The first use of a radioactive substance for detect
ing brain tumors was carried out by Moore in 1948
(1 ). Fluorescein was labeled with radioiodine and
detected with a hand-held G-M counter. During the
past 20 years technological developments have led
to the use of o9mTc and automatic scanners and
cameras. But as yet, the reasons for the localization
of the radioactive compound in the lesion have not
been found (2—5).
This paper shows that vascularity of the neoplasm
is correlated with positive brain scans in the series
studied and probably is a factor in the unusual per
meabiity of the disrupted blood-brain barrier.
MATERIALS
Records of patients who had brain scans at the
Nuclear Medicine Unit of The University of Michi
gan were analyzed. Tissue sections were available
for review in 107 cases and these formed the basis
of the study. All the tissues were accessioned at the
University of Michigan Department of Pathology
from 1964 to 1967 and were obtained by craniotomy
following brain scanning except for one autopsy
case.
METHODS
Scan reading. The brain scans were done with four
commercially available scanners (Picker Magna
scanners III and V and Ohio-Nuclear Models 54H
and 54F) . The agent used was ‘°THg-chlormerodrin
in a dose of 15 ,@Ci/kg (to a maximum of 1,050
@@Ci) or oomTc@pertechnetate in a standard dose of
10 mCi. The brain scans were put into three groups,
positive, negative and equivocal, using the initially
reported clinical reading.
Histopathologic grading for vascularity. Routine
sections cut 4—6 microns thick and stained with
hematoxylin and eosin were used. Degree of vas
cularity was judged according to the size and num
ber of patent arterial and venous channels present
in the most vascular area of the lesion without
knowledge of the brain-scan readings. The degree
of vascularity was then compared to that of sur
rounding tissue whenever normal brain or meninges
were included in the surgical specimen. If only the
lesion was available, its vascularity was compared
to that of routine necropsy sections of a similar area.
The following grading system was used:
1+ less vascular than normal brain
2+ vascularity similar to that of normal brain
3+ definitely more vascular than normal brain
4+ highly vascular
5+ extremely vascular
All lesions were examined on two different oc
casions to try to ensure consistent evaluations. The
same microscope was used throughout to maintain
an unvarying field size and to keep the magnifica
tion constant.
Received July 5, 1968; revision accepted Feb. 28, 1969.
For reprints contact: Manfred H. Soiderer, Dept. of
Pathology, Univ. of Michigan Medical Center, Ann Arbor,
Mich. 48104.
C Present address : Nuclear Medicine Division, Mary's
Help Hospital, Daly City, California 94015.
RAW DATA
Scan Readings
(c@
5
28
47 TOTALS
21
6
5+
4+
3+
2+
1+
TiSSUE
VASCULARfl@V
READINGS
IS 8 74 1107
TOTALS
. OnsCos. @)ç@.I5.S4p@OOI
0 AIIQICIVIOSS VOSCIIIOfNolfon.otions
FIG.1. Plotofdatain 3 X 5 tableaccordingtobrain-scan
reading and tissuevascularity reading. Each small black dot rep
resents one case. Small circles represent cases with vascular
malformations, which as a group seem to be apart from main dis
tribution. Inspection of raw data suggests a positive correlation.
Using heavy lines, data were condensed into 2 X 2 table with
four groups.Chi-square value was @3.6with p < 0.01. It may be
inferred from this figure that there is significant positive corre
lation between tissue vascularity and visualization on brain scan.
Volume 10, Number 8 553
TISSUE VASCULARITY IN POSITIVE
AND NEGATIVE BRAIN SCANS
Howard J. Cohn* and Manfred H. Soiderer
Wayne County General Hospital, Eloise, Michigan and University of Michigan, Ann Arbor, Michigan
