Abstract
BACKGROUND
Myxopapillary ependymomas (MPEs) are rare intradural extramedullary tumors that often involve the conus and cauda equina. Current guidelines recommend resection as first-line management, but, despite a 92.4% overall 10-year survival rate, MPEs often have a high potential for postoperative morbidity with persisting or worsened neurological function. While several studies have identified potential radiological prognostic factors, none have specifically explored their relation to neurological outcomes.
OBSERVATIONS
The authors retrospectively reviewed 16 MPE cases surgically treated within their institution between 2010 and 2021, evaluating preoperative/postoperative clinical data, surgical variables, and preoperative radiological features. Univariable analysis was performed using ANOVA for continuous variables and the chi-square test for categorical variables across neurological outcome groups. Multivariable modeling of neurological outcomes was done using logistic regression utilizing only radiological variables with a p value < 0.2. None of the radiological predictors reached statistical significance. The presence of preoperative weakness was the only significant predictor of a good neurological outcome within this model, while conus involvement was a significant predictor of urinary dysfunction at the final follow-up.
LESSONS
Tumor size, conus involvement, and preoperative weakness are prognostic factors that may affect the morbidity profile of patients following resection.
Keywords: myxopapillary ependymoma, radiological features, neurological outcome
ABBREVIATIONS: MPE = myxopapillary ependymoma
Ependymomas are neuroepithelial tumors that arise from ependymal cells that line the cerebral ventricles and spinal cord central canal.1 Myxopapillary ependymomas (MPEs) are rare and primarily occur at the conus medullaris, cauda equina, and filum terminale.2 The estimated incidence of MPE is 1.00 case per million within the American population, with peak incidences in young and middle-aged adults.3 As of the 2021 WHO classification, despite their slow-growing, benign biology, MPEs are grade 2 due to their rates of recurrence and metastatic CSF dissemination.4–6
The National Comprehensive Cancer Network guidelines for spinal ependymomas recommends resection as first-line management, followed by potential adjuvant therapy depending on postoperative staging.7 Because MPEs are predominantly found within the lumbosacral region, patients commonly present with chronic low back pain and lumbar radiculopathy.4,8 The overall 10-year survival rate is 92.4%, but as these tumors often encase the cauda equina roots and are difficult to fully resect, they have a high potential for postoperative morbidity with persisting or worsened neurological function.9 A cohort study of 101 patients described the most common presenting symptom to be pain (90%), followed by sphincter dysfunction (41%) and motor weakness (35%).10 Interestingly, the clinical evolution of these symptoms revealed different ratios of neurological improvement, stability, and deterioration postoperatively, potentially indicating tumor characteristics that may affect the presence and persistence of these symptoms.
While several studies have identified radiological factors that may have prognostic significance to predict survival and oncological outcomes, including tumor size, involvement of adjacent anatomical structures, and violation of tumor capsule integrity, none have specifically explored their relation to neurological outcomes.10–14 In order to determine how tumor characteristics may play a role in a patient’s clinical presentation and recovery, our investigation examined the association between various radiological features and postoperative neurological outcomes of MPE cases surgically treated in our institution.
Study Description
Patient Selection and Clinical Data Collection
We retrospectively reviewed consecutive cases of MPE surgically treated at Mount Sinai Hospital from 2010 to 2021. Inclusion criteria included patients older than 18 years of age with pathologically confirmed MPE and availability of preoperative contrast-enhanced MRI studies. Preoperative clinical data were collected by attending neurosurgeons, senior residents, and senior physician assistants and included duration of symptoms, presence of motor weakness, paresthesias, and urinary symptoms. We used the McCormick Scale and Frankel grading system, two 5-point scoring systems evaluating both motor and sensory function, to categorize pre- and postoperative neurological function.15,16 The McCormick Scale scores are as follows: score I for neurologically intact at the time of evaluation, score II for mild deficit yet functionally independent, score III for moderate deficiency and limited functionality, score IV for severe motor or sensory deficit, and score V for paraplegia or quadriplegia. Similarly, the Frankel grading system assesses spinal cord function following spinal cord injury, where grade A indicates no motor or sensory function below the level of the lesion, grade B indicates no motor function but some preserved sensation, grade C indicates some motor function without practical application, grade D indicates useful motor function, and grade E indicates normal motor and sensory function with or without the presence of abnormal reflexes. Postoperative clinical data collected included tumor recurrence, the patient’s hospital length of stay, and discharge disposition.
Surgical variables included number and levels of laminectomies performed, intraoperative complications, resection of the filum terminale, capsule integrity, and piecemeal compared with en bloc resection. Lastly, we collected data regarding further oncological care including the use of radiotherapy for each case. Institutional review board approval was obtained.
Radiological Characteristics
Preoperative MRI studies were evaluated and screened for multiple radiological features. Tumor enhancement patterns were documented and categorized as either a homogeneous or heterogeneous enhancement pattern based on the final radiology read. Involvement of the conus medullaris and/or sacrum, tumor multifocality, and evidence of prior hemorrhaging were also noted. Tumor volumetric data, including total preoperative volume, maximum tumor height along the sagittal plane, and maximum tumor width along the axial plane, were also collected. Tumor preoperative volume was calculated using the volume equation of a cylinder taking into consideration tumor height and squared radius multiplied by pi.
Statistical Analysis
Statistical analysis was performed using RStudio (version 2024.12.1+563, Posit). Univariable analysis was performed using ANOVA for continous variables and the chi-square test for categorical variables across three postoperative neurological outcome groups: improved, stable, and worsened. Multivariable modeling of neurological outcomes was done using logistic regression. Only radiological variables with a p value < 0.2 were utilized for multivariable analysis. This multivariable model compared a group of patients with good (improved or stable) neurological outcomes to a group of patients with poor (worsened) outcomes.
Results
Clinical Characteristics and Neurological Baselines
Sixteen patients met inclusion criteria; 9 (56.2%) were male and 7 (43.8%) were female. The mean age at the time of surgery was 41.6 years, and the mean symptom duration was 13.9 months. Fifteen (93.8%) patients experienced paresthesia or numbness, whereas motor weakness and urinary dysfunction were experienced by 3 (18.8%) patients each. The McCormick score was either I (8 patients, 50%) or II (8 patients, 50%) on preoperative neurological examination. Similarly, the Frankel grade was either E (8 patients, 50%) or D (8 patients, 50%) on preoperative examination (Table 1). All 8 patients who had a McCormick score of I also had a Frankel grade of E, while the other 8 with a McCormick score of II had a Frankel grade of D. The mean length of follow-up was 23.9 months.
TABLE 1.
Summary of preoperative neurological function
| Value | |
|---|---|
| Age at surgery, yrs | 41.6 ± 16.2 |
| Sex, male/female | 9/7 |
| Symptom duration, mos | 13.9 ± 13.1 |
| Preop symptoms | |
| Motor weakness | 3 (18.8) |
| Paresthesia or numbness | 15 (93.8) |
| Urinary dysfunction | 3 (18.8) |
| Preop McCormick grade | |
| I | 8 (50.0) |
| II | 8 (50.0) |
| III | 0 (0.0) |
| IV | 0 (0.0) |
| V | |
| Preop Frankel grade | |
| A | 0 (0.0) |
| B | 0 (0.0) |
| C | 0 (0.0) |
| D | 8 (50.0) |
| E | 8 (50.0) |
Values are given as number of patients (%) or mean ± SD.
Univariable Analysis of Radiological Characteristics
We performed univariable analysis of preoperative radiological features and postoperative neurological outcomes. Four (25%) patients had improved neurological function after surgery, while 7 (43.8%) remained stable and 5 (31.2%) patients had a decline in neurological function (Fig. 1). All MPEs demonstrated contrast enhancement, with 9 (56.2%) heterogeneously enhancing and 7 (43.8%) homogeneously enhancing; however, the enhancement pattern was not associated with postoperative neurological outcomes (p = 0.85). Multifocal disease at presentation was seen in 2 (12.5%) patients with stable neurological function, in 3 (18.8%) patients with worsened neurological function, and in none of the patients who improved postoperatively (p = 0.15). A similar effect was seen with intratumoral hemorrhage, which was observed in 1 (6.3%) patient who was stable and in 2 (12.5%) patients who declined neurologically after surgery (p = 0.28). Sacral involvement was seen in 1 (6.3%) patient who was stable postoperatively and in 2 (12.5%) patients who were worse postoperatively (p = 0.50). Conus involvement was observed twice, both (12.5%) observed in patients who had worsened neurological function (p = 0.08).
FIG. 1.
Preoperative MPE tumor volumes and postoperative neurological (neuro) outcomes of each patient.
Only variables directly representing tumor size were independently associated with neurological outcomes. Tumors with greater mean heights were observed in patients with worsened outcomes (60.9 cm) compared to those with stable (29.8 cm) and improved (35.4 cm) function (p = 0.02). There was a nonsignificant trend toward greater mean axial tumor width in patients who had worsened neurological function (19.4 cm) compared to improved (15.2 cm) or stable (13.7 cm) patients (p = 0.09). Small mean tumor volumes were observed in patients who improved (6.5 cm3) or were stable (5.9 cm3) neurologically compared to those who declined (19.3 cm3) (p = 0.01) (Table 2).
TABLE 2.
Univariable analysis of radiographic features and neurological outcomes
| Total | Improved | Stable | Worsened | p Value | |
|---|---|---|---|---|---|
| Multifocal | 5 | 0 | 2 | 3 | 0.15 |
| Conus involvement | 2 | 0 | 0 | 2 | 0.08 |
| Sacral involvement | 3 | 0 | 1 | 2 | 0.50 |
| Intratumor hemorrhage | 3 | 0 | 1 | 2 | 0.28 |
| Mean max height, cm | 40.9 | 35.4 | 29.8 | 60.9 | 0.02 |
| Mean max width, cm | 15.9 | 15.2 | 13.7 | 19.4 | 0.09 |
| Mean vol, cm3 | 10.6 | 6.5 | 5.9 | 19.3 | 0.01 |
| Enhancement pattern | 0.85 | ||||
| Homogeneous | 7 | 2 | 3 | 2 | |
| Heterogeneous | 9 | 2 | 4 | 3 |
Values are given as number of patients unless otherwise indicated. Boldface type indicates statistical significance.
Multivariable Analyses
We performed multivariable modeling of neurological outcomes using the same radiological variables from univariable analysis. Given the limited sample size, we restricted the model construction to include predictors with a p value < 0.2 from univariable analysis (multifocality, conus involvement, maximum height, maximum width, and volume). In this model, 11 patients with good neurological outcomes were compared to a group of 5 patients with poor outcomes (Fig. 2A). None of the radiological predictors reached statistical significance.
FIG. 2.
Multivariable logistic regression models for good versus poor outcome. Only radiological variables with p value < 0.2 from univariable analysis were included. Error bars are not included. The only predictors that reached statistical significance are in red. A: Multivariable logistic regression model using only radiological predictors. B: Multivariable logistic regression model of good versus poor outcome using radiological predictors and preoperative neurological examination. C: Multivariable logistic regression model of urinary dysfunction at the final follow-up versus normal urinary function using radiological predictors and preoperative urinary dysfunction status.
Given that the model, which only utilized pertinent radiological factors, did not demonstrate significant predictors, we next sought to improve the model by including preoperative neurological factors such as the presence of preoperative weakness and urinary dysfunction. Interestingly, the only significant predictor of a good neurological outcome in this model was the presence of preoperative weakness (Fig. 2B).
Lastly, we attempted to model urinary dysfunction as a final outcome using radiological and clinical predictors. This model compared patients with urinary dysfunction at the final follow-up to those without it. The model utilized tumor volume, preoperative urinary dysfunction, multifocality, maximum width, maximum height, and involvement of the conus (Fig. 2C). Conus involvement was a significant predictor of urinary dysfunction at the final follow-up.
Surgical and Oncological Outcomes
Surgical and oncological outcomes are summarized in Table 3. Piecemeal resection was performed in 10 (62.5%) patients, while 6 (37.5%) patients underwent an en bloc resection. All 4 (25%) patients who had improved neurological function were treated with piecemeal resection. Twelve (75%) patients were discharged home, and 4 (25%) were discharged to acute rehabilitation facilities. There were no perioperative complications. Two (12.5%) patients received adjuvant radiation therapy following tumor resection, and 2 (12.5%) patients had radiological recurrence during follow-up.
TABLE 3.
Surgical and oncological outcomes
| Value | |
|---|---|
| No. of laminectomies | 2.8 ± 0.8 |
| Fusion | 1 (6.3) |
| Resection technique | |
| En bloc | 6 (37.5) |
| Piecemeal | 10 (62.5) |
| Complications | 0 (0.0) |
| Hospital length of stay, days | 9.4 ± 12.4 |
| Discharge disposition | |
| Home | 12 (75.0) |
| Acute rehab | 4 (25.0) |
| Radiographic recurrence | 2 (12.5) |
| Radiation therapy | 2 (12.5) |
Values are given as number of patients (%) or mean ± SD.
Informed Consent
The necessary informed consent was obtained in this study.
Discussion
Observations
MPEs pose particular challenges to neurosurgeons, given the difficulty of achieving a gross-total resection and the relatively high recurrence rate, but excellent long-term survival outcomes make minimizing postoperative morbidity and persisting or worsened neurological deficits essential.17,18 Preoperative contrast-enhanced MRI is a crucial study for surgical planning and preoperative counseling. Radiological studies have been done to differentiate MPEs from schwannomas and other intradural tumors; however, there are no studies that explore the association between imaging characteristics and neurological outcomes.19,20 In this paper, we take a detailed look at 16 MPE cases managed in our instituition to identify any radiological features that may be utilized prognostically to determine postoperative neurological outcome.
Univariable analysis demonstrated that tumor metrics directly related to size were most prognostic of operative risk and neurological outcome. We observed that maximum height and volume significantly differed across neurological outcome cohorts (improved, stable, and worsened) and were markedly higher in the group with worsened neurological outcomes. Similar findings in previous studies were also observed for maximum height, particularly in tumors that extend more than 2 vertebrae.21 While not significant, there was also a trend toward increased tumor width and worsened neurological outcome.
Conus involvement and multifocality were metrics not directly related to tumor size that appeared important from the univariable analysis. While not statistically significant, there was a trend toward worsened outcomes in patients with tumors involving the conus or those with multifocal spread. Both patients with conus involvement had a worsened postoperative examination, and 3 of 5 patients with multifocal tumors had a worsened neurological outcome. These findings correlate to current literature, hypothesizing that MPE tumors, due to their predominant location at the terminal end of the spinal cord, may surround and envelop the nerves of the cauda equina, leading to unique surgical challenges.22
Multivariable logistic regression models yielded largely inconclusive findings. While metrics directly related to tumor size were independently associated with worsened neurological outcomes, neither tumor height, width, nor volume reached statistical significance across models. Interestingly, the presence of preoperative weakness was associated with a good neurological outcome. Of the 3 patients who presented with preoperative weakness, 2 improved while 1 worsened. Those who improved presented with radicular weakness in a single myotome that improved after resection. A previous study assessing the time to diagnose intraspinal tumors and potential factors contributing to long delays highlighted that pain and paresis in the back and upper or lower extremities were the most common initial symptoms present at diagnosis.23 Patients with these symptoms are commonly seen for outpatient evaluation, likely due to the symptoms impacting daily activities, increasing the chance of an earlier detection of an intraspinal tumor compared to nonspecific, slowly progressive symptoms such as erectile dysfunction or sphincter dysfunction.23 This conclusion is supported by our finding when modeling urinary dysfunction as a final outcome, which showed that conus involvement was significantly associated with a poorer outcome.
This correlation of larger tumor sizes leading to poorer prognoses accords with the current literature, which emphasizes that smaller MPE size facilitates gross-total resection, while larger tumors, particularly with involvement of the sacral canal, yield poorer functional outcomes at the time of discharge.11,24 Our study demonstrates that while tumor size and location can raise the morbidity profile of resection, the presence of preoperative deficits should not be a deterrent to operative management. In prior studies, patients with preoperative radicular pain were reported to have either improved or stable neurological outcomes, while asymptomatic patients reported worse outcomes, potentially highlighting the effects of delayed clinical presentation and definitive tumor diagnosis.25,26 Tumors involving the conus involve considerable risk with resection, and while our study demonstrated an association with postoperative urinary dysfunction, more debilitating morbidity from paraplegia is also possible. Our finding is comparable to previous work demonstrating that patients with conus-involved MPE are more likely to experience urinary dysfunction both pre- and postoperatively.12
There are several limitations involved with this study. The study design of this paper is retrospective in nature and had different attending physicians involved in each case, increasing variability. Most notably, the cohort is small given the rarity of this tumor type. To accommodate for this, we included a descriptive analysis to emphasize the presence of identified features that may affect postoperative morbidity. Future studies of MPE conducted prospectively with multi-institutional participation have the potential to reveal additional findings. It is also worth noting that half of the patients within the cohort achieved the highest possible preoperative McCormick and Frankel scores, potentially creating a ceiling effect. To address this, potential use of a more granular scaling system, such as the Medical Research Council system for muscle strength grading, should be considered to provide a better evaluation and resolution of neurological deficit postoperatively in future studies.
Lessons
Our study, which retrospectively evaluates a small case series within our institution, highlights that MPE tumor size metrics obtained from preoperative radiological imaging were most prognostic of operative risk and neurological outcome. Additionally, the presence of preoperative weakness was associated with a good neurological outcome, while urinary dysfunction in association with radiological conus involvement yielded poor outcomes. Despite the potential that these factors may characterize the postoperative neurological outcome following resection, they should not act as a deterrent for operative management given the excellent long-term survival outcomes for these patients.
Disclosures
Dr. Margetis reported training support from Stryker, Globus, and Medtronic.
Author Contributions
Conception and design: Margetis, Nichols. Acquisition of data: Nichols. Analysis and interpretation of data: Seenarine, Nichols. Drafting the article: Seenarine, Nichols, Houten. Critically revising the article: Margetis, Seenarine, Houten. Reviewed submitted version of manuscript: Margetis, Seenarine. Approved the final version of the manuscript on behalf of all authors: Margetis. Study supervision: Margetis.
Supplemental Information
Previous Presentations
The abstract of this paper was presented in poster format at the 2025 Congress of Neurological Surgeons (CNS) Annual Meeting, Los Angeles, CA, October 11–15, 2025.
Correspondence
Konstantinos Margetis: Mount Sinai Health System, New York, NY. konstantinos.margetis@mountsinai.org.
References
- 1.Rudà R Bruno F Pellerino A Soffietti R.. Ependymoma: evaluation and management updates. Curr Oncol Rep. 2022;24(8):985-993. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Bhutada AS Adhikari S Cuoco JA Hoggarth AR Patel VM Olasunkanmi AL.. Development and validation of a predictive nomogram for patients with myxopapillary ependymoma: a surveillance, epidemiology, and end results (SEER) retrospective cohort analysis. Glob Spine J. 2025;15(4):1905-1913. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Bates JE Choi G Milano MT.. Myxopapillary ependymoma: a SEER analysis of epidemiology and outcomes. J Neurooncol. 2016;129(2):251-258. [DOI] [PubMed] [Google Scholar]
- 4.Tabor JK, Ryu B, Schneider D.Multifocal lumbar myxopapillary ependymoma presenting with drop metastasis: a case report and review of the literature. Spinal Cord Ser Cases. 2022;8(1):43. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Straus D Tan LA Takagi I O’Toole JE.. Disseminated spinal myxopapillary ependymoma in an adult at initial presentation: a case report and review of the literature. Br J Neurosurg. 2014;28(5):691-693. [DOI] [PubMed] [Google Scholar]
- 6.Khan NR VanLandingham M O’Brien T Boop FA Arnautović K.. Primary seeding of myxopapillary ependymoma: different disease in adult population? Case report and review of literature. World Neurosurg. 2017;99:812.e21-812.e26. [DOI] [PubMed] [Google Scholar]
- 7.National Comprehensive Cancer Network.. Central Nervous System Cancers. Version 4.2024. Accessed February 13, 2025. https://www.nccn.org/professionals/physician_gls/pdf/cns.pdf [Google Scholar]
- 8.Leeper H Felicella MM Walbert T.. Recent advances in the classification and treatment of ependymomas. Curr Treat Options Oncol. 2017;18(9):55. [DOI] [PubMed] [Google Scholar]
- 9.Weber DC, Wang Y, Miller R.Long-term outcome of patients with spinal myxopapillary ependymoma: treatment results from the MD Anderson Cancer Center and institutions from the Rare Cancer Network. Neuro Oncol. 2015;17(4):588-595. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Montero AS, Tran S, Amelot A.Clinical characteristics and long-term surgical outcome of spinal myxopapillary ependymoma: a French cohort of 101 patients. J Neurooncol. 2021;152(3):491-499. [DOI] [PubMed] [Google Scholar]
- 11.Marchesini N Tommasi N Faccioli F Pinna G Sala F.. Cauda equina ependymomas: surgical treatment and long-term outcomes in a series of 125 patients. J Neurosurg Spine. 2021;36(3):452-463. [DOI] [PubMed] [Google Scholar]
- 12.Helal A Mikula AL Laack NN Krauss WE Clarke MJ.. Myxopapillary ependymomas; proximity to the conus and its effect on presentation and outcomes. Surg Neurol Int. 2021;12:429. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Abdulaziz M, Mallory GW, Bydon M.Outcomes following myxopapillary ependymoma resection: the importance of capsule integrity. Neurosurg Focus. 2015;39(2):E8. [DOI] [PubMed] [Google Scholar]
- 14.Hutuca I Egervari KL Merkler D Vargas MI.. The many faces of myxopapillary ependymomas. AJNR Am J Neuroradiol. 2025;46(3):583-588. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Samuel N, Tetreault L, Santaguida C.Clinical and pathological outcomes after resection of intramedullary spinal cord tumors: a single-institution case series. Neurosurg Focus. 2016;41(2):E8. [DOI] [PubMed] [Google Scholar]
- 16.van Middendorp JJ Goss B Urquhart S Atresh S Williams RP Schuetz M.. Diagnosis and prognosis of traumatic spinal cord injury. Glob Spine J. 2011;1(1):1-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Ellison DW, Aldape KD, Capper D.cIMPACT-NOW update 7: advancing the molecular classification of ependymal tumors. Brain Pathol. 2020;30(5):863-866. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Louis DN, Perry A, Wesseling P.The 2021 WHO Classification of Tumors of the Central Nervous System: a summary. Neuro Oncol. 2021;23(8):1231-1251. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Song C, Kim HS, Lee JH.Development of a novel prediction model for differential diagnosis between spinal myxopapillary ependymoma and schwannoma. Sci Rep. 2024;14(1):149. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Adelved A Tötterman A Hellund JC Glott T Madsen JE Røise O.. Radiological findings correlate with neurological deficits but not with pain after operatively treated sacral fractures. Acta Orthop. 2014;85(4):408-414. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Gembruch O, Chihi M, Haarmann M.Surgical outcome and prognostic factors in spinal cord ependymoma: a single-center, long-term follow-up study. Ther Adv Neurol Disord. 2021;14:17562864211055694. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Kucia EJ Maughan PH Kakarla UK Bambakidis NC Spetzler RF.. Surgical technique and outcomes in the treatment of spinal cord ependymomas: part II: myxopapillary ependymoma. Neurosurgery. 2011;68(1)(Suppl Operative):90-94. [DOI] [PubMed] [Google Scholar]
- 23.Jellema K Overbeeke JJ Teepen HL Visser LH.. Time to diagnosis of intraspinal tumors. Eur J Neurol. 2005;12(8):621-624. [DOI] [PubMed] [Google Scholar]
- 24.Dougherty MC, Sandhu MRS, Teferi N.Surgical outcomes and risk factors for recurrence of myxopapillary ependymoma: a single-center experience. J Neurosurg Spine. 2023;39(4):548-556. [DOI] [PubMed] [Google Scholar]
- 25.Fridley JS Syed S Niu T Leary OP Gokaslan ZL.. Presentation of spinal cord and column tumors. Neurooncol Pract. 2020;7(suppl 1):i18-i24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Kakutani K, Kanda Y, Yurube T.The identification of risk factors for symptomatic spinal metastasis onset: a prospective cohort study of 128 asymptomatic spinal metastasis patients. Cancers (Basel). 2023;15(4):1251. [DOI] [PMC free article] [PubMed] [Google Scholar]