Institutional members access full text with Ovid®

Using Higher Isodose Lines for Gamma Knife Treatment of 1 to 3 Brain Metastases Is Safe and Effective

Shiue, Kevin BS*; Barnett, Gene H. MD, MBA‡,§; Suh, John H. MD‡,¶; Vogelbaum, Michael A. MD, PhD*,‡; Reddy, Chandana A. MS; Weil, Robert J. MD‡,§; Angelov, Lilyana MD‡,§; Neyman, Gennady PhD; Chao, Samuel T. MD‡,¶

doi: 10.1227/NEU.0000000000000289
Research-Human-Clinical Studies

BACKGROUND: Higher isodose lines (IDLs) in Gamma Knife (GK) Perfexion treatment of brain metastases (BMet) could result in lower local control (LC) or higher radiation necrosis (RN) rates, but reduce treatment time.

OBJECTIVE: To assess the impact of the heterogeneity index (HI) and conformality index (CFI) ion local failure (LF) for patients treated with GK for 1 to 3 BMet.

METHODS: From an institutional review board—approved database, 320 patients with 496 BMet were identified, treated for 1 to 3 BMet from July 2007 to April 2011 on GK Perfexion. Cox proportional hazards regression was used to analyze significance of HI, CFI, IDL, dose, tumor diameter, recursive partitioning analysis class, tumor radioresistance, primary, smoking history, metastasis location, and whole-brain radiation therapy (WBRT) history with LF and RN.

RESULTS: Median follow-up by lesion was 6.8 months (range, 0-49.6). The series median survival was 14.2 months. Per RECIST, 9.5% of lesions failed, 33.9% were stable, 38.3% partially responded, 17.1% responded completely, and 1.2% could not be assessed. The 12-month LC rate was 87.3%. On univariate analysis, a dose less than 20 Gy (hazard ratio [HR]: 2.940, P < .001); tumor size (HR: 1.674, P < .001); and cerebellum/brainstem location vs other (HR: 1.891, P = .043) were significant for LF. Non-small cell lung cancer (HR: 0.333, P = .0097) was associated with better LC. On multivariate analysis, tumor size (HR: 1.696, P < .001) and cerebellum/brainstem location vs other (HR: 1.959, P = .033) remained significant for LF. Variables not significant for LF included CI, IDL, and HI.

CONCLUSION: Our study of patients with 1 to 3 BMet treated with GK demonstrated no difference in LC or RN with varying HI, indicating that physicians can treat to IDL at 70% or higher IDL to reduce treatment time without increased LF or RN.

ABBREVIATIONS: BMet, brain metastases

CFI, conformality index

CI, confidence interval

CR, complete response

GK, Gamma Knife

HI, heterogeneity index

HR, hazard ratio

IDL, isodose line

LC, local control

LF, local failure

MVA, multivariate analysis

NSCLC, non-small cell lung cancer

PR, partial response

RECIST, Response Evaluation Criteria in Solid Tumors

RN, radiation necrosis

RTOG, Radiation Therapy Oncology Group

SD, stable disease

SRS, stereotactic radiosurgery

UVA, univariate analysis

*Case Western Reserve University School of Medicine, Cleveland, Ohio;

Rose Ella Burkhardt Brain Tumor and Neuro-Oncology Center;

§Department of Neurosurgery, Neurological Institute;

Department of Radiation Oncology, Taussig Cancer Institute, Cleveland Clinic, Cleveland, Ohio

Correspondence: Samuel T. Chao, MD, Cleveland Clinic Lerner College of Medicine, Department of Radiation Oncology, Desk T28, Cleveland Clinic, 9500 Euclid Avenue, Cleveland, OH 44195. E-mail:

Received October 02, 2013

Accepted January 02, 2014

Copyright © by the Congress of Neurological Surgeons