Article Text

Clinical and MRI discordance in a case of delayed radiation myelopathy
  1. St James's University Hospital, Beckett Street, Leeds LS9 7TF, UK
  1. Dr Paul Maddison paul{at}

Statistics from

Delayed radiation myelopathy (DRM) is a rare complication of radiotherapy, especially when the total dose delivered to the spinal cord is less than 50 Gy.1 From the limited data reported on the radiological features of DRM, typical MRI changes of cord swelling and gadolinium enhancement on T1 weighted images and increased intramedullary signal on T2 weighted images often correlate well with progression of neurological symptoms. We report a case of DRM in which the clinical features remained static at a level corresponding only to the lower end of the abnormal intramedullary lesion seen on MRI, without radiological evidence of blood-brain barrier breakdown, suggesting that pathological changes to the spinal cord after radiation are not always displayed as visible changes on MRI.

A previously well 71 year old woman presented in December 1997 to the ear, nose and throat department with hoarseness and difficulty swallowing, having discovered a lump in the right side of her neck. Laryngoscopy showed a right pyriform fossa tumour, histologically confirmed as a lymphoepithelioma after biopsy. Staging neck CT showed extensive tumour in the right pyriform sinus and lateral pharyngeal wall with associated lymphadenopathy extending to the thoracic inlet.

Between April and May 1998, she received chemotherapy comprising 5-fluorouracil and cis-platinum. Radiotherapy was started in June 1998 with fields covering the pyriform sinus and neck, anterior supraclavicular fossa, and posterior neck. The calculated dose delivered to the cervical spinal cord was 42.5 Gy in 2 Gy equivalent fractions.

Over the next 4 months, she improved clinically and on review in October 1998 there were no signs of tumour recurrence, confirmed on laryngoscopy. She remained well until November 1998 when she began to notice progressive numbness and weakness spreading proximally up both legs over a period of 6 weeks, such that she was unsteady when walking. Soon thereafter, she developed urinary incontinence and constipation. At this time, cervical and thoracic spine MRI (figure) disclosed cervical cord swelling with a spindle shaped cavity of low signal intensity on T1 weighted images and high signal intensity on T2 weighted images extending from C4 to the upper end of T1. There was no cord enhancement after the administration of gadolinium.

Sagittal T2 weighted MRI showing a spindle shaped area of increased signal extending from C4 to T1.

On neurological review in January 1999, she was incontinent of urine and faeces and was able to mobilise only with the aid of a walking frame. There was generalised wasting in both legs with mild reduction in strength of hip flexion, knee flexion, and extension, but no detectable weakness in the arms. There was a loss of spinothalamic sensation below the level of T5 bilaterally, with a subjective feeling of altered sensation in both hands.

Over the ensuing 4 months, her level of disability progressed such that she was unable to walk or transfer independently. Further examination disclosed flaccid paralysis of both legs and minor weakness of the intrinsic muscles of both hands. There was reduced perception of pain and temperature sensation below the level of C8. She commenced hyperbaric oxygen therapy given on a daily basis for 3 weeks but there was no symptomatic recovery.

After a recurrence of her original swallowing difficulties, she was found to have a tumour recurrence in her pyriform fossa at biopsy, with associated lymphadenopathy. No further radiotherapy was given. At last review in February 2000, she had no movement in her lower limbs and marked weakness of the intrinsic muscles of both hands, with preserved power in proximal upper limb muscles. Sensation above the level of C8 was intact.

Diagnostic criteria of DRM as outlined by Pallis et al 2 stated that the spinal cord should be included in the irradiated area; the main neurological lesion should be within segments of the cord exposed to radiation; and that other causes for the neurological disorder have been excluded.

Despite the almost syrinx-like appearance of radiation myelopathy on MRI, the clinical presentation of DRM is usually one of progressive rostral spread of sensorimotor symptoms, where the final clinical level usually corresponds to a level well within the boundaries of the length of abnormal signal seen on MRI, typically within 1 year of symptomatic onset.3 4 All but one of the reported cases of MRI features of DRM in the literature have shown ring-like enhancement after gadolinium administration, suggesting breakdown of the blood-brain barrier.3 4 As yet, there are insufficient data on precise clinical sensorimotor cord levels in relation to MRI appearances (and possible gadolinium enhancement) in the studies of DRM currently published to enable prediction of clinical outcome based on radiological findings at presentation. Our patient's symptoms and signs plateaued at a level which would indicate an upper level of spinal cord damage at C8/T1, whereas MRI showed extensive cord changes from C4 to the upper end of T1. There was no gadolinium enhancement, but T2 weighted images disclosed evidence of marrow change in the vertebral bodies of C1 to T3, suggesting that the lesion was within the irradiated area. Pathological studies in DRM have shown white matter changes of demyelination and malacia, often affecting the lateral and posterior aspects of the cord, with associated vascular changes of fibrinoid necrosis, thrombosis, perivascular oedema, and haemorrhage.1

The high signal T2 weighted MRI findings with cord swelling resemble those of syringomyelia. However, almost all patients with cervical syrinx formation develop marked sensorimotor symptoms or signs in the upper limbs, and a clinical picture of isolated spastic paraparesis is unusual.5 The main areas of cord damage after radiotherapy would seem to be the lateral corticospinal tracts and posterior columns from radiological-pathological studies.6 Peripheral ring enhancement of the cord after gadolinium administration suggests that the major focus of cord damage or at least blood-brain barrier breakdown does not lie within the central cord regions. Consequently, the central cord swelling seen on T2 weighted images may well represent oedema, produced in response to altered vascular permeability after radiation damage to the vascular endothelium. In direct intrinsic cord compression, this spindle shaped region of cord oedema has limited correlation with clinical signs, as would be suggested by our case. The presence of gadolinium ring enhancement of the cord, missing in our case, serves as a more reliable marker of structural cord damage, but even in its absence, the diagnosis of DRM can be made in a patient with ascending sensorimotor symptoms within months of radiation therapy, where the only clue to radiation damage to the cord is central cord swelling and vertebral body marrow change with increased signal on T2 weighted images. Follow up MRI shows cord atrophy only in the region of previous gadolinium enhancement, correlating with clinical findings.7 Our case has shown that the ultimate level of disability cannot be judged from the extent of the T2 weighted central cord changes seen on initial MRI.


Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.