muscles should be easy to see. It becomes even easier to see when you scroll from one image to the next in your “stack”. It is sometimes difficult to find the border (indicated by the three arrows) between adductor magnus and semimembranosus (SEMI_M).

Quadratus femoris and the obturator internus/gemelli complex

It can be very difficult to differentiate quadratus femoris from the “obturator internus/gemelli complex”. It is sometimes difficult to know where the inferior gemellus ends and the quadratus femoris starts. As a general rule, if the femoral neck is still completely visible, then quadratus femoris is not (yet) present. Based upon experience, you will however get better at working this out.

Below: the obturator internus (OI) is shown with its tendon and the gemelli. The head of femur (HOF) is also marked. OI is on the internal aspect of the obturator foramen – where you can see the bulge where OI is marked. However, it is typically not possible to know where the OI stops and the gemelli start. So we have usually measured these muscles together.

Below: the femoral neck (NOF) is still complete in the image, so the “lump” below the femoral neck is still most likely the inferior gemellus

Below: here the femoral neck is no longer visible and – if you were able to scroll up and down in this data set – you can be convinced that you are now looking at quadratus femoris and not the inferior gemellus. Note that the distal part of OI is still visible. The arrows mark the border between quadratus femoris and obturator externus.

The distal end of quadratus femoris, obturator externus and adductor magnus

Once the lesser trochanter is no longer visible in the image, then quadratus femoris is usually gone. Deciding where obturator externus ends and when adductor magnus starts is often not clear. To ensure consistency, we implement the rule: once the lesser trochanter is no longer visible, then this is where adductor magnus starts. The slice beforehand is the last measurement of obturator externus. In some data sets, it would make sense to combine these muscles

Below: The lesser trochanter is quite prominent. Obturator externus (OE) is easily differentiated from the adductor brevis (ADD_B) and quadratus femoris.

Below: here the lesser trochanter is more or less gone. The distal end of quadratus femoris can still be seen (more obvious if you are able to scroll through the data set). Is the muscle mass between quadtratus femoris and ADD_B obturator externus (OE) or adductor magnus (ADD_M)? To ensure consistency, we would apply the rule here that this is the start of ADD_M, since the lesser trochanter is more or less not visible any more. Again, you will have to decide for your own data whether it makes sense to try to separate OE and ADD_M.

Pectineus, adductor brevis and adductor longus

Differentiating these muscles at their origins can be tricky. If you have good quality MR-data, such as the images here, then you won’t have too much of a problem. However, it is not always so easy.

Below: here you can see the start of pectineus on the superior pubic ramus.

Below: the adductor brevis (ADD_B) starts in a small “triangle” wedged between pectineus and obturator externus (OE).

Below: as you move more distally, pectineus “moves” laterally on these axial images. Adductor brevis (ADD_B) “grows” in its wedge between pectineus and obturator externus (OE). The adductor longus (ADD_L) can be seen starting “on top” of adductor brevis.

Below: The adductor longus (ADD_L) is getting “bigger”. Pectineus has “moved” even more laterally towards its insertion on the pectineal line below the lesser trochanter. The border between pectineus and the iliopsoas (arrows) is difficult to see on this image, but becomes more obvious if you can scroll through the images.

Below: even more distally, it is easy to see the different parts of the adductors

Lower-leg

A few points:

• The lower leg is (usually, after a bit of practice) fairly easy to measure on MR-images. It will of course be affected by image contrast, resolution and what kind of sequence you have, however.
• Soleus is the probably the trickiest to differentiate from the other muscles. This is mainly because if you measure soleus before you measure flexor hallucis longus, gastrocnemius lateralis, peroneals and popliteus, you will often accidentally include parts of these muscles with soleus.
• To help to explain the image measurement, you can download an example data set - to be opened in ImageJ - from here. (Note: The DICOM header has been stripped from file). I used this example data set for generating the images below.
• The process I use involves "drawing" the region of interest for each muscle on the image once I have measured it. Then you can use this in later meaurements to know where you have already measured.

I measure the lower-leg in the following order:

1) Popliteus: this is (should be) very easy to see in the example data. The key is to just measure it first so that you don't mix it up with soleus. Or, if you are not interested in popliteus, maybe mark it some way on the image to ensure you don't accidentally include it.

2) Flexor hallucis longus: scroll down to the end of the muscle and then follow it back up. You should be able to follow the border of the muscle higher up to where it originates.

The image below shows slice 45 of the example dataset. The red arrows show the border between flexor hallucis longus and soleus

3) Peroneals: these are pretty easy to differentiate when you start proximally and move distally. Just as some points knowing where the border between peroneals and soleus is can be difficult

The image below is slice 31 of the example data set. The red arrows show the border between soleus and peroneals. If you follow this up and down the muscle in the example data set it should be pretty obvious.

4) Soleus: the real trick now is differentiating soleus from gastronemius medialis. Proximally, it can be VERY hard to tell what is soleus and what is gastrocnemius lateralis.

The main thing to note anatomically is that soleus starts just below the proximal fibular head. So, this is a useful landmark to work out where you should start.

Image below is slice 9 of the example data set. The traced area is where I would do my first measurement of soleus. If you follow it distally, you will see how it "expands" into soleus.

Proximally, you could argue that it is some of it is soleus. But, they are hard to differentiate. For the sake of consistency, and to stick to what we see in anatomy texts, I start just below the head of the fibula where I can be fairly certain. Anything above this point will be included with gastrocnemius lateralis.

5) Gastronemius medialis: the fascial border between soleus and gastrocnemius medialis is almost always very easy to see.

6) Gastrocnemius lateralis: once you have soleus done, this becomes easy.

7) Tibialis posterior: The border between FDL and tibialis posterior can be difficult to see. Depending on your image quality, you may not be able to see it (solution here = pool the muscles).

This image is slice 32 in the example data set. The red arrows show the fascial border between tibialis posterior and flexor digitorum longus. You can follow this proximally and distally.

8) Flexor digitorum longus: Once tib.post and the other muscles are done, this should be easy.

9) Anterior tibial muscles: Its up to you to decide whether to measure tibialis anterior with or without extensor digitorum longus. I can be hard to differeniate the two and I'd suggest just to pool them.