High-precision Radiotherapy of Motor Deficits Due to Metastatic Spinal Cord Compression

This is a multinational, multicenter study (single arm) supplemented by a comparison to a historical, propensity score matched control group (superiority study).

A total of 44 patients (40 patients + 10% for potential drop-outs) who will receive modern high-precision radiotherapy (RT) treatment for the metastases on their vertebral bodies are planned to take part in the clinical study. The historical control group treated with 5x4 Gy of conventional RT between 2001 and 2016 consists of more than 500 patients The primary aim of this study is to investigate the local progression-free survival (LPFS) of metastatic spinal cord compression (MSCC) after 5x5 Gray (Gy) of high-precision RT and to demonstrate that this regimen results in significantly better 6-month LPFS than conventional RT with 5x4 Gy. For the high-precision RT volumetric modulated arc therapy (VMAT), stereotactic body radiotherapy (SBRT) is allowed for treatment of patients with involvement of only one vertebra, if the required constraints can be met. LPFS is defined as freedom from progression of motor deficits during RT and freedom from an in-field recurrence of MSCC (i.e. freedom from motor deficits due to a recurrence of MSCC in the previously irradiated parts of the spine) following RT. The LPFS rate will be assessed 6 months after the end of RT.

In patients treated with RT alone for MSCC, conventional RT with 10x3 Gy in 2 weeks results in similar motor function but significantly better LPFS than conventional RT with 5x4 Gy in 1 week. Since patients with MSCC are often significantly impaired, a RT regimen with an overall treatment time of only 1 week would be preferable if it resulted in similar LPFS as 10x3 Gy in 2 weeks. This may be achieved with 5x5 Gy in 1 week, since the equivalent dose in 2 Gy fractions (EQD2) with respect to tumor cell kill of 5x5 Gy and 10x3 Gy are similar. The tolerance dose of the spinal cord, 5x5 Gy can be safely administered with high-precision RT such as VMAT (or SBRT). Therefore, the present study investigates the LPFS after high-precision RT with 5x5 Gy in 1 week. To demonstrate superiority the patients of this study will be compared to a historical control group receiving conventional RT with 5x4 Gy in 1 week. If superiority regarding LPFS can be shown for high-precision RT with 5x5 Gy, patients with MSCC would benefit from this regimen, since they can achieve high LPFS rates with an RT regimen lasting only 1 week (5x5 Gy) instead of 2 weeks (10x3 Gy). This study aims to make a significant contribution to the most appropriate RT schedule for patients with MSCC.

In accordance with a previous study assessing local control of MSCC, the following patient characteristics will be recorded to allow adequate comparison with the historical, propensity-score matched control group:

• Age (2 groups, depending on median age)
• Gender
• Type of primary tumor (breast cancer vs. prostate cancer vs. myeloma/lymphoma vs. lung cancer vs. other tumors)
• Interval from tumor diagnosis to MSCC (≤15 months vs. >15 months)
• Number of involved vertebrae (1-2 vs. ≥3)
• Other bone metastases at the time of RT (no vs. yes)
• Visceral metastases at the time of RT (no vs. yes)
• Time developing motor deficits prior to RT (1-7 days vs. 8-14 days vs. >14 days)
• Ambulatory status prior to RT (no vs. yes)
• Eastern Cooperative Oncology Group (ECOG) performance score (1-2 vs. 3-4) Study arm: 5x5 Gy of high-precision RT in 1 week Historical control: 5x4 Gy of conventional RT in 1 week Follow-up directly and at 1, 3 and 6 months following RT RT is administered as high-precision RT with 25.0 Gy in 1 week, i.e. with 5.0 Gy per fraction on 5 days per week (representing an EQD2 of 43.8 Gy for radiation myelopathy). An EQD2 of 45 Gy is estimated to be associated with a risk of radiation-related myelopathy of 0.03% and is therefore considered safe. The clinical target volume (CTV) includes the vertebral and soft tissue tumor as seen on the planning computed tomography and diagnostic MR-imaging, the spinal canal, the width of the involved vertebrae, and half a vertebra above and below those vertebrae involved by MSCC. The planning target volume (PTV) should include the CTV plus 0.8 cm and should be covered by the 95%-isodose. The maximum relative dose allowed to the spinal cord is 101.5% of the prescribed dose (representing an EQD2 of 44.9 Gy for radiation myelopathy). This maximum dose is estimated to be associated with a risk of radiation-related of <0.03% and is, therefore, also considered safe. Both the EQD2 of the prescribed dose (41.7 Gy) and the EQD2 of the maximum dose (43.8 Gy) are well below the tolerance dose of bone. In accordance with the Quantitative Analyses of Normal Tissue Effects in the Clinic (QUANTEC) data, the mean doses EQD2 for esophagus, heart and lung must be <34 Gy, <26 Gy and ≤7 Gy, respectively. Taking into account a radiation regimen of five fractions, the corresponding mean doses per fraction are 4.5 Gy, 3.8 Gy and 1.54 Gy, respectively. MSCC may affect single or multiple spinal sites. All sites need to be treated with high-precision RT following the dose prescriptions and constraints given above.

It is recommended that the patients receive concomitant treatment with dexamethasone during the period of radiotherapy if indicated.

Quality assurance plan:

Monitoring: The Centre for Clinical Trials Lübeck will conduct clinical on-site monitoring at the German sites according to Good Clinical Practice and written standard operating procedures (SOPs) to ensure the patients' rights and safety as well as the reliability of trial results.

The frequency of monitoring visits per site will be defined depending on the recruitment rate and the quality of data.

Patient Registration:

The patients were assigned 2 code numbers, the number of the contributing center plus a patient identification number starting with 001 and chronologically ongoing.

Coordination and supervision of the inclusion of the patients will be performed by the Department of Radiation Oncology of the University of Lübeck, Germany.

Sample size calculation:

The primary goal of this study is to assess high-precision RT with 5x5 Gy in 1 week with respect to 6-month LPFS and to demonstrate that this rate is superior to conventional RT with 5x4 Gy with respect to LPFS of MSCC (hypothesis).

With respect to tumor cell kill, the EQD2 of 5x5 Gy is similar to the EQD2 of 10x3 Gy (31.3 Gy vs. 32.5 Gy) and higher than the EQD2 of 5x4 Gy (23.3 Gy) . In a previous prospective non-randomized study, the 6-month LPFS rates were 86% after longer-course RT and 67% after short-course RT, respectively (p=0.034). In that study, 95 of 117 patients (81%) in the longer-course RT group had received 10x3 Gy, and 91 of 114 patients (80%) in the short-course RT group had received 5x4 Gy. Thus, assuming - for the present study - that conventional RT with 5x4 Gy in fact results in a 6-month LPFS rate of 67%, an increase by roughly 20 percentage point is considered to be clinically minimal relevant and even appears to be realistic when applying high-precision RT with 5x5 Gy.

The sample size is chosen to obtain prospective data that can be interpreted on its own and to allow for comparison with historical data:

• A sample size of at least 40 eligible patients is needed to estimate the probability of LPFS at 6 month with adequate precision, based on the following assumptions:
• 6-month LPFS can be assumed to be 87%
• 6-month LPFS is to be estimated with a precision of plus/minus 20 percentage points expressed as the half length of the associated two-sided confidence interval with a confidence coefficient of 95%.
• The power -i.e. the probability to obtain this precision, should be at least 80%
• Confidence limits are used as the approach for sample size. Assuming that roughly 10% of enrolled patients will not be eligible for efficacy analysis due to early lost-to-follow-up or due to premature discontinuation of high-precision RT, a total of 44 patients should be enrolled in the prospective part of the trial.
• The aim of the confirmatory study is to compare the prospectively collected data with a historical, propensity-score matched cohort collected up to the time of data analysis. Assuming for simplicity and conservative power calculation that this comparison could be conducted with a simple Pearson-Chi-Square test using a two-sided significance level of 5% (10%), a power of 79% (86%) is reached, if 40 patients are treated with high-precision RT and roughly 400 patients of the historical control group qualify for Propensity-Score adjusted comparison and assuming that the expected 6-month LPFS are 87% and 67%, respectively. Taking into account that the more sophisticated propensity-score adjusted statistical analysis will increase statistical power, the power for treatment arm comparison reached with 40 eligible patients in the prospective part of the study can be assumed to be at least 80%.

Primary endpoint:

The primary endpoint (LPFS of MSCC) will be evaluated based on the sample size calculations. The evaluation will be performed in those patients, who are available for assessment of the primary endpoint and have received at least 80% of the planned RT dose. Due to the open nature of the study, a blind data review is not necessary. The safety population includes all patients who have received at least 1 fraction of high-precision RT.

LPFS is defined as freedom from progression of motor deficits during RT and freedom from an in-field recurrence of MSCC (i.e. freedom from motor deficits due to a recurrence of MSCC in the previously irradiated parts of the spine) following RT. Deterioration of motor deficits during RT will be counted as LPFS of 0 months. Freedom from an in-field recurrence following RT will be referenced from the last day of RT. LPFS rates will be calculated for each potential prognostic factor using the Kaplan-Meier method. Differences between the Kaplan-Meier curves will be calculated with the log-rank test (univariate analysis). Factors that achieve significance (p<0.05) or show a trend (p<0.06) on univariate analysis will be additionally included in a multivariate analysis (Cox proportional hazards model).

Secondary endpoints:

The comparison of the high-precision RT group and the conventional RT group with respect to their effect on motor function and sensory function (improvement, no further progression, deterioration) will be evaluated with the ordered logit model adjusted for propensity score, because the data for the impact of RT on motor and sensory function are ordinal (-1 = deterioration, 0 = no further progression, 1 = improvement). An improvement or deterioration of the motor function will be defined as a change of at least 1 point on a 5-point scale. Post-RT ambulatory and sphincter dysfunction rates will be compared between both groups with the Chi-square test stratified for propensity score. Effect of RT on motor function, post-RT ambulatory rates, sensory function and sphincter dysfunction will be evaluated directly after RT and at 1, 3 and 6 months following RT in those patients who are alive and available for assessment. P-values of <0.05 are considered significant.

Overall survival (OS) will be counted from the last day of RT. OS rates will be calculated with the Kaplan-Meier-method. Differences between Kaplan-Meier curves will be analyzed with the log-rank test. Again, p-values of <0.05 are considered significant. Factors that achieve significance or show a trend (p<0.06) on univariate analysis will be included in a multivariate analysis (Cox proportional hazards model). The analysis of OS will be conducted in the intent-to-treat population.

Data regarding pain, quality of life and toxicity will only be assessed in the study group.