Department of Neurosurgery, Johns Hopkins Hospital, Baltimore, Maryland

See the Editorial in this issue, pp 53 55. J Neurosurg Spine 11:56 63, 2009 Outcome following decompressive surgery for different histological types of metastatic tumors causing epidural spinal cord compression Clinical article Ka i s o r n L. Ch a i c h a n a, M.D., Co u rt n e y Pe n d l e t o n, B.S., Da n i e l M. Sc i u b b a, M.D., Je a n-pa u l Wo l i n s k y, M.D., a n d Zi ya L. Go k a s l a n, M.D. Department of Neurosurgery, Johns Hopkins Hospital, Baltimore, Maryland Object. Metastatic epidural spinal cord compression (MESCC) is a relatively common and debilitating complication of metastatic disease that often results in neurological deficits. Recent studies have supported decompressive surgery over radiation therapy for patients who present with MESCC. These studies, however, have grouped all patients with different histological types of metastatic disease into the same study population. The differential outcomes for patients with different histological types of metastatic disease therefore remain unknown. Methods. An institutional database of patients undergoing decompressive surgery for MESCC at an academic tertiary-care institution between 1996 and 2006 was retrospectively reviewed. Patients with primary lung, breast, prostate, kidney, or gastrointestinal (GI) cancer or melanoma were identified. Fisher exact and log-rank analyses were used to compare pre-, peri-, and postoperative variables and survival for patients with these different types of primary cancers. Results. Twenty-seven patients with primary lung cancer, 26 with breast cancer, 20 with prostate cancer, 21 with kidney cancer, 13 with GI cancer, and 7 with melanoma were identified and categorized. All of these patients were followed up for a mean ± SD of 10.8 ± 3.8 months following surgery. Patients with primary lung and prostate cancers were typically older than patients with other types of primary cancers. Patients with prostate cancer had the shortest duration of symptoms and more commonly presented with motor deficits, while patients with breast cancer more commonly had cervical spine involvement and compression fractures. For all histological types, > 90% of patients retained the ability to ambulate following surgery. However, the group with the highest percentage of patients who regained ambulatory function after decompressive surgery was the lung cancer group. Patients with breast or kidney cancer and those with melanoma had the highest median duration of survival following decompressive surgery. Conclusions. The present study identifies differences in presenting symptoms, operative course, perioperative complications, long-term ambulatory outcomes, and duration of survival for patients with lung, breast, prostate, kidney, and GI cancers and melanoma. This understanding may allow better risk stratification for patients with MESCC. (DOI: 10.3171/2009.1.SPINE08657) Ke y Wo r d s breast cancer kidney cancer lung cancer walking prostate cancer metastasis spine Me ta s tat i c cancer causes ~ 500,000 deaths in the US each year. 2 Among osseous sites, the spinal column represents the most common location for metastatic deposits. 1 These spinal deposits can invade the epidural space and compress the spinal cord, leading to motor weakness, sensory deficits, severe pain, and/or paralysis. 7,8,15 In fact, MESCC eventually affects 5 10% of patients with cancer, and the most common primary cancers are lung, breast, prostate, GI, and kidney cancers and melanoma. 3,10,20 Abbreviations used in this paper: GI = gastrointestinal; IQR = interquartile range; MESCC = metastatic epidural spinal cord compression. Treatment for MESCC has historically consisted of corticosteroid agents and radiotherapy. 5,15 However, with recent advances in neuroimaging, surgical techniques, and spinal instrumentation, direct decompressive surgery has become the standard treatment for metastatic lesions from solid primary tumors. 6,14,17 In fact, this approach has recently been shown to be superior to radiotherapy in preserving neurological function in patients with metastatic spine tumors that are not significantly radiosensitive. 17 These studies, however, have grouped all patients with different histological types of metastatic disease into the same study population. 6,14,17 The differential outcomes for patients with different histological types of metastatic disease therefore remain relatively unknown. 56 J Neurosurg: Spine / Volume 11 / July 2009

Metastatic epidural spinal cord compression This study was designed to analyze and compare the outcomes for patients with different histological types of MESCC (lung, breast, prostate, kidney, GI cancers, and melanoma) who were treated at the same institution. Understanding these factors could promote the development of new therapeutic and surgical strategies, as well as aid in clinical decision-making tailored for specific metastatic histological types. Methods J Neurosurg: Spine / Volume 11 / July 2009 Patient Selection A total of 219 cases involving patients who underwent surgery for MESCC at an academic tertiary-care institution between 1996 and 2006 were retrospectively reviewed. Patients at least 18 years old with a tissue-proven diagnosis of a solid primary tumor (not of CNS origin) and MR imaging evidence of spinal cord displacement from its normal position in the spinal canal by an epidural mass were eligible for inclusion in this study. The MESCC had to be restricted to a single area; thus patients with multiple, discrete compressive lesions were excluded. Patients with concomitant brain metastases or cauda equina or spinal root compression were also excluded. Since the goal of this study was to evaluate the outcomes for patients with the most common types of metastatic spine tumors, patients with lung, breast, prostate, kidney, and GI cancers and melanoma were identified. Information recorded for each patient included demographic characteristics, clinical presentation, comorbidities, medications prescribed and administered prior to the day of surgery, preoperative MR imaging findings, intraoperative recordings including somatosensory or motor evoked potential monitoring changes, pathological findings, and postoperative neurological function. Mechanical pain was defined as patient-reported pain that worsened with movement. Radicular pain was pain that followed a dermatomal distribution. Local pain was back pain confined to the region of the spine affected by metastatic disease, and did not include diffuse, nonlocalized pain. Sensory deficits were defined as sensory losses to light touch, temperature, and/or pain confined to a spinal level and/or a nerve root distribution. Patients with multiple presenting symptoms were categorized according to all of their symptoms. On neuroimages, a pathologic vertebral compression fracture was defined as collapse of the vertebral body secondary to tumor involvement. Compression was defined as anterior, lateral, or posterior on the basis of the location in which the spinal cord was compressed from its normal position in the spinal canal. The extent of resection (gross-total vs subtotal) was determined by evaluating pre- and postoperative MR images. By definition, resection was considered gross-total if there was no evidence of residual tumor on postoperative MR images. Additionally, spinal recurrence was defined as local recurrence directly adjacent to the resected lesion. Perioperative mortality was defined as death within 30 days of surgery. In general, the aim of the surgery was to circumferentially decompress the spinal cord. The particular approach was dependent on the location of the tumor and the patient s circumstances. An anterior approach was defined as approaching the spine via the vertebral body first. A posterior approach was defined as approaching the spine via the posterior elements. In some cases involving multidirectional compression, a combined anteriorposterior approach was used. Ambulatory Status and Survival Outcomes The patient s ability to walk, even if a cane or a walker was needed, at the last follow-up visit was used as the primary end point to minimize observer bias and errors associated with retrospective patient classification. 17 This end point has also been demonstrated to be a critical quality-of-life indicator, and accurate assessment of this basic functional measurement was uniformly included in all clinical documentation. 13 The ability to walk at last follow-up included both patients who maintained their preoperative ambulatory status and those who regained the ability to walk postoperatively. The patients who regained the ability to walk are those who were unable to walk preoperatively, but could walk at the last follow-up visit. Survival analysis was conducted using the Social Security Death Index database. 23 Patients whose deaths were unconfirmed were classified as lost to follow-up at the time of the last clinic visit. Statistical Analysis Summary data were presented as means ± SDs for parametric data and as medians with IQRs for nonparametric data. Comparison of percentages between each of the histological groups was performed by means of the Fisher exact test (JMP 7, SAS Institute). Survival as a function of time after resection was expressed using estimated Kaplan-Meier plots (GraphPad Prism 5, GraphPad Software). Survival curves were compared using log-rank analysis (GraphPad). Probability values < 0.05 were considered statistically significant. Results Patient Population The cohort summary data, as well as the comparison between patients with primary lung, breast, prostate, kidney, and GI cancers and melanoma, are outlined in Table 1. Of the original 219 consecutive patients, a total of 114 met the inclusion/exclusion criteria for this study during the reviewed period; the histological type of the primary cancer was lung in 27 of these patients, breast in 26, prostate in 20, kidney in 21, GI in 13, and melanoma in 7. Among the 27 patients with lung cancer, 4 had small cell and 23 had non small cell lung cancer. Chemotherapy and radiation therapy had failed in 2 of the 4 patients with small cell lung cancer; the 2 others underwent emergency surgery because of acute neurological decline. Among the 13 patients with GI cancer, 8 patients had colon cancer, 3 had cholangiocarcinoma, and 2 had appendiceal carcinomas. Of the 114 patients, 63 (55%) were men, and the mean age at the time of surgery for the entire cohort was 58 ± 12 57

K. L. Chaichana et al. TABLE 1: Summary of preoperative characteristics in 114 patients with MESCC stratified by histological type of primary cancer* Characteristic Lung (27 pts) Breast (26 pts) Prostate (20 pts) Kidney (21 pts) GI (13 pts) Melanoma (7 pts) mean age in yrs ± SD sex male female comorbidities smoking DM CAD COPD presenting Sx mean duration in mos inability to walk pain mechanical local radicular motor sensory incontinence prior treatment prior resection radiation treatment chemotherapy preop steroids radiographic findings anterior compression lateral compression posterior compression spinal cord location cervical cervicothoracic thoracic thoracolumbar lumbar compression/wedge fracture extracranial extraspinal metastases 62 ± 11 17 (63) 10 (37) 15 (56) 7 (26) 1.0 7 (26) 20 (74) 3 (11) 14 (52) 15 (56) 18 (67) 2 (7) 15 (56) 16 (59) 9 (33) 12 (44) 8 (30) 7 (26) 3 (11) 18 (67) 2 (7) 6 (22) 12 (44) 11 (41) 56 ± 13 26 (100) 2.4 17 (65) 3 (12) 10 (38) 16 (62) 17 (65) 2 (8) 18 (69) 13 (50) 10 (38) 14 (54) 8 (31) 11 (42) 3 (12) 16 (62) 2 (8) 15 (58) 10 (38) 63 ± 10 20 (100) 3 (15) 0.6 6 (30) 13 (65) 9 (45) 14 (70) 12 (60) 3 (15) 15 (75) 9 (45) 11 (55) 3 (15) 14 (70) 6 (30) 11 (55) 55 ± 11 17 (81) 1 (5) 1.2 5 (24) 14 (67) 8 (38) 10 (48) 1 11 (52) 9 (43) 9 (43) 8 (38) 6 (29) 1 (5) 1 5 (24) 7 (33) 5 (24) 53 ± 8 6 (46) 1.8 9 (69) 6 (46) 5 (38) 8 (62) 9 (69) 6 (46) 4 (31) 8 (62) 59 ± 8 2.0 4 (57) 4 (57) 4 (57) * Values represent numbers of patients (%) unless otherwise indicated. Abbreviations: CAD = coronary artery disease; COPD = chronic obstructive pulmonary disease; DM = diabetes mellitus; pts = patients. Values are significantly higher (p < 0.05) than the value(s) for more than 1 other relatively common histological types of MESCC. years. The median duration of symptoms was 0.5 months (IQR 0.2 1.8 months). Prior to surgery, 26 (23%) of 114 patients were unable to walk as a result of motor weakness in 23 patients and intractable pain in 3. Sixty-three patients (55%) presented with motor symptoms, 72 (63%) with sensory symptoms, and 9 (8%) with bowel and/or bladder incontinence. Additionally, 78 patients (68%) presented with pain; 15 (13%) had mechanical pain, 18 (16%) had local pain, and 51 (45%) had radicular pain. Radiographically, the spinal cord was primarily compressed anteriorly in 56 patients (49%), laterally in 28 (25%), and posteriorly in 33 (29%). The tumor had a cervical component in 29 patients (25%), thoracic components in 75 (66%), and lumbar components in 29 (25%). In 45 patients (39%) there was radiographic evidence of a pathologic compression fracture of the vertebral body. Additionally, 49 patients (43%) had extracranial extraspinal metastases at the time of presentation. Preoperative Variables Patients with lung cancer and MESCC were significantly older than patients with breast (p = 0.04), kidney 58 J Neurosurg: Spine / Volume 11 / July 2009

Metastatic epidural spinal cord compression (p = 0.006), or GI (p = 0.006) cancer. Likewise, patients with prostate cancer and MESCC were significantly older than patients with breast (p = 0.03), kidney (p = 0.02), or GI (p = 0.006) cancer. Not surprisingly, the percentage of patients with prostate cancer who were male (100%) was significantly higher than the percentage of patients with lung (p = 0.003), breast (p < 0.0001), or GI (p = 0.002) cancer or melanoma (p = 0.003) who were male. Likewise, the percentage of patients with breast cancer who were female (100%) was significantly higher than the percentage of patients with lung (p < 0.0001), prostate (p < 0.0001), kidney (p < 0.0001), or GI (p < 0.0001) cancer or melanoma (p = 0.04) who were female. Furthermore, the percentage of patients with lung cancer who were smokers (56%) was significantly higher than the percentage of patients with breast (p < 0.0001), prostate (p = 0.002), kidney (p = 0.006), or GI (p = 0.05) cancer. The percentage of patients with chronic obstructive pulmonary disease was also higher in the group of patients with lung cancer than in the patients with breast (p = 0.01), prostate (p = 0.02), or kidney (p = 0.05) cancer. Patients with prostate cancer had a shorter duration of symptoms prior to surgery than patients with lung (p = 0.04), breast (p = 0.004), and kidney (p = 0.04) cancers. Additionally, the percentage of patients who presented with motor deficits was higher in the group of patients with prostate cancer than in patients with GI cancers (p = 0.04). More patients with lung (p = 0.04) and GI (p = 0.03) cancer received preoperative chemotherapy than patients with prostate cancer. With respect to radiographic findings, the percentage of patients with cervical MESCC was higher in patients with breast cancer than in those with lung cancer (p = 0.04). The percentage of patients with compression fractures was higher in the group of patients with breast cancer than in those with prostate cancer (p = 0.04). No other clinical, imaging, operative, or pathological variables were found to be significantly different between the histological cohorts, including the ability to ambulate preoperatively. Surgical Outcomes The outcomes following surgery for patients with MESCC and lung, breast, prostate, kidney, or GI cancer or melanoma are outlined in Table 2. Fifty-five patients (48%) underwent surgery at > 2 spinal levels; anterior, posterior, and combined anterior-posterior approaches were used in 40 (35%), 34 (30%), and 40 (35%) of the patients, respectively. Gross-total resection of the lesion was performed in 22 patients (19%), and 98 (86%) underwent spinal fusion. In the perioperative period, wound dehiscence developed in 11 patients (10%), 3 (3%) incurred postoperative CSF leaks requiring operative intervention, and 1 (1%) sustained an epidural hematoma requiring operative intervention. Three patients (3%) died during the perioperative period; death was due to pulseless electrical activity arrest, pulmonary embolism, and myocardial infarction in 1 patient each. At a mean (± SD) of 10.8 ± 3.8 months following surgery, 14 (12%) of the patients had undergone additional J Neurosurg: Spine / Volume 11 / July 2009 surgery, 31 (27%) had undergone postoperative radiation therapy, and 28 (25%) had undergone postoperative chemotherapy. Following resection, 91 (80%) patients were able to ambulate at last follow-up. Of the 88 (77%) patients able to ambulate prior to surgery, 81 (92%) maintained the ability to walk and 7 (8%) lost the ability to walk postoperatively. This loss of ambulatory status was due to disease progression in all 7 patients 5 patients developed motor weakness and 2 patients developed intractable pain preventing the ability to walk. Of the 26 patients unable to walk prior to surgery, 7 (27%) regained the ability to walk. At last follow-up, 82 (72%) of 114 patients had died. The median survival for all patients was 7.0 months, with 6-, 12-, 18-, and 24-month survival rates of 50, 38, 30, and 20%, respectively. Of the 27 patients with primary lung cancer, 22 (81%) patients had died at the time of this writing; the median duration of postoperative survival was 4.3 months, with 6-, 12-, 18-, and 24-month survival rates of 18, 18, 12, and 0%. Of the 26 patients with breast cancer, 13 (50%) had died; the median duration of postoperative survival was 21 months, with 6-, 12-, 18-, and 24-month survival rates of 78, 73, 58, and 41%. Of the 20 patients with prostate cancer, 16 (80%) had died; the median duration of postoperative survival was 3.8 months, with 6-, 12-, 18-, and 24-month survival rates of 31, 12, 6, and 6%. Of the 21 patients with kidney cancer, 16 (76%) had died; the median duration of postoperative survival was 19.8 months, with 6-, 12-, 18-, and 24-month survival rates of 74, 53, 47, and 24%. Of the 13 patients with GI cancer, 11 (85%) had died; the median duration of postoperative survival was 5.1 months, with 6-, 12-, 18-, and 24-month survival rates of 25, 17, 8, and 8%. Of the 7 patients with melanoma, 4 (57%) had died; the median duration of postoperative survival was 40.9 months, with 6-, 12-, 18-, and 24-month survival rates of 64, 64, 64, and 32%. Perioperative and Postoperative Variables The percentage of patients who underwent additional surgery was lower in the group of patients with lung cancer than in those with prostate cancer (p = 0.02), melanoma (p = 0.04), or kidney cancer (p = 0.004). The percentage of patients who underwent postoperative radiation therapy was higher in the group with breast cancer than in the patients with lung (p = 0.04) or kidney (p = 0.05) cancer. The percentage of patients treated with chemotherapy postoperatively was lower in the group of patients with GI cancers than in those with breast cancer (p = 0.03). Importantly, the percentage of patients who regained ambulatory function following surgery was higher in the group with lung cancer and MESCC than in patients with breast (p = 0.07), prostate (p = 0.14), or kidney (p = 0.21) cancer and the differences trended toward but did not reach statistical significance. Patients with primary breast cancer lived significantly longer after surgery than patients with primary lung (p = 0.002), prostate (p = 0.004), or GI (p = 0.01) cancer. Patients with primary kidney cancer lived significantly longer than patients with lung (p = 0.001), prostate (p = 0.006), or GI (p = 0.02) cancer. Patients with melanoma lived significantly longer than patients with lung (p = 0.0006), prostate (p = 0.03), 59

K. L. Chaichana et al. TABLE 2: Summary of long-term surgical outcomes in 114 patients with MESCC stratified by histological type of primary cancer Characteristic surgical approach anterior posterior anterior-posterior >2 spinal levels >2 laminectomies monitoring change fusion gross-total resection complications periop mortality wound dehiscence CSF leak epidural hematoma postop treatment additional surgeries radiation therapy chemotherapy Lung (17 pts) 7 (26) 9 (33) 11 (41) 17 (63) 17 (63) 22 (81) 8 (30) 3 (11) Breast (26 pts) 11 (42) 8 (31) 7 (27) 13 (50) 10 (38) 22 (85) 3 (12) 2 (8) Prostate (20 pts) 8 (40) 8 (40) 8 (40) 7 (35) 18 (90) 3 (15) Kidney (21 pts) 6 (29) 11 (52) 8 (38) 7 (33) 1 (5) 20 (95) 1 (5) GI (13 pts) 5 (38) 4 (31) 4 (31) 6 (46) 10 (77) Melanoma (7 pts) 6 (86) * 12 (46)* 11 (42) 6 (30) 6 (29) 5 (24) * spinal recurrence 2 (7) ambulatory outcome ambulatory preop ambulatory postop maintained ambulation regained ambulation 20 (74) 24 (89) 19 (95) 4 (57) 21 (81) 21 (81) 19 (90) 14 (70) 14 (70) 13 (93) 1 (17) 16 (76) 16 (76) 15 (94) 1 (20) 12 (92) 11 (85) 11 (92) median survival (months) 4.3 21.0 3.8 19.8 5.1 40.9 4 (80) 1 (50) * Value is significantly (p < 0.05) different from the value(s) for more than 1 of the other relatively common histological types of MESCC. Percentages were calculated on the basis of the number of patients in each group who were unable to walk at presentation. Trends towards significance. Value is significantly higher (p < 0.05) than the value(s) for more than 1 of the other relative common histological types of MESCC. or GI cancer (p = 0.05). Of note, there was a trend toward longer survival in patients with melanoma compared with patients with kidney cancer (p = 0.07), but the difference did not reach statistical significance. No other outcome variables were found to be significantly different between the histological cohorts. Discussion In this study of 114 patients with MESCC, 27 presented with lung cancer, 26 with breast, 20 with prostate, 21 with kidney, 13 with GI primary cancers, and 7 with melanoma. Patients with primary lung or prostate cancer were typically older than patients with other types of primary cancers. Patients with prostate cancer had the shortest average duration of symptoms prior to presentation and more commonly presented with motor deficits, whereas patients with breast cancer more commonly had cervical spine involvement and compression fractures. For all histological types, > 90% of patients retained the ability to ambulate following surgery. However, the percentage of patients who regained ambulatory function following decompressive surgery was highest in the group of patients with primary lung cancer. Patients with primary breast or kidney cancer and those with melanoma had the longest duration of survival, with median survival times > 12 months. Metastatic epidural spinal cord compression is a common and debilitating complication of cancer, in which the lesion invades the epidural space and compresses the spinal cord. 5,8,15 This compression can lead to neurological deficits and compromise ambulatory function. 6,17 Recently, direct decompressive surgery has become the standard treatment for MESCC in patients with solid primary tumors because of its increased efficacy over conventional radiotherapy in preserving neurological function. 6,14,17 Patchell et al. 17 found that patients who underwent decompressive surgery, as compared with those who received radiation therapy alone, had a greater likelihood of walking after treatment, retaining the ability to 60 J Neurosurg: Spine / Volume 11 / July 2009

Metastatic epidural spinal cord compression walk at last follow-up, and experiencing a reduced need for postoperative steroid and analgesic therapy. Among patients undergoing decompressive surgery, preoperative ability to walk and the presence of compression fractures were independently associated with the ability to walk at last follow-up. 6 Furthermore, patients in whom symptoms were present for < 48 hours and/or who underwent postoperative radiation therapy had an increased chance of regaining the ability to walk. 6 These previous studies, despite demonstrating the efficacy of decompressive surgery for MESCC, did not analyze the differential neurological outcomes for patients with different types of primary cancers. 4,6,11,14,16,17,21,24 Patients with different types of primary cancers were all considered together in the same study population. It is difficult to conceive, however, that patients with different types of primary cancers will have similar presenting symptoms and neurological outcomes. In fact, it is well known that some tumors, including metastatic myeloma, lymphomas, and germ cell tumors, are more radiosensitive than most solid tumors. 9 Furthermore, some studies have found that patients with prostate and lung cancer along with MESSC have decreased survival times as compared with patients with other types of solid primary cancers and MESCC. 18,25,28 Nevertheless, the differential presenting symptoms and neurological outcomes for patients with different types of MESCC remain poorly understood. The primary cancers with the greatest propensity to metastasize and cause spinal cord compression are lung, breast, prostate, kidney, and GI cancers and melanoma. 3,10,20 In this case series, patients with lung or prostate cancer were significantly older than patients with other types of primary cancers. Not surprisingly, in the group of patients with lung cancer there was a significantly higher fraction of patients who were smokers and/ or had a history of chronic obstructive pulmonary disorder. Importantly, patients with prostate cancer had the highest prevalence of preoperative motor deficits and the shortest average duration of preoperative symptoms prior to surgical intervention. This may mean that patients with prostate cancer may potentially warrant neurosurgical intervention on a more urgent basis than patients with other primary cancers. Because preoperative neurological function often predicts postoperative neurological outcome in MESCC, 6,14,17 patients with prostate cancer should be closely monitored for the development of motor deficits. Additionally, patients with breast cancer had the highest incidence of vertebral compression fractures. This may be due to the nature of this type of tumor, or the fact that osteoporosis is more common in female patients. 19 Since vertebral compression fractures have been associated with poor ambulatory outcomes, 6 patients with breast cancer may be at increased risk of losing the ability to walk as compared with patients with other types of cancer. With respect to the extent of surgery, the betweengroups differences were not statistically significant; however, patients with lung cancer had more extensive surgeries than the other patients. A greater percentage of patients in this group had surgery that involved more than J Neurosurg: Spine / Volume 11 / July 2009 2 spinal levels and/or required laminectomies spanning 2 or more levels. Patients with lung cancers may therefore have more extensive disease requiring more extensive surgery. Perioperatively, patients with GI cancer had the highest rate of incidence of wound dehiscence requiring operative intervention. The incidence rate was not significantly different from that observed in the other cohorts, but patients with GI cancer may have a greater degree of impaired wound healing. Likewise, although the difference was not statistically significant, patients with primary lung cancer had the highest rate of incidence of CSF leaks requiring operative intervention. This finding may be a consequence of these patients requiring more extensive surgeries. With respect to ambulatory function, decompressive surgery has been shown to provide better results than radiation therapy. 5,8,15 In the present study, regardless of tumor histology, the overwhelming majority of patients were able to maintain ambulation at last follow-up. However, the percentage of patients able to walk was lowest in the groups with prostate cancer and melanoma, with respect to both pre- and postoperative ambulatory function. Although the difference between these groups and the others was not statistically significant, this finding may mean that patients with these tumors have worse ambulatory outcomes. Finally, the rate of recovery of ambulatory function was highest in patients with lung cancer, with the difference trending toward significance, and these patients may therefore benefit most from decompressive surgery. This finding is particularly interesting given that the prognosis of metastatic lung cancer is usually extremely grim and the expected duration of survival is short; thus, patients may not be offered major spine surgery. However, such patients, when treated with decompressive surgery, showed substantial ability to maintain and/or regain ambulation, which is a large contributor to quality of life. 12 Surgery may therefore optimize ambulatory function in these patients with lung cancer despite their overall short duration of survival. Patients with metastatic disease are considered to have poor survival in general, and patients with spinal metastases may have particularly poor survival rates. 7,26,27 Prior studies have reported median durations of survival ranging from 10 to 30 months in patients with metastatic spine disease, including those with and without epidural cord compression as well as those with diverse histological types. 22,24 In this study, patients who had prolonged survival were those with primary breast and kidney cancers and melanoma, with median postoperative duration of survival > 18 months. In patients with primary lung, prostate, and GI cancers, on the other hand, the median duration of survival was < 6 months. Based on these survival times, it intuitively makes sense that patients with primary breast or kidney cancer or melanoma and MESSC would benefit from surgical decompression. For patients with primary lung, prostate, or GI cancer, some clinicians would argue that aggressive surgical decompression is not warranted given the poor survival times in these patient groups. However, it can also be argued that surgery may maximize these patients quality of life despite their reduced survival. 61

K. L. Chaichana et al. Limitations of This Study The goal of this study was to identify clinical differences between patients presenting with the most common types of MESCC. The primary limiting factor was the number of patients in each histological cohort, which prevented drawing definitive conclusions using multivariate models. However, to our knowledge, the current study is, among published studies, the one with the largest number of patients per cohort who met the inclusion criteria originally established by Patchell et al. 17 in their randomized prospective clinical trial. We used these criteria in order to identify the patients at highest risk of spinal cord related paralysis. Patient selection and differences in surgical skill and judgment between surgeons and over time also inherently limit this study. We included patients with neuroimaging evidence of spinal cord compression who underwent surgery performed at a single institution by our dedicated spine team; we excluded patients with very radiosensitive tumors, concomitant brain metastases, and/or cauda equina or spinal root compression. Therefore, the results of this study do not necessarily apply to patients with very radiosensitive tumors, metastatic involvement limited to osseous structures, or brain metastases. Nevertheless, we feel this study adds to the work of previous studies by demonstrating differences among patients with different primary cancers and MESCC. This is important because the majority of previous clinical studies include all patients with MESCC, regardless of primary pathology, in the same study population. This study is also inherently limited by its retrospective design, and, as a result, it is not appropriate to infer direct causal relationships. We opted to use ambulation as the primary outcome to minimize observer bias and errors associated with retrospective patient classification. This end point is also important because it has been demonstrated to be a critical quality-of-life indicator, and accurate assessment of this basic functional measure was uniformly included in all clinical documentation. Duration of survival was also assessed to ascertain how surgery affects survival for patients with metastatic spine disease. Duration of survival for patients with metastatic disease is poor, which often makes clinicians wonder if aggressive treatment is even warranted. It should be noted, however, that many patients, and even clinicians, would argue that prolonged functional independence is more important than prolonged survival with functional impairment. We believe our findings offer useful insights into the care of patients with MESCC. Nevertheless, prospective and higher-volume studies are needed to provide better data to guide clinical decision-making. Conclusions Recent studies have supported the efficacy of decompressive surgery at maintaining the ability to walk for patients with MESCC from solid primary tumors. These studies, however, have grouped patients with distinct types of primary cancers all into the same study population. Studies identifying differences between patients with different primary cancers remain few and limited. The present study identifies differences in presenting symptoms, operative course, perioperative complications, long-term ambulatory outcomes, and duration of survival for patients with lung, breast, prostate, kidney, and GI cancers and melanoma. This understanding may allow for better risk stratification for patients with MESCC. Disclaimer The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper. References 1. Aaron AD: The management of cancer metastatic to bone. JAMA 272:1206 1209, 1994 2. American Cancer Society: Cancer Facts and Figures 2005. Atlanta: American Cancer Society, 2005 3. Barron KD, Hirano A, Araki S, Terry RD: Experiences with metastatic neoplasms involving the spinal cord. Neurology 9: 91 106, 1959 4. Bauer HC: Posterior decompression and stabilization for spinal metastases. Analysis of sixty-seven consecutive patients. J Bone Joint Surg Am 79:514 522, 1997 5. Byrne TN: Spinal cord compression from epidural metastases. N Engl J Med 327:614 619, 1992 6. Chaichana KL, Woodworth GF, Sciubba DM, McGirt MJ, Witham TJ, Bydon A, et al: Predictors of ambulatory function after decompressive surgery for metastatic epidural spinal cord compression. Neurosurgery 62:683 692, 2008 7. Chen CH, Tzai TS, Huang SP, Wu HC, Tai HC, Chang YH, et al: Clinical outcome of Taiwanese men with metastatic prostate cancer compared with other ethnic groups. Urology 72:1287 1292, 2008 8. Findlay GF: Adverse effects of the management of malignant spinal cord compression. J Neurol Neurosurg Psychiatry 47:761 768, 1984 9. Finn MA, Vrionis FD, Schmidt MH: Spinal radiosurgery for metastatic disease of the spine. Cancer Control 14:405 411, 2007 10. Gerszten PC, Welch WC: Current surgical management of metastatic spinal disease. Oncology (Williston Park) 14:1013 1024, 2000 11. Hatrick NC, Lucas JD, Timothy AR, Smith MA: The sur - g ical treatment of metastatic disease of the spine. Radiother Oncol 56:335 339, 2000 12. Ibrahim A, Crockard A, Antonietti P, Boriani S, Bunger C, Gasbarrini A, et al: Does spinal surgery improve the quality of life for those with extradural (spinal) osseous metastases? An international multicenter prospective observational study of 223 patients. Invited submission from the Joint Section Meeting on Disorders of the Spine and Peripheral Nerves, March 2007. J Neurosurg Spine 8:271 278, 2008 13. Jain NB, Sullivan M, Kazis LE, Tun CG, Garshick E: Factors associated with health-related quality of life in chronic spinal cord injury. Am J Phys Med Rehabil 86:387 396, 2007 14. Klimo P Jr, Thompson CJ, Kestle JR, Schmidt MH: A metaanalysis of surgery versus conventional radiotherapy for the treatment of metastatic spinal epidural disease. Neuro Oncol 7: 64 76, 2005 15. Loblaw DA, Perry J, Chambers A, Laperriere NJ: Systematic review of the diagnosis and management of malignant extradural spinal cord compression: the Cancer Care Ontario Practice Guidelines Initiative's Neuro-Oncology Disease Site Group. J Clin Oncol 23:2028 2037, 2005 16. Moore AJ, Uttley D: Anterior decompression and stabilization 62 J Neurosurg: Spine / Volume 11 / July 2009

Metastatic epidural spinal cord compression of the spine in malignant disease. Neurosurgery 24:713 717, 1989 17. Patchell RA, Tibbs PA, Regine WF, Payne R, Saris S, Kryscio RJ, et al: Direct decompressive surgical resection in the treatment of spinal cord compression caused by metastatic cancer: a randomised trial. Lancet 366:643 648, 2005 18. Rompe JD, Hopf CG, Eysel P: Outcome after palliative posterior surgery for metastatic disease of the spine evaluation of 106 consecutive patients after decompression and stabilisation with the Cotrel-Dubousset instrumentation. Arch Orthop Trauma Surg 119:394 400, 1999 19. Saarto T, Vehmanen L, Blomqvist C, Elomaa I: Ten-year follow-up of 3 years of oral adjuvant clodronate therapy shows significant prevention of osteoporosis in early-stage breast cancer. J Clin Oncol 26:4289 4295, 2008 20. Schaberg J, Gainor BJ: A profile of metastatic carcinoma of the spine. Spine 10:19 20, 1985 21. Schoeggl A, Reddy M, Matula C: Neurological outcome following laminectomy in spinal metastases. Spinal Cord 40: 363 366, 2002 22. Sioutos PJ, Arbit E, Meshulam CF, Galicich JH: Spinal metastases from solid tumors. Analysis of factors affecting survival. Cancer 76:1453 1459, 1995 23. Social Security Administration: Social Security Death In - dex Database. http://ssdi.rootsweb.ancestry.com/cgi-bin/ ssdi.cgi [Accessed 16 April 2009] 24. Sundaresan N, Rothman A, Manhart K, Kelliher K: Surgery for solitary metastases of the spine: rationale and results of treatment. Spine 27:1802 1806, 2002 25. Sundaresan N, Steinberger AA, Moore F, Sachdev VP, Krol G, Hough L, et al: Indications and results of combined anteriorposterior approaches for spine tumor surgery. J Neurosurg 85: 438 446, 1996 26. Thatcher N, Chang A, Parikh P, Rodrigues Pereira J, Ciuleanu T, von Pawel J, et al: Gefitinib plus best supportive care in previously treated patients with refractory advanced nonsmall-cell lung cancer: results from a randomised, placebocontrolled, multicentre study (Iressa Survival Evaluation in Lung Cancer). Lancet 366:1527 1537, 2005 27. van der Linden YM, Dijkstra SP, Vonk EJ, Marijnen CA, Leer JW: Prediction of survival in patients with metastases in the spinal column: results based on a randomized trial of radiotherapy. Cancer 103:320 328, 2005 28. Wise JJ, Fischgrund JS, Herkowitz HN, Montgomery D, Kurz LT: Complication, survival rates, and risk factors of surgery for metastatic disease of the spine. Spine 24:1943 1951, 1999 Manuscript submitted September 29, 2008. Accepted January 23, 2009. Address correspondence to: Kaisorn L. Chaichana, M.D., De partment of Neurosurgery, Johns Hopkins Hospital, 600 North Wolfe Street, Meyer 8-161, Baltimore, Maryland 21287. email: kaisorn@ jhmi.edu. J Neurosurg: Spine / Volume 11 / July 2009 63