Cancer Therapy Vol 2, 423-428, 2004

1Department of Radiation Oncology, University Medical Center, Utrecht, The Netherlands

*Correspondence: SY El Sharouni, MD, University Medical Center Utrecht, Dept. of Radiation Oncology, Q 00.118, Heidelberglaan 100, 3584 CX Utrecht, The Netherlands; Fax +31 30 2581226; e-mail: S.Y.ElSharouni@azu.nl

Key words: radiation myelopathy, intramedullary metastases, MRI, neurological complications

Abbreviations: biologically effective dose, (BED); cerebral spinal fluid, (CSF); computerized tomography, (CT); linear-quadratic, (LQ); Magnetic Resonance Imaging, (MRI)

Radiation myelopathy is one of the complications most feared in radiotherapy. It was first described by Ahlbom (1941). The clinical picture consists of central motor neuron signs, sometimes without loss of sensibility (Goldwein, 1987). Diagnosis can be difficult with a paraneoplastic syndrome, herpes zoster or myelopathy as a result of chemotherapy and epidural / intramedullary of leptomeningeal metastases as differential diagnosis.

Radiation myelopathy is preceded by a latency period, during which the patient is asymptomatic. Symptoms begin as paresthesia and/or inability to perceive pain and/or temperature. In literature the latency period has been reported to last from 1 month to 6 years. Typically, the median latency is 6 months. About three out of four patients develop myelopathy within 18 months after radiation treatment (Goldwein, 1987; Michikawa et al, 1991; Koehler et al, 1996).

Diagnosis is in fact a diagnosis per exclusionem. The three criteria for myelopathy as a result of radiation treatment according to Pallis et al, (1961) are:

1. the spinal cord must have been included in the radiation field
2. the main neurological lesion must be within the segments of cord exposed to radiation and
3. metastases or other primary spinal cord lesions must be ruled out as the cause of the neurological disorder.

Recommended for further diagnostics are lumbar puncture and myelography, as well as CT- (computerized tomography) and MRI-imaging (Magnetic Resonance Imaging) (Goldwein, 1987). Analysis of the cerebral spinal fluid (CSF) might reveal an elevated protein. The myelogram in some cases demonstrates mild cord atrophy. Myelography is only advised in case of contra-indications for MRI, because of the superior imaging quality of the latter. CT/MRI imaging is also important to rule out the presence of intramedullary tumour (Schiff and O'Neill, 1996). Other causes of myelopathy could be excluded by means of MRI.

Pathological changes of the myelum are described in a number of studies. Both early and late changes can be distinguished (Schultheiss et al, 1988; Van Daal et al, 1989; Van der Kogel, 1997).

A bimodal distribution of latencies has been described (Goldwein, 1987) with peaks occurring at 13 months and 26 months after irradiation. Patients with shorter intervals have, in general, been treated with higher radiation doses. This supports the hypothesis of two mechanisms of damage to the spinal cord: white matter necrosis at higher doses and vascular damage at lower doses. Patients undergoing a reirradiation showed a shorter latency period, as did paediatric patients.

Other causes of progressive myelitis must be excluded, most notably extramedullary metastases, intramedullary tumour, necrotizing carcinomatous myelopathy and hypertrophy of the posterior vertebral facets of the laminae as a result of arthritis.

We present a patient with combined radiation myelopathy and intramedullary metastases.

Our patient was a female, aged 47, without a medical history. In August 1996 she developed complaints of pain in the right shoulder. She also coughed, producing viscous sputum and there was some shortness of breath while exercising. Finally, she developed fever and night sweat. At the time of presentation she smoked two packs of cigarettes a day, no consumption of alcohol. No abnormalities were found at physical examination.

A chest X-ray showed a large density in the upper right lobe of the lung. A CT-scan of the thorax showed a large mass extending into the mediastinum and a second mass ventrally, with probable destruction of the rib, explaining the complaints of pain.

A bronchoscopy showed a tumour, just in front of the ostium of the upper right lobe, causing a stricture to 60%, with submucosal growth.

Pathological findings showed a large-cell squamous cell carcinoma.

She received palliative irradiation to the lung, to a dose of 39 Gy, in fractions of 3 Gy, administered by two opposing AP/PA fields. The target area included the right upper lobe, the ipsilateral hilus and the mediastinum. As a consequence the thoracic vertebrae Th 3 / Th9 were also included (Figure 1).

Nine months later the patient experienced paraesthesia in her left leg and difficulty in stair climbing especially with her left leg. Within a few weeks these complaints extended to her right leg and urine incontinence developed. On neurological examination hyperreflexia of both legs was found but at this time no sensory level at the trunk could be determined. MRI imaging of the thoracal spine showed intramedullary metastases at vertebrae Th5 / Th6. She was irradiated to a dose of 20 Gy in fractions of 4 Gy. The target area was Th4 / Th7, irradiated in one PA field, the dose was given at a depth of 6 cm. This was a reirradiation because the irradiated part of the myelum was partially included in the prior treatment (Figure 2). Despite reirradiation the symptoms got worse resulting in progressive paresis of both legs. Two months after the last radiation treatment she developed complaints of numbness of the right arm, without loss of strength.

Figure 1. First irradiation field Figure 2. Second irradiation field, re-irradiation

Neurological investigation showed numbness in dermatome C8 with hyper-abduction of the right arm but no pareses of the arms. Loss of sensibility at the trunk (delineated at Th6 on both sides), hyperreflexia of the legs and finally a Babinski reflex on both sides were noted.

Imaging of the cervico-thoracical spine by means of MRI showed abnormal signal in the myelum at C7 / Th5. In the cervical (not irradiated) part of the spine, ring-like enhancing lesions were seen, arousing suspicion of radiation myelopathy (Figure 3). In the thoracic (irradiated) part of the spine intramedullary metastases were suspected (Figure 4).

The complaints were rapidly progressive; paresis in the legs became complete. There was total urine incontinence.

Finally she died within three months after reirradiation. Autopsy was performed.

The autopsy showed some surprising findings. In the reirradiated area intramedullary metastases were found, histologically corresponding with a squamous cell carcinoma (Figure 5). These metastases were found at the level of Th7. At the levels C7 / Th1, corresponding with the neurological findings, changes were seen, but not based on tumour. The found abnormalities mainly consisted of vascular changes with subtotal necrosis and hyaline changes of the vessels (Figure 6). ). These findings match radiation myelopathy.

Figure 3. MRI cervical spine, showing suspected myelopathy Figure 4. MRI thoracal spine, showing intramedullary metastases

Figure 5. Intramedullary metastases of squamous cell carcinoma in re-irradiated area

Figure 6. Vascular changes with subtotal necrosis and hyaline changes of the vessels at level C7 / Th1.

It is remarkable that both MRI and pathological investigation showed injury outside the irradiated area of both treatments. We were confronted with radiation myelopathy outside the irradiated volume. A contradiction in terms?

Early changes in radiation myelopathy notably consist of focal areas with spongious demyelination, axonal swelling and the absence of observable vascular injury. Late changes are characterised by involvement of the entire myelum with a preference for the white matter, as well as vascular injury (Schultheiss et al, 1988; Ang and Stephens, 1994). Most studies report vascular injury as the primary pathophysiological cause of radiation myelopathy. This seems to be the explanation for the damage outside the irradiated volume as seen with this patient. Vascular changes as a result of the formation of hyaline membranes, necrosis and thrombosis can migrate into the course of a blood vessel. In our case, however, we observed a short interval between the second treatment and the complaints.

Recent MRI studies in Japan have mentioned a number of characteristics that can almost be conclusive for myelopathy as a result of radiotherapy: 1) swelling of the spinal cord on T1-weighted images and an intramedullary high-intensity area on T2-weighted images (usually attributed to spinal cord oedema, since the cord size diminishes following steroid administration on subsequent MR images) and 2) a ring-like enhancing lesion with gadolinium on T1-weigthed images, resulting from blood-brain barrier breakdown (Michikawa 1991; Yasui 1992). In our case swelling of the spinal cord as well as a ring-like enhancing lesion were indeed found. The three criteria for myelopathy as a result of radiation treatment according to Pallis et al, (1961) are:

1. The spinal cord must have been included in the radiation field;
2. The main neurological lesion must be within the segments of cord exposed to radiation;
3. Metastases or other primary spinal cord lesions must be ruled out as the cause of the neurological disorder.

In our patient described, the spinal cord level of neurological disturbances was located inside as well as outside the irradiated area.

For every fractionation scheme a Biologically Effective Dose (BED) can be calculated using the Linear-Quadratic (LQ) model (Barendsen, 1982). The incidence of a biological effect according to the LQ model is I = n (a d+b d²), and the BED that can be derived from this linear-quadratic equation is nd(1+d/(a /b )); n is the number of fractions, d is the fraction dose, and a /b is a tissue parameter with values of about 10 Gy for acute reacting tissues and a /b is about 2 Gy for late reacting tissues.

A reasonable estimate of radiation myelopathy risk is 5% at 60 Gy in 30 fractions. This myelum tolerance dose results in a BED_tot value of 60(1+2/2) = 120 Gy (with a /b is 2 Gy). The retreatment tolerance dose that can be derived from reirradiated experimental animals and human patients is about 130% of the BED_tot (Stewart, 1994). A prerequisite is that the interval is at least 6 to 9 months and the first treatment dose has not exceeded about 65% of the BED_tot value. The total BED for reirradiation then is 156 Gy.

With our patient the first treatment BED₁ = 39(1+(3/2)) = 97.5 Gy (this is 81% of the BED_tot). The BED₂ of the second treatment is 5*4(1+(4/2) = 60 Gy. This sums up to 157.5 Gy. This BED value of 157.5 Gy seems equivalent to the retreatment tolerance dose. However, the dose of the first treatment delivered to the myelum was 81% of the tolerance dose, and repair of induced damage was probably not complete between the first and second treatment. Experimental data with animals show that the tolerance dose after reirradiation is inversely proportional to the first dose. For instance, after an initial dose of 50% of the BED_tot and a sufficient time interval of at least 6 months, the retreatment tolerance dose would be approximately 80% of the BED_tot, while after an initial dose close to full tolerance (90% of the BED_tot), only a maximum of 40% of the BED_tot would be left for retreatment. Moreover there probably remained substantial radiation-induced damage after the first (high) dose in our patient. Thus the myelopathy observed in our patient could also well be radiation induced.

The diagnostic pitfall in this case is the erroneous assumption that the complaints of this patient can be attributed to the known intramedullary metastases at Th7, which was confirmed by autopsy. However, this level does not correspond with the eventual clinical findings, which point to a localisation at the cervico-thoracal part of the myelum. These findings, as we have shown by MRI and pathological investigations are most probably caused by radiation myelopathy.

We recommend that in case of complaints after (re-) irradiation, accurate imaging is performed, preferably by means of MRI with gadolinium. One should always be aware of radiation myelopathy, even if the level is not corresponding with the irradiated part of the myelum.

The retreatment dose of 20 Gy in 5 fractions of 4 Gy could better have been reduced, with respect to the relatively high first dose and short interval time, to 3 fractions, or to a scheme with a lower fraction dose, e.g. 5x3 Gy. The BED of this alternative scheme is about 37 Gy, significantly lower than the BED of 60 Gy.

One of the three criteria of myelopathy as result of radiation according to Pallis et al: the main neurological lesion must be within the segments of cord exposed to radiation, is not obligatory. This criterion should be read as: 'the main neurological lesion must be in or near to the segments of the cord exposed to radiation'.

To date we have not found a resembling case in literature.

Robbert JHA Tersteeg

.