|Home | About | Journals | Submit | Contact Us | Français|
This study investigated the clinical benefit of using hypofractionated stereotactic body radiotherapy (SBRT) to manage spinal metastases in patients with cancer and to reduce cancer-related symptoms.
Cancer patients (n=149) with mechanically stable, non–cord-compressing, spinal metastases (n=166) were treated by SBRT in a phase I/II study. Patients received a total dose of 27–30 Gy, typically in three fractions. Symptoms were measured repeatedly by the Brief Pain Inventory (BPI) and the M. D. Anderson Symptom Inventory (MDASI). The primary endpoint was to establish the safety, feasibility, and efficacy of using a CT-on-Rails or Trilogy Stereotactic Spine Radiation Therapy system to treat spinal and paraspinal tumors and to document pain relief and toxicity associated with such treatment. Symptom outcomes were estimated according to protocol using descriptive analysis and ordinal regression modeling. This is the final report for the completed enrollment and follow-up.
The median follow-up time was 15·9 (interquartile range 9·5–30·3) months and the mean was 20·9 (SD=17·1) months. The actuarial tumor progression-free survival rates at one year and two years post-SBRT were 80·5% and 72·4%, respectively. Patients reported significant MDASI pain reduction (p=0·00003) during the six months post-SBRT. Patients reporting no pain from bone metastases on the BPI increased from 39/149 (26·2%) before SBRT to 55/102 (53·9%) six months post-SBRT (p<0·0001). BPI pain reduction from baseline to four weeks post-SBRT was clinically meaningful (effect size=0·47, p<0·01). These improvements were accompanied by significant reduction in opioid use during the six months post-SBRT (p<0·05) and a significant reduction in MDASI symptom interference with daily life (p<0·01).. Only a few instances of nonneurological grade 3 toxicities occurred (one report each of nausea, vomiting, diarrhea, fatigue, dysphagia, neck pain, diaphoresis, two reports of pain associated with severe tongue edema and trismus, and 3 reports of noncardiac chest pain). No grade 4 toxicities occurred.
SBRT is an effective primary or salvage treatment of mechanically stable spinal metastasis. Significant reduction in patient-reported pain and other symptoms was evident six months post-SBRT, along with satisfactory progression-free survival and no late spinal cord toxicities.
Almost forty percent of all patients with cancer develop spinal metastases during the course of their disease.1, 2 Inadequately treated spinal metastases may lead to pain and neurological complications, including metastatic epidural spinal cord compression. As a result, patients may experience severe symptom burden and diminished health-related quality of life (HRQOL).3–5
Palliative radiotherapy effectively controls pain for patients with spinal metastases4; however, higher-dose radiotherapy may be required for durable tumor control and for prevention of bony destruction of the spinal column, which results in spinal instability. The spinal cord’s sensitivity to radiation generally precludes high radiation doses to the spine or its reirradiation using conventional techniques.6 Accordingly, new techniques have been developed to optimize radiation dose delivery to the bone metastasis while sparing the spinal cord. Stereotactic body radiotherapy (SBRT), an emerging technique, utilizes image guidance to deliver high-dose radiation precisely, creating a steep dose gradient at the interface between spinal cord and tumor. This approach increases the therapeutic window by lowering risk for spinal cord myelopathy.6–8 Delivered in high doses and one to five fractions, spinal SBRT is available on various platforms, some of which include computed tomography (CT)-image guided stereotaxy. SBRT can be used in combination with or in lieu of surgery and allows patients to avoid possible perioperative risk factors, such as general anesthesia, bleeding, infection, or hospitalization.
Patients with late-stage cancer who are considering various therapy options often are not informed about the symptom-reduction benefits associated with a treatment. In a preliminary report of a prospective phase I/II trial of SBRT, we detailed the safety, efficacy, and patterns of failure for SBRT using results from a subset of patients (n=63) with spinal metastases who were followed for up to 50 months.2 In the present analysis of the entire patient cohort, we investigated the symptom reduction benefit of spinal SBRT during the six months posttreatment, and clinical benefit for up to two years. The primary endpoint was to establish the safety, feasibility, and efficacy of using a CT-on-Rails or Trilogy Stereotactic Spine Radiation Therapy system to treat spinal and paraspinal tumors and to document pain relief and toxicity associated with such treatment. We hypothesized that, for patients with mechanically stable spinal metastases, SBRT would be a clinically effective therapy for both tumor control (evidenced by radiographic depiction of tumor progression) and symptomatic improvement (evidenced by patient-reported outcomes).
The phase I/II trial was approved by the institutional review board of The University of Texas MD Anderson Cancer Center. Patients were accrued between November 6, 2002 and January 20, 2011, and all provided written informed consent before enrolling. Eligibility requirements included a diagnosis of cancer (excluding multiple myeloma), a Karnofsky performance status score of ≥40, and a magnetic resonance imaging (MRI) scan documenting spinal or paraspinal metastasis within four weeks of enrollment. Acceptable indications included oligometastatic disease arising from a known primary tumor, failure of previous conventional external beam radiotherapy or surgery, residual tumor after surgery, medical inoperability, and refusal to undergo surgery. A maximum of two distinct noncontiguous spinal metastases were allowed. Paraspinal tumors along the cervical, thoracic, or lumbar spine were included. The tumor could involve the vertebral column, but did not have to, nor did it need to enter the spinal canal. Patients receiving bisphosphonates or hormonal therapy were not excluded. Patients with mechanically unstable spine or epidural spinal cord compression were excluded; however, patients with previously documented spinal cord compression that had been decompressed and stabilized were eligible for enrollment. Patients were excluded if they had a pacemaker, were unable to undergo MRI, or had received systemic radiotherapy (strontium 89) or cytotoxic chemotherapy within 30 days of enrollment or spinal external beam radiotherapy within three months of enrollment.
We repeatedly measured pain at the metastatic sites treated with SBRT via the Brief Pain Inventory (BPI).9 The BPI assesses pain right now and at its worst, its least, and on average in the past 24 hours, on a 0–10 scale. At the same time, we also measured general symptom burden via the M. D. Anderson Symptom Inventory (MDASI).10 The MDASI assesses 13 common cancer-related symptoms, including pain, and six symptom interference items, each rated on a 0–10 scale within a 24-hour recall period. A composite interference score was computed as the mean of all six MDASI interference items for all patients. For the MDASI assessment in this study, patients were instructed to rate their pain, but not their specific pain at the spine site. The BPI and MDASI are well validated in patients with various types of cancer.9, 10 The Medical Outcomes Study 12-Item Short-Form Health Survey (SF-12) was administered as an HRQOL measure.11
Patient-reported symptom outcomes (BPI, MDASI, and SF-12) were collected in the clinic pre-SBRT (baseline) and at the three-month and six-month post-SBRT assessments; assessments at two weeks, four weeks, and two months were completed by the patient at home and returned by U.S. mail, with a reminder call from a study nurse.
MRI scans of the region treated were performed at three, six, nine, 12, 18, and 24 months post-SBRT and then every six months thereafter, as standard care. Lesions were classified as “progressive” if larger than at the previous assessment"stable” if unchanged, or “smaller” by radiologists who were central nervous system specialists. These assessments were not time blinded. The radiologists’ reports were subsequently discussed by a multidisciplinary spine tumor board.
History, neurological exam results, and McCormick functional classification12 were obtained at baseline and at each follow-up visit. Patient clinical data, including age, sex, tumor volume, and diagnosis, were collected prospectively upon enrollment. Baseline and longitudinal pain medication, Karnofsky performance status, and metastatic tumor evaluation post-SBRT also were recorded. Opioid use was documented as morphine equivalents.
All patients underwent intensity-modulated, near-simultaneous, CT-guided SBRT using a BodyFix stereotactic body frame immobilization system (Elekta), consisting of a whole-body vacuum cushion, carbon fiber base plate, and plastic fixation sheet. Stereotactic localization and target-positioning frames were used (Integra-Radionics). Patients received a total dose of 27–30 Gy, delivered to most patients in three fractions given every other day, with 10-Gy radiation volume received by the spinal cord limited to 0·01 cc. Gross target volume encompassed the lesion as visualized on the pretreatment CT scan. The clinical target volume encompassed the gross target volume and surrounding vertebral body (including superior and inferior endplates and any existing paraspinal component), along with all additional spinal structures deemed to be at risk for recurrence, such as the pedicle, lamina, and posterior elements. In patients with postsurgical metallic artifacts near the area of interest, intrathecal contrast injection with iohexol (Omnipaque, Amersham Health) was performed 30–60 minutes before CT image acquisition to assist with accurate spinal cord delineation. Spinal MRI was performed within four weeks of study enrollment and every three months thereafter. Baseline MRI was fused in many instances to assist in target delineation.
Descriptive statistics—mean, median, standard deviation (SD), and proportions—are used to describe patient and clinical characteristics. The patient subject, not the tumor, is the unit of analysis. All reported p values are two-tailed and considered significant if <0·05 for univariate testing. To account for multiple comparisons in the symptom and HRQOL outcomes, which required eight modelings, we adjusted the individual type I error to be 0·05/8=0·006 to maintain a conservative family-wise error rate of 0·05.
Using pain cutpoints established by Serlin et al.,13 we categorized ratings of the BPI “pain at its worst” item as “no pain” (0 on the BPI’s 0–10 scale), “mild pain” (1–4), “moderate pain” (5–6), and “severe pain” (7–10). We tracked proportions of patients in each of these categories over time during the study. Concordance between BPI “pain at its worst” ratings and MDASI “pain at its worst” ratings, both scored on a 0–10 scale in the past 24 hours, was examined using paired t-tests. Effect sizes were calculated to estimate the magnitude of change in BPI and MDASI ratings between baseline and four weeks posttreatment.14, 15 For patient-reported outcomes measures, effect sizes are clinically meaningful at approximately one-half SD or higher, the level often used in distribution-based methods of determining meaningful differences.16
Lowess curves,17 which represent a smoothed estimate of average MDASI symptom severity and interference as a function of time, were constructed from baseline to six months post-SBRT.
Ordinal regression models18 and generalized linear mixed models were fitted to examine symptom development trends for the five most-severe MDASI symptoms and the symptom-interference component score from baseline to six months post-SBRT. Individual symptom scores were treated as ordinal responses. Independent variables included weeks from start of therapy, age, sex, tumor volume of spinal metastasis at baseline, type of primary cancer, disease progression status at the sixth month post-SBRT based on radiographic (spinal MRI) results, opioid use, and Karnofsky performance status at baseline.
Actuarial freedom-from-tumor-progression and survival curves from date of enrollment were generated using the Kaplan-Meier method. Patient survival information was obtained from a retrospective medicalrecords review. Toxicity was graded by the patient’s treatment team according to the National Cancer Institute Common Toxicity Criteria for Adverse Events, version 2·0.19
Statistical analysis was conducted using Statistical Package of the Social Sciences version 17.0 (SPSS, Chicago, IL, USA) and SAS version 9.2 (SAS Institute, Cary, NC, USA).
This study is registered with ClinicalTrials.gov, number NCT00508443. The researchers from the Department of Symptom Research are partially funded by a grant (R01 CA026582, Reducing the Symptom Burden Produced by Aggressive Cancer Therapies; PI: Charles S. Cleeland, PhD) from the National Cancer Institute of the National Institutes of Health. Neither the NCI nor the NIH had any role in the study design, data collection, analysis, interpretation, or preparation of the report. The authors were responsible for the design of the trial. XSW had final responsibility for the decision to submit for publication.
Patient demographic and clinical characteristics are given in table 1. Of the 184 patients approached and consented, 35 did not provide symptom data and thus became inevaluable. The remaining 149 patients with 166 spinal metastases at cervical, thoracic, or lumbar vertebral levels were included in this analysis. Seventeen of 149 patients had two distinct spinal metastasis sites treated in the same session, and 34/149 (22·8%) were receiving bisphosphonates at enrollment. Spinal MRIs were performed for 142/149 patients (95·3%) at the six-month follow-up.
At the time of analysis, 40/149 patients (26·8% of the sample) were still living, with a median follow-up time of 15·9 (range 1·0–91·6; interquartile range 9·5–30·3]) months and mean 20·9 (SD=17·1) months. The median overall survival time was 23 months post-SBRT (95% CI 18·6–27·2), with one-year and two-year actuarial survival rates of 71·9% and 48·8%, respectively. Tumor progression was seen in 41/149 patients (27·5%) and occurred at a median 13 (range <1–101) months, based on MRI scans. The actuarial tumor progression-free survival rates based on MRI scans at six months, one year, and two years post-SBRT were 86·1% (95% CI 79·4%–90·7%), 80·5% (95% CI 72·9%–86·1%), and 72·4% (95% CI 63 ·1%–79·7%), respectively.
The prevalence of spinal pain according to BPI “pain at its worst” ratings is presented in figure 1. We observed significant reductions in the severity of patient-reported pain between baseline and four weeks posttreatment (from 3·4 to 2·1 on the BPI’s “pain at its worst” item on a 0–10 scale; effect size=0·47, p<0·001), and between baseline and six months post-SBRT (from 3·4 to 1·7; effect size=0·64, p<0·001). The proportion of patients reporting no spine pain on the BPI increased significantly between baseline and four weeks post-SBRT, from 39/149 (26·2%) to 43/109 (39·4%) (p<0·05). This improvement continued throughout the study, with 53/120 (44·2%) reporting no pain at three months (p<0·05) and 55/102 (53·9%) reporting no pain at six months (p<0·0001). Further, a significant decrease in the percentage of patients with moderate-to-severe BPI spine pain (rated ≥5 on the 0–10 scale) was noted from baseline to four weeks (p<0·01), two months (p<0·0001), and six months (p<0·01) post-SBRT.
No differences in mean pain severity between BPI (metastatic bone pain) and MDASI (general pain) ratings were found at any time point other than the six-month assessment (significant at p=0·022; table 2). We noted significant reduction in opioid use from baseline to three months and baseline to six months post-SBRT (both p<0·05).
During the six months of observation, the five most-severe MDASI symptoms were fatigue, pain, disturbed sleep, drowsiness, and distress. The average severity of these symptoms over time is illustrated in figure 2 using Lowess curves, with the week of SBRT completion shown as week 0. We observed clinically meaningful reductions in MDASI pain ratings between baseline and four weeks posttreatment (from 3·4 to 2·1 on the MDASI’s 0–10 scale; effect size 0·47). The Lowess curves in figure 2 also illustrate that symptom interference lessened over time.
Table 3 presents p-values from the ordinal regression modeling of MDASI symptom severity and SF-12 physical and mental health component scores, adjusted for week from start of therapy, patient’s age, sex, baseline tumor volume of spinal metastasis, type of cancer, disease progression status, opioid use over time, and baseline Karnofsky performance status; columns present results for each symptom model and rows present the independent variables. Patients reported significant pain reduction (p=0·00003) six months post-SBRT, accompanied by significant reduction in multiple symptoms, including disturbed sleep, drowsiness, sadness (all p<0·0001), fatigue, distress, lack of appetite, nausea, and difficulty remembering (all p<0·05).
Patients whose lesions were categorized as progressive at the six-month follow-up examination (19/149; 12·8%) also reported significantly more-severe MDASI pain (p<0·0001), fatigue (p=0·01), and drowsiness (p=0·00008) than did patients with stable or smaller lesions. Patients who had received opioids during the six months post-SBRT reported more severe MDASI pain, fatigue (all p<0·0001), disturbed sleep, distress, and drowsiness (all p<0·001) than did patients not using opioids. Ordinal regression modeling showed that a composite score of all six interference items decreased significantly at each successive assessment during the six months post-SBRT (p<0·0001).
Mild toxicities were documented during the study, including grade 1 and 2 transient numbness and tingling, nausea, and vomiting. Grade 3 toxicities manifested as nausea (n=1), vomiting (n=1), diarrhea (n=1), fatigue (n=1), noncardiac chest pain (n=3), dysphagia (n=1), neck pain (n=1), diaphoresis (n=1), and pain associated with severe tongue edema and trismus (n=2). No grade 4 toxicities were reported, and we observed no radiation-related spinal cord myelopathy during the study.
This study incorporated validated single-symptom (BPI) and multisymptom (MDASI) assessment tools to measure patient-reported outcomes for pain and other symptoms in patients with metastatic spine lesions treated with SBRT. Complementary to our previous publication documenting the safety, effectiveness, and patterns of failure of spinal SBRT,2 we here demonstrate significant reduction in the severity of pain and consistent reduction in other patient-reported symptoms and symptom interference six months after spinal SBRT, along with satisfactory progression-free survival and no late spinal cord toxicities.
In a retrospective review, Sheehan et al.20 found that pain was the most common presenting symptom from spinal metastases. Our results indicate that SBRT is an effective treatment for metastatic spinal pain in patients with late-stage cancer. Between baseline and four weeks post-SBRT, we observed medium, but clinically meaningful, effect sizes for pain reduction as reported on both the BPI “pain at its worst” and MDASI pain items, along with significant increase in the number of patients reporting complete pain relief as early as four weeks post-SBRT. Significant improvement in effect size and BPI pain ratings relative to pre-SBRT levels appeared even stronger at six months (effect size=0·64, p<0·001). This result is not surprising, because six months post-SBRT, only two patients had tumor progression, high pain severity, and were receiving opioid therapy. The effectiveness of SBRT for both tumor control and pain control was further evidenced by a reduction over time in the use of strong opioids, a standard of care for managing severe pain.
From a methodological viewpoint, patient-reported data from two instruments using the same 0–10 pain-severity rating scale and 24-hour recall period (the BPI and the MDASI, see table 2) were not generally significantly different in this cohort of patients with advanced cancer. This result suggests that researchers can use either scale in clinical studies when pain at its worst is the outcome of interest and the same type of pain is expected. Even so, the MDASI measures a broader array of critical symptoms and has various modules tailored to specific treatments and types of cancer.
This study demonstrates that pain reduction and functional improvement may also be reflected in the reduction of associated symptoms.21 Using a sensitive multiple-symptom assessment tool (the MDASI) with a highly selective but comprehensive set of symptom items,22, 23 we were able not only to prospectively identify pain and other major symptoms in a cohort of patients who received spinal SBRT, but also to illustrate how multiple symptoms improved over time after treatment. Frequent measurement allowed us to capture the quick response to SBRT and its durability over time. The significant reduction in the severity of multiple symptoms besides metastatic pain suggests that spinal SBRT therapy produces minimal symptom burden and toxicity for patients with late-stage cancer. Renal cell carcinoma metastatic tumors to spine are among the most difficult to control and therefore are most frequently referred for spinal SBRT. We observed no significant difference between renal and other cancers on any MDASI symptom.
Fatigue was consistently the most severe symptom over time, potentially as a result of both advanced disease and continuous use of opioids. Although fatigue showed a trend of improvement over the six months post-SBRT (p=0·037), the physical and mental health component scores of the SF-12 remained essentially constant in follow-up. These results are consistent with those of Degen et al,24 who observed significant pain reduction four weeks after spinal SBRT that was durable to one year, but no statistically significant change in physical or mental well-being data from the SF-12. Baseline Karnofsky performance status was significantly related to total interference and physical well-being data from the SF-12; see table 3.
The efficacy and safety of SBRT are supported by the study’s structured schedule with defined assessment intervals and follow-up serial spinal MRIs, which permitted close evaluation of the procedure. MRI scans, obtained for 95·3% of study participants, revealed a progression-free survival rate of 80·4% one year post-SBRT. This result is comparable with results reported by other investigators8 and indicates that spinal SBRT is an efficacious primary or salvage treatment of metastatic tumors of the spine. In this study, SBRT produced no radiation-related spinal cord myelopathy; the numbness reported by some patients was likely caused by pre-SBRT chemotherapy. Only a few instances of non-neurological grade 3 toxicities (nausea, vomiting, diarrhea, fatigue) and no grade 4 toxicities occurred.
Patient compliance and missing data can be an issue in prospective studies. One of the benefits of ordinal regression modeling is that it better handles random missing data from a longitudinal study. In the current study, 24 patients did not return their paper-and-pencil symptom assessment results to us by mail at the two-week assessment time point; nonetheless, most of them contributed symptom data at subsequent time points. Having study staff conduct symptom surveys over the phone is one possibility for reducing the rate of random missing patient-reported outcomes assessments.
One limitation of the study is the lack of a control arm against which to measure the effect of SBRT on symptom development. However, it is well accepted that SBRT has a quantifiable clinical effect on tumor growth with an accompanying reduction in pain at the radiation site, as was evidenced in our previous report in a subset of this cohort of patients with stage IV cancer.2 One strength of this study is that multiple symptoms were assessed simultaneously and longitudinally for each patient and compared with baseline, with each patient serving as his or her own control. Tumor progression six months post-SBRT was significantly correlated with more-severe pain, as expected, suggesting a true (non-placebo) palliative effect. An even higher level of evidence could be provided by a randomized controlled trial comparing stereotactic radiation with conventional radiation. Such a trial is currently ongoing with the Radiation Therapy Oncology Group 06-31 study.
In reviewing patient records, we found that 16 patients were positive for adrenal metastasis and 12 were positive for brain metastasis at enrollment. For this reason, although the protocol designated data collection up to 24 months, in the current study we did not use patient-reported outcomes data beyond six months, after which the symptom-reduction benefit from SBRT could be confounded by increased pain from rapid disease progression at or near end of life in this patient cohort with very advanced cancer. Further study of this data is warranted to determine the entire profile of pain and other symptoms, from beyond six months post-SBRT to near the time of patient death.
The role of SBRT in treating mechanically stable spinal metastases without spinal cord compression is continuing to evolve in an era in which new technologies and treatments are being highly scrutinized. Most patients with spinal metastases still can benefit from conventional palliative radiation therapy. Surgery or vertebral augmentation with cement should be considered for stabilizing patients with mechanically unstable spines before proceeding to radiation therapy. Nonetheless, the current study provides additional data that support the clinical benefit of SBRT for carefully selected patients and suggests that SBRT reliably halts the progression of disease, reduces patient symptoms, and leads to improved functioning in daily life—thus demonstrating both symptomatic and clinical benefit. This work also highlights the importance of integrating patient-reported symptom assessments with clinical outcome evaluations to fully demonstrate the benefit of SBRT in patients with metastatic spinal disease.
The authors acknowledge contributions of Patricia Grossman, RN Department of Radiation Oncology who coordinated the study and data collection. Denise Leblanc, Department of Radiation Oncology collected data for and maintained the study database. Jeanie F. Woodruff, ELS, Department of Symptom Research at MD Anderson, provided medical editing, for which she received no additional compensation other than her salary.
The researchers from the Department of Symptom Research are partially funded by a grant (R01 CA026582, Reducing the Symptom Burden Produced by Aggressive Cancer Therapies; PI: Charles S. Cleeland, PhD) from the National Cancer Institute of the National Institutes of Health. Neither the NCI nor the NIH had any role in the study design, data collection, analysis, interpretation, or preparation of the report.
Funding support was provided by CSC.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
This work was partly presented at the plenary session of the 77th Annual Meeting of the American Association of Neurological Surgeons in San Diego, CA, May 6, 2009.
ContributorsXSW, LDR, ASS, JNY, SSA, CSC, and ELC conceived and designed the study. XSW and IG performed the literature search. IG, PKA, HJS, and ELC collected the data. Data analysis and interpretation was performed by XSW, LDR, IG, PL, PKA, US, CSC, CLC, HJS, DCW and ELC. XSW, US, CSC, LDR, HJS and ELC wrote the manuscript.
Conflicts of interest
LDR has received teaching honoraria from Medtronic and Stryker. The remaining authors declare no conflicts of interest. The corresponding author states that she had full access to all the data in the study and had final responsibility for the decision to submit for publication.
Xin Shelley Wang, Department of Symptom Research, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Boulevard, Unit 1450, Houston, TX 77030.
Laurence D. Rhines, Department of Neurosurgery, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Boulevard, Unit 442, Houston, TX 77030.
Almon S. Shiu, Department of Radiation Physics, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Boulevard, Unit 94, Houston, TX 77030.
James N. Yang, Department of Radiation Physics, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Boulevard, Unit 94, Houston, TX 77030.
Ugur Selek, Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Boulevard, Unit 97, Houston, TX 77030.
Ibrahima Gning, Department of Symptom Research, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Boulevard, Unit 1450, Houston, TX 77030.
Ping Liu, Department of Biostatistics, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Boulevard, Unit 1411, Houston, TX 77030.
Pamela K. Allen, Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Boulevard, Unit 1202, Houston, TX 77030.
Syed S. Azeem, Department of Neurosurgery, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Boulevard, Unit 442, Houston, TX 77030.
Paul D. Brown, Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Boulevard, Unit 0097, Houston, TX 77030.
Hadley J. Sharp, Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Boulevard, Unit 97, Houston, TX 77030.
David C. Weksberg, Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Boulevard, Unit 97, Houston, TX 77030.
Charles S. Cleeland, Department of Symptom Research, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Boulevard, Unit 1450, Houston, TX 77030.
Eric L. Chang, Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Boulevard, Unit 97, Houston, TX 77030.