Validation and simplification of a score predicting survival in patients irradiated for metastatic spinal cord compression
Abstract
BACKGROUND:
Based on an analysis of 1852 retrospectively evaluated patients with metastatic spinal cord compression (MSCC), a scoring system was developed to predict survival. This study was performed to validate the scoring system in a new data set.
METHODS:
The score included 6 prognostic factors: tumor type, interval between tumor diagnosis and MSCC, other bone or visceral metastases, ambulatory status, and duration of motor deficits. Scores ranged between 20 and 45 points, and patients were initially divided into 5 groups: those with 20 to 25 points, those with 26 to 30 points, those with 31 to 35 points, those with 36 to 40 points, and those with 41 to 45 points. To facilitate the clinical use of the score, the patients were regrouped into 3 groups: those with 20 to 30 points, those with 31 to 35 points, and those with 36 to 45 points. In this study, data of 439 new patients were included who were divided into the same prognostic groups as in the preceding study.
RESULTS:
In this study, the 6-month survival rates were 7% (for those with 20-25 points), 19% (for those with 26-30 points), 56% (for those with 31-35 points), 73% (for those with 36-40 points), and 90% (for those with 41-45 points), respectively (P < .0001). After regrouping, the 6-month survival rates were 14% (for those with 20-30 points), 56% (for those with 31-35 points), and 80% (for those with 36-45 points), respectively, in this study (P < .0001).
CONCLUSIONS:
In the current study, the difference in 6-month survival between the prognostic groups was found to be as significant as in the preceding study. Thus, this scoring system was considered valid to estimate survival of MSCC patients. The system could have been simplified by including only 3 instead of 5 prognostic groups. Cancer 2010. © 2010 American Cancer Society.
Approximately 5% to 10% of all cancer patients develop metastatic spinal cord compression (MSCC) during their lifetime.1 The majority of MSCC patients are treated with radiotherapy alone. Each treatment session may be associated with discomfort for these debilitated patients during transportation to the radiation oncology department and patient positioning on the treatment couch. Thus, a short-course of radiotherapy with no more than 5 radiation sessions would be the best option, in particular for patients with a short survival time. However, a considerable proportion of MSCC patients live longer than a few months and may live long enough to experience a recurrence of MSCC in the irradiated spinal region. The rate of such in-field recurrences is significantly higher after short-course radiotherapy than after longer radiation programs such as 10 × 3 grays (Gy) in 2 weeks or 20 × 2 Gy in 4 weeks.2, 3 A scoring system that allows one to estimate the survival of MSCC patients would help the physician select the appropriate radiation regimen for the individual patient (ie, single-fraction radiotherapy, multifraction short-course radiotherapy, or longer course radiotherapy). We have developed just such a scoring system based on a multivariate analysis of survival in 1852 MSCC patients.4 The scoring system included the 6 prognostic factors that were found to be significant for survival on the multivariate analysis. These factors were tumor type, interval between tumor diagnosis and MSCC, other bone metastases at the time of radiotherapy, visceral metastases at the time of radiotherapy, ambulatory status before radiotherapy, and duration of motor deficits before radiotherapy. The total score ranged between 20 and 45 points, and the patients were divided into 5 groups according to their score. The difference between the 5 groups was highly significant. However, this scoring system was not validated until this report. Furthermore, 5 groups appear too many to allow a quick use of the score during clinical routine, in particular if radiotherapy has to be administered urgently outside the regular working time.
The current study included 439 patients and aimed to validate our scoring system. Furthermore, a simplification of the previous scoring system has been performed by reducing the number of prognostic groups from 5 to 3.
MATERIALS AND METHODS
In the current study, 439 patients (265 patients from a previous prospective study3 plus 174 additional patients who have been prospectively followed since January 2008) were included. The patients had been treated with radiotherapy alone.
These 439 patients represent the validation group for the scoring system that has been previously developed to estimate the survival of MSCC patients.4 The scoring system to be validated was based on the retrospective analysis of 1852 patients. It included the following 6 prognostic factors that were found to be significantly associated with survival on the multivariate analysis of those 1852 patients: type of primary tumor (breast cancer vs prostate cancer vs myeloma/lymphoma vs lung cancer vs other tumors), interval between tumor diagnosis and MSCC (≤15 months vs >15 months), presence of other bone metastases at the time of radiotherapy, presence of visceral metastases at the time of radiotherapy, pretreatment ambulatory status (ambulatory vs nonambulatory), and time of developing motor deficits before radiotherapy (1-7 days vs 8-14 days vs >14 days) (Table 1). For each of these 6 prognostic factors, a separate score was calculated by dividing the 6-month survival rate by 10. The total score included in the scoring system represents the sum of all the scores (rounded values) from the 6 prognostic factors. The total scores ranged between 20 and 45 points. Five groups were formed according to the total score based on the 6-month survival rates for each score: 20 to 25 points (Group A), 26 to 30 points (Group B), 31 to 35 points (Group C), 36 to 40 points (Group D), and 41 to 45 points (Group E). The 5 groups were compared for survival using the Kaplan-Meier method.5 The Kaplan-Meier curves were compared using the log-rank test. The difference was significant with a P < .0001.
| Prognostic Factor | Score |
|---|---|
| Type of primary tumor | |
| Breast cancer | 8 |
| Prostate cancer | 7 |
| Myeloma/lymphoma | 9 |
| Lung cancer | 3 |
| Other tumors | 4 |
| Other bone metastases at the time of RT | |
| Yes | 5 |
| No | 7 |
| Visceral metastases at the time of RT | |
| Yes | 2 |
| No | 8 |
| Interval from tumor diagnosis to MSCC, mo | |
| ≤15 | 4 |
| >15 | 7 |
| Ambulatory status before RT | |
| Ambulatory | 7 |
| Nonambulatory | 3 |
| Time of developing motor deficits before RT, d | |
| 1-7 | 3 |
| 8-14 | 6 |
| >14 | 8 |
- RT indicates radiotherapy; MSCC, metastatic spinal cord compression.
The scoring system, which was developed on the basis of the retrospective analysis of 1852 patients, was applied in the same way to the 439 patients in the current series. Five prognostic groups were formed in accordance with the previously developed scoring system.
To simplify the scoring system, the 5 prognostic groups were replaced by 3 groups: 20 to 30 points (Group I), 31 to 35 points (Group II), and 36 to 45 points (Group III). The results for both the preceding cohort and the present cohort were generated with the 3-group system to confirm its utility.
RESULTS
In the current study, the 6-month survival rates were 11% for the 97 patients with a total score of 20 to 25 points, 20% for those 140 patients with a score of 26 to 30 points, 48% for those 162 patients with a score of 31 to 35 points, 72% for those 141 patients with a score of 36 to 40 points, and 93% for those 112 patients with a score of 41 to 45 points (P lt;.0001) (Fig. 1). The 6-month survival rates in the preceding series of 1852 patients who formed the basis for developing the survival score were 4%, 11%, 48%, 87%, and 99%, respectively (P < .0001). The comparisons of each of the different prognostic groups (Groups A to E) of this prospective series with the prognostic groups (Groups A to E) of the preceding retrospective study did not reveal a significant difference. The P values were .45 for the comparison of both Groups A, .10 for the comparison of Groups B, .40 for the comparison of Groups C, .24 for the comparison of Groups D, and .57 for the comparison of Groups E, respectively.
Kaplan-Meier curves of the 5 groups from the current study are shown with respect to survival (Group A: 20-25 points; Group B: 26-30 points; Group C: 31-35 points; Group D: 36-40 points; and Group E: 41-45 points). Pts indicates patients.
After reduction of the number of prognostic groups in the current study, the 6-month survival rates of the 3 prognostic groups [20-30 points [Group I; N = 237], 31-35 points [Group II; N = 162], and 36-45 points [Group III; N = 253]) were 16%, 48% and 81%, respectively (P < .0001) (Fig. 2). The corresponding 6-month survival rates in the preceding study were 9%, 48%, and 93%, respectively (P < .0001). Again, the comparison of each of the different prognostic groups of both series did not reveal a significant difference. The P values were .16 for the comparison of both Groups I, .40 for the comparison of Groups II, and .15 for the comparison of Groups III.
Kaplan-Meier curves of the 3 newly designed groups of patients (Pts) in the current study are shown with respect to survival (Group I: 20-30 points; Group II: 31-35 points; and Group III: 36-45 points).
DISCUSSION
The duration of survival for patients with MSCC varies considerably. The majority of the patients have a short survival of only a few months, whereas other MSCC patients may live for years.2, 6 It is important in oncology to individualize cancer care to the needs of each patient. For MSCC patients with a poor survival prognosis, short-course radiotherapy with 1 to 5 fractions administered in a week or less is considered appropriate. Short-course radiotherapy is associated with less discomfort for the patients than longer course radiotherapy programs because of fewer trips to the radiation oncology department and fewer, mostly painful, positionings on the treatment couch. Furthermore, a shorter radiation regimen uses less of a patient's limited lifespan and reduces the cost of therapy.7 MSCC patients with a relatively favorable survival prognosis often live long enough to develop a recurrence of MSCC in the previously irradiated spinal region. Because longer course radiotherapy has been reported to be associated with fewer recurrences than short-course radiotherapy, longer course radiotherapy appears to be the better option for this subset of patients.2, 3 In addition, more prolonged, higher dose radiotherapy regimens were associated with significantly longer survival in patients with favorable prognostic factors (scores of ≥36 in our retrospective analysis).4 It is still not clear whether longer course radiotherapy with a radiation dose >30 Gy in 10 fractions or 40 Gy in 20 fractions may improve the patient's ambulatory function. Thus, randomized trials are required that investigate a potential benefit of such a dose escalation. Furthermore, randomized trials should be performed to define the role of high-precision radiotherapy techniques for the treatment of MSCC.
A scoring system that allows one to estimate the survival of MSCC patients can help to select the most appropriate radiation regimen for the individual patient. Such a scoring system including 5 prognostic groups has been developed on a large retrospective series of 1852 patients.4 However, that score has not yet been validated. The current study included 439 new MSCC patients. The 6-month survival rates in the current series were not found to be significantly different from those of the preceding study. This finding demonstrates the validity of the previously developed score. However, one may consider 5 prognostic groups too complicated for daily routine, in particular if the patient is presented to the radiation oncologist outside the regular working hours (ie, at night or on weekends). Therefore, we decided to simplify the scoring system by reducing the number of prognostic groups from 5 to 3. The 6-month survival rates of these groups have been determined in both series (ie, in the primary series of 1852 patients and in the current series of 439 patients). Again, the survival rates of the 3 prognostic groups were not found to be significantly different in both series, which demonstrated the validity also of the scoring system with the new grouping.
In conclusion, the previously developed scoring system proved valid in another series of patients. It appears simpler for daily clinical routine to use 3 instead of 5 prognostic groups. Patients in the prognostic Groups I (20-30 points) should be treated with short-course radiotherapy because their 6-month survival rates are low. Patients in the prognostic Group III (36-45 points) should receive longer course radiotherapy because they are at a higher risk to develop a recurrence of MSCC in the previously irradiated spinal region because they live significantly longer. In patients in the prognostic Group II, the treatment decision should be left to the treating physician who is able to assess other less tangible factors such as Karnofsky performance status and comorbidity. Because this score has been developed and validated in patients treated with radiotherapy alone, it should be used only in patients not receiving decompressive surgery.
CONFLICT OF INTEREST DISCLOSURES
The authors made no disclosures.
