In regards to decision making for reirradiation of a recurrent intramedullary spinal cord metastasis

1 Discussion

The optimal treatment of an intramedullary lesion within a previously irradiated spinal cord is a difficult clinical dilemma. For select patients who survive long enough for intramedullary disease to recur after initial palliative therapy, the management of recurrent disease is far more complicated. The radiation oncologist must carefully balance the competing risks of myelopathy with the potential toxicity of a progressive tumor. Though the optimal dose and fractionation for hypofractionated radiotherapy continues to be studied (1,2), clinical studies have shown that the actual incidence of myelopathy and other toxicities following reirradiation of vertebral metastases with standard 3D conformal radiotherapy (3DCRT) is low (Table 1). Of 13 surveyed studies covering 546 patients and 601 treatments, only three studies reported toxicity, occurring in seven patients (3-15). There was only one report of grade 4 neurotoxicity (15). Local control subsequent to reirradiation was promising in all studies, from 61% to 100%.

Table 1.

Reirradiation of vertebral body metastases outcomes from the literature.

Study	# Patients / treatments	Prior Radiation Data (median values)		Reirradiation Data (median values)		Inter-RT Interval (Months) (median)	Follow-up (months)	Toxicity	Local Control
		Dose (Gy)	# Fractions	Dose (Gy)	# Fractions
Schiff et al. 1995	54 / 54	30	10	22	11	9.1	4.7	None	92.5%
Sminia et al. 2002	8 / 8	30.1	7	21	7	30	12.2	None	NR
Milker-Zabel et al. 2003	18 / 19	38	NR	39.6	n/a	17.7	12.3	None	94.7%
Mahan et al. 2005	8 / 8	30	13.5	30	10-15	n/a	15.2	None	100%
Rades et al. 2005	62 / 62	20 (n=28) 8 (n=34)	5 (20Gy) 1 (8Gy)	15-20 (n=28) 8 (n=34)	1, 5	6	8	None	85%
Saghal et al. 2009	25 / 37	36	14	24	3	11	7	None	69%
Choi et al. 2010	42 / 51	40	NR	20	2	19	7	Grade 4 Neurotox. (n=1)	73%
Sterzing et al. 2010	36 / 36	36.3 (mean)	20	34.8 (mean)	11	17.5	7.5	None	1-year 76% 2-year 63%
Damast et al. 2010	94 / 97	30	NR	20 (n=42) 30 (n=55)	5	18 for 20 Gy 25 for 30 Gy	12.1	None	1-year 66%
Mahadevan et al. 2011	60 / 81	30	10	24-30	3-5	20	12	Radiculopathy (n=3) LE weakness (n=1)	93%
Garg et al. 2011	59 / 63	30	NR	27, 30	3, 5	n/a	13	Grade 3 Neurotox. (n = 2)	76%
Navarria et al. 2012	31 / 31	30	10	30	11	17	10	None	61%
Chang et al. 2012	49 / 54	39.2	NR	27	3	24.5	17.3	None	79%

Treatment of vertebral tumors with stereotactic body radiotherapy (SBRT) is an area of active investigation, with hopes that precisely delivered high doses could lead to more rapid pain relief and prolonged cancer control. Studies have demonstrated the usefulness of this technique in treating paraspinal tumors with high rates of local control and complete pain response (16). In SBRT for vertebral metastases, the use of highly conformal radiotherapy, stereotaxis, and patient immobilization allows for a steep dose gradient between the vertebral target and the adjacent spinal cord segment, reducing the reirradiation exposure to that segment (17).

This is juxtaposed against the treatment of an intramedullary tumor, where no meaningful separation exists between the tumor and normal cord. Additionally, the highest dose regions (so called “hot spots”) are likely to occur within the GTV or PTV, which is situated within the cord in intramedullary tumor reirradiation versus in the bone in the case of vertebral metastasis treatment. Consequently, reirradiation of an intramedullary tumor is significantly more difficult and more fraught with risk when compared to that of a vertebral metastasis. The few small studies that have examined SBRT for intramedullary metastases have reported relatively favorable results, though. The Stanford experience reviewed nine patients treated with a median 21Gy in 3 fractions, and with a median survival of 4.1 months, no neurologic toxicities were noted (18). Henry Ford Hospital published their experience of 9 patients with 11 intradural and intramedullary metastases. The median marginal dose was 14Gy in 1 fraction. Median survival was 8 months, 80% had an improvement in symptoms, and no toxicities were observed (19). Lastly, the University of Pittsburgh reported on a single patient with an intramedullary C5 metastasis, initially treated with conventional irradiation to 30Gy in 10 fractions. Several weeks thereafter he was noted to have continued pain and paresthesias, and was treated with SBRT to 15Gy in 3 fractions. The patient was followed for 26 months and had an improvement in pain, albeit with continued paresthesias but no treatment related toxicity (20).

Ultimately, the decision to deliver reirradiation with standard fractionated 3DCRT as opposed to SBRT depends on the relative risk of myelopathy from the relatively uniform irradiation of a larger volume of the spinal cord using standard fractionated 3DCRT compared to the irradiation of a smaller volume of cord immediately adjacent to the tumor with more heterogeneous higher doses of radiation with SBRT. This point is illustrated in Figure 1, which compares three 4Gy fractions of 3DCRT to a single fraction of 10Gy with SBRT, two retreatment regimens considered in our patient. Given that the spinal cord is the quintessential example of an organ with functional subunits arrayed in series, injury to a discrete segment of this organ can result in derangement in function of the organ as a whole. Therefore, it stands to reason that SBRT is theoretically more capable of irreparably injuring even a small number of functional subunits. However, mouse models indicate that increasing length of spinal cord irradiated results in a higher potential for myelopathy (21), arguing for the treatment of the smallest volumes possible, no matter the approach used.

Figure 1.

A comparison of dose volume histograms for dose to the spinal cord, showing the theoretical difference between a single high dose of radiotherapy (solid line) and multifraction standard irradiation (dashed line).

Importantly, the human spinal cord exhibits significant recovery from radiation injury in a time and dose dependent fashion. In a review of 40 patients treated with reirradiation, Nieder et al. did not observe myelopathy when a cumulative spinal BED₂ (assuming an α/β of 2Gy) of ≤ 150Gy₂ (2Gy equivalents) was delivered at least six months after an initial course with BED₂ < 102Gy₂. Moreover, they proposed a risk stratification system that places patients into low (0%), intermediate (33%), and high risk (90%) of myelopathy (22). Ang et al. underscored the importance of time between radiotherapy courses in a primate reirradiation study, wherein several recovery models were constructed. Based on their data, they suggested that 50%, 60%, and 65-70% of occult spinal injury is recovered within 1, 2, and 3 years after radiation to 44Gy (23).

2 Conclusion

The optimal treatment of previously irradiated intramedullary tumors presents a difficult clinical dilemma, and the potential toxicity of treatment must be balanced with the toxicity of disease progression. We feel that the selective use of highly conformal techniques such as SBRT can be employed for this purpose in experienced centers when treating a patient with a relatively favorable prognosis and good performance status, given that the interval between radiotherapy courses is greater than six months (preferably greater than a year), that the initial treatment was less than 102 Gy₂, and that the cumulative cord segment dose is kept beneath 150Gy₂. In patients who do not fit these criteria, more conventional means of palliation should be pursued, including standard fractionated radiotherapy or hospice care, as appropriate.