Ex vivo high resolution MR imaging of the human lip with a dedicated low field system

ARTICLE

Almost all organs and tissues, in many species, have been explored by magnetic resonance imaging (MRI) since its inception. However, to our knowledge, there is no published matter specifically regarding magnetic resonance (MR) images of lips. The size of the different tissues lying in the thickness of the lips requires successful high resolution MR images. This is the same requirement as for skin MRI [1, 2]. To this end, we use a low-field (0.1 T) dedicated MRI system. 0.1 T is an uncommon field value for MR imaging near a microscopic scale although it has already been used for skin imaging [3]. We show the MR images obtained, one being correlated with a histological section, and discuss the different contrasts obtained.

Materials and methods

We used a dedicated low-field MRI system, based on a water-cooled variable electromagnet (Drusch, Poissy, France) creating a main magnetic field of 0.1 T [4]. Maximum gradient strength is 20 mT/m (rising time of 0.5 ms). The orientation of B₀, perpendicular with respect to the sample access, allows the use of solenoidal transmit-receive radio-frequency (rf) coils. In order to achieve a good filling factor with respect to the sample, an rf coil of 2.3 cm diameter, 1.7 cm length and 6 turns made with 2 mm diameter copper wire was built. The rf coil was tuned at 4.32 MHz with 3 fixed non-magnetic capacitors (American Technical Ceramics, New York, USA) in parallel (total capacitance of 1,890 pF). The imaging software running on a PC-based spectrometer (MR 3030, SMIS, Guilford, Great-Britain) has already been developed in our laboratory [5].

A portion (between the midline and the angle of the mouth) of the inferior lip of a male cadaver was set apart. The cadaver was embalmed by femoral cannulation using a solution containing phenic acid, formol, alcohol, glycerin, chloral hydrate and conserved in alcohol solution. After MR imaging, it was processed to 5 µm thick paraffin sections, stained with trichrome technique or hematoxylin and eosine. Digitized photographs of the histological sections were obtained for further comparison with MR images.

All MR images were acquired with 3D-FLASH sequences, T₁ and T₂* weighted (see details of the sequences in figure legends). Asymmetric acquisition matrix was used due to the shape of the sample and to reach a good compromise between signal to noise ratio and resolution.

Both histological and MR images were processed on a Macintosh Centris 650 (Apple, Cupertino, USA) using the public domain NIH Image program (developed at the US National Institutes of Health and available on the Internet at http://rsb.info.nih.gov/nih-image/). In particular, scaling of the histological sections was done. This allowed us to take the exact measurement of structures and to compare them with measurements made on MR images. In the following, no indication means that the values given were measured on histological sections and were identical with values found on corresponding MR slices.

Results

Both skin, red lip and oral mucosa are classically described by their 3 constituent layers [6]. The superficial and thin layer is a stratified squamous epithelium. The epidermis disappears on the red lip and the oral mucosa [7]. The dermis is divided into the papillary layer, close to the epithelium, and the deeper reticular layer. The dermis of the red lip is characterized by a typical micro circulation [8]. Below the epithelium of the oral mucosa, a thick lamina propria, in continuity with the dermis, contains labial glands in a loose connective tissue. The last layer is the hypodermis, or the deeper part of the lamina propria. In the skin, minimal fat between the musculus orbicularis oris and dermal planes is characteristic of the structure of the lip, although there are close attachments between the skin and perioral muscles [9]. The vermilion border is a particular kind of tissue joining the skin and mucosa, outside and inside. At the limit between the epidermis and red lip (a transitional zone), the epithelium becomes thicker and progressively less keratinized [7], the orbicular muscle arises in the deep surface of dermis.

In Figure 1, we present an histological section compared with the corresponding MR image (T₂* weighted) made at the same location. The epidermis is too thin (about 40 µm) to be visible on the MR image. Dermis (1 mm thick) along with epithelium and the superficial layer of lamina propria of oral mucosa (2 mm thick) both gave high signal intensity relative to the other structures. This is also the case of the vermilion border. The classic distinction between papillary and reticular dermis is not quite evident in the lip. Deep parts of lamina propria of oral mucosa and hypodermis have similar appearances of irregular structures at low signal intensity. Most muscular fibres could be found in the hypodermis, marked by lines of high intensity, and labial glands are present in the lamina propria. Musculus orbicularis oris (about 2.5 mm thick on histological section vs 2 mm on MR images) gives a high, uniform signal.

In Figure 2, more T1 weighted yield almost an absence of contrast between structures easily visible in Figure 1. However this allows the clear delineation of the inferior arteria labialis (internal diameter about 550 µm) and particularly of a perilabial gland vascular network (diameter about 450 µm, measured on MR image) which could be followed on 4 contiguous slices.

Discussion

We show that high resolution images of lips are able to fit histological sections made at the same location even at a low field of 0.1 T. The relative obliquity between MR slices and histological ones, and the thickness of MR images could lead to difficulty in finding exactly the same location for the same structure. This is particularly the case of labial glands, which are hard to identify on MR image (only one is clearly identifiable out of 3) in Figure 1.

The outer layer of mucosa gives a high signal similar to the second skin layer (dermis). Likewise hypodermis and deep parts of lamina propria are of similar low signal intensity. These appearances may be explained by similar tissular composition, but also by similar mechanical function (gliding zones). When following the periphery of the MR image, the continuous white layer corresponds successively to the dense connective tissue of the dermis in the cutaneous part of the lip, to the red lip and finally to the epithelium and superficial part of the lamina propria of oral mucosa.

These results are different from previous studies regarding the MR appearance of skin: notably the lip dermis appears like a high signal intensity layer when compared with skin dermis studies made at a higher field (1.5 T) [1, 2]. Due to the high collagen content of these layers, a very short T₂ is expected, but the chemical content of skin, notably collagen content, may also vary with location; and vascular structures [8, 10] are different between calf, heel (location of skin images in [1, 2]) and lips. In particular, the hypodermis has less fat content than in other locations [9].

Measurements on histological sections accord well with those on MR images except for the musculus orbicularis oris. The thinner appearance on MR images may be related to partial volume effect at the boundary of the muscle bundles.

Lips, explored by ultrasound, are a subject of interest for fetal malformation diagnosis [11] or dental investigation [12], for example. Their exploration could take place in decisions in reconstructive surgery [9], or in dermatology. The ability to make precise measurements on MR images acquired with variable obliquity could lead to potential surgical applications, for facial rejuvenation or reanimation.

CONCLUSION

Acknowledgements

The authors thank Mr. VETTER for his technical assistance and Apple Computer France for its grant.

REFERENCES

1. Richard DS, Querleux B, Bittoun J, Idy-Peretti I, Jolivet O, Cermakova E, Lévêque JL. In vivo proton relaxation times analysis of the skin layers by MRI. J Invest Dermatol 1991; 97: 120-5.

2. Song HK, Wehrli FW, Ma J. In vivo MR microscopy of the human skin. Magn Reson Med 1997; 37: 185-91.

3. Yassine MM, Darasse L, Saint-Jalmes H, Sauzade M, Querleux B, Lévêque JL. In vivo skin imaging at 0.1 Tesla. Book of abstracts: Sixth annual Meeting of the Society of Magnetic Resonance in Medicine, Vol. 1. Berkeley, CA:SMRM, 1987: 466.

4. Gries P, Constantinesco A, Ortlieb C, Pittard S, Brunot B, Vennart W, Demangeat J, Ellis R. Développement d'un imageur de RMN à 0,1T pour la main et le poignet. J Med Nucl Biophys 1990; 14: 343-8.

5. Arbogast-Ravier S, Xu F, Choquet P, Brunot B, Constantinesco A. Dedicated low-field MRI: a promising low cost technique. Med Biol Eng Comput 1995; 33: 735-9.

6. Bloom W, Fawcett DW. A textbook of histology. Philadelphia: W.B. Saunders Company, 1968: 479-90.

7. Krstic RV. Human microscopic anatomy. Berlin: Springer Verlag, 1991: 162-3.

8. Wolfram-Gabel R, Sick H. Microvascularization of the lips in the fetus and neonate. Cells Tissues Organs 1999; 164: 102-11.

9. Gosain AK, Amarante MT, Hyde JS, Yousif NJ. A dynamic analysis of changes in the nasolabial fold using magnetic resonance imaging: implications for facial rejuvenation and facial animation surgery. Plast Reconstr Surg 1996; 98: 622-36.

10. Crouzet C, Fournier H, Papon X, Hentati N, Cronier P, Mercier Ph. Anatomy of the arterial vascularization of the lips. Surg Radiol Anat 1998; 20: 273-8.

11. Sivan E, Chan L, Mallozzi-Eberle A, Reece EA. Sonographic imaging of the fetal face and the establishment of normative dimensions for chin length and upper lip width. Am J Perinatol 1997; 14: 191-4.

12. Prabhu NT, Munshi AK. Ultrasonographic observation of the circumoral musculature: an in vivo study. J Clin Pediatr Dent 1995; 19: 195-203.