Journal of Clinical Neuroscience xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Journal of Clinical Neuroscience journal homepage: www.elsevier.com/locate/jocn
Clinical Study
Outcomes and patterns of care in adult skull base chordomas from the Surveillance, Epidemiology, and End Results (SEER) database Pamela S. Jones a, Manish K. Aghi b, Alona Muzikansky c, Helen A. Shih d, Fred G. Barker II a, William T. Curry Jr. a,⇑ a
Department of Neurosurgery, Stephen L. Harris Center for Chordoma Care, Massachusetts General Hospital, 55 Fruit Street, Yawkey 9E-9026, Boston, MA 02114, USA Department of Neurosurgery, University of California, San Francisco, 5 San Francisco, CA, USA Biostatistics Center, Massachusetts General Hospital, Boston, MA, USA d Department of Radiation Oncology, Massachusetts General Hospital, Boston, MA, USA b c
a r t i c l e
i n f o
Article history: Received 29 January 2014 Accepted 8 February 2014 Available online xxxx Keywords: Chordoma Population study Radiation therapy Skull base tumors
a b s t r a c t This study aims to demonstrate survival rates and treatment patterns among patients with chordomas of the skull base using a large population database. Patients with cranial chordomas between 1973 and 2009 were identified from the USA Surveillance, Epidemiology, and End Results (SEER) public use database. Kaplan–Meier analysis was used to examine the effect of surgery and radiation on overall survival. We identified 394 patients with histologically-confirmed cranial chordomas. Median survival was 151 months. Most patients (89.09%) underwent surgery. Less than half (44.92%) received radiation after diagnosis. Patients who underwent surgical resection survived significantly longer than those who did not undergo resection, regardless of other treatments (151 versus 81 months, p < 0.001). Ten year survival was lower among patients receiving radiation (44.8% versus 61.4%, p = 0.66). Surgery predicted better overall survival by univariate analysis (hazard ratio [HR] 0.603, p = 0.0293); younger age at diagnosis (HR 1.028, p < 0.001), and later year of diagnosis (HR 0.971, p = 0.0027) were prognostic of improved survival in a multivariate model. In subgroup analysis of patients with documented tumor size, smaller tumor size (HR 1.021, p = 0.0067), younger age (HR 1.031, p = 0.001), and treatment within a higher volume registry (HR 0.490, p = 0.0129) predicted improved survival. Surgical intervention offers survival benefit for cranial chordomas. Findings of decreased survival in patients receiving radiation may be associated with selection. Studies examining surgical extent of resection data and radiation details are needed to determine the impact of radiotherapy. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction Chordoma is a rare cancer arising from notochord remnants, with approximately 30–40% occurring at the skull base, mostly arising from the clivus [1,2]. Although typically slow-growing, these tumors are locally destructive and frequently recur following treatment. Proximity to critical structures such as the brainstem, cranial nerves, and major intracranial vessels combined with the high recurrence rate renders these histologically low-grade neoplasms clinically malignant, and chordoma management is a challenge for patients and health care providers alike [3]. Due to the rarity of the disease, current treatment paradigms are largely based on single-center retrospective studies, and to our knowledge no prospective comparative studies exist in the ⇑ Corresponding author. Tel.: +1 617 724 0884; fax: +1 617 724 0887. E-mail address: [email protected] (W.T. Curry Jr.).
literature [4]. Typically, surgical resection is the first-line treatment, and radiation may be used for residual tumor, recurrent tumor, cases considered unresectable, or to the operative site following a gross total resection. Little consensus exists over whether radiation should be integrated into standard therapy following surgery whether or not there is evident residual tumor [2,5–7]. Extensive surgical resection is technically challenging and morbid, although advances in technology, such as the use of endoscopy, neuronavigation, and neuromonitoring have facilitated greater access to clival chordomas with improved safety [7–9]. There is also a more widespread capacity to deliver high-dose conformal radiation, including proton beam radiotherapy, intensity modulated radiotherapy, photon-based radiosurgery and hypofractionated radiotherapy [10,11]. In this analysis we describe treatment and survival trends of patients with skull base chordoma over a 36 year period using data obtained from the USA Surveillance, Epidemiology, and End Results
http://dx.doi.org/10.1016/j.jocn.2014.02.008 0967-5868/Ó 2014 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Jones PS et al. Outcomes and patterns of care in adult skull base chordomas from the Surveillance, Epidemiology, and End Results (SEER) database. J Clin Neurosci (2014), http://dx.doi.org/10.1016/j.jocn.2014.02.008
2
P.S. Jones et al. / Journal of Clinical Neuroscience xxx (2014) xxx–xxx
(SEER) registry. We paid particular attention to outcomes and trends in radiation utilization.
strategies were compared using chi-squared analysis. Calculations were performed using SAS version 9.3 (SAS Institute, Cary, NC, USA).
2. Methods 3. Results 2.1. SEER program 3.1. Demographics The SEER registry, a database maintained by the USA National Cancer Institute, collects incidence and survival data from population-based cancer registries. Data on chordoma patients diagnosed between 1973 and 2009 were available from 18 population-based cancer registries across the USA that together represent approximately 25% of the USA population. The 18 registries include nine entire states and nine metropolitan regions. These registry populations are comparable to the overall USA population on a variety of socioeconomic parameters, including poverty indices, high school graduation rates, percentage in rural areas, and percentage of foreign-born residents. The database contains information on patient demographics, primary cancer site, histology, methods of diagnostic confirmation, treatment regimens (including surgery and radiation therapy), and year of death. Institutional Review Board approval was not required for this study, as the SEER database is free of any sensitive patient information or identifiers. An overview of the SEER database is available at http://seer.cancer.gov/about/.
Between 1973 and 2009, 422 patients were diagnosed with skull base chordoma in the SEER database. Of these, 394 patients satisfied the criteria for histologically-confirmed disease and were included in our analysis. The median age at diagnosis was 46 years, with an age range of 0 to 92, and a higher proportion of patients were male (55.6%). The majority of patients (220 patients, 55.8%) were diagnosed before the median diagnosis year 2002. Cases occurred across 16 registries, of which nine were considered lowvolume (118 patients) and seven were considered high-volume (276 patients). The results are summarized in Table 1. 3.2. Tumor characteristics Information regarding tumor size was available in 196 of 394 patients. Mean diameter was 3.7 cm (range 0.2 to 8.5 cm). Information regarding histologic subtype of chordoma was not available.
2.2. Chordoma case definition 3.3. Treatment We compiled chordoma cases diagnosed between 1973 and 2009. Patients were identified through the World Health Organization’s International Classification of Diseases for Oncology, Version 3 (ICD-O-3) morphology code for chordoma (9370–9373). Primary sites of disease presentation were compiled according to ICD-O-3 topography codes and were limited to skull and face bones (410); meninges (700); any structure in the brain (710–9, 751, 753); connective and soft tissue of head, face and neck (490); or any head structure (760). We restricted our analysis to patients with histologic confirmation of chordoma.
Most patients (351 of 394, 89.1%) underwent some type of surgical resection beyond biopsy. The SEER data only specified whether surgery was performed, and did not reliably specify extent of skull base tumor resection. We made the assumption that patients who were not coded as having had surgery underwent biopsy alone. Ultimately, 196 patients were treated with surgery alone (49.7%), 155 patients were treated with surgery and radiation (39.3%), and 22 patients treated with radiation alone (5.5%). Twenty-one patients received no treatment (5.3%). Radiation utilization did not increase over time (chi-squared analysis, p = 0.7),
2.3. Treatment Treatment types were limited to surgery and radiation. The SEER surgery code of 0 was used to indicate that surgery was performed. For the purpose of our analysis, all surgical cases were considered to have undergone an attempt at tumor resection. All other codes were considered cases where only a diagnostic biopsy had been performed. Adequate information on extent of resection was not available from the database. The SEER radiation codes 0, 6, 7, 8, 9 represented no radiation treatment. We further categorized registries by number of skull base cases in order to assess for potential volume-outcome effect, whereby better outcomes are observed at treatment centers with high volume. We looked at numbers of both skull base chordomas and chondrosarcomas (ICD-O-3 code 9220–9239) treated by each registry to assess for volume of skull base cases. A registry was considered to contain a high-volume treatment center if it contained 40 or more histologically-confirmed skull base chordomas and chondrosarcomas during the study period. 2.4. Statistical analysis The primary outcome of interest was overall survival. Survival analyses were performed using Kaplan–Meier product limit estimators and Cox univariate and multivariate proportional hazards models. Predictors of overall survival in univariate analysis (p < 0.05) were included in the multivariate model. Treatment
Table 1 Characteristics of 394 patients with chordoma tumors of the skull base registered in the SEER database (1973–2009) Patients, n Mean age, years (± SD) Median age, years Males
394 45.3 (± 19.2) 46 219 (55.6)
Race White African American Asian Native American
333 (85.4) 13 (3.3) 42 (10.8) 2 (0.5)
Registry Low volume (2002
2002 median year (1999 mean) 220 (50.8) 174 (44.2)
Surgical data available No resection Resection
43 (10.9) 351 (89.1)
Radiation therapy Median survival, months 5 year survival, % 10 year survival, %
175 (44.9) 151 72.2 54.8
Data presented as number of patients (%) unless otherwise indicated. SD = standard deviation, SEER = Surveillance, Epidemiology, and End Results.
Please cite this article in press as: Jones PS et al. Outcomes and patterns of care in adult skull base chordomas from the Surveillance, Epidemiology, and End Results (SEER) database. J Clin Neurosci (2014), http://dx.doi.org/10.1016/j.jocn.2014.02.008
3
P.S. Jones et al. / Journal of Clinical Neuroscience xxx (2014) xxx–xxx
and there were no significant differences in the rates of radiation use amongst registries (p = 0.07).
age at diagnosis (HR 1.027, 95% CI 1.018–1.037, p < 0.0001), and later year of diagnosis (HR 0.970, 95% CI 0.952–0.988, p = 0.0015) predicted better overall survival (Table 3).
4. Outcomes 4.2. Tumor size subgroup 4.1. All patients The 10 year overall survival rate for all patients by Kaplan– Meier analysis was 54.7% and median overall survival was 151 months. Median overall survival was 220 months for patients treated with surgery only, 107 months for patients treated with surgery and adjuvant radiation, 42 months for radiation only, and 162 months for no treatment (Fig. 1). The log-rank test was used to compare survival between treatment groups. Median survival for all patients treated with surgery was significantly increased compared to those without surgery, other than biopsy (151 versus 81 months, p = 0.03). Median survival for patients treated with radiation (with or without resection) was 94 months, considerably shorter than for those who did not receive radiation (220 months, p = 0.06). Patients treated with postoperative adjuvant radiation also had decreased median survival compared to patients treated with surgery alone (220 versus 107 months, p = 0.193). The 10 year survival rate for patients treated with surgery and radiation was 47.2% versus 62.4% for those treated with surgery only (Fig. 2). By univariate analysis, factors predicting better overall survival included patients who underwent surgical resection (hazard ratio [HR] 0.603, 95% confidence interval [CI] 0.382–0.950, p = 0.0293), had younger age at diagnosis (HR 1.027, 95% CI 1.018–1.037, p < 0.0001), and had later year of diagnosis (HR 0.970, 95% CI 0.952–0.988, p = 0.0015) (Table 2). In a multivariate model, younger
We separately analyzed 196 patients with a tumor size recorded at the time of diagnosis, with a median age of 46 years (range 2 to 85). Table 4 shows the breakdown of mean tumor size according to the four different treatment strategies. The largest mean tumor size, 4.4 cm, was that of the patient group treated with radiation only. There was no significant difference between mean tumor size for each treatment group (p = 0.6). The 10 year overall survival rate for this subgroup population by Kaplan–Meier analysis was 58.7%, and the median overall survival was 146 months. By log-rank comparison, median survival for patients treated with surgery and radiation in this subgroup was non-significantly increased compared to those treated with surgery alone (259 versus 146 months, p = 0.10). Ten year survival for patients treated with both surgery and radiation was 63.3% versus 54.8% for patients treated with surgery alone (Fig. 2). Only nine patients in this subgroup were treated with radiation as a single modality. By univariate analysis, variables predicting improved overall survival were registries with higher numbers of skull base chordomas, smaller tumor size, and younger age at diagnosis (Table 5). The multivariate model demonstrated the same three predictors for better overall survival (Table 6). For this subgroup, we again observed no increasing trend of radiation use over time on chisquared analysis (p = 0.6) and no significant difference in rates of radiation use among registries (p = 0.07).
100 90 80 70 60 50 40 30 20 10 0 0
50
100
150
200
250
300
350
400
Fig. 1. Kaplan–Meier survival plot for all patients with chordomas based on treatment options of surgery (solid line), radiation (short-long dashed line), and surgery and radiation (short dashed line).
Please cite this article in press as: Jones PS et al. Outcomes and patterns of care in adult skull base chordomas from the Surveillance, Epidemiology, and End Results (SEER) database. J Clin Neurosci (2014), http://dx.doi.org/10.1016/j.jocn.2014.02.008
4
P.S. Jones et al. / Journal of Clinical Neuroscience xxx (2014) xxx–xxx
100 90 80 70 60 50 40 30 20 10 0 0
50
100
150
200
250
300
Fig. 2. Kaplan–Meier survival plot for patients with known tumor size based on treatment options of surgery and radiation (dashed line) compared to surgery alone (solid line).
5. Discussion
Table 2 Univariate analysis of all variables with respect to survival Variable
Parameter estimate
SE
Hazard ratio (95% Cl)
p value
Radiation Yes No
0.31385 Reference
0.16903 –
1.369 (0.983–1.906) –
0.0633 –
Surgery Yes No
0.50603 Reference
0.23223 –
.603 (0.382–0.950) –
0.0293 –
Registry Small Large
Reference 0.3106
– 0.17157
– 0.733 (0.524–1.026)
– 0.0702
Sex Male Female
Reference 0.04086
– 0.16879
– 1.042 (0.748–1.450)
0.8087
Age By year
0.2703
0.00490
1.027 (1.018–1.037)
Contents lists available at ScienceDirect
Journal of Clinical Neuroscience journal homepage: www.elsevier.com/locate/jocn
Clinical Study
Outcomes and patterns of care in adult skull base chordomas from the Surveillance, Epidemiology, and End Results (SEER) database Pamela S. Jones a, Manish K. Aghi b, Alona Muzikansky c, Helen A. Shih d, Fred G. Barker II a, William T. Curry Jr. a,⇑ a
Department of Neurosurgery, Stephen L. Harris Center for Chordoma Care, Massachusetts General Hospital, 55 Fruit Street, Yawkey 9E-9026, Boston, MA 02114, USA Department of Neurosurgery, University of California, San Francisco, 5 San Francisco, CA, USA Biostatistics Center, Massachusetts General Hospital, Boston, MA, USA d Department of Radiation Oncology, Massachusetts General Hospital, Boston, MA, USA b c
a r t i c l e
i n f o
Article history: Received 29 January 2014 Accepted 8 February 2014 Available online xxxx Keywords: Chordoma Population study Radiation therapy Skull base tumors
a b s t r a c t This study aims to demonstrate survival rates and treatment patterns among patients with chordomas of the skull base using a large population database. Patients with cranial chordomas between 1973 and 2009 were identified from the USA Surveillance, Epidemiology, and End Results (SEER) public use database. Kaplan–Meier analysis was used to examine the effect of surgery and radiation on overall survival. We identified 394 patients with histologically-confirmed cranial chordomas. Median survival was 151 months. Most patients (89.09%) underwent surgery. Less than half (44.92%) received radiation after diagnosis. Patients who underwent surgical resection survived significantly longer than those who did not undergo resection, regardless of other treatments (151 versus 81 months, p < 0.001). Ten year survival was lower among patients receiving radiation (44.8% versus 61.4%, p = 0.66). Surgery predicted better overall survival by univariate analysis (hazard ratio [HR] 0.603, p = 0.0293); younger age at diagnosis (HR 1.028, p < 0.001), and later year of diagnosis (HR 0.971, p = 0.0027) were prognostic of improved survival in a multivariate model. In subgroup analysis of patients with documented tumor size, smaller tumor size (HR 1.021, p = 0.0067), younger age (HR 1.031, p = 0.001), and treatment within a higher volume registry (HR 0.490, p = 0.0129) predicted improved survival. Surgical intervention offers survival benefit for cranial chordomas. Findings of decreased survival in patients receiving radiation may be associated with selection. Studies examining surgical extent of resection data and radiation details are needed to determine the impact of radiotherapy. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction Chordoma is a rare cancer arising from notochord remnants, with approximately 30–40% occurring at the skull base, mostly arising from the clivus [1,2]. Although typically slow-growing, these tumors are locally destructive and frequently recur following treatment. Proximity to critical structures such as the brainstem, cranial nerves, and major intracranial vessels combined with the high recurrence rate renders these histologically low-grade neoplasms clinically malignant, and chordoma management is a challenge for patients and health care providers alike [3]. Due to the rarity of the disease, current treatment paradigms are largely based on single-center retrospective studies, and to our knowledge no prospective comparative studies exist in the ⇑ Corresponding author. Tel.: +1 617 724 0884; fax: +1 617 724 0887. E-mail address: [email protected] (W.T. Curry Jr.).
literature [4]. Typically, surgical resection is the first-line treatment, and radiation may be used for residual tumor, recurrent tumor, cases considered unresectable, or to the operative site following a gross total resection. Little consensus exists over whether radiation should be integrated into standard therapy following surgery whether or not there is evident residual tumor [2,5–7]. Extensive surgical resection is technically challenging and morbid, although advances in technology, such as the use of endoscopy, neuronavigation, and neuromonitoring have facilitated greater access to clival chordomas with improved safety [7–9]. There is also a more widespread capacity to deliver high-dose conformal radiation, including proton beam radiotherapy, intensity modulated radiotherapy, photon-based radiosurgery and hypofractionated radiotherapy [10,11]. In this analysis we describe treatment and survival trends of patients with skull base chordoma over a 36 year period using data obtained from the USA Surveillance, Epidemiology, and End Results
http://dx.doi.org/10.1016/j.jocn.2014.02.008 0967-5868/Ó 2014 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Jones PS et al. Outcomes and patterns of care in adult skull base chordomas from the Surveillance, Epidemiology, and End Results (SEER) database. J Clin Neurosci (2014), http://dx.doi.org/10.1016/j.jocn.2014.02.008
2
P.S. Jones et al. / Journal of Clinical Neuroscience xxx (2014) xxx–xxx
(SEER) registry. We paid particular attention to outcomes and trends in radiation utilization.
strategies were compared using chi-squared analysis. Calculations were performed using SAS version 9.3 (SAS Institute, Cary, NC, USA).
2. Methods 3. Results 2.1. SEER program 3.1. Demographics The SEER registry, a database maintained by the USA National Cancer Institute, collects incidence and survival data from population-based cancer registries. Data on chordoma patients diagnosed between 1973 and 2009 were available from 18 population-based cancer registries across the USA that together represent approximately 25% of the USA population. The 18 registries include nine entire states and nine metropolitan regions. These registry populations are comparable to the overall USA population on a variety of socioeconomic parameters, including poverty indices, high school graduation rates, percentage in rural areas, and percentage of foreign-born residents. The database contains information on patient demographics, primary cancer site, histology, methods of diagnostic confirmation, treatment regimens (including surgery and radiation therapy), and year of death. Institutional Review Board approval was not required for this study, as the SEER database is free of any sensitive patient information or identifiers. An overview of the SEER database is available at http://seer.cancer.gov/about/.
Between 1973 and 2009, 422 patients were diagnosed with skull base chordoma in the SEER database. Of these, 394 patients satisfied the criteria for histologically-confirmed disease and were included in our analysis. The median age at diagnosis was 46 years, with an age range of 0 to 92, and a higher proportion of patients were male (55.6%). The majority of patients (220 patients, 55.8%) were diagnosed before the median diagnosis year 2002. Cases occurred across 16 registries, of which nine were considered lowvolume (118 patients) and seven were considered high-volume (276 patients). The results are summarized in Table 1. 3.2. Tumor characteristics Information regarding tumor size was available in 196 of 394 patients. Mean diameter was 3.7 cm (range 0.2 to 8.5 cm). Information regarding histologic subtype of chordoma was not available.
2.2. Chordoma case definition 3.3. Treatment We compiled chordoma cases diagnosed between 1973 and 2009. Patients were identified through the World Health Organization’s International Classification of Diseases for Oncology, Version 3 (ICD-O-3) morphology code for chordoma (9370–9373). Primary sites of disease presentation were compiled according to ICD-O-3 topography codes and were limited to skull and face bones (410); meninges (700); any structure in the brain (710–9, 751, 753); connective and soft tissue of head, face and neck (490); or any head structure (760). We restricted our analysis to patients with histologic confirmation of chordoma.
Most patients (351 of 394, 89.1%) underwent some type of surgical resection beyond biopsy. The SEER data only specified whether surgery was performed, and did not reliably specify extent of skull base tumor resection. We made the assumption that patients who were not coded as having had surgery underwent biopsy alone. Ultimately, 196 patients were treated with surgery alone (49.7%), 155 patients were treated with surgery and radiation (39.3%), and 22 patients treated with radiation alone (5.5%). Twenty-one patients received no treatment (5.3%). Radiation utilization did not increase over time (chi-squared analysis, p = 0.7),
2.3. Treatment Treatment types were limited to surgery and radiation. The SEER surgery code of 0 was used to indicate that surgery was performed. For the purpose of our analysis, all surgical cases were considered to have undergone an attempt at tumor resection. All other codes were considered cases where only a diagnostic biopsy had been performed. Adequate information on extent of resection was not available from the database. The SEER radiation codes 0, 6, 7, 8, 9 represented no radiation treatment. We further categorized registries by number of skull base cases in order to assess for potential volume-outcome effect, whereby better outcomes are observed at treatment centers with high volume. We looked at numbers of both skull base chordomas and chondrosarcomas (ICD-O-3 code 9220–9239) treated by each registry to assess for volume of skull base cases. A registry was considered to contain a high-volume treatment center if it contained 40 or more histologically-confirmed skull base chordomas and chondrosarcomas during the study period. 2.4. Statistical analysis The primary outcome of interest was overall survival. Survival analyses were performed using Kaplan–Meier product limit estimators and Cox univariate and multivariate proportional hazards models. Predictors of overall survival in univariate analysis (p < 0.05) were included in the multivariate model. Treatment
Table 1 Characteristics of 394 patients with chordoma tumors of the skull base registered in the SEER database (1973–2009) Patients, n Mean age, years (± SD) Median age, years Males
394 45.3 (± 19.2) 46 219 (55.6)
Race White African American Asian Native American
333 (85.4) 13 (3.3) 42 (10.8) 2 (0.5)
Registry Low volume (2002
2002 median year (1999 mean) 220 (50.8) 174 (44.2)
Surgical data available No resection Resection
43 (10.9) 351 (89.1)
Radiation therapy Median survival, months 5 year survival, % 10 year survival, %
175 (44.9) 151 72.2 54.8
Data presented as number of patients (%) unless otherwise indicated. SD = standard deviation, SEER = Surveillance, Epidemiology, and End Results.
Please cite this article in press as: Jones PS et al. Outcomes and patterns of care in adult skull base chordomas from the Surveillance, Epidemiology, and End Results (SEER) database. J Clin Neurosci (2014), http://dx.doi.org/10.1016/j.jocn.2014.02.008
3
P.S. Jones et al. / Journal of Clinical Neuroscience xxx (2014) xxx–xxx
and there were no significant differences in the rates of radiation use amongst registries (p = 0.07).
age at diagnosis (HR 1.027, 95% CI 1.018–1.037, p < 0.0001), and later year of diagnosis (HR 0.970, 95% CI 0.952–0.988, p = 0.0015) predicted better overall survival (Table 3).
4. Outcomes 4.2. Tumor size subgroup 4.1. All patients The 10 year overall survival rate for all patients by Kaplan– Meier analysis was 54.7% and median overall survival was 151 months. Median overall survival was 220 months for patients treated with surgery only, 107 months for patients treated with surgery and adjuvant radiation, 42 months for radiation only, and 162 months for no treatment (Fig. 1). The log-rank test was used to compare survival between treatment groups. Median survival for all patients treated with surgery was significantly increased compared to those without surgery, other than biopsy (151 versus 81 months, p = 0.03). Median survival for patients treated with radiation (with or without resection) was 94 months, considerably shorter than for those who did not receive radiation (220 months, p = 0.06). Patients treated with postoperative adjuvant radiation also had decreased median survival compared to patients treated with surgery alone (220 versus 107 months, p = 0.193). The 10 year survival rate for patients treated with surgery and radiation was 47.2% versus 62.4% for those treated with surgery only (Fig. 2). By univariate analysis, factors predicting better overall survival included patients who underwent surgical resection (hazard ratio [HR] 0.603, 95% confidence interval [CI] 0.382–0.950, p = 0.0293), had younger age at diagnosis (HR 1.027, 95% CI 1.018–1.037, p < 0.0001), and had later year of diagnosis (HR 0.970, 95% CI 0.952–0.988, p = 0.0015) (Table 2). In a multivariate model, younger
We separately analyzed 196 patients with a tumor size recorded at the time of diagnosis, with a median age of 46 years (range 2 to 85). Table 4 shows the breakdown of mean tumor size according to the four different treatment strategies. The largest mean tumor size, 4.4 cm, was that of the patient group treated with radiation only. There was no significant difference between mean tumor size for each treatment group (p = 0.6). The 10 year overall survival rate for this subgroup population by Kaplan–Meier analysis was 58.7%, and the median overall survival was 146 months. By log-rank comparison, median survival for patients treated with surgery and radiation in this subgroup was non-significantly increased compared to those treated with surgery alone (259 versus 146 months, p = 0.10). Ten year survival for patients treated with both surgery and radiation was 63.3% versus 54.8% for patients treated with surgery alone (Fig. 2). Only nine patients in this subgroup were treated with radiation as a single modality. By univariate analysis, variables predicting improved overall survival were registries with higher numbers of skull base chordomas, smaller tumor size, and younger age at diagnosis (Table 5). The multivariate model demonstrated the same three predictors for better overall survival (Table 6). For this subgroup, we again observed no increasing trend of radiation use over time on chisquared analysis (p = 0.6) and no significant difference in rates of radiation use among registries (p = 0.07).
100 90 80 70 60 50 40 30 20 10 0 0
50
100
150
200
250
300
350
400
Fig. 1. Kaplan–Meier survival plot for all patients with chordomas based on treatment options of surgery (solid line), radiation (short-long dashed line), and surgery and radiation (short dashed line).
Please cite this article in press as: Jones PS et al. Outcomes and patterns of care in adult skull base chordomas from the Surveillance, Epidemiology, and End Results (SEER) database. J Clin Neurosci (2014), http://dx.doi.org/10.1016/j.jocn.2014.02.008
4
P.S. Jones et al. / Journal of Clinical Neuroscience xxx (2014) xxx–xxx
100 90 80 70 60 50 40 30 20 10 0 0
50
100
150
200
250
300
Fig. 2. Kaplan–Meier survival plot for patients with known tumor size based on treatment options of surgery and radiation (dashed line) compared to surgery alone (solid line).
5. Discussion
Table 2 Univariate analysis of all variables with respect to survival Variable
Parameter estimate
SE
Hazard ratio (95% Cl)
p value
Radiation Yes No
0.31385 Reference
0.16903 –
1.369 (0.983–1.906) –
0.0633 –
Surgery Yes No
0.50603 Reference
0.23223 –
.603 (0.382–0.950) –
0.0293 –
Registry Small Large
Reference 0.3106
– 0.17157
– 0.733 (0.524–1.026)
– 0.0702
Sex Male Female
Reference 0.04086
– 0.16879
– 1.042 (0.748–1.450)
0.8087
Age By year
0.2703
0.00490
1.027 (1.018–1.037)
