Metastatic skull base chordoma: A systematic review

Abstract Objective/Hypothesis To investigate the clinical features, management strategies and outcomes for patients with metastatic primary skull base chordomas. Study Design Systematic review. Methods A systematic search through Pubmed/Medline, Web of Science, and EBSCOhost (CINAHL) was conducted without restriction on dates. After study screening and full‐text assessment, two authors independently extracted all data using a pre‐established abstraction form. Results Forty cases were included from 38 studies. The average age (standard deviation [SD]) of the sample at presentation was 28.5 (23.3) and was equally distributed across genders. The average time (SD) between initial diagnosis to local recurrence was 40.1 (60.3) months. The average time (SD) from primary tumor detection to the diagnosis of metastatic disease was 55.2 (49.0) months. The most common subsite for metastatic spread were the lungs (32.5%). Of the 33 patients with data on outcomes, 48.5% were found to have expired by the time of publication. The median overall survival was estimated to be 84 months (95% confidence interval [CI] 62.3–105.7). Conclusions The most common subsites for metastatic spread of skull base chordoma were the lungs and bone. Overall survival for patients in the current cohort was a median of 84 months, with no significant differences noted when stratifying by the extent of surgery or the site of metastases. Level of Evidence 3a

including the brainstem, internal carotid arteries, cavernous sinuses, and several cranial nerves. 4,5 For these reasons, chordomas present a unique challenge to the skull base community.
The frequency with which chordomas metastasize has been estimated to be anywhere between 3% and 48%. 6 However, these initial studies were conducted prior to 1980, and it is likely that the paucity of reported skull base lesions were a function of limitations in screening and technology. 7,8 There have been fewer investigations studying chordoma metastatic patterns in the 21st century, and these investigations have predominately examined neoplasms in the spine. 9 Additionally, outcome data on local recurrences, distant metastases and survival of metastatic skull base (SB) chordomas are also limited. Furthermore, practical guidelines for the surgical and oncologic management of metastatic SB chordoma have yet to be defined. For instance, it is still unclear whether patients with distant SB chordoma metastases should be treated with primary resection, surgical removal of the metastasis, directed radiotherapy, or limited to palliative therapies.
The present study is a systematic review designed to evaluate clinical outcomes of patients with SB chordoma that developed distant metastatic disease.
Following PICOS criteria (population, intervention, comparison, outcomes, study design), the current research question was devel- Several search phrases including, "skull base," "chordoma" or "metastasis," and several Boolean operators including "AND" or "OR" were combined in various permutations. This process and search parameters are fully described by Figure 1 and Appendix S1. The reference management software, Rayyan QCRI, was used to remove duplicates and to screen the initial group of studies. Authors Torbjoern Nielsen and Hannah Bulosan separately screened studies based on titles and abstracts through a blinded process. Author Kurtis Young reviewed and resolved all conflicts between the two aforementioned authors, and finalized all included studies through full-text assessment. Inclusion criteria required that patients were diagnosed with primary skull base chordoma that subsequently metastasized either on presentation or throughout the course of the study. Additionally, included articles must have either been a case report, case series, or F I G U R E 1 PRISMA flow diagram for search and review strategy observational study with individual case data. Non-English articles were excluded if they lacked an English language abstract or if that abstract did not have information on demographics, clinical presentation, and management/outcomes.

| Data extraction
Authors Torbjoern Nielsen and Hannah Bulosan independently extracted data from the included articles through an abstraction form that was created beforehand, and authors Kurtis Young and Christian T. Ogasawara independently verified all extracted data.
Data on study ID (publication year and author names), baseline demographics (age and gender), clinical features (presenting symptoms, primary site, recurrence data, and metastatic details), management (surgery, radiotherapy, and chemotherapy), and outcomes (overall survival, causes of death) were extracted. Meta-analysis was not performed due to the level of evidence of the included studies (IV-V), as determined following the 2011 Oxford Centre for Evidence-Based Medicine guidelines. 10

| Data analysis
All analyses were performed utilizing SPSS software version 26 (IBM Corp., Armonk, New York). Means and standard deviations were calculated for continuous variables. The intervals between SB chordoma onset and death (OS curve) were approximated through Kaplan-Meier curves, with significance being assessed through the log-rank test.
Calculations were deemed to be statistically significant using an α of 0.05.

| RESULTS
The systematic database search initially yielded 214 titles, with 53 studies that were further screened through full-text assessment.
Here, five and three studies were excluded for not featuring patients with metastatic disease or primary skull base tumors, respectively.
Additionally, there were several studies (5) that were excluded for insufficient information. Two titles were found to be journal supplements and were excluded. Finally, there were 38 studies included in this review, with a total of 40 included cases.  Study dates ranged

| Quality appraisal
All included studies were case reports or case series and were subsequently classified as level V or IV evidence, respectively. The overall quality of the cases included were deemed to be moderate to high, as further detailed in Appendix S2. All studies reported on demographic data (100.0%), surgical management (100.0%), and local recurrences and metastases (100.0%). The majority of included cases featured information on symptoms associated with the primary tumor (77.5%), but many of the remaining patients were initially treated at different institutions and had unavailable prior medical history. Regarding outcomes, most patients (82.5%) presented with follow-up data.

| Demographic information and clinical characteristics
The average age (SD) on diagnosis of the primary tumor was 25.5 (21.7) years. The cohort was evenly distributed across genders. These data are more completely depicted in Table 1. Of the 31 patients with data on the symptoms at the time of primary tumor presentation, the most common reported findings were headaches (38.7%), neck pain or rigidity (32.3%), dysphagia or pharyngeal mass (29.0%), and vision impairment (22.6%). Less common symptoms included nausea (9.7%), vomiting (9.7%), seizure/epilepsy (6.5%), and gait impairment (6.5%).

| Metastatic and locally recurrent disease management
All of the cases included in this study developed metastatic disease, and 37.5% of the sample experienced distant spread to more than one metastatic site. The lungs were the most common subsite involved overall (32.5%) and were the most common subsite for isolated metastatic spread, accounting for 15% of the sample. There   This systematic review had several limitations. First, the data were limited to case reports and small series, with nonsystematic reporting of the data. Additionally, since all cases featured SB chordomas that eventually metastasized, some data (including those on recurrences or the extent of surgery) may be inherently skewed. Furthermore, finer granularity data on the specific areas of clival involvement were not available in most of the included investigations, preventing additional analyses. Finally, the data pertaining to specific surgical approaches and adjuvant therapies were highly variable, precluding further subgroup analysis through the Kaplan-Meier method.

| Survival outcomes
However, this study is not without any strengths. These included a multiple author systematic database search strategy, study screening, and data extraction and synthesis. Additionally, this study identifies a specific area within the skull base field in need of further research.
Most importantly, this is the first systematic review to investigate a rare subset of exceedingly aggressive skull base malignancies. Further multi-institutional prospective research is needed to fully elucidate the clinical patterns and outcomes of this challenging skull base malignancy.

| CONCLUSION
Pulmonary and spinal metastasis were the most common sites of metastatic spread of SB chordoma. Overall survival was a median of 84 months, but no significant differences in survival outcomes were noted when stratifying the sample by the extent of surgery or the site of metastases. Overall survival was found to be lower in the current metastatic subgroup compared with more general SB chordoma populations.