cannot reliably detect lesions less than 2 cm in size
99m .
and Tc -MDP bone scanning cannot r eliably distinguish tumour involvement from benign disease. As such,
radionuclide scanning of oropharyngeal squamous carcinoma 9 9 rn
has no current role in diagnosis or staging. Tc (v) DMSA requires further evaluation.
183
The Hypopharynx
Squamous cell carcinoma of the hypopharynx often presents late with advanced local disease and cervical metastases. In the past,tumours have been evaluated using dire c t endoscopy and contrast studies but these techniques on their own have distinct disadvantages since they can only infer deep-seated abnormalities f rom changes in surface
contour. The commonest tumours are those affecting the post- c r i c o i d region and the p y r i f o r m sinus. As these
areas are often difficult to assess e n d o s c o p i c a l l y , CAT and MRI have added a new dimension to their evaluation.
CAT does not reveal mucosal detail, so the best w a y to evaluate the hypopharyngeal mucosa is at endoscopy. CAT can identify mucosal abnormalities but these may be due to tumour, fibrosis, oedema, haemorrhage or
inflammation. CAT complements endoscopy and biopsy since it reveals direct extension and can identify cartilage invasion (Mancuso and Hanafee, 1985, p 253). Accurate e v a l u a t i o n of the pyriform sinus can be improved by perfo r m i n g phonation scans (Gamsu et al, 1981), and the asse s s m e n t of laryngeal involvement and vocal cord m obility is discussed in subsequent sections. MRI is superior to
CAT in identifying soft tissue definition and tumour extension and sensitivity can be increased by using a surface coil (Lufkin et al, 1986b), In some institutions surface coil MR imaging is the investigation of choice to evaluate the hypopharynx (Lufkin et al, 1986b).
Ultrasound imaging of the hypopharynx has not been 6 7
evaluated. The uptake of Ga -Citrate has b e e n reported in hypopharyngeal carcinoma (Kornblutt et al, 1974;
Smith et a l f 1975; Teates et al, 1980). However, it cannot identify lesions less than 2 cm and since, it is less accurate than endoscopy compared with CAT or MRI, it has no role
in either diagnosis or staging.
The Larynx
All patients with squamous carcinoma of the larynx should have a direct laryngoscopy and biopsy. The
subsequent radiological evaluation of the larynx has undergone significant changes in the last d e c a d e and conventional techniques of plain film tomo g r a p h y and
laryngography have b een supplemented by CAT and MRI since, combined with direct endoscopy, they are more accurate
185
(Archer et al, 1981; Lewis and Carter, 1987) and permit precise tumour localisation so that the surgeon can decide between radiotherapy and partial or total laryngectomy.
CAT provides helpful information a bout areas that may be hidden from visual inspection by b ulky tumours, such as the subglottis, the apex of the p yriform sinus and the
laryngeal ventricle, and sensitivity is increased by
performing phonation scans (Gamsu et al, 1981). It reveals deep extension and helps to clarify suspected tumour
extension in submucosal laryngeal structures where o verlapping structures prevent a full two-dimensional evaluation. Minor mucosal abnormalities will not be seen on CAT, but these are much better evaluated at laryngoscopy. Caution must be used when diagnosing c a r t i l a g e invasion on CAT due to the random distribution of c a l c i f ication and ossification w h i c h occurs within normal cartilage
(Archer et al, 1983). Although CAT may b e of value in the assessment of vocal cord fixation (Mancuso et al, 1980b; Mancuso and Hanafee, 1985), it cannot defi n e a transition zone from the true to the false cords and spread of tumour from the true to the false cords (or vice-versa) can be difficult to assess (Castelijns, 1987, p 30) . Lastly, CAT may over-estimate tumour size due to o e dema and inflammation, or under-estimate tumour size due to failure to identify microscopic disease (Silverman et al, 1984).
The demonstration of normal and normal variant
anatomy using MRI can be improved by using a surface coil (Castelijns et al, 1985; McArdle et al, 1986).
Employing standard head and body coils, the structures of the neck are particularly difficult to image using MRI
(Castelijns, 1987). Head coils will not cover the middle and inferior regions of the neck and body coils are
inefficient due to the small size of the region of interest. The application of specially designed surface coils has
solved these problems, and the d e monstration of normal laryngeal anatomy using an MRI surface coil is superior to that obtained with CAT.
MRI using a surface coil has potential advantages over CAT although patients w i t h difficulty in swallowing, or with excessive coughing, are unsuitable due to motion artefacts
(Castelijns, 1987). In addition, due to the long scanning times, it is not yet possible to image the larynx while performing phonation manoeuvres w i t h o u t getting motion artefacts.
By using frontal images, MRI can show the cranio- caudal extension of a tumour, particularly subglottic extension, and the relationship b e t w e e n the caudal margin of the tumour and the upper border of the cricoid is more
i
[
\
t.
187
clearly d e m o n strated than on CAT. On sagittal images, infiltration of the base of the tongue is well visua l i s e d so that the distance between the caudal margin of a
supraglottic tumour and the anterior commissure c a n be assessed. A lthough it was thought initially that MRI was no better than CAT in predicting laryngeal cartilage
invasion (Rothberg et al, 1986), it has now been shown to be superior (Castelijns et al, 1987b-c). MRI of the larynx is the investigation of choice, when available, in the
diagnostic imaging of laryngeal squamous carcinoma
(Castelijns, 1987, p 150). However, if MRI fails due to motion artefacts, claustrophobia, metallic implants, or
is impossible to p e r f o r m due to a pacemaker or surgical clips, CAT may still be necessary. If imaging laryngeal carcinoma can be accomplished w i t h o u t the use of intravenous contrast, then this alone w o u l d make MRI the investigation of choice
(Mancuso and Hanafee, 1985).
Ultrasound has b e e n used to evaluate the larynx and has been used p r i n c i p a l l y to evaluate tumour invas i o n of the thyroid cartilage. Although sensitive images c a n be obtained regarding c artilage thickness and integrity, they are non-specific and require close clinical
correlation (Miskin et al, 1978). However, u l t r a s o u n d laryngeal high resolution imaging may be of value in isolated cases.
It can identify tumour invasion into the thyroid gland, the internal jugular v e i n and the carotid artery. Where CAT or MRI findings of cartilage destruction are
equivocal, u l t r a s o u n d c a n effectively delineate cartilage integrity in lesions and cartilage destruction in
lesions (Rothberg et al, 1986).