Multivariate Analyses of Survival Time

1

• Gamma Knife Radiosurgery for Brain Metastases
  from Primary Breast Cancer
• Norbert Kased, Devin K. Binder, David A.
  Larson,, Jean L. Nakamura, Kim Huang,
  Mitchel S. Berger,
• Michael W. McDermott, Penny K. Sneed,
• Department of Radiation Oncology, University of
  California, San Francisco
• Department of Neurological Surgery,
  University of California, San Francisco

• Purpose
• Stereotactic radiosurgery (SRS), with or without
  whole brain radiotherapy (WBRT), is commonly used
  to treat brain metastases from breast cancer.
• We reviewed our SRS experience with brain
  metastases from breast cancer at the University
  of California, San Francisco in order to
  determine how size and number of metastases and
  omission of WBRT affected median survival time
  (MST) and freedom from progression (FFP).

• No association was found between number of
  treated brain metastases and survival time in
  patients with newly-diagnosed (p 0.16) or
  recurrent brain metastases (p 0.11) (Table 2).

Figure 1 Day of SRS with 16-20 Gy to 7 mets
Figure 2 Response 5 mo. after SRS. Repeat GK
for 2 small new mets 3 yr later. Never had WBRT.
SRS -/ WBRT for SRS for newly-dxd brain
mets recurrent brain mets
Table 2

• The overall MSTs were 16.3 months among the 93
  newly-diagnosed patients and 11.7 months for the
  76 patients with recurrent brain metastases.
• Among patients with newly-diagnosed brain
  metastases, omission of upfront WBRT did not
  significantly affect MST (p 0.15 Table 1,
  Figure 3), brain FFP (p 0.83 Table 1, Figure
  4), or freedom from new brain metastases (p
  0.87) (Table 1).
• There was no association between number of brain
  metastases (1 vs 2 vs 3 vs 4-6 vs gt6) and brain
  FFP for patients treated with SRS alone initially
  (p 0.48) or for those treated with SRS
  upfront WBRT (p 0.27).

• Methods and Materials
• We retrospectively reviewed records of all
  patients treated with SRS from 1991-2005 to
  identify patients with brain metastases from
  breast cancer who underwent SRS with or without
  WBRT for newly-diagnosed or recurrent metastases.
  
• MST and FFP were calculated using the
  Kaplan-Meier method measuring survival from the
  date of diagnosis of brain metastases in patients
  with newly-diagnosed brain metastases. All other
  endpoints were measured from the date of SRS.
• Prognostic factors were evaluated with the
  log-rank test and Cox proportional hazards model.

On multivariate analysis,

• Longer survival time was significantly
  associated with smaller total target volume by
  quartile for newly-diagnosed (p 0.03) and
  recurrent brain metastases (p 0.001).
• Neither upfront WBRT (p 0.21 HR 1.36 favoring
  SRS alone) nor number of brain metastases (1 vs 2
  vs 3 vs 4-6 vs gt6) were prognostic factors for
  survival (p 0.43 or 0.85).

Table 1
Multivariate Analyses of Survival Time
Newly-dxd brain mets Recurrent brain mets
Parameter p-value Parameter p-value RPA
class 0.18 RPA class 0.09 Number of
mets 0.43 Number of mets 0.85 Total target
volume 0.03 Total target volume 0.001 SRS vs.
SRSWBRT 0.21 (1 vs 2 vs 3 vs 4-6 vs. gt6)
(by quartile)

• Results
• From 1991-2005, 176 patients underwent SRS for
  brain metastases from breast cancer.
• Excluding seven patients who had surgical
  resection of a brain metastasis prior to SRS, 93
  patients had SRS with or without WBRT for
  newly-diagnosed brain metastases and 76 patients
  had SRS for recurrent brain metastases after
  prior WBRT.
• The median patient age at the time of SRS was 50
  years (range, 25-83 years) and the median KPS was
  90 (50-100) 17 of patients were RPA class 1,
  79 class 2, and 5 class 3.

• Conclusions
• In this SRS series, the omission of WBRT did not
  seem to significantly affect MST or FFP in
  patients with newly-diagnosed brain metastases
  from primary breast cancer.
• MST was not affected by the number of
  newly-diagnosed or recurrent brain metastases
  treated.
• Although known prognostic factors were adjusted
  for using multivariate analysis, selection biases
  may have influenced results.

Figure 3
Figure 4
Survival time (months)
Brain FFP (months)