CORRELATION BETWEEN ACTUAL SURVIVAL AND TOKUHASHI AND TOMITA SCORES IN SPINE METASTASES

Objective: To evaluate the accuracy of the scores of Tokuhashi and Tomita and the actual survival of patients with vertebral metastases. Methods: A retrospective assessment of 45 patients with spinal metastases. Thirty-one patients underwent surgical treatment and adjuvant therapy and 14 received conservative treatment (chemotherapy/radiotherapy) or palliative/supportive, depending on the scores of Tokuhashi and Tomita. Results: In the study, 80% of patients were female and the mean age was 57.8 years (SD=11.3 years). The most frequent primary tumors were breast and prostate (68.9%). The accuracy of Tokuhashi scale was 53.4% and the Tomita, 64.5%. The concentration of Tomita range of correct classification was in the category of survival > 12 months (57.8%), while the Tokuhashi scale presented some adjustment in the other categories, < 6 months (15.6%) and 6 to 12 months (2.2%). The histological type of the primary tumor was the only variable that statistically influenced the survival time of patients (p<0.001), and patients with lung or liver tumor (most aggressive) presented a risk of death 9.89 times higher than patients with primary tumors of breast or prostate (less aggressive) (95% CI: 3.10 to 31.57). Conclusion: The Tokuhashi and Tomita scores showed good accuracy with respect to the actual survival of patients with tumor metastasis in the spine.


INTRODUCTION
The spine is the most common site of bone metastasis.Between 30% and 70% of cancer patients will have evidence of metastasis in the spine in post-mortem exams. 1 The most common tumors that metastasize to the spine are breast, lung, kidney, prostate, thyroid, melanoma, lymphoma, and colorectal.Most metastases occur in the thoracic spine (70%), followed by the lumbar spine (20%) and the cervical spine (10%).Multiple non-contiguous metastases are found in 10% to 38% of cases. 1 Spinal metastases can cause significant clinical problems for patients, including pain and neurological symptoms.Pain can be a direct effect, caused by a local or indirect inflammatory response, or due to instability and/or a fracture caused by the actual metastasis.Similarly, the neurological symptoms of medullary or radicular compression are caused directly by the tumor mass, or indirectly by the fracture or deformity resulting from the instability.Around 10% of patients with some type of tumor develop neurological compression due to metastases to the spine. 2 The incidence of metastases to the spine is increasing.Factors such as population aging and improvements in medical treatment for cancer, which increase patient survival time, are contributing to the development of metastatic disease in higher numbers of patients. 3ith advances in chemotherapy, radiation therapy, and hormone therapy, the life expectancy of patients has increased.Progress has been made in surgical techniques, together with advances in technology, enabling more effective surgical treatment of spinal metastases. 4he role of the surgeon in metastatic tumors of the spine is always a subject of discussion, because the surgeon can improve mechanical instability, medullary compression, and pain, but there are still doubts as to the surgeon's role in increasing survival times.In the past, decompression techniques without stabilization resulted in worse outcomes than radiation therapy.Thus, one might think that radiation therapy is the preferred option when compared to surgery for certain types of cancers.However, recent evidence has shown that modern surgery (including anterior and posterior approaches with stabilization) generates better results than isolated radiation therapy, and that the quality of life of those patients also increases. 5,6hen opting for surgical treatment, we must keep in mind that in most patients with metastatic spinal tumors, life expectancy governed by the site of the primary tumor and by staging, generally around 1 to 2 years of survival time, since tumor metastasis in and of itself indicates an already advanced stage of the disease.Therefore, surgery should not have a negative impact on the remaining quality of life.The rate of surgical complications can be high (20-30%) and this should be taken into account in selecting the treatment.This is especially applicable in complex surgeries such as en bloc resections, which are associated with an increase in morbidity and mortality as compared to simpler procedures, such as palliative resections.Although surgery is currently considered the treatment of choice for spinal metastases, more evidence is necessary to define the role and the indications of the various surgical techniques and the newer, more radical treatments available. 2everal survival scoring systems have been developed, among them those proposed by Tokuhashi et al 7,8 and by Tomita et al. 9 These scores are tools used to help choose the best treatment for patients with spinal metastases, based on average survival time.
The objective of this study was to evaluate the accuracy between the Tokuhashi and Tomita scores and the actual survival time in patients with spinal metastases.

MATERIALS AND METHODS
Following approval by the Ethics Committee of the Universidade Estadual de Campinas (UNICAMP), Campinas, SP, Brazil, 45 patients with metastases to the spine, who were in follow-up at the Spinal Surgery Clinic of the Department of Orthopedics at UNICAMP, were evaluated retrospectively.Of these 45 patients, 31 had undergone surgery and adjuvant chemotherapy and or radiation therapy and 14 had received only adjuvant (chemotherapy and/or radiation therapy) or palliative/supporting treatment, in accordance with the Tokuhashi and Tomita scores.
Patients with metastasis to the spine who were evaluated by radiographs, axial computed tomography, and magnetic resonance of the entire spine, as well as bone scintigraphy, and magnetic resonance of the chest, abdomen, and skull, were included in this study.These exams were standardized in order to diagnose and stage the patients.Patients with incomplete medical records, and in whom the Tokuhashi and Tomita scores had not been calculated, were excluded.The minimum follow-up time was one year.
The Tokuhashi score was developed to indicate the type of treatment in metastatic lesions of the spine, based on the following criteria: 1) General condition of the patient, according to Karnofsky and Young 10 : poor, 0 points; moderate, 1 point; and good, 2 points; 2) Number of extraspinal metastases: greater than or equal to three, 0 points; two, 1 point; and one, 2 points; 3) Number of spinal metastases: greater than or equal to three, 0 points; two, 1 point; and one, 2 points; 4) Resectability of metastases in vital organs: non--resectable, 0 points; resectable, 1 point; and absent, 2 points; 5) Site of the primary tumor: lung and stomach, 0 points; kidney, liver, and uterus, 1 point; and thyroid, prostate, breast, rectum, others, and unidentified, 2 points; 6) Medullary neurological compromise: total, 0 points: incomplete, 1 point; absent, 2 points.A score of from 0 to 2 is given for each of the six parameters for a maximum of 12 points. 7The authors later modified the scoring system, changing the site of the primary tumor parameter, which now ranges from 0 to 5 points, due to its relevance in the prognosis of these patients. 8fter this modification, the new score was: lung, osteosarcoma (which was included in the list, despite being a primary bone tumor), stomach, bladder, esophagus, and pancreas, 0 points; liver, ureter, and unidentified, 1 point; other, 2 points; kidney and uterus, 3 points; rectum, 4 points; and thyroid, breast, prostate, and carcinoid, 5 points.In this study, the revised Tokuhashi score was used.(Table 1) Based on these indices, Tokuhashi et al 8 define the prognosis and the treatment option for these patients, respectively, as follows: a) 0 to 8 points, prognosis of up to 6 months, conservative or palliative treatment in isolated cases; 9 to 11 points, prognosis greater than 6 months, palliative treatment or excisional surgery in cases of a single lesion and without metastases to vital organs; 12 to 15 points, prognosis of more than 1 year, treatment with excisional surgery.(Table 2) The scoring system of Tomita et al 9 assigns points based on three prognostic factors: 1) Degree of malignity of primary tumor in terms of growth (slow, 1 point; moderate, 2 points; and rapid, 4 points); 2) Presence of visceral metastasis (without metastasis, 0 points; treatable, 1 point; untreatable, 4 points); and 3) Presence of bone metastasis (solitary or isolated, 1 point; multiple, 2 points).The total may range from 2 to 10 points.(Table 3) According to Tomita et al 9 , the treatment strategy will be: a) 2 to 3 points, wide or marginal excision for long term local control; b) 4 to 5 points, marginal or intralesional excision for midterm local control; c) 6 to 7 points, palliative surgery for short term control; d) 8 to 10 points, non-surgical treatment.(Table 4) In the study of Tomita et al 9 , patients with scores of 2 or 3 have and average survival time of 38.2 months, scores of 4 or 5 have an average survival time of 21.5 months, 6 and 7 points a survival time of 10.1 months, and patients with a score of 8 to 10 points have and average survival time of 5.3 months.Therefore, to standardize the Tomita scale in relation to the Tokuhashi scale, patients with total points of between 2 and 5 have an average survival time of more than 12 months; those with 6 or 7 points have an average survival time of 6 to 12 months; and those with a total of between 8 and 10 points have a survival time of less than 6 months.

RESULTS
Forty-five patients with metastatic disease in the spine were selected.The patients were classified into three groups using the Tokuhashi and Tomita scales: Group 1 (< 6 months of survival), Group 2 (6 to 12 months of survival), and Group 3 (> 12 months of survival), and were followed up to evaluate their actual survival times.Thus, it was possible to verify whether the scales developed by Tokuhashi et al 8 and by Tomita et al 9 were accurate in terms of the patients' actual survival times, and whether patient characteristics such as sex, age, site of the primary tumor, and surgery influenced the survival of patients with metastatic tumors of the spine.
The personal and diagnostic characteristics were described using summary measurements (average, standard deviation) by age and absolute and relative frequencies for sex, histological type of primary tumor, and surgical or non-surgical treatment.Nine patients presented neurological deficits, four of them with metastasis of breast cancer (Frankel D); one with epidermoid carcinoma of the uterus (Frankel D), one with prostrate carcinoma (Frankel C); one with gastric carcinoma (Frankel C), and two with lung carcinoma (Frankel B).
The accuracies for the respective scales with the real survival groups were calculated, with confidence intervals of 95%.
The average survival time of the patients was estimated using the Kaplan-Meier function 11 , by sex, age range, site of the primary tumor, and surgical or non-surgical treatment.Comparing the categories using the log-rank test, 11 the average survival time was calculated, as it was impossible to calculate the mean survival time due to the low number of deaths in certain categories.The Hazard Ratio (HR) was estimated using Cox's bivariate regression model, with the respective confidence intervals of 95%, and the risk of death among the categories was estimated using Cox's multivariate regression.
The tests were conducted with a level of significance of 5%.Table 5 shows that most of the patients in the study were female (80%), with an average age of 57.8 years (SD = 11.3 years), and the primary tumors were divided according to aggressiveness.The least aggressive (breast and prostate) corresponded to 68.9%, the most aggressive (lung and liver) to 13.3%, and the other tumors to 17.8% of the patients selected.Approximately 69% of the patients underwent surgery.
Table 6 shows that the accuracy of the Tokuhashi scale was 53.4% and that of the Tomita scale was 64.5%, indicating slightly better accuracy for the Tomita scale than for the Tokuhashi scale.

2-3 points
Wide or marginal excision, long term control.
The concentration of matches with the Tomita scale was in the survival > 12 months category (57.8%), while the Tokuhashi scale presented some matches in the other categories -< 6 months (15.6%) and 6 to 12 months (2.2%).Figures 1-4 suggest that the histological type of the primary tumor (breast and prostate) was the only variable among those evaluated that influenced patient survival.
Table 7 shows that the histological type of the primary tumor was the only variable that influenced the survival time of the patients statistically (p < 0.001), and that patients with lung or liver tumors (more aggressive) had a 9.89 times greater risk of death than patients with breast or prostate (less aggressive) primary tumors (CI95% : 3.10 to 31.57).

DISCUSSION
The spine is the most common site of tumor metastasis to the bones.][14] The tumors that most commonly metastasize to the spine are those of the breast, lung, kidney, prostate, thyroid, melanoma, lymphoma, Table 6.Description of real survival of the patients according to the scales and the accuracy results of the scales.6][17] The presence of bone metastases to the spine often indicates that the disease at the primary site in incurable; however, with better adjuvant therapies, patients with metastases are living for longer periods following diagnosis. 18ome prognostic classifications are used to guide the treatment of patients affected by metastatic disease of the spine in terms of the best therapeutic option.Among them we cite Tokuhashi et al, 7,8 Sioutos et al, 19 Van der Linden et al, 20 Tomita et al, 9 and Bauer et al. 21,22 In this study, the modified classification developed by Tokuhashi et al 8 and the classification developed by Tomita et al 9 were used.
Tokuhashi et al 8 described a prognostic evaluation system for metastatic tumors of the spine based on six variables: site of the primary tumor, presence or absence of paralysis, Karnofsky clinical performance status, number of extraspinal bone metastases, number of spinal metastases, and number of visceral metastases.
These six factors are evaluated together, producing a value ranging from 0 to 15 points, with zero indicating a poor prognosis and 15 a good prognosis (Table 2).It is interesting to note that Tokuhashi et al 8 consider neurological deficit to be a significant prognostic factor of survival.
Tomita et al 9 studied several prognostic factors for tumor metastases in order to describe a system based on three factors: the growth rate of the primary tumor, the number of bone metastases, and the number of visceral metastases.(Table 3) These three factors are evaluated together, generating a score value from 2 to 10 points for a good to a bad prognosis, respectively.(Table 4) The site of the primary tumor is considered to be the most important prognostic factor in both the Tomita and Tokuhashi scales.According to Tomita et al 9 , metastases of breast, prostate, and thyroid cancers permit longer survival.According to the modified Tokuhashi scale, less aggressive tumors such as those of the breast, prostate, thyroid, and carcinoid tumors score 5 points.More aggressive tumors, like those of the lung, osteosarcoma, esophagus, and pancreas, score zero points, indicating a worse prognosis for patient survival. 8his study corroborates this information, given that the patients with breast or prostate metastases, considered a less aggressive histological type, survived longer than those with more aggressive tumors, such as tumors of the lung and liver.(Table 7 and Figure 3) In another study, Enkaoua et al 23 report that patients with spinal metastasis in which the site of origin of the primary tumor is unknown have a worse prognosis for survival.They also state that neurological deficit should not be considered a prognostic factor of survival because it can be resolved with decompression, and is related to the speed of growth of the tumor.However, Tokuhashi et al 9 consider neurological deficit to be an isolated prognostic factor to be considered, and it is therefore attributed a score in the evaluation scale described by those authors.
In our evaluation, nine patients had neurological deficits, four of them from metastasis of breast cancer (Frankel D), one from epidermoid carcinoma of the uterus (Frankel D), one from prostate carcinoma (Frankel C), one from gastric carcinoma (Frankel C), and two from carcinoma of the lung (Frankel B).The patients with metastasis from adenocarcinoma of the lungs died prior to any neurological recovery.The other patients recovered from the neurological deficit.
Zou et al 24 report a differentiation in relation to the analysis of the Tomita and Tokuhashi scores.According to this study, the score

Figure 1 .
Figure 1.Kaplan-Meier function of patient survival by sex.

Figure 2 .
Figure 2. Kaplan-Meier function of patient survival by age range.

Figure 3 .
Figure 3. Kaplan-Meier function of patient survival by primary tumor.
older CORRELATION BETWEEN ACTUAL SURVIVAL AND TOKUHASHI AND TOMITA SCORES IN SPINE METASTASES

Figure 4 .
Figure 4. Kaplan-Meier function of patient survival by surgical status.

Table 5 .
Description of the study patient characteristics.

Table 2 .
Prognosis and treatment according to Tokuhashi et al.

Table 3 .
Scale of Tomita et al.

Table 4 .
Prognosis and treatment according to the scale of Tomita et al.