those having the disease in a diseased population (true positive, TP). The
specificity of the test is the probability of being able to identify correctly
those who do not have the disease in a healthy population (true negative,
TN). By these definitions, it is obvious that a given test may not identify all
patients correctly whether or not they have the disease. This results in false-
positives (FP) in the healthy group and false-negatives (FN) in the diseased
group.
It should be noted that when sensitivity is assessed for a diseased popu-
lation or specificity for a healthy group, the disease or healthy status of the
group must be assessed by an established standard diagnostic test. This test
is called the “gold standard” and is considered the best method available
for comparison. As determined by the gold standard in a group of N
patients, if NPis the total number of persons with disease, and NAis the
total number of persons without disease, then the results of the new test
can be summarized in Table 4.2. The following parameters can be obtained
from data in Table 4.2.
(4.10)
(4.11)
(4.12)
(4.13)
(4.14)
Problem 4.6
In a group of 1000 patients, 840 patients (Group A) had brain tumor and
160 patients (Group B) did not have tumor by biopsies. The SPECT study
diagnosed 780 patients having tumor in Group A and 15 patients having
tumor in Group B. Calculate the sensitivity, specificity, accuracy, positive
predictive value, and negative predictive value of the SPECT study.
Negative predictive valueTN
TN FN
=
+
Positive predictive valueTP
TP FP
=
+
AccuracyTP TN
TP TN FP FN
=
+
+++
SpecificityTN
TN FP
TN
NA
=
+
=
SensitivityTP
TP FN
TP
NP
=
+
=
42 4. Statistics of Radiation Counting
Table4.2. Distribution of data obtained by a new
diagnostic test.
Test result Disease present Disease absent Totals
+ TP FP N+
− FN TN N−
TOTALS NP NA N