Abstract
Background/Aim: Sacral chordoma is a rare primary bone neoplasm associated with high morbidity. The aim of this study is to identify demographic and clinicopathological characteristics of this tumor and evaluate their impact on survival outcomes. Patients and Methods: The Surveillance, Epidemiology and End Results (SEER) database collecting data between 2000 and 2018 was searched for all cases of sacral chordoma. We analyzed demographic aspects, cancer stage and treatment patterns. Overall survival was calculated using the Kaplan-Meier method and compared between subgroups using the log-rank test. A multivariate Cox hazard regression analysis was conducted to identify independent predictors of overall survival. Results: Four hundred and forty-two patients were identified with a mean age of 62.7 years. Most tumors presented regional invasion at diagnosis (43.2%). Mean overall survival was 124.7 months. No significant difference in terms of overall survival was found between surgery alone and surgery associated with radiotherapy. Both options provided a significantly increased survival than radiotherapy alone. Age of less than 50 years or between 50 and 69 correlated significantly with improved survival. Conclusion: Age and stage at diagnosis impact significantly survival outcomes. Surgery remains the mainstay treatment with the highest overall survival. Its association with radiotherapy is currently questionable and needs further research.
With a reported incidence of less than one per million persons per year, chordoma are rare primary bone tumors arising from remnants of the notochord (1). They develop mainly along the axial skeleton with the three most reported localization being the skull base, the sacrum and the mobile spine. This neoplasm affects mainly males around their fifth and sixth decade of life. It has a malignant behavior characterized by a slow growth with local invasion and late onset metastasis (2-5). Surgery maintains a primordial role in oncological therapy of the chordoma with wide margin resection being considered the mainstay of treatment. Sacral localization is associated with high morbidity and surgical treatment appears to be challenging because of adjacent structure that could be invaded or in contact with the tumor. Because of late onset symptoms that can often be trivialized such as lower back pain or radiculopathy, the sacral chordoma is often locally advanced at diagnosis (6). Local pain seems to be the most frequently reported symptom (4, 7). Wide surgical margins appear to be an important operative strategy, as local recurrence rate between 47% and 66% have been observed (8-10).
The Surveillance, Epidemiology and End Results (SEER) program is an incidence and survival database extracted from different United States of America (USA) cancer registries that covers up to 35% of the country population (11). Those registries are considered representatives of the demographics of the entire US population (11).
Numerous studies have been conducted using the SEER Registry to analyze prognostic factors of diverse tumors including chordomas, but to our knowledge none focused specifically on the sacral chordoma (2, 12, 13). Because this localization induces real therapeutical challenges, the aim of this study is to analyze prognostic factors, demographics and clinicopathological characteristics specifically for the sacral localization of chordoma. By choosing a selection interval between 2000 and 2018, we aim to analyze the most updated data which include the latest medical advancements that can be compared to the previous literature to guide the oncological treatment.
Patients and Methods
Patient selection. A cohort of patients was extracted from the SEER database in April 2021 using the SEER*stat software provided by the National Cancer Institute (NCI). The study included patients from 18 registries covering the period from 2000 to 2018. Selection criteria for inclusion in the cohort were the use of the dedicated codes of the International Classification of Oncology (ICD-O-3) for histological morphology 9370/3: Chordoma, Not Otherwise Specified (NOS) and primary topographic site C41.4: Pelvic bones, sacrum, coccyx and associated joints (14).
Demographic and clinicopathologic analysis. Demographic and clinicopathologic analysis were conducted using the “case listing” option of the software. Variables such as age at diagnosis (<50; 50-69; ≥70), sex, race, marital status, stage at diagnosis (localized; regional; distant), tumor size (longest axis), treatments (none; surgery; lymph node dissection; radiotherapy) and survival in months were defined. Some of them were created using the “merge function” to avoid variables with more than 4 categories.
Survival analysis. Survival analysis was conducted using the case listing of the “survival session” of SEER*stat software and processed through SPSS software (version 26). Life Tables and survival curves were generated using the Kaplan-Meier method and assuming the subjects were alive at the study cut-off in December 2018. The comparison between the different subgroups of each variable was done using the Log-Rank test. We conducted a multivariate analysis to identify predictors of overall survival using a Cox hazard regression analysis. A p-Value <0.05 was considered statistically significant.
Variable selection. Age at diagnosis is studied as a three-category variable: less than 50, 50 to 69 and more than 70 years old. This choice is due to a low incidence before the fifth decade of life with an improved survival as reported by Pan et al. (12). We decided to separate the third category of patients older than 70 years because of the reported lower survival of geriatric patients associated to our aim to at specifically identify outcomes for people between 50 to 69 years as this age interval is close to the mean age of diagnosis of chordoma and patients in this group are relatively fit and remain good “surgical candidates”.
The ‘stage at diagnosis’ variable was generated by merging four different SEER variables reporting the initial staging as localized, regional or distant (details of the merging function are available in the Annex). Manual review was done to ensure no discordant value existed between the four variables used and the merged one. Localized disease included in situ tumors and was defined as a neoplasm confined by the bone cortex or a tumor extending beyond the cortex but limited to the periosteum (without effraction of the periosteum). Regional disease was defined by direct extension beyond the periosteum (adjacent bone, cartilage or adjacent skeletal muscles) or/and by the presence of positive regional lymph nodes. Distant disease is defined by distant lymph nodes, further contiguous extension, metastasis or extension to the skin. However, for cases before the 1st January 2001, skin extension was considered regional. Medical files reporting localized, regional or distant diseases without further details were directly coded as reported (15).
Results
For population demographic and clinicopathological characteristics analysis, 442 patients were identified (Table I). Age at diagnosis ranged from 11 to 98 years old with a mean diagnosis age at 62.7 years (median: 64). The number of tumor diagnosis increases with patient age (Figure 1). Most of the population was male (61.1%). White race was the most represented (83.3%) followed by black (4.3%) and other races such as American Indian/Alaska Native or Asian/Pacific Islander (11.3%). Most patients were married (55.7%). There seems to be an increase of diagnosed cases in the latest years as depicted in Figure 2 with 57.9% of the cases diagnosed between 2010-2018. Sacral chordoma appears to be diagnosed at already locally advanced stage in 43.2% of the cases, which show regional invasion, while 38.7% of the patients present localized diseases, and only 10.9% present with distant disease. Tumor size (longest axis) was reported in 330 patients with a median of 80 mm and a range from 8 to 989 mm. Mean tumor size is 90.3 mm but this value might be underestimated as the highest range because tumor size over 989 mm were reported as 989 mm. The mainstay of treatment appears to be surgery (44%), sometimes associated with radiotherapy (21.3%). Radical excision or resection of lesion (36.4%) is the most frequent surgery reported, followed by partial resection (12%). A minority of patients however did not receive surgical treatment, 16.1% of which received radiotherapy alone and 14.9% did not receive any treatment at all. Most of those choices were justified by a surgical therapy not recommended (74.8%) while patient refusal was low (6%). Chemotherapy did not appear to be a popular therapeutic option with only five percent of the patients receiving it.
Demographic and clinicopathological characteristics of the study population.
Distribution of sacral chordoma diagnosis according to patients’ age.
Number of sacral chordoma cases diagnosed each year.
Three hundred forty-six patients were identified for survival analysis. The mean overall survival time was 124.7 months (m) (median: 116 m). Five years overall survival rate was 72.9 % (95% at 1 year; 87.6% at 2 years; 83.1% at 3 years; 78% at 4 years) (Figure 3). Table II offers a global vision of the mean and median survival times according to different variables.
Sacral chordoma overall survival curve and survival curves according to diagnostic year, race, and marital status.
Mean and median survival of patients with 95% confidence intervals.
Gender was not significantly associated with improved survival (p=0.420) (Figure 4). A tumor diagnosis at an age equal or superior to 70 was associated with significantly lower survival than the two groups of less than 50 and between 50 to 69 (p<0.05), while no significant difference was found between those two groups (p=0.126) (Figure 4). No significant survival difference was noted between races (Figure 3). Patients with the married status presented improved survival upon those who were not (p<0.05) (Figure 3). A more recent diagnosis did not significantly improved survival compared to diagnosis before 2010 (Figure 3). Localized and regional disease did not differ in survival, but both were significantly associated with improved survival compared to distant disease (p<0.05) (Figure 4). The groups including surgery alone or associated with radiotherapy were superior to those with radiotherapy alone or no treatment (p<0.05). No significant difference was noted between surgery alone and surgery combined with radiotherapy, or between no therapy at all and radiotherapy alone (Figure 4).
Sacral chordoma survival curves according to gender, age, stage at diagnosis and treatment sequence.
In multivariate Cox-Regression (Table III), only an age younger than 50 (B –1.2; p<0.05) or between 50 to 69 years old (B –0.7; p<0.05) were identified as significant prognostics factors associated with improved survival.
Cox multivariate regression.
Discussion
This study offers updated data on the sacral chordoma, collected in the largest cohort population available between 2000 and 2018. This allows to identify trends in demographics and treatments to establish optimal therapeutic strategies and identify emerging risk factors and predictors of survival after comparison with previous data.
Dedifferentiated and chondroid chordoma were not included in the study as their behavior is different and incidence is extremely low. Dedifferentiated chordoma are described in less than 10% of the chordoma and are believed to present a more aggressive evolution with higher metastatic rates and lower survival than conventional chordomas (16, 17). The chondroid chordoma is a controversial histo-pathological subtype of chordoma. Some authors advocate a distinction between chondroid and conventional chordomas as their behavior might differ (18, 19). In the SEER database, only eight Chondroid chordoma and eight dedifferentiated chordoma were found between 2000 and 2018. We deliberately decided to exclude them as their behavior is thought to be different from conventional chordoma and might bias our results.
Regarding the demographic characteristics of our population, we found similar results to the previous literature with most of the population being white male around their fifties or sixties (4, 5, 7, 10, 20, 21). Also, distribution of the age at diagnosis between 2000 and 2018 didn’t differ from the situation between 1973 to 2014, as shown in Figure 1, which displays data close to the ones reported by Pan et al. (12).
An age at diagnosis over 70 years was associated with a significantly lower overall survival. These results are in line with the studies of Varga et al., that found lower survival over 65 years old, and Pan et al., that found lower survival over 60 years (7, 12). This result might be explained by the higher number of comorbidities presented by patients in these age groups, which would make them less capable of tolerating heavy surgeries such as “en bloc” resection of the tumor.
A trend toward a raise of the number of cases diagnosed was observed through the years, but without a significant evolution in terms of survival. This augmentation might be explained by an expansion of the population included in the registry through the years, but also by the overall increase of the ageing population in the U.S.A. associated with the tumor tendency to affect people over fifty years (22).
Interestingly, people of black ethnicity seem to be minoritarian in our study. This observation might be explained by a lower life expectancy and more difficulties of accessing healthcare resources, meaning that chordoma might not be diagnosed because of earlier death caused by other pathologies (23). However, no significant survival difference was noted between races. This result might be influenced by a low proportion of the different ethnicities in the study population.
Lower survival in unmarried patients observed in this study is a known fact that is appliable to other cancers and chronic disease such as cardiovascular pathologies. A meta-analysis by Wang et al. suggested that this phenomenon might be linked to lower socio-economic status or an already bad health (somatic or psychological comorbidities) (24).
When looking at the initial stage of the disease, we can observe that most of the cases present localized disease (38.7%) with sometimes regional invasion (43.2%), but only 10.9% presented with distant disease. Those results are in line with previous reports on the disease. In survival analysis, distant disease was associated with a significant lower survival. The median tumor size of 8 cm reported in this study highlights the fact that the disease is often diagnosed late because of its pauci-symptomatic evolution and this could explain why the majority of cases present regional invasion at diagnosis (6). Angelini et al. found that tumor volume was positively correlated with increased risk of loco-regional recurrence and survival (9).
Surgery remains a mainstay of oncological therapy with 65.9% of patients receiving it, in association with radiotherapy in 21.3% of the cases. We found that surgical treatment with or without radiotherapy was significantly associated with better overall survival than radiotherapy alone. However, no significant difference was noted between the use of radiation combined with surgery and surgery alone. Radiotherapy as a single therapy is associated with a low survival similar to the group that received no surgery and no radiation. This might be explained by an indication bias since most of the patients receiving radiotherapy only were considered unfit for a curative surgery and the goal of this therapy was mainly palliative. This explanation is supported by the fact that for 74.8% of the patients that were not treated with surgery; the surgical act was not recommended.
Wide surgical margins have an important impact in local control, but even with adequate margins, muscular invasion increases the risk of recurrence (5, 8, 9). Many authors suggest that radiotherapy could have a positive impact in terms of survival and local recurrence rate in patients where R0 surgical margins might not be achieved (10, 25-28). More recent literature suggests that “modern” radiotherapy protocols such as carbon ion radiotherapy and proton radiotherapy, can be a valuable alternative treatment of choice, even in those patients considered unfit for surgery (29-32). Chemotherapy is not supported as an alternative of treatment compared to surgery or radiotherapy, but molecular targeted therapy seems to emerge as an alternative in case of advanced diseases (33). Furthermore, some molecular markers might help identify tumors with lower prognosis and guide therapeutic strategies (34, 35).
However, this study presents the limitations which are common for the large population analysis such as the SEER analysis. The main limitation is the use of merged categories necessary to simplify survival analysis because of the multiples values that some SEER variables can assume. This approach limits fine analysis between different modalities of treatment. For example, SEER only allows to group data of all type of surgeries within the surgery subgroup, which may include treatments ranging from local excision to hemipelvectomy. Analysis of the initial surgical margins or the necessity of surgery for recurrence is also not possible. Evaluation of these factors might produce diverse survival outcomes, as surgical margins, relapse surgery, tumor volume and initial local invasion are indicated as important recurrence prognostic factor (5, 8, 9, 36, 37). Furthermore, another limitation is the lack of precision of certain SEER variables such as the primary site where no distinction is made between sacral levels, coccyx, or pelvic. This lack of precision is also seen in tumor size report that limits itself to 989 mm, meaning that tumor over this size would be categorized as 989 mm and influence the mean value by underestimating it.
Conclusion
Sacral chordoma is a slowly evolving malignant neoplasm that affects mainly white men around their fifties and sixties with a mean overall survival of 125 months. The outcome is mainly affected by the age and stage at diagnosis with surgery being the mainstay of treatment. Current role of modern radiotherapies as adjuvant or stand-alone treatment is currently questionable despite an increasing number of studies highlighting their benefits in selected patients.
Footnotes
Authors’ Contributions
Matteo Scampa extracted study data and conducted statistical analysis. Matteo Scampa and Carlo M. Oranges wrote the original article. Enrico Tessitore, Dennis E. Dominguez, Didier Hannouche, Nicolas C. Buchs and Daniel F. Kalbermatten offered proof-reading and helped in the redaction of the article by suggesting corrections and by bringing their expertise. Carlo M. Oranges ensured coordination and supervision of the research project. All Authors approved the final version of the manuscript.
Conflicts of Interest
None.
- Received October 28, 2021.
- Revision received November 30, 2021.
- Accepted December 10, 2021.
- Copyright © 2022 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.