RADIOGRAPHIC DETECTION OF DEFECTS OF THE NASAL BOUNDARIES

TOBIAS SCHWARZ, MA, DR. MED. VET., MARTIN SULLIVAN, BVMS, PHD, KLAUS HARTUNG, PROF. DR. MED. VET.

The sensitivity of conventional radiography for lesions of the cribriform plate, naso-orbital wall, lateral nasal wall and hard palate was investigated in 13 canine cadaver heads by creating measured defects in these structures. The location of the perforations were marked with a thin copper wire and the radiographic appearance of the defects was evaluated retrospectively by a single reviewer. Despite demarcation cribriform plate destruction of 2 mm could not be detected. Defects of 3 mm were detected in only 2 heads, 4 mm defects in 1 further head, an oblong 4 x 10 mm defect in 7 heads and in the remaining 3 heads only a 10 x 10 mm defect became visible as such. The naso-orbital wall had to be destroyed in its whole vertical length for detection of a defect in ventrodorsal or dorsoventral views. Therefore conventional radiography is of low diagnostic value for these lesions. Defects of 2 mm in the lateral nasal wall and the hard palate could be detected confidently in all heads indicating high sensitivity of conventional radiography. Soft tissue opacification did not alter the detectability of any nasal border structure lesion. Veterinary Radiology & Ultrasound, Vol. 41, No. 2, 2000, p p 226-230.

Key words: lamina cribrosa, cribriform plate lesions, conventional radiography, canine.

Introduction

ASAL BOUNDARIES such as the cribriform plate, lateral N nasal wall, hard palate and naso-orbital wall are key structures in the assessment of the extent of chronic canine nasal disease. This is important for prognosis and treatment planning. It is useful to define the sensitivity of conven- tional radiography for detecting defects in nasal boundaries for those clinicians who do not have access to cross- sectional imaging modalities. It is also important to have a clearer understanding of the capabilities of conventional ra- diography for future comparative studies of different imag- ing modalities in nasal disease.

The radiographic anatomy of the cribriform plate (Lamina cribosa) was described for different skull types and projections.' The purpose of this study was to investigate the sensitivity of conventional radiography to detect defects of the cribriform plate, naso-orbital wall, hard palate and the lateral nasal wall.

Material and Methods

Thirteen formalin-fixed canine cadaver heads from a pre- vious study' were used. The canine heads were classified anatomically using skull indices as described previously.

From the Department of Veterinary Clinical Studies, Faculty of Veteri- nary Medicine, University of Glasgow, Bearsden Road, Glasgow G61 1 QH, Scotland (Schwarz, Sullivan) and the Department of Radiology, Clinic for Equine Diseases, Veterinary Faculty, Free University Berlin, Oertzenweg 19B, 14163 Berlin, Germany (Hartung).

Address correspondence and reprint requests to Dr. Schwarz at the ad- dress above.

Received May 19, 1999; accepted for publication September 17, 1999.

Five heads were represented in the intermediate mesati- dolichocephalic skull range (group I ) and 4 each heads each in the mesaticephalic (group 2) and intermediate brachy- mesaticephalic skull ranges (group 3). No extreme brachycephalic or dolichocephalic skull types were repre- sented in this study.

Lesions of Nasal Border Structures

The eyes and orbital muscles were partially removed to allow defects to be created in the naso-orbital wall. The brain was removed to allow defects to be created in the cribriform plate. The hard palate, the lateral nasal wall, the cribriform plate and the naso-orbital wall were perforated bilaterally with holes of increasing size (2 , 3, and 4 mm) using a drill and bit. The skulls were sequentially radio- graphed until a defect was visible. The perforations were always created slightly asymmetrically to avoid superimpo- sition of bilateral defects on lateral radiographs. All defects were marked with thin copper wire for radiography. The cribriform plate perforations were subsequently joined to- gether with a milling head resulting in a perforation with an area of 4 x 10 mm. This corresponded to approximately 50% of the plate area. Lastly, the perforation was enlarged to 10 x 10 mm, corresponding to 75-100% of the plate, depending on the size of the dog.

In the naso-orbital wall the 4 mm holes were enlarged by milling a 4 x 10 mm vertical shaft which corresponded to approximately 50% of the vertical extent of the naso-orbital wall. Finally, this shaft was lengthened to the total vertical extension of the naso-orbital wall.

Radiographic projections are as previously described.

226

VOL. 41, No. 2 NASAL BOUNDARY DEFECTS 227

For each hole size dorsoventral (DV), dorsol0"rostal- ventrocaudal oblique (D10"R-VCdO), dorso10"caudal- ventrorostal oblique (D1O"Cd-VRO), ventrodorsal (VD), ventro10"rostral-dorsocaudal oblique (V10"R-DCdO), ventro1O"caudal-dorsorostral oblique (V1O"Cd-DRO) and lateral radiographs were made. In one head of each group rostrocaudal projections of the nose were taken for further assessments of defects in the cribriform plate.

Sojl Tissue Opacification

Soft tissue opacification of the right nasal cavity was mimicked by injecting ultrasonographic gel into the perfo- rations until it appeared at the nostril.

Radiographic Assessment

Radiographs with wire marked perforations were as- sessed retrospectively by a single reviewer for the presence of lesions using the following criteria. Radiographic evi- dence of a defect in the cribriform plate was assumed if interruption of the bone-opaque stripe was present at least unilaterally (for 2, 3, and 4 mm holes) on at least one of the DVND and associated oblique views. Lateral views were reviewed for cribriform plate destruction but not included in the assessment as interruptions of the plate can be seen in normal heads in this view.' To avoid false positive results, areas of the cribriform plate that had a decrease in bone opacity but without clear interruption, were disregarded. Destruction of the naso-orbital wall, the lateral nasal wall and the hard palate was assumed if a radiolucent halo was present around the wire.

Results

Lesions of' the Cribrijorm Plate

The smallest detectable lesion of the cribriform plate was 3 mm in diameter (2 of the 13 heads). In one head the 4 mm lesion, in seven heads the 4 x 10 mm defect and in three heads only the 10 x 10 mm lesion were the smallest detect- able defects (Fig. 1). On lateral views large holes became visible as osteolucencies (Fig. 2). The defects were best seen when the nasal cavity was slightly overexposed (all views) .

In group 1 and 2, 10" vertically tilted views did not im- prove the detection of defects. However, in short-nosed dogs of group 3 superimposition of frontal bone on the cribriform plate masked the cribriform plate and the lesions on the DV or VD view. A 10" vertically tilted view was necessary to identify the true cribriform plate and defects. Apart from that there was no significant difference in the detectability of the lesion between the different groups of dogs. Although defects were created bilaterally, one lesion was always more pronounced than the contralateral one on an individual view (Fig. 1). Views of different obliquity highlighted the contralateral lesion. The appearance of le-

FIG. 1. Close-up VD view of the cribriform plate in a mesaticephalic canine head with a bilateral 4 mm perforation. The diameter of both per- forations is marked by wire but an osteolucent defect is only visible on the left (arrow).

sions in mirror projections (VD = DV, DR-VCdO = VCd- DRO, VR-DCdO = DCd-VRO) was similar.

The diagnostic value of the rostro-caudal view for the cribriform plate was evaluated in one head of each group. The sieve-like structure of the cribriform plate contributed minimally to the trabecular pattern of the image, as after total removal of the cribriform plate the pattern was un- changed. Lesions were not visible as such (Fig. 3).

Lesions of the Lateral Nasal Wall and the Hard Palate

Perforations of 2 mm in the lateral nasal wall were always visible in lateral and vertical projections, except if the ca- nine tooth overlaid the defect. Perforations of 2 mm in the hard palate were always visible in vertical projections (Fig. 4) but not in lateral views.

FIG. 2. Close-up lateral view with bilateral 4 mm perforations of the cribriform plate (straight wires), naso-orbital walls (circular wires) and hard palate (wavy wires) in a mesaticephalic canine head. Only one defect is visible in the cribriform plate (arrow) whereas all other lesions are invisible.

228

A

SCHWARZ ET AL.

B

2000

FIG. 3. (A) Rostrocaudal view of the nose. Vertical and horizontal wires positioned within the rostral cranial cavity mark position and extent of the cribriform plate (arrowheads). A trabecular pattern is present bilateral. (B) Rostrocaudal view of the nose after total destruction of the cribriform plate. Vertical and horizontal wires mark position and extent of the area of plate destruction (arrowheads). Trabecular pattern is still present bilateral. This pattern is created by nasal conchae. Marked soft tissue opacification on the right does not obscure turbinate pattern.

Radiographic Description of the Naso-Orbital Wall and Naso-Orbital Wall Lesions

The naso-orbital wall could only be localized radio- graphically by marking wires. The wall was a part of a tangential line in VD/DV projections marking the lateral border of the caudal nasal cavity (Fig. 5). The main part of this very prominent radiographic structure was formed by the ventrally situated lateral wall of the choanae and dor- sally by the cranial cavity. This was proven in one dog by removal of these structures and subsequent radiography. The naso-orbital wall had to be destroyed in almost all dogs (10 of the 13 heads) along its entire vertical length (4 x 20 mm shaft) before a defect became detectable on vertical views. In 3 heads the defect became visible with a 4 x 10 mm shaft. Oblique VD views (V10"R-DCd0, V1O"Cd- DRO) were needed to see the defect as they separated su- perimposed bone structures of choanal and cranial cavity walls (Fig. 5). In DV views the mandible was partially superimposed on the area of interest. On the lateral view the defect was visible, but only as an indistinct osteolucency in the caudal nasal region.

Soft Tissue Opacijication

Moderate opacification of the affected nasal cavity was achieved by filling the cavity with ultrasound gel. The tur- binate and ethmoid nasal pattern was similar to the contra-

lateral unfilled nasal cavity. The soft tissue opacification did not hinder the detectability of nasal border structure defects (Fig. 5). Soft tissue opacification appeared most pronounced on the rostrocaudal view but did not obscure the turbinate nasal pattern (Fig. 3).

Discussion To our knowledge there is only one study where the

sensitivity of conventional radiography at detecting defects of the cribriform plate was assessed.2 In that study the crib- riform plate was penetrated with holes of 3 and 5 mm and conventional radiography was compared to linear tomogra- phy and CT. No details were given on the detectability of individual 3 and 5 mm defects. In our study the reviewer had the benefit of wire demarcation of all nasal border struc- ture lesions in order to evaluate the potential of the radio- graphic technique rather than interpretative performance.

In a study including eight patients with neoplastic de- struction of the cribriform plate and cerebral invasion diag- nosed by CT, cribriform plate destruction was detected by conventional radiography in only Other authors de- scribed CT detectable osteolytic changes in the rostral cra- nial cavity of seven patients, but corresponding osteolucen- cies could be confirmed by conventional radiography in only Similar disparity was reported in another study.5 In most of the dogs in these studies aggressive nasal disease was advanced.

VOL. 41, No. 2 NASAL BOUNDARY DEFECTS 229

RC. 4. VD view of the nasal cavity with 2 mm perforations in the lateral nasal wall and hard palate (white arrows) marked by wires. Lesions are visible as circular osteolucent defects. A bilateral 4 mm perforation of the cribriform plate is not visible on the right and only as an area of slightly decreased opacity on the left (black arrow).

In our study it appeared that lesions of the cribriform plate have to be large before they become radiographically detectable. In the majority of the specimens at least half of the plate area had to be destroyed before a defect became detectable, confirming that conventional radiography has a low diagnostic value for evaluation of potential cerebral involvement of aggressive nasal disease. Superimposition of overlying bone structures in short nosed dogs and changes in appearance of cribriform plate lesions in views of different obliquity caused further limitations on the de- tection of plate defects. Due to the hemispherical shape of the cribriform plate only one tangential aspect of this wall is visible on each single view and only those lesions which are situated in that highlighted region will be clearly detect- able. In one study it was suggested that visualization of the

FIG. 5. Close-up of a V1OoR-DCd0 view of the nasal cavity of an intermediate mesati-dolichocephalic canine head. The naso-orbital wall is part of a tangential line marking the lateral border of the caudal nasal cavity (a-a). Due to obliquity separate bone structures partially split and allow identification of a 4 x 20 mm vertical shaft in the naso-orbital wall on the left side (black arrow). On the right a similar wire marked perforation (black arrowhead) is not visible due to the slightly more caudal position, where the bone structures superimpose. The cribriform plate was perfo- rated by a 10 x 10 mm hole. Only a remnant on the right lateral side is still visible (white arrow). Wires mark location and extent of the cribriform plate perforation (white arrowheads). Mild soft tissue opacification in the right nasal cavity.

cribriform plate was possible in the rostrocaudal view of the nose.6 We found the rostrocaudal projection of the nose to be insensitive for cribriform plate destruction and doubt that the cribriform plate is identifiable as such on this view.

It should be pointed out that it is not the cribriform plate itself which is of primary diagnostic interest but the cerebral involvement which is assumed in instances of cribriform plate destruction. Due to the sieve-like structure of the plate,

230 SCHWARZ ET AL. 2000

neoplasms may even cause cerebral microinvasion by pen- etration via the olfactory nerves. Conventional radiography has no potential to detect such lesions.

The detectability of 2 mm defects of the hard palate and the lateral nasal wall supports the relatively high sensitivity of conventional radiography for lesions in these boundary structures. Nasal lesions of that size can occur in aggressive forms of fungal rhinitis. The radiographic picture of “de- structive rhinitis” which is typically described includes punctate, hyperlucent lesions in the nasal cavities.738 The multiple lesions can be very small giving an osteoporotic appearance. These lesions represent supporting bone os- teolucencies or destruction of the conchae7 and our study confirms that conventional radiography is capable to detect such small lesions in the nasal cavity.

The naso-orbital wall is a purely topographic term which has not been described as such in systematic veterinary anatomic literature. Although this wall is a composite of different skull bones, it is its entirety which is of clinical relevance and was therefore investigated as such in this study. The extremely thin wall (in humans termed Lamina pupyrucea) is bordered by much thicker bone structures ventrally and dorsally in the canine skull. These structures are superimposed on it in vertical views and make a diag-

nosis of lesions of the naso-orbital wall almost impossible. Assessment of a lesion which is only detectable if extending to the whole length of the naso-orbital wall under experi- mental conditions must be regarded as of very low or no diagnostic value under clinical circumstances. Conventional radiography is therefore inappropriate for assessing orbital extension of aggressive nasal diseases.

Apart from the specific superimposition of the frontal bone on the cribriform plate in short-nosed heads, there was no difference between the different groups of skull confor- mation for the detectability of nasal boundary lesions.

In an experimental study investigating the effects of soft tissue opacification on the turbinate pattern it was found that a loss of the characteristic parallel pattern in the rostra1 part of the nose is primarily a result of silhouetting of maxillary conchae by fluid, whereas a loss of turbinate bone structures must be caused by destruction.’ None of the nasal border lesions in our study disappeared due to soft tissue opacifi- cation. In addition, there was no obvious loss in the turbi- nate pattern between the filled and empty nasal cavities. This demonstrates that despite the presence of soft tissue opacity associated with aggressive nasal disease such as neoplasia and fungal rhinitis, defects in nasal borders are still detectable.

REFERENCES

1 Schwarz T, Sullivan M, Hartung K Radiographic anatomy of the cribriform plate (Lamina crzbrosa) Vet Radiol Ultrasound 2000, 41

2 Berry CR, Koblik PD Evaluahon of survey radiography, linear to- mogrdphy and computed tomography for detecting experimental lesions of the cribriform plate in dogs Vet Rahol 1990,31.146-154

3 Schwarz T The role of conventional radiography and computed to- mography (cat-scan) In the &agnostic of clinical rhinitis in dogs with specific consideration of nasal tumonrs and fungal rhinitis, Berlin Veteri- nary Thesis, Egelshach - Fiankfurt - St Peter Port Haensel-Hohenhausen German Academic Writings Vol 2418, 1997

4 Thrall DE, Robertson ID, McLeod DA, Heidner GL, Hoopes PJ, Page RL A comparison of radiographic and computed tomographic findings in 31 dogs with malignant nasal cavity tumors Vet Radiol 1989,30 59-66

5 Park RD, Beck ER, LeCouteur RA Comparison of computed tomog-

raphy and radiography for detecting changes induced by malignant nasal neoplasia in dogs. J Am Vet Med Assoc 1992;201:1720-1724.

6. Kirberger RM, Fourie S. An investigation into the usefulness of a rostrocaudal nasal radiographic view in the dog. 1 lth IVRA meeting, Je- rusalem, Israel, July 27-August 1, 1997; Vet Radiol Ultrasound 1997;38: 489.

7. Gibbs C, Lane JG, Denny HR. Radiologic features of intra-nasal lesions in the dog: a review of 100 cases. J Small Anim Pract 1979;20 5 15-535.

8. Sullivan M, Lee R, Jakovlevic S, Sharp NJH. The radiological fea- tures of aspergillosis of the nasal cavity and frontal sinuses in the dog. J Small Anim Pract 1986;27:167-180.

9. Schmidt M, Voorhout G. Radiography of the canine nasal cavity: significance of the presence or absence of the trabecular pattern. Vet Radiol Ultrasound 1992;33:83-86.