1977, British Journal of Radiology, 50, 687-698 VOLUME 50 NUMBER 598
OCTOBER 1977
The British Journal of Radiology Cranial chordomas By B. E. Kendall, F.R.C.P., F.R.C.R., D.M.R.D., and B. C. P. Lee, M.R.C.P., F.R.C.R. The National Hospitals for Nervous Diseases, London WC1N 3BG (Received April, 1977) ABSTRACT
The clinical and radiological features of 44 cranial chordomas are recorded. Classical features of a mid-line partially calcified tumour, destroying the clivus and causing a soft tissue mass in the sphenoid sinus or nasopharynx were present in about half of the cases. The high incidence of unilateral bone erosion which occurred in nearly a third of the cases, the diagnostic significance of erosion of the tip of the clivus and odontoid peg and the not infrequent sclerotic bone reaction to the tumour, are stressed. Increased uptake of " T c m occurred in the five cases in which gamma encephalography was performed. Computed transmission tomography was valuable not only in showing the intracranial extension of the tumours and changes in the ventricular system, but also in outlining the extent of the mass within the skull base. Chordomas tend to displace the dura before transgressing it so that the subarachnoid space is usually patent adjacent to quite large tumours allowing the margins of any intracranial or spinal extension to be outlined at encephalography or myelography. Displacements of basal vessels are helpful in confirmation of the site and extent of the tumour. Narrowing of arteries by compression is not infrequent but encasement is rare. Pathological circulation which has not been considered to be a feature of chordomas was shown in six cases.
MATERIAL
We reviewed all the available records of patients who had been diagnosed as cranial or cranio-cervical chordomas from The National Hospitals for Nervous Diseases and The Middlesex Hospital, London. From these records we selected the patients with unequivocal histological evidence of chordomas within the cranium who had adequate X-ray films for analysis. RESULTS
The various features of our cases are set out in Tables I-VIII. In addition radioisotope encephalography using 99Tcm was performed in five cases (11%). These were large tumours and all showed easily visible uptake in a mass adjacent to the skull base.
TABLE I Chordomas are uncommon tumours; the cranial variety which constitute about 40% of all chordomas (Utne and Pugh, 1955; Firooznia et al, 1976) form under 1 % of all intracranial tumours. Nevertheless, there have been several publications dealing with the clinical and plain film features (Grood and Moffie, 1954; Higginbottom et al, 1967; Allen and Kerr, 1968; Phelps et al, 1970; Windeyer, 1959) and others which also discuss contrast studies (Dahlin and MacCarty, 1952; Moya, 1960; Kamrin et al, 1964; Plant and Blatt, 1967; Wood and Himaldi, 1950). We recently analysed the clinical and radiological abnormalities in a large series of cranial chordomas. In general our findings were similar to those previously reported but there were a few unusual and poorly documented features, which we consider to be worth recording. Computed transmission tomograms (C.T.T.) were performed in 11 cases and the value of C.T.T. in assessing the tumours will be discussed.
Total number of patients 44 Males 28 (63.6%) Females 16 (36.4%) Average age at onset of symptoms 37.9 years (range 3-63 years) Time between onset of symptons and presentation 3.45 years (range 0.2-16 years)
TABLE II PRESENTING CLINICAL FEATURES IN 44 CRANIAL CHORDOMAS
687
Cranial nerve palsies Cranial nerve palsies + facial pain Cranial nerve palsies + conductive deafness Cranial nerve palsies + paresis or ataxia of limbs Cranial nerve palsies + paresis of limbs + epilepsy Cranial nerve palsies + facial neck pain + paresis of limbs Tetraparesis + neck pain Nasal blockage + epistaxis Amenorrhoea + headache + dementia Headache
17 (39%) 4 (9%) 1 (2.3%) 13 (30%) 1 (2.3%) 2(4.5%) 2(4.5%) 2(4.5%) 1 (2.3%) 1 (2.3%)
VOL.
50, No. 598 B. E. Kendall and B. C. P. Lee TABLE III SYMPTOMS AND SIGNS IN 44 CRANIAL CHORDOMAS
Cranial nerve palsies
Unilateral
Bilateral
Upper (I-VI) 16(36%) Lower (VII-XII) 14(32%) Upper and Lower 7(16%) Total 37 (84%)
7(16%) 6(14%) 6(14%) 19(43%)
9 (20%) 8(18%) 1 (2.3%) 18(41%)
Paralysis or ataxia of limbs Facial or neck pain Facial spasm Nasal airway obstruction + epistaxis Conductive deafness Headache Epilepsy Amenorrhoea Dementia
18(41%) 8(18%) 2(4.5%) 3 (6.8%) 2 (4.5%) 7(16%) 2(4.5%) 1 (2.3%) 1 (2.3%)
TABLE IV ABNORMALITIES SHOWN ON PLAIN FILMS OR TOMOGRAMS OF SKULL
Bone erosion only Bone erosion + soft tissue mass Bone erosion + sclerosis Bone erosion + sclerosis + soft tissue mass Bone erosion + calcification Bone erosion + calcification + soft tissue mass Bone erosion + sclerosis + calcification + soft tissue mass Calcification alone No abnormality shown Summary Erosion of bone Sclerosis of bone Calcification in tumour Soft tissue mass in sphenoidal sinus or naso-pharynx Abnormalities bilateral Abnormalities unilateral
12(27%) 12(27%) 2 (4.5%) 2(4.5%) 9 (20%) 4 (9%) 1 (2.3%) 1 (2.3%) 1 (2.3%) 42(95%) 5(11%) 15(34%) 19(43%) 26 (66%) 14(32%)
DISCUSSION
Clinical features (Tables I—III) The age range in our series at onset of symptoms (3-63 years) is well within that at which chordomas have previously presented, which varies from the seventh month of fetal life (Hennig, 1900) to 82 years (Plant and Blatt, 1967). However, the youngest of our patients is exceptional because less than a dozen chordomas have been recorded before puberty (Becker et al., 1975). The average age in our series (38 years) is identical with that obtained by averaging previously recorded series (Hass, 1934; Mabrey, 1935). Males are more frequently involved
(64% in our series) though the precise proportion varies between 67% (Faust et al., 1944) and 50% (Utne and Pugh, 1955; Schechter et al., 1974). The length of history prior to presentation to a specialist varies markedly, usually more than one month and less than five years, the average being under two years (Schechter et al., 1974). The longer interval in our series reflects the relatively minor symptoms occurring early in the course of some of our cases. The clinical findings (Tables II and III) are most frequently accounted for by compression by the tumour of the adjacent cranial nerves and brain stem or upper spinal cord. They depend on the site of origin and direction of growth of the mass (Danziger et al., 1974). The tumour may arise at any point along the body of the sphenoid, basi-occiput or upper cervical vertebral bodies. It usually commences close to the mid-line and frequently remains almost symmetrical but may be entirely or predominantly unilateral. Thus, upper or lower cranial nerves may be involved bilaterally or unilaterally and the tumour must be considered in the differential diagnosis, not only of suprasellar and pre-pontine tumours, but also of parasellar, cerebello-pontine angle, jugular foramen and foramen magnum syndromes. Tumour induced pain, which was a presenting feature in eight (18%) of our series, is a well recognized and relatively common symptom of chordomas. It may be a deep ache, apparently due to bone destruction, or a neuralgia in the distribution of an involved nerve, mostly frequently the trigeminal. In one case in this series, which was recently reported in detail (Perkins, 1973), the pain was unusual and merits further comment. This was a 34-year-old man, who presented with 12 months' history of occipital and nuchal pain only present ten minutes after a moderate consumption of alcohol lasting for about two hours. For eight months the pain had become continuous, but exacerbated by alcohol and straining. For two months there had been progressive ataxia, dysphagia and facial numbness. Examination revealed fifth and seventh cranial nerve palsies, minimal left-sided spasticity and mild cerebellar ataxia. Plain X rays and tomograms showed extensive destruction of the clivus with sclerosis adjacent to the sphenoidal air sinus. Bilateral carotid angiograms were normal but there was posterior displacement of the vertebral and basilar arteries (Fig. 1). Since pain induced by alcohol had not been previously recorded from a chordoma and is a well known feature of Hodgkin's disease (Brewin, 1966) which is also a cause of bone sclerosis, this was the preferred diagnosis prior to surgery. The
688
TABLE V DISTRIBUTION OF BONE EROSION IN 44 CRANIAL CHORDOMAS
Case
Clivus
23 (52%)
Axis
Pit. fossa
Orbit
6(14%) 15 (34%) 11 (25%)
tumour, which was extending into the right cerebello-pontine angle, from the right side of the clivus and jugular foramen, was a typical chordoma. Despite the common occurrence of a soft tissue mass in the naso-pharynx, nasal airway obstruction is an unusual complaint (Bhalia, 1965; Hingorani, 1970). It was present in three of our patients and was the predominant feature of two of them. In one of these cases the tumour extended throughout the nasal cavity and despite treatment it eventually destroyed the right orbit and antrum, finally causing
death from severe epistaxis. Conductive deafness due to involvement of the Eustachian tube or middle ear occurred in three of our cases and was a presenting symptom in one of them in whom a mass, presumably part of the chordoma, was visible through the ear drum. Headache occurred in seven patients (16%). Dementia was an early feature in only one case though it did occur late in the course of two others and was associated with hydrocephalus (Greenwald etal., 1957).
689
VOL.
50, No. 598 B. E. Kendall and B. C. P. Lee TABLE VI
FINDINGS ON COMPUTED TRANSMISSION TOMOGRAPHY
Number of cases examined 11(25%) Bone erosion + 3 intracranial mass Bone erosion + intracranial mass + calcification in tumour + opaque 2 sphenoid sinus Bone erosion + expansion + calcification in tumour 1 mass Bone erosion + 1 expansion Bone erosion + opaque sphenoid sinus 1 1 Bone erosion alone Normal (tumour below 2 level of sections) Summary Bone erosion Expansion of bone Soft tissue mass Calcification in tumour Opaque sphenoid sinus Intravenous contrast medium given Lateral ventricles dilated
9 2 6 3 3
FIG. 1. Case 1. Right vertebral angiogram. There is marked posterior displacement of the vertebral and basilar arteries. The clivus is eroded but there is sclerosis of the part adjacent to the sphenoidal air sinus.
5 Enhancement shown, 3 2 Third ventricle displaced in two cases Fourth ventricle displaced in two others
TABLE VII Cases (%) Pneumo-encephalography Positive contrast myelography Pneumo-ventriculography Positive contrast ventriculography
9(20.5%) 6(13.6%) 6(13.6%) 1 (2.3%)
Normal
Extra-axial masses
0 1 06.7%) 0 0
9(100%) 5(83.3%)
Basal mass
6(100%) 1 (100%)
TABLE VIII ANGIOGRAPHY
Carotid angiograms Vertebral angiograms
Number performed
(total cases %)
Basal mass (%)
Extradural displacements (%)
Vessel narrowing (%)
Pathological circulation (%)
Normal (%)
29
(66%)
21 (72%)
8 (28%)
4(13.8%)
4(13.8%)
9 (27.6%)
23
(52%)
22(95.5%)
0
2 (8.7%)
1 (4.5%)
600
OCTOBER 1977
Cranial chordomas
FIG. 2. Case 2. Anteroposterior tomogram. There is erosion ( ^ ) of the inferior surface of the petrous bone bordering the left jugular foramen, extending up to the inferior border of the internal auditory canal and of the part of the occipital bone forming the medial border of the jugular foramen. Similar appearances occur with glomous tumours.
Plain films and tomograms
In general the abnormalities noted were similar to those reported by other authors (Table IV). The predominant feature (95%) was well denned bone erosion usually in the skull base; its distribution is shown in Table V. The incidence in our cases is slightly greater than previously recorded (83% Utne and Pugh, 1955; 87% Schechter et al, 1974) which may be due to the longer interval between onset of symptoms and presentation. Classically the bone changes in chordomas involve the clivus and are fairly symmetrical but in our series the clivus was affected on only 52% and there was a relatively high incidence (32%) of unilateral plain film changes which correlated well with the cases presenting with unilateral cranial nerve palsies. Erosion of the superior orbital fissure (Rischbieth and Bull, 1958) may simulate the classical appearances of an intracavernous aneurysm. Destruction of the petrous apex may resemble that caused by fifth nerve neuromas and extension around the jugular foramen (Phelps et al., 1970) may be like that of a glomus tumour (Fig. 2) though the margins are usually less well defined with the latter. Erosion of the pituitary fossa most frequently involves the floor (Fig. 3). The whole sella or only part of it may be destroyed, but expansion similar to that of a pituitary tumour is extremely rare, having only been described in a few cases of ectopic chordoma arising within the sella (Werner, 1960; Belza, 1966; Mathews and Wilson, 1974).
FIG. 3. Case 3. Lateral tomogram. The body of the sphenoid including the floor and most of the anterior wall of the pituitary fossa and the adjacent part of the basi-occiput are partially destroyed. There is partially calcified tumour within the sphenoid sinus and the floor of the sinus is slightly depressed.
A feature present in four of our patients, which has not been stressed in the literature and which is suggestive of chordoma (Danziger et ah, 1974), is erosion of the tip of the clivus, especially in
691
VOL.
50, No. 598 B. E. Kendall and B. C. P. Lee
association with destruction of the top of the odontoid (Fig. 4). In cases with the tumour situated predominantly in the upper cervical spine this may be the only plain film evidence of intracranial extension. In other cases destruction of the second cervical vertebra was extensive. A soft tissue mass within the sphenoidal sinus or naso-pharynx (43%) and calcification in the tumour (34%) are within the range of incidence previously recorded (Fig. 3). Calcification may be dense and obvious or faint and spotty and sometimes only shown on tomography (Lindgren and Di Chiro, 1951; Schisano and Tovi, 1962). Similar densities may be caused by incomplete destruction of bone, but this usually retains some trabeculation. Sclerosis of bone, which was evident in five (11%) of our patients, is a controversial feature of these tumours. Some authors, for example, Zingesser and Schechter (1964), Di Chiro and Anderson (1965), Plant and Blatt (1967) and Schechter et al. (1974) state that sclerosis does not occur or was not present in any of their cases, and others, for example Danziger et al. (1974) do not mention it. However, du Boulay (1965), Epstein (1974) and Lim (1975) describe and illustrate it in some of their cases. Sclerosis is much more usual in meningiomas and is well recognized in naso-pharyngeal carcinomas and secondary deposits involving the skull base. Although relatively uncommon in cranial chordomas it is important to realize that it occurs; in two of our cases with
FIG. 4. Case 4. Mid-line lateral tomogram. The inferior part of the body of the sphenoid and the basi-occiput are sclerotic. There is new bone formation on the inferior surface of the sphenoid. The tip of the clivus (AA) a n d of the odontoid process (A) a r e eroded and there is slight soft tissue thickening along the posterior wall of the nasopharynx.
otherwise typical features the diagnosis was dismissed because sclerosis was present (Fig. 4). Expansion of bone, which is common with sacral chordomas, has been said not to occur in cranial lesions (Epstein, 1974); it was present, but not a dominant feature (Fig. 3) in two of our cases. In one patient skull films and tomograms were normal. This was a 49-year-old man who presented in 1958 with a two-year history of progressive loss of vision and one month of intermittent headache. Examination revealed bi-temporal hemianopia and diminished visual acuity in both eyes. A pneumoencephalogram showed a mass extending from the region of the sella to elevate the anterior part of the floor of the third ventricle. A histologically typical chordoma lifting the diaphragma sellae and displacing the optic chiasm anteriorly was partially excised. Vision improved but in 1961 it deteriorated and there was onset of left trigeminal pain. Suprasellar calcification was now evident for the first time and pneumo-encephalography showed a large suprasellar mass which was again excised. Symptoms were relieved temporarily, but recurred and despite further surgery the patient was eventually certified blind in 1965. Radioisotope encephalography
There is little information on gamma encephalographic findings in chordomas. Increased uptake was present in all the five of our cases (Fig. 5) and in the one case of the series of Schechter et al. (1974) in which isotope scanning was performed. Single positive cases are referred to by Mishkin and Mealey (1969), Becker et al. (1975), and Mamo et al.
FIG. 5. Case 5. Gamma encephalogram. " T c m right lateral projection. There is increased uptake adjacent to the clivus in a large chordoma.
692
OCTOBER 1977
Cranial chordomas (1975). Chordomas are not mentioned in most books dealing with brain scintigraphy including the recent treatise by Penning and Front (1975). Computed transmission tomography (Hounsfield, 1973; Ambrose, 1973; Paxton and Ambrose, 1974) C.T.T. was abnormal in all cases in which the sections passed through the tumour (Table VI). The plane of the sections, more or less parallel to the base of the skull, is ideally suited to show the extent of bone erosion (Fig. 6) which is sometimes difficult to determine on plain skull films. It is best shown with a wide window set at a relatively high level (e.g. W400. L100). The margins of the erosion are usually well defined, suggesting pressure by the tumour which usually grows mainly by expansion. Expansion of bone or distraction of residual fragments by the mass of the tumour was present in two cases (Fig. 7). The soft tissue mass of the tumour extending into the skull (Fig. 8) (six cases) or a
sphenoidal sinus (Fig. 6) (three cases) was of slightly increased attenuation. Intracranial calcification is shown on C.T.T. when it is present in too small a concentration to be shown on plain X rays. It was evident in three of our cases,
FIG. 8. Case 8. (A) C.T.T. at level of outlet of pituitary fossa (L+18, W40). There is a lobulated extra-axial mass of slightly increased attenuation in the left side of the suprasellar cisterns and posterior to the upper part of the clivus extending to both petrous apices displacing the brain stem posteriorly and widening the cerebello-pontine angle cisterns. (B) Similar level after bolus intravenous injection of 60 ml of Conray 420. There is a minor degree of enhancement in the tumour.
..-if:.
FIG. 6. Case 6. C.T.T. at level of body of sphenoid and petrous bones. Window level +18 EMI units, width 400 EMI units (L +18, W400). There is destruction of the left petrous apex and left side of the body of the sphenoid bone. The sphenoidal air sinus is of soft tissue attenuation indicating involvement by tumour.
FIG. 9. Case 9. (A) C.T.T. at level of auditory meati (L+100, W400). There is destruction sparing only a tiny fragment (•
OCTOBER 1977
The British Journal of Radiology Cranial chordomas By B. E. Kendall, F.R.C.P., F.R.C.R., D.M.R.D., and B. C. P. Lee, M.R.C.P., F.R.C.R. The National Hospitals for Nervous Diseases, London WC1N 3BG (Received April, 1977) ABSTRACT
The clinical and radiological features of 44 cranial chordomas are recorded. Classical features of a mid-line partially calcified tumour, destroying the clivus and causing a soft tissue mass in the sphenoid sinus or nasopharynx were present in about half of the cases. The high incidence of unilateral bone erosion which occurred in nearly a third of the cases, the diagnostic significance of erosion of the tip of the clivus and odontoid peg and the not infrequent sclerotic bone reaction to the tumour, are stressed. Increased uptake of " T c m occurred in the five cases in which gamma encephalography was performed. Computed transmission tomography was valuable not only in showing the intracranial extension of the tumours and changes in the ventricular system, but also in outlining the extent of the mass within the skull base. Chordomas tend to displace the dura before transgressing it so that the subarachnoid space is usually patent adjacent to quite large tumours allowing the margins of any intracranial or spinal extension to be outlined at encephalography or myelography. Displacements of basal vessels are helpful in confirmation of the site and extent of the tumour. Narrowing of arteries by compression is not infrequent but encasement is rare. Pathological circulation which has not been considered to be a feature of chordomas was shown in six cases.
MATERIAL
We reviewed all the available records of patients who had been diagnosed as cranial or cranio-cervical chordomas from The National Hospitals for Nervous Diseases and The Middlesex Hospital, London. From these records we selected the patients with unequivocal histological evidence of chordomas within the cranium who had adequate X-ray films for analysis. RESULTS
The various features of our cases are set out in Tables I-VIII. In addition radioisotope encephalography using 99Tcm was performed in five cases (11%). These were large tumours and all showed easily visible uptake in a mass adjacent to the skull base.
TABLE I Chordomas are uncommon tumours; the cranial variety which constitute about 40% of all chordomas (Utne and Pugh, 1955; Firooznia et al, 1976) form under 1 % of all intracranial tumours. Nevertheless, there have been several publications dealing with the clinical and plain film features (Grood and Moffie, 1954; Higginbottom et al, 1967; Allen and Kerr, 1968; Phelps et al, 1970; Windeyer, 1959) and others which also discuss contrast studies (Dahlin and MacCarty, 1952; Moya, 1960; Kamrin et al, 1964; Plant and Blatt, 1967; Wood and Himaldi, 1950). We recently analysed the clinical and radiological abnormalities in a large series of cranial chordomas. In general our findings were similar to those previously reported but there were a few unusual and poorly documented features, which we consider to be worth recording. Computed transmission tomograms (C.T.T.) were performed in 11 cases and the value of C.T.T. in assessing the tumours will be discussed.
Total number of patients 44 Males 28 (63.6%) Females 16 (36.4%) Average age at onset of symptoms 37.9 years (range 3-63 years) Time between onset of symptons and presentation 3.45 years (range 0.2-16 years)
TABLE II PRESENTING CLINICAL FEATURES IN 44 CRANIAL CHORDOMAS
687
Cranial nerve palsies Cranial nerve palsies + facial pain Cranial nerve palsies + conductive deafness Cranial nerve palsies + paresis or ataxia of limbs Cranial nerve palsies + paresis of limbs + epilepsy Cranial nerve palsies + facial neck pain + paresis of limbs Tetraparesis + neck pain Nasal blockage + epistaxis Amenorrhoea + headache + dementia Headache
17 (39%) 4 (9%) 1 (2.3%) 13 (30%) 1 (2.3%) 2(4.5%) 2(4.5%) 2(4.5%) 1 (2.3%) 1 (2.3%)
VOL.
50, No. 598 B. E. Kendall and B. C. P. Lee TABLE III SYMPTOMS AND SIGNS IN 44 CRANIAL CHORDOMAS
Cranial nerve palsies
Unilateral
Bilateral
Upper (I-VI) 16(36%) Lower (VII-XII) 14(32%) Upper and Lower 7(16%) Total 37 (84%)
7(16%) 6(14%) 6(14%) 19(43%)
9 (20%) 8(18%) 1 (2.3%) 18(41%)
Paralysis or ataxia of limbs Facial or neck pain Facial spasm Nasal airway obstruction + epistaxis Conductive deafness Headache Epilepsy Amenorrhoea Dementia
18(41%) 8(18%) 2(4.5%) 3 (6.8%) 2 (4.5%) 7(16%) 2(4.5%) 1 (2.3%) 1 (2.3%)
TABLE IV ABNORMALITIES SHOWN ON PLAIN FILMS OR TOMOGRAMS OF SKULL
Bone erosion only Bone erosion + soft tissue mass Bone erosion + sclerosis Bone erosion + sclerosis + soft tissue mass Bone erosion + calcification Bone erosion + calcification + soft tissue mass Bone erosion + sclerosis + calcification + soft tissue mass Calcification alone No abnormality shown Summary Erosion of bone Sclerosis of bone Calcification in tumour Soft tissue mass in sphenoidal sinus or naso-pharynx Abnormalities bilateral Abnormalities unilateral
12(27%) 12(27%) 2 (4.5%) 2(4.5%) 9 (20%) 4 (9%) 1 (2.3%) 1 (2.3%) 1 (2.3%) 42(95%) 5(11%) 15(34%) 19(43%) 26 (66%) 14(32%)
DISCUSSION
Clinical features (Tables I—III) The age range in our series at onset of symptoms (3-63 years) is well within that at which chordomas have previously presented, which varies from the seventh month of fetal life (Hennig, 1900) to 82 years (Plant and Blatt, 1967). However, the youngest of our patients is exceptional because less than a dozen chordomas have been recorded before puberty (Becker et al., 1975). The average age in our series (38 years) is identical with that obtained by averaging previously recorded series (Hass, 1934; Mabrey, 1935). Males are more frequently involved
(64% in our series) though the precise proportion varies between 67% (Faust et al., 1944) and 50% (Utne and Pugh, 1955; Schechter et al., 1974). The length of history prior to presentation to a specialist varies markedly, usually more than one month and less than five years, the average being under two years (Schechter et al., 1974). The longer interval in our series reflects the relatively minor symptoms occurring early in the course of some of our cases. The clinical findings (Tables II and III) are most frequently accounted for by compression by the tumour of the adjacent cranial nerves and brain stem or upper spinal cord. They depend on the site of origin and direction of growth of the mass (Danziger et al., 1974). The tumour may arise at any point along the body of the sphenoid, basi-occiput or upper cervical vertebral bodies. It usually commences close to the mid-line and frequently remains almost symmetrical but may be entirely or predominantly unilateral. Thus, upper or lower cranial nerves may be involved bilaterally or unilaterally and the tumour must be considered in the differential diagnosis, not only of suprasellar and pre-pontine tumours, but also of parasellar, cerebello-pontine angle, jugular foramen and foramen magnum syndromes. Tumour induced pain, which was a presenting feature in eight (18%) of our series, is a well recognized and relatively common symptom of chordomas. It may be a deep ache, apparently due to bone destruction, or a neuralgia in the distribution of an involved nerve, mostly frequently the trigeminal. In one case in this series, which was recently reported in detail (Perkins, 1973), the pain was unusual and merits further comment. This was a 34-year-old man, who presented with 12 months' history of occipital and nuchal pain only present ten minutes after a moderate consumption of alcohol lasting for about two hours. For eight months the pain had become continuous, but exacerbated by alcohol and straining. For two months there had been progressive ataxia, dysphagia and facial numbness. Examination revealed fifth and seventh cranial nerve palsies, minimal left-sided spasticity and mild cerebellar ataxia. Plain X rays and tomograms showed extensive destruction of the clivus with sclerosis adjacent to the sphenoidal air sinus. Bilateral carotid angiograms were normal but there was posterior displacement of the vertebral and basilar arteries (Fig. 1). Since pain induced by alcohol had not been previously recorded from a chordoma and is a well known feature of Hodgkin's disease (Brewin, 1966) which is also a cause of bone sclerosis, this was the preferred diagnosis prior to surgery. The
688
TABLE V DISTRIBUTION OF BONE EROSION IN 44 CRANIAL CHORDOMAS
Case
Clivus
23 (52%)
Axis
Pit. fossa
Orbit
6(14%) 15 (34%) 11 (25%)
tumour, which was extending into the right cerebello-pontine angle, from the right side of the clivus and jugular foramen, was a typical chordoma. Despite the common occurrence of a soft tissue mass in the naso-pharynx, nasal airway obstruction is an unusual complaint (Bhalia, 1965; Hingorani, 1970). It was present in three of our patients and was the predominant feature of two of them. In one of these cases the tumour extended throughout the nasal cavity and despite treatment it eventually destroyed the right orbit and antrum, finally causing
death from severe epistaxis. Conductive deafness due to involvement of the Eustachian tube or middle ear occurred in three of our cases and was a presenting symptom in one of them in whom a mass, presumably part of the chordoma, was visible through the ear drum. Headache occurred in seven patients (16%). Dementia was an early feature in only one case though it did occur late in the course of two others and was associated with hydrocephalus (Greenwald etal., 1957).
689
VOL.
50, No. 598 B. E. Kendall and B. C. P. Lee TABLE VI
FINDINGS ON COMPUTED TRANSMISSION TOMOGRAPHY
Number of cases examined 11(25%) Bone erosion + 3 intracranial mass Bone erosion + intracranial mass + calcification in tumour + opaque 2 sphenoid sinus Bone erosion + expansion + calcification in tumour 1 mass Bone erosion + 1 expansion Bone erosion + opaque sphenoid sinus 1 1 Bone erosion alone Normal (tumour below 2 level of sections) Summary Bone erosion Expansion of bone Soft tissue mass Calcification in tumour Opaque sphenoid sinus Intravenous contrast medium given Lateral ventricles dilated
9 2 6 3 3
FIG. 1. Case 1. Right vertebral angiogram. There is marked posterior displacement of the vertebral and basilar arteries. The clivus is eroded but there is sclerosis of the part adjacent to the sphenoidal air sinus.
5 Enhancement shown, 3 2 Third ventricle displaced in two cases Fourth ventricle displaced in two others
TABLE VII Cases (%) Pneumo-encephalography Positive contrast myelography Pneumo-ventriculography Positive contrast ventriculography
9(20.5%) 6(13.6%) 6(13.6%) 1 (2.3%)
Normal
Extra-axial masses
0 1 06.7%) 0 0
9(100%) 5(83.3%)
Basal mass
6(100%) 1 (100%)
TABLE VIII ANGIOGRAPHY
Carotid angiograms Vertebral angiograms
Number performed
(total cases %)
Basal mass (%)
Extradural displacements (%)
Vessel narrowing (%)
Pathological circulation (%)
Normal (%)
29
(66%)
21 (72%)
8 (28%)
4(13.8%)
4(13.8%)
9 (27.6%)
23
(52%)
22(95.5%)
0
2 (8.7%)
1 (4.5%)
600
OCTOBER 1977
Cranial chordomas
FIG. 2. Case 2. Anteroposterior tomogram. There is erosion ( ^ ) of the inferior surface of the petrous bone bordering the left jugular foramen, extending up to the inferior border of the internal auditory canal and of the part of the occipital bone forming the medial border of the jugular foramen. Similar appearances occur with glomous tumours.
Plain films and tomograms
In general the abnormalities noted were similar to those reported by other authors (Table IV). The predominant feature (95%) was well denned bone erosion usually in the skull base; its distribution is shown in Table V. The incidence in our cases is slightly greater than previously recorded (83% Utne and Pugh, 1955; 87% Schechter et al, 1974) which may be due to the longer interval between onset of symptoms and presentation. Classically the bone changes in chordomas involve the clivus and are fairly symmetrical but in our series the clivus was affected on only 52% and there was a relatively high incidence (32%) of unilateral plain film changes which correlated well with the cases presenting with unilateral cranial nerve palsies. Erosion of the superior orbital fissure (Rischbieth and Bull, 1958) may simulate the classical appearances of an intracavernous aneurysm. Destruction of the petrous apex may resemble that caused by fifth nerve neuromas and extension around the jugular foramen (Phelps et al., 1970) may be like that of a glomus tumour (Fig. 2) though the margins are usually less well defined with the latter. Erosion of the pituitary fossa most frequently involves the floor (Fig. 3). The whole sella or only part of it may be destroyed, but expansion similar to that of a pituitary tumour is extremely rare, having only been described in a few cases of ectopic chordoma arising within the sella (Werner, 1960; Belza, 1966; Mathews and Wilson, 1974).
FIG. 3. Case 3. Lateral tomogram. The body of the sphenoid including the floor and most of the anterior wall of the pituitary fossa and the adjacent part of the basi-occiput are partially destroyed. There is partially calcified tumour within the sphenoid sinus and the floor of the sinus is slightly depressed.
A feature present in four of our patients, which has not been stressed in the literature and which is suggestive of chordoma (Danziger et ah, 1974), is erosion of the tip of the clivus, especially in
691
VOL.
50, No. 598 B. E. Kendall and B. C. P. Lee
association with destruction of the top of the odontoid (Fig. 4). In cases with the tumour situated predominantly in the upper cervical spine this may be the only plain film evidence of intracranial extension. In other cases destruction of the second cervical vertebra was extensive. A soft tissue mass within the sphenoidal sinus or naso-pharynx (43%) and calcification in the tumour (34%) are within the range of incidence previously recorded (Fig. 3). Calcification may be dense and obvious or faint and spotty and sometimes only shown on tomography (Lindgren and Di Chiro, 1951; Schisano and Tovi, 1962). Similar densities may be caused by incomplete destruction of bone, but this usually retains some trabeculation. Sclerosis of bone, which was evident in five (11%) of our patients, is a controversial feature of these tumours. Some authors, for example, Zingesser and Schechter (1964), Di Chiro and Anderson (1965), Plant and Blatt (1967) and Schechter et al. (1974) state that sclerosis does not occur or was not present in any of their cases, and others, for example Danziger et al. (1974) do not mention it. However, du Boulay (1965), Epstein (1974) and Lim (1975) describe and illustrate it in some of their cases. Sclerosis is much more usual in meningiomas and is well recognized in naso-pharyngeal carcinomas and secondary deposits involving the skull base. Although relatively uncommon in cranial chordomas it is important to realize that it occurs; in two of our cases with
FIG. 4. Case 4. Mid-line lateral tomogram. The inferior part of the body of the sphenoid and the basi-occiput are sclerotic. There is new bone formation on the inferior surface of the sphenoid. The tip of the clivus (AA) a n d of the odontoid process (A) a r e eroded and there is slight soft tissue thickening along the posterior wall of the nasopharynx.
otherwise typical features the diagnosis was dismissed because sclerosis was present (Fig. 4). Expansion of bone, which is common with sacral chordomas, has been said not to occur in cranial lesions (Epstein, 1974); it was present, but not a dominant feature (Fig. 3) in two of our cases. In one patient skull films and tomograms were normal. This was a 49-year-old man who presented in 1958 with a two-year history of progressive loss of vision and one month of intermittent headache. Examination revealed bi-temporal hemianopia and diminished visual acuity in both eyes. A pneumoencephalogram showed a mass extending from the region of the sella to elevate the anterior part of the floor of the third ventricle. A histologically typical chordoma lifting the diaphragma sellae and displacing the optic chiasm anteriorly was partially excised. Vision improved but in 1961 it deteriorated and there was onset of left trigeminal pain. Suprasellar calcification was now evident for the first time and pneumo-encephalography showed a large suprasellar mass which was again excised. Symptoms were relieved temporarily, but recurred and despite further surgery the patient was eventually certified blind in 1965. Radioisotope encephalography
There is little information on gamma encephalographic findings in chordomas. Increased uptake was present in all the five of our cases (Fig. 5) and in the one case of the series of Schechter et al. (1974) in which isotope scanning was performed. Single positive cases are referred to by Mishkin and Mealey (1969), Becker et al. (1975), and Mamo et al.
FIG. 5. Case 5. Gamma encephalogram. " T c m right lateral projection. There is increased uptake adjacent to the clivus in a large chordoma.
692
OCTOBER 1977
Cranial chordomas (1975). Chordomas are not mentioned in most books dealing with brain scintigraphy including the recent treatise by Penning and Front (1975). Computed transmission tomography (Hounsfield, 1973; Ambrose, 1973; Paxton and Ambrose, 1974) C.T.T. was abnormal in all cases in which the sections passed through the tumour (Table VI). The plane of the sections, more or less parallel to the base of the skull, is ideally suited to show the extent of bone erosion (Fig. 6) which is sometimes difficult to determine on plain skull films. It is best shown with a wide window set at a relatively high level (e.g. W400. L100). The margins of the erosion are usually well defined, suggesting pressure by the tumour which usually grows mainly by expansion. Expansion of bone or distraction of residual fragments by the mass of the tumour was present in two cases (Fig. 7). The soft tissue mass of the tumour extending into the skull (Fig. 8) (six cases) or a
sphenoidal sinus (Fig. 6) (three cases) was of slightly increased attenuation. Intracranial calcification is shown on C.T.T. when it is present in too small a concentration to be shown on plain X rays. It was evident in three of our cases,
FIG. 8. Case 8. (A) C.T.T. at level of outlet of pituitary fossa (L+18, W40). There is a lobulated extra-axial mass of slightly increased attenuation in the left side of the suprasellar cisterns and posterior to the upper part of the clivus extending to both petrous apices displacing the brain stem posteriorly and widening the cerebello-pontine angle cisterns. (B) Similar level after bolus intravenous injection of 60 ml of Conray 420. There is a minor degree of enhancement in the tumour.
..-if:.
FIG. 6. Case 6. C.T.T. at level of body of sphenoid and petrous bones. Window level +18 EMI units, width 400 EMI units (L +18, W400). There is destruction of the left petrous apex and left side of the body of the sphenoid bone. The sphenoidal air sinus is of soft tissue attenuation indicating involvement by tumour.
FIG. 9. Case 9. (A) C.T.T. at level of auditory meati (L+100, W400). There is destruction sparing only a tiny fragment (•
