85. The neurology stability epidural compression assessment: a new score to establish the need for surgery in spinal metastases

Abstract

BACKGROUND CONTEXT Management of spinal metastases has always been challenging. Traditional scores like those proposed by Tokuhashi, Tomita or Bauer have guided decision-making in management of patients with spinal metastases for many years, although some limits were already described by their authors. First, patients are often treated in an emergent setting without a diagnosis and/or a global assessment of the disease. Second, subclinical metastases were not taken into account. Further, and more importantly, the decision for or against surgery was based on survival prognostication alone, but a precise prediction appeared to be limited. In the last two decades, targeted and biologic therapies dramatically changed survival prognosis in metastatic patients, making these scores unreliable. At the same time, the development of stereotactic radiosurgery (SRS) and minimally invasive surgery (MIS) techniques imposed a true paradigm shift. It appeared progressively clear that surgical indications for spinal metastases could not rely anymore on prognostication of survival, but needed to consider functional recovery/preservation and local control as targets to pursue a palliative goal. The algorithm proposed by Boriani and Gasbarrini published in 2008 first focused on functional targets of spinal metastases surgery. In 2013, Laufer et al developed the NOMS framework, incorporating new technological tools, surgical techniques and advances in radiosurgery and systemic treatments. PURPOSE The aim of this study was to translate new evidence-based frameworks in a practical and reliable score for surgeons, radiation oncologists and oncologists, that could establish the need for surgery in the evaluation of a patient with spinal metastases even in the absence of a histological diagnosis and regardless of the available technology. STUDY DESIGN/SETTING This is the validation of a score through a retrospective study. A detailed retrospective evaluation was performed. Every patient who received a surgical procedure or conservative treatment after neurosurgical evaluation for spinal metastases at the authors' institution from January 2015 to May 2019 was considered. Patients were usually treated according to the NOMS framework principles and individual patient preferences. Data recorded for each case included: sex, age, type of tumor, time of occurrence of the spinal metastases, spinal level of the lesion, American Society of Anesthesiologists (ASA) score, performance status according to the Eastern Cooperative Oncology Group (ECOG), neurologic evaluation, Spinal Instability Neoplastic Score (SINS), grade of epidural compression according to the Epidural Spinal Cord Compression Scale (ESCC), mechanical pain evaluation, type of treatment, type of surgical treatment if performed, necessity of surgery after initial conservative management, and neurological and axial/radicular pain evaluation at 3 months (at least) follow-up. To reduce the incidence of events considered as exclusion criteria, a maximum of 6 months of follow-up (when available) was evaluated. PATIENT SAMPLE Inclusion criteria were: (1) a diagnosis of a malignant neoplasm with a treated location of spinal metastases, including both solid and hematopoietic tumors; (2) a multidisciplinary evaluation by neurosurgeons, radiotherapists and oncologists of the patient with available pre-treatment imaging (MRI and CT scan); (3) availability of at least a 3-month follow-up with recorded data for all variables included in the analysis (described below); (4) ASA score 1 and/or SINS score >6; (5) occurrence, after the treatment and during the follow-up, of other spinal metastases with ESCC > 1 and/or SINS score >6; (6) pre-existing conditions or occurrence after treatment of adverse events that could influence the neurological status of the patient and/or of bony adverse events that could condition or influence the evaluation of axial/radicular pain of the patient; and (7) occurrence after treatment and during the follow-up of postsurgical and/or postradiation and/or postsystemic treatment complications and/or systemic adverse events that could impair the global evaluation of the patient. Clinical variables at diagnosis included: (A) neurological status - at diagnosis, the classification identified different conditions: absence of deficits, non-motor pure radicular pain caused by direct compression of the tumor on the root, motor radicular impairment or resistant mechanical radicular pain caused by foramen collapse, complete cord damage for 72 hours, incomplete spinal cord impairment, and cauda equina syndrome; (B) axial/radicular pain - it was defined as mechanical pain, which is movement related and should be distinguished in these patients from biologic pain due to inflammation, which presents in the evenings and early mornings. Radicular pain, even if belonging to the sphere of neurological status, could be considered as a typical symptom of instability when present upon standing because of the collapse of the neural foramen. Mechanical pain has been investigated according the WHO Pain Ladder. Pharmacological therapy before and after treatment at follow-up was evaluated to objectively define a clinical stability, improvement or worsening OUTCOME MEASURES Patient analysis considered the agreement between the NSE score (see Methods) and the undertaken management. The two groups of patients with and without agreement were separately evaluated according to both neurological and axial pain status before and after the treatment at follow-up (3 and 6 months). The same analysis was made for patients in the grey zone (NSE score of 3 or 4) who received a surgical or nonsurgical treatment. Neurological status: at follow-up, improvement was registered when the patient recovered at least 1 point in at least 1 limb according to the standard Medical Research Council (MRC) scale for muscle strength. Worsening was registered in cases where the patient lost 1 point in at least 1 limb according to the standard MRC scale for muscle strength. Axial/radicular pain: Patients were considered improved, stable or worsened. Associations between the clinical evaluation criteria and the age of patients, type of tumor, level of surgery, SINS score, ESCC score, and type of neurological deficit were also investigated in the group of agreement. METHODS Three main items were identified for the score: clinical neurological status, stability of the spine, and epidural compression (NSE score). Patients were considered suitable for surgery if ASA score 3 were excluded from the present study. Neurological status was graded between 0 and 5. Zero points were given for the absence of deficits or the presence of a complete clinical cord damage for more than 72 hours. One point was given in cases of non-motor pure radicular pain. Three points were given in cases of motor radicular impairment or resistant mechanical radicular pain. Four points were given in cases of complete cord injury for less than 72 hours. Five points were given in cases of incomplete spinal cord impairment or cauda equina syndrome. This classification is inspired by the Thoracolumbar Injury Classification and Severity (TLICS) score for thoracolumbar spine trauma, but adding a distinction between an acute onset of complete cord impairment, potentially still able to recover, and a stable deficit. Another issue was the difference between a pure radicular pain, due to the direct invasion of the foramen, a motor radicular impairment and a mechanical resistant radicular pain (usually with SINS >6), caused by the dynamic collapse of the vertebral body and thus of the foramen in standing position. Stability was evaluated according to the SINS score. Zero points were given in cases of a stable spine (SINS score between 0 and 6). Three points were given in cases of a potentially unstable spine (SINS score between 7 and 12). Five points were given in cases of an unstable spine (SINS between 13 and 18). Epidural compression was classified according with the ESCC scale also known as the Bilsky scale. Zero points were given in cases of ESCC 0, 1a, 1b. One point was given in cases of ESCC 1c. Three points were given in cases of high grade spinal cord compression (ESCC 2 or 3). In patients with a total NSE score of 0,1, or 2, surgery was not considered as a valid or useful option. Patients with a total NSE score of 3 or 4 identified a grey zone in which both surgery or radiation/systemic treatments alone could be considered, depending above all on the type of tumor (if known), availability of tools like SRS, and clinical and general status. With a total NSE score of 5 or more (maximum 13), surgery was considered mandatory, regardless of tumor histology. Descriptive statistics were reported as a median, mean and standard deviation for continuous variables or frequency and percentage for categorical variables. Comparisons of proportions were performed with chi-square test for categorical variables and, when needed (>20% of values <=5 and/or presence of values <1), with Cramer's Phi and V coefficients to verify association between variables. Statistical significance was defined with a p-value <=0.05. All statistical analyses were performed using SPSS Statistics software (IBM SPSS Statistics for Windows, Version 24.0; IBM Corp., Armonk, New York, USA). RESULTS A total of 283 patients were reviewed but only 145 of them (91 M, 54 F) were included in the study after implementation of the inclusion and exclusion criteria. The most common reason for exclusion was the lack of all needed data for the analysis (82/138). The absence of a 3-month follow-up was recorded in 18 cases but 16 of them were graded as ASA >3 and/or ECOG >2. Mean age was 63.8 years (range 22 to 68). The most common types of tumor were non–small cell lung cancer (NSCLC), breast, prostate and myeloma (respectively 22.1%, 20.7%, 11.7% and 9.7%). The most common location was the thoracic spine (47.6%). Patients were mainly evaluated as ASA 2 and ECOG 2 (respectively 61.4% and 46.2%). Before treatment 42.8% of patients had a neurological impairment. According to the SINS score, 31.7% had no instability, 55.9% were potentially unstable, and 12.4% were frankly unstable. A total number of 69 patients had high grade ESCC. Mechanical axial/radicular pain was registered in 114 patients (78.6%). Patients underwent surgery in 64.8% of cases. Fixation followed by posterolateral or followed by circumferential decompression were the most commonly performed surgical procedures (30.9% and 30.9% respectively). Among patients with neurological impairment before the procedure, a median improvement of 2 points (mean 1.61, SD 0.67) according to the MRC scale of at least one limb was registered at 3 months, and similar results were reported at 6 months (median 2, mean 1.57, SD 0.68). In 88.3% of patientsm there was agreement between the treatment and the NSE score. In cases of nonagreement, the preference of the patient was the most common reason for a nonsurgical management. At last follow-up, respectively 89.6% and 82% of patients did not show neurologic or mechanical pain worsening. In the agreement group, a strong and statistically significant association was recorded with improvement or preservation of neurological functions and mechanical pain (p <0.001) at 3 and 6 months. On the other hand, in the non-agreement group, there was no statistically significant association between preoperative neurologic status and mechanical pain at the 3-month follow-up (p 0.486 and 0.063 respectively) and at 6 months (p 0.343 and 0.858 respectively). In mechanical pain analysis, because of the small number of patients, patients with clinical improvement after deficits or stable were considered together. Considering patients in the grey zone, who received either surgery and postoperative radiotherapy or radiotherapy alone, the follow-up showed no neurological worsening (p 0.007) and improvement/stability of mechanical pain compared to the evaluation before treatment (p 0.001 at 3 and 6 months). In the agreement group, associations were found: between SINS score and the absence of mechanical pain worsening (both clinically stable and improved patients) at the last follow-up (p 0.003); and between the grade of ESCC and the absence of clinical worsening for neurological status (but not for pain) at last follow-up (p 0.03 and 0.13 respectively). CONCLUSIONS Functional outcomes of the study group showed that the proposed NSE score could represent a practical and reliable tool to establish the need for surgery. Agreement between the score and the performed treatments resulted in better clinical outcomes, when compared with patients without agreement. Further validation is needed with a larger number of patients and to assess reproducibility among surgeons, radiation oncologists, and oncologists. FDA DEVICE/DRUG STATUS This abstract does not discuss or include any applicable devices or drugs. Management of spinal metastases has always been challenging. Traditional scores like those proposed by Tokuhashi, Tomita or Bauer have guided decision-making in management of patients with spinal metastases for many years, although some limits were already described by their authors. First, patients are often treated in an emergent setting without a diagnosis and/or a global assessment of the disease. Second, subclinical metastases were not taken into account. Further, and more importantly, the decision for or against surgery was based on survival prognostication alone, but a precise prediction appeared to be limited. In the last two decades, targeted and biologic therapies dramatically changed survival prognosis in metastatic patients, making these scores unreliable. At the same time, the development of stereotactic radiosurgery (SRS) and minimally invasive surgery (MIS) techniques imposed a true paradigm shift. It appeared progressively clear that surgical indications for spinal metastases could not rely anymore on prognostication of survival, but needed to consider functional recovery/preservation and local control as targets to pursue a palliative goal. The algorithm proposed by Boriani and Gasbarrini published in 2008 first focused on functional targets of spinal metastases surgery. In 2013, Laufer et al developed the NOMS framework, incorporating new technological tools, surgical techniques and advances in radiosurgery and systemic treatments. The aim of this study was to translate new evidence-based frameworks in a practical and reliable score for surgeons, radiation oncologists and oncologists, that could establish the need for surgery in the evaluation of a patient with spinal metastases even in the absence of a histological diagnosis and regardless of the available technology. This is the validation of a score through a retrospective study. A detailed retrospective evaluation was performed. Every patient who received a surgical procedure or conservative treatment after neurosurgical evaluation for spinal metastases at the authors' institution from January 2015 to May 2019 was considered. Patients were usually treated according to the NOMS framework principles and individual patient preferences. Data recorded for each case included: sex, age, type of tumor, time of occurrence of the spinal metastases, spinal level of the lesion, American Society of Anesthesiologists (ASA) score, performance status according to the Eastern Cooperative Oncology Group (ECOG), neurologic evaluation, Spinal Instability Neoplastic Score (SINS), grade of epidural compression according to the Epidural Spinal Cord Compression Scale (ESCC), mechanical pain evaluation, type of treatment, type of surgical treatment if performed, necessity of surgery after initial conservative management, and neurological and axial/radicular pain evaluation at 3 months (at least) follow-up. To reduce the incidence of events considered as exclusion criteria, a maximum of 6 months of follow-up (when available) was evaluated. Inclusion criteria were: (1) a diagnosis of a malignant neoplasm with a treated location of spinal metastases, including both solid and hematopoietic tumors; (2) a multidisciplinary evaluation by neurosurgeons, radiotherapists and oncologists of the patient with available pre-treatment imaging (MRI and CT scan); (3) availability of at least a 3-month follow-up with recorded data for all variables included in the analysis (described below); (4) ASA score 1 and/or SINS score >6; (5) occurrence, after the treatment and during the follow-up, of other spinal metastases with ESCC > 1 and/or SINS score >6; (6) pre-existing conditions or occurrence after treatment of adverse events that could influence the neurological status of the patient and/or of bony adverse events that could condition or influence the evaluation of axial/radicular pain of the patient; and (7) occurrence after treatment and during the follow-up of postsurgical and/or postradiation and/or postsystemic treatment complications and/or systemic adverse events that could impair the global evaluation of the patient. Clinical variables at diagnosis included: (A) neurological status - at diagnosis, the classification identified different conditions: absence of deficits, non-motor pure radicular pain caused by direct compression of the tumor on the root, motor radicular impairment or resistant mechanical radicular pain caused by foramen collapse, complete cord damage for 72 hours, incomplete spinal cord impairment, and cauda equina syndrome; (B) axial/radicular pain - it was defined as mechanical pain, which is movement related and should be distinguished in these patients from biologic pain due to inflammation, which presents in the evenings and early mornings. Radicular pain, even if belonging to the sphere of neurological status, could be considered as a typical symptom of instability when present upon standing because of the collapse of the neural foramen. Mechanical pain has been investigated according the WHO Pain Ladder. Pharmacological therapy before and after treatment at follow-up was evaluated to objectively define a clinical stability, improvement or worsening Patient analysis considered the agreement between the NSE score (see Methods) and the undertaken management. The two groups of patients with and without agreement were separately evaluated according to both neurological and axial pain status before and after the treatment at follow-up (3 and 6 months). The same analysis was made for patients in the grey zone (NSE score of 3 or 4) who received a surgical or nonsurgical treatment. Neurological status: at follow-up, improvement was registered when the patient recovered at least 1 point in at least 1 limb according to the standard Medical Research Council (MRC) scale for muscle strength. Worsening was registered in cases where the patient lost 1 point in at least 1 limb according to the standard MRC scale for muscle strength. Axial/radicular pain: Patients were considered improved, stable or worsened. Associations between the clinical evaluation criteria and the age of patients, type of tumor, level of surgery, SINS score, ESCC score, and type of neurological deficit were also investigated in the group of agreement. Three main items were identified for the score: clinical neurological status, stability of the spine, and epidural compression (NSE score). Patients were considered suitable for surgery if ASA score 3 were excluded from the present study. Neurological status was graded between 0 and 5. Zero points were given for the absence of deficits or the presence of a complete clinical cord damage for more than 72 hours. One point was given in cases of non-motor pure radicular pain. Three points were given in cases of motor radicular impairment or resistant mechanical radicular pain. Four points were given in cases of complete cord injury for less than 72 hours. Five points were given in cases of incomplete spinal cord impairment or cauda equina syndrome. This classification is inspired by the Thoracolumbar Injury Classification and Severity (TLICS) score for thoracolumbar spine trauma, but adding a distinction between an acute onset of complete cord impairment, potentially still able to recover, and a stable deficit. Another issue was the difference between a pure radicular pain, due to the direct invasion of the foramen, a motor radicular impairment and a mechanical resistant radicular pain (usually with SINS >6), caused by the dynamic collapse of the vertebral body and thus of the foramen in standing position. Stability was evaluated according to the SINS score. Zero points were given in cases of a stable spine (SINS score between 0 and 6). Three points were given in cases of a potentially unstable spine (SINS score between 7 and 12). Five points were given in cases of an unstable spine (SINS between 13 and 18). Epidural compression was classified according with the ESCC scale also known as the Bilsky scale. Zero points were given in cases of ESCC 0, 1a, 1b. One point was given in cases of ESCC 1c. Three points were given in cases of high grade spinal cord compression (ESCC 2 or 3). In patients with a total NSE score of 0,1, or 2, surgery was not considered as a valid or useful option. Patients with a total NSE score of 3 or 4 identified a grey zone in which both surgery or radiation/systemic treatments alone could be considered, depending above all on the type of tumor (if known), availability of tools like SRS, and clinical and general status. With a total NSE score of 5 or more (maximum 13), surgery was considered mandatory, regardless of tumor histology. Descriptive statistics were reported as a median, mean and standard deviation for continuous variables or frequency and percentage for categorical variables. Comparisons of proportions were performed with chi-square test for categorical variables and, when needed (>20% of values <=5 and/or presence of values <1), with Cramer's Phi and V coefficients to verify association between variables. Statistical significance was defined with a p-value <=0.05. All statistical analyses were performed using SPSS Statistics software (IBM SPSS Statistics for Windows, Version 24.0; IBM Corp., Armonk, New York, USA). A total of 283 patients were reviewed but only 145 of them (91 M, 54 F) were included in the study after implementation of the inclusion and exclusion criteria. The most common reason for exclusion was the lack of all needed data for the analysis (82/138). The absence of a 3-month follow-up was recorded in 18 cases but 16 of them were graded as ASA >3 and/or ECOG >2. Mean age was 63.8 years (range 22 to 68). The most common types of tumor were non–small cell lung cancer (NSCLC), breast, prostate and myeloma (respectively 22.1%, 20.7%, 11.7% and 9.7%). The most common location was the thoracic spine (47.6%). Patients were mainly evaluated as ASA 2 and ECOG 2 (respectively 61.4% and 46.2%). Before treatment 42.8% of patients had a neurological impairment. According to the SINS score, 31.7% had no instability, 55.9% were potentially unstable, and 12.4% were frankly unstable. A total number of 69 patients had high grade ESCC. Mechanical axial/radicular pain was registered in 114 patients (78.6%). Patients underwent surgery in 64.8% of cases. Fixation followed by posterolateral or followed by circumferential decompression were the most commonly performed surgical procedures (30.9% and 30.9% respectively). Among patients with neurological impairment before the procedure, a median improvement of 2 points (mean 1.61, SD 0.67) according to the MRC scale of at least one limb was registered at 3 months, and similar results were reported at 6 months (median 2, mean 1.57, SD 0.68). In 88.3% of patientsm there was agreement between the treatment and the NSE score. In cases of nonagreement, the preference of the patient was the most common reason for a nonsurgical management. At last follow-up, respectively 89.6% and 82% of patients did not show neurologic or mechanical pain worsening. In the agreement group, a strong and statistically significant association was recorded with improvement or preservation of neurological functions and mechanical pain (p <0.001) at 3 and 6 months. On the other hand, in the non-agreement group, there was no statistically significant association between preoperative neurologic status and mechanical pain at the 3-month follow-up (p 0.486 and 0.063 respectively) and at 6 months (p 0.343 and 0.858 respectively). In mechanical pain analysis, because of the small number of patients, patients with clinical improvement after deficits or stable were considered together. Considering patients in the grey zone, who received either surgery and postoperative radiotherapy or radiotherapy alone, the follow-up showed no neurological worsening (p 0.007) and improvement/stability of mechanical pain compared to the evaluation before treatment (p 0.001 at 3 and 6 months). In the agreement group, associations were found: between SINS score and the absence of mechanical pain worsening (both clinically stable and improved patients) at the last follow-up (p 0.003); and between the grade of ESCC and the absence of clinical worsening for neurological status (but not for pain) at last follow-up (p 0.03 and 0.13 respectively). Functional outcomes of the study group showed that the proposed NSE score could represent a practical and reliable tool to establish the need for surgery. Agreement between the score and the performed treatments resulted in better clinical outcomes, when compared with patients without agreement. Further validation is needed with a larger number of patients and to assess reproducibility among surgeons, radiation oncologists, and oncologists.