Prognostic Factors and Treatment of Spinal Astrocytomas

A Multi-institutional Cohort Analysis

Yingjie Zou, MD; James Sun, BA; Yangying Zhou, BS; Harrison Xiao Bai, MD; Xiangyan Huang, MD; Ranjith Babu, MD; Alessandro Landi, MD; Kap Sum Foong, MD; Zishu Zhang, MD; John H. Woo, MD; Yongguang Tao, PhD; Xuejun Li, MD, PhD; Xiangqi Tang, MD, PhD; Bo Xiao, MD, PhD; Paul J. Zhang, MD; Li Yang, MD, PhD


Spine. 2018;43(10):E565-E573. 

In This Article


Spinal astrocytomas are rare intramedullary spinal cord tumors with a neuroepithelial origin that can potentially lead to severe neurologic deterioration, decreased function, poor quality of life, or even death.[8] Given its low incidence, it is unlikely that randomized prospective trials will ever be completed. Analysis of our multi-institutional cohort consisted of 94 patients from four institutions between 1976 and 2016. At the same time, literature data consisting of 339 individual patients from 72 published studies was analyzed to make comparison with the results of our multi-institutional cohort analysis.


Previous large studies of strictly spinal astrocytoma analyzing age as a prognostic factor have reported heterogeneous results.[10,12–16] In the largest institutional study of spinal astrocytoma consisting of 136 consecutive patients treated at Mayo Clinic from 1962 to 2005, Minehan et al[13] showed that younger age predicted better outcome after accounting for tumor grade, diagnosis after 1984, symptom duration, extent of resection, and postoperative RT. In the study by Minehan et al, patients younger than 20 years were more likely to have a lower-grade tumor (pilocytic vs. infiltrative) than older patients. However, in the study by Adams et al,[10] where the authors reviewed the SEER database of 135 patients from 1973 to 2007, adult patients had longer survival than pediatric group (16 vs. 9-month median survival). Of note, only high-grade tumors (WHO grade III and IV) were included in the analysis of the study by Adams et al,[10] which may have accounted for the contradictory result.

Several smaller studies demonstrated no prognostic significance of age for survival, but were likely underpowered.[14–16] Our results demonstrated that age was a significant predictor of OS. Specifically, adult patient (age ≥ 18) had shorter OS than pediatric patients (age <18).

Clinical Symptom

Paresthesia as a presenting symptom was found to be associated with shorter OS on multivariate analysis for our multi-institutional cohort. Similarly, Fakreddine et al[14] included 83 patients treated at MD Anderson and found that patients with more than one neurologic deficit (pain, motor, or sensory) had significantly worse PFS (P = 0.0496) for infiltrative patients which may correspond to the extent of tumor infiltration. However, nonsignificant results for the relationship between clinical symptoms and survival were also reported.[13]

World Health Organization Grade

We found that WHO grade was a significant prognostic factor for both PFS and OS. Specifically, patients with higher-grade tumors (WHO grades III and IV) had worse survival than those with lower-grade tumors (WHO grades I and II). Most studies reported similar results.[13,14,16–18] Minehan et al[13] showed patients with pilocytic tumors (WHO grade I) survived significantly longer than those with infiltrative astrocytomas (WHO grades II–IV) (39.9 vs. 1.85 yr, P < 0.001). Kim et al[17] demonstrated a significant difference in survival between low- and high-grade spinal astrocytomas (184 vs. 8 mo, P < 0.05). These results hold true for pediatric patients as well.[16,18]

Extent of Surgical Resection

The extent of surgical resection on the outcome of spinal astrocytomas is currently controversial.[11,12] More extensive surgical resection can result in additional morbidities that are ultimately detrimental for the patients.[13] Conceivably, for infiltrative spinal astrocytomas (WHO grade II–IV), surgery is more likely to cause damage to vital neuronal axons.[13] The development of new intraoperative microsurgical techniques and imaging over the past decade has significantly improved patient outcomes after surgical resection.[5]

Adams et al[10] revealed that extent of resection had a significant association with OS on both univariate (P = 0.041) and multivariate analyses (P = 0.054). A median OS of 24 months was observed for cases undergoing GTR, whereas STR, biopsy, and no surgery groups had a median survival of 11, 10, and 9 months, respectively.[10] However, in the study of 136 patients by Minehan et al,[13] minimal extent of surgery was a positive predictor OS in the multivariate analysis. These contradictory results can be explained by the long period of the time over which the study recruited patients (1962–2005) since surgical techniques have likely improved over the years.[9] A common limitation of the two studies cited above was the lack of data on PFS. Several smaller studies including PFS as an outcome have also revealed contradictory results.[14,16] In the MD Anderson study by Fakreddine et al,[14] less than GTR showed a trend toward worse PFS among pilocytic patients (P = 0.07), but was not associated with OS (P = 0.35). In infiltrative tumors (WHO grades II–IV), extent of resection was not associated with either PFS or OS.[14] Another study of 73 pediatric patients from 13 French treatment centers found that the extent of tumor resection did not have a significant impact on patient's survival in the univariate analysis.[16]

Our multi-institutional cohort analysis demonstrated that GTR trended toward longer PFS and OS, whereas the literature cohort showed that GTR was associated with longer PFS but not OS when compared to STR.

Functional Status

A risk that has long served as a deterrent to aggressive surgical resection in treatment of spinal astrocytomas and other intramedullary lesions has been postoperative deterioration in functional status. Contrary to previous work,[19] we find that the change in McCormick grade among our cohort was significant: in approximately a quarter of cases, a patient's McCormick grade increased after operation. We note that in only one case did the score increase by more than one on the McCormick scale. Surprisingly, subgroup analysis did not uncover a bias toward change in functional status by extent of resection. Even in the group, which did not receive surgical treatment, patients experienced a decline in neurological function in 5 of 23 cases (21.7%). Although the risk of motor deficits caused by aggressive resection continues to present a challenge to successful management of spinal astrocytomas, the conscientious neurosurgeon must also consider whether inaction might also result in neurological decline. Further study is needed to identify which factors are driving susceptibility to worsening functional status in spinal astrocytoma.


The effect of chemotherapy on survival was not analyzed in many studies either because the cohort size was too small[13,16] or because the relevant data were not available.[10,15] We find insufficient evidence in our cohort and in the literature to make a strong claim about the efficacy of drug treatment. Thus, the role of chemotherapy in management of spinal astrocytomas remains poorly defined.

For RT, previous studies have tended to report nonsignificance with respect to survival.[14,16,17] Although our survey of the literature agrees with the lack of notable effect, our multi-institutional cohort analysis showed that RT was associated with longer OS after controlling for other covariates. Patients who received RT in our cohort were much more likely to carry tumors of grades III or IV than those who did not receive RT (70.4% vs. 27.8%, respectively). Importantly, the contrast between the univariate result of worse OS given RT and the multivariate result of improved OS indicates the sensitivity of results to adequate capture of patient and disease characteristics. Based on these data, adjuvant RT appears to be effective in improving survival outcomes for high-grade tumors.