Stereotactic Radiosurgery in Treating Patients With Greater Than 3 Melanoma Brain Metastases

A Phase II Trial to Determine Local Control and Neurocognitive Preservation After Initial Treatment With Stereotactic Radiosurgery (SRS) for Patients With >3 Melanoma Brain Metastases

This phase II trial studies how well stereotactic radiosurgery works in treating patients with melanoma that has spread to more than 3 places in the brain. Stereotactic radiosurgery is a specialized radiation therapy that delivers a single, high dose of radiation directly to the tumor and may cause less damage to normal tissue.

PRIMARY OBJECTIVES: I. To determine local control of brain metastases at 4 months after initial treatment with stereotactic radiosurgery (SRS) in patients with > 3 melanoma brain metastases (MBM). II. To determine cognitive decline at 4 months defined as a significant decline (>= 5 point decrease from baseline based on the reliable change index) in the Hopkins Verbal Learning Test-Revised (HVLT-R) Total Recall after initial treatment with SRS versus whole brain radiation therapy (WBRT) in patients with > 3 MBMs. SECONDARY OBJECTIVES: I. To determine local tumor control and distal tumor control in the brain at 1, 4, 6, 9 and 12 months post-treatment. II. To determine overall survival in treated patients. III. To assess the pattern of neurocognitive change in memory at 1, 4, 6, 9, and 12 months post-treatment as well as executive function, attention, processing speed and upper extremity fine motor dexterity. IV. To evaluate composite neurocognitive function scores in treated patients. V. To assess the pre-treatment factors of age, Karnofsky performance scale (KPS), extra-cranial disease, BRAF-V600E mutation status in the predictive determination of local and distal control and neurocognitive outcome in each treatment arm. VI. To assess the correlation between number of lesions and total volume of intracranial disease and neurocognitive outcome in each treatment arm. VII. To document post-treatment adverse side effects in treated patients. VIII. Evaluate the time to initiation of systemic therapy from completion of radiation treatment. IX. Evaluate the duration/number of cycles of systemic chemotherapy given following radiation treatment. CORRELATIVE STUDIES: I. To determine if apolipoprotein E (Apo E) (i.e., Apo E2, Apo E3, and Apo E4) genotyping may prove to be a predictor of radiation induced neurocognitive decline (or neuro-protection). II. To determine if inflammatory markers (i.e., IL-1, IL-6, and TNF-alpha) may prove to be predictors of radiation induced neurocognitive decline. III. To determine if hormone and growth factors (i.e., glucocorticoids [e.g., cortisol], gonadal steroids [e.g., estradiol, testosterone, progesterone], growth hormone, human chorionic gonadotropin (hCG), insulin-like growth factor-1 [IGF-1], and neuronal growth factor [NGF]) may prove to be a predictor of radiation induced neurocognitive decline. IV. To assess whether baseline and post-radiation fludeoxyglucose F-18 (FDG)-positron emission tomography (PET)/computed tomography (CT) scans can predict for neurocognitive decline. OUTLINE: Patients undergo SRS on day 1. After completion of study treatment, patients are followed up for 12 months.

Clinical Stage IV Cutaneous Melanoma AJCC v8, Metastatic Malignant Neoplasm in the Brain, Metastatic Melanoma, Pathologic Stage IV Cutaneous Melanoma AJCC v8

Stereotactic External Beam Irradiation, stereotactic external-beam radiation therapy, stereotactic radiation therapy, Stereotactic Radiotherapy, stereotaxic radiation therapy, stereotaxic radiosurgery

Time to local failure will be estimated using the product-limit estimator of Kaplan and Meier, and log-rank test will be used for comparison of local failure rate in patients treated with stereotactic radiosurgery (SRS) to null hypothesis with respect to the time to local failure. Patients who are lost to follow-up or who die from distant disease before having local failure will be censored. Local control rates at 4 months may be estimated with 95% confidence intervals using Kaplan-Meier method.

The baseline Listening Vocabulary Levels Test-Revised (HVLT-R) score will be compared to the HVLT-R score in patients surviving 4 months. Preservation of function is defined as improvement of HVLT-R score or decline by 4 points or less. Failure is defined as decline by 5 or more points. Time to neurocognitive decline will be estimated using the product-limit estimator of Kaplan and Meier, and log-rank test will be used for comparison of neurocognitive decline rate in patients treated with SRS to null hypothesis with respect to the time to neurocognitive decline. Patients who are lost to follow-up or who die before having neurocognitive decline will be censored. Rates of neurocognitive decline at 4 months may be estimated with 95% confidence intervals using Kaplan-Meier method.

Will be estimated using the product-limit estimator of Kaplan and Meier. Cox proportional hazards regression will be used to model overall survival as a function of age, Karnofsky performance status, extra-cranial disease, and BRAF mutation status. Will model time to local failure, time to distal failure, and time to neurocognitive decline using competing risk regression when death without events is considered as a competing risk.

Will use descriptive statistics and boxplots to summarize and illustrate the neurocognitive function score at each assessment time. Will similarly summarize and illustrate the change from baseline in neurocognitive function score. Will also fit the neurocognitive data with a general linear model including the baseline score as covariates to assess differences in neurocognitive scores over time (to 4 months) for those patients that are alive and progression-free at 4 months. We will also model the data with mixed effects regression including baseline HVLT-R, time, number of lesions, extra-cranial disease, and a patient specific random effect. Will use logistic regression methods to model the logit of the probability of neurocognitive decline as function of ApoE (i.e., Apo E2, Apo E3, Apo E4) genotyping, inflammatory markers, hormone growth factors.

Will use descriptive statistics to summarize the number of cycles of systemic chemotherapy given following radiation treatment for each treatment arm.

Inclusion Criteria: All patients with histologic proof of malignant melanoma. Histologic confirmation may be from the primary tumor site, or from another metastatic site (systemic lymph node, etc). Cytology-alone is not an acceptable method of diagnosis Greater than 3 presumed melanoma brain metastases on contrast-enhanced brain MRI scan obtained no greater than 4 weeks prior to study registration Patients must sign informed consent indicating that they are aware of the investigational nature of this study in keeping with the policies of the hospital Patients must have Karnofsky performance status (KPS) >= 70 Patients must be eligible to have all lesions treated as determined by the study radiation oncologist Creatinine clearance > 30 ml/min Platelets > 50,000 Patients should have normal coagulation (international normalized ratio [INR] < 1.3) and be able to withhold anticoagulation/antiplatelet medications a minimum of 24 hours prior to radiosurgery treatment (or until INR normalizes), on the day of treatment and 24 hours after radiosurgery treatment has concluded Patients can be undergoing concurrent systemic therapy, such as temozolomide, at the discretion of their treating oncologist Exclusion Criteria: Patients are excluded if they have been treated with whole brain radiotherapy within the prior 3 months Patients are excluded if they have a history of metastatic cancer in addition to melanoma or a history of uncontrolled non-metastatic cancer. Patients with localized squamous cell carcinoma and/or basal cell carcinoma are not excluded Patients are excluded if there is radiographic or cerebrospinal fluid (CSF) evidence of leptomeningeal disease Female patients of childbearing age are excluded if they are pregnant as determined with a serum beta HCG no greater than 14 days prior to study registration, or breast-feeding. (The exclusion is made because gadolinium may be teratogenic in pregnancy) Patients are excluded if there is any history of gadolinium allergy Patients are excluded if they are unable to obtain a magnetic resonance imaging (MRI) scan for any other reason