R. Oommen et al., CORRELATION OF SERUM PROSTATE-SPECIFIC ANTIGEN LEVELS AND BONE-SCINTIGRAPHY IN CARCINOMA PROSTATE, British journal of radiology, 67(797), 1994, pp. 469-471
Serum prostate specific antigen (PSA) has been suggested as an accurat
e means of monitoring prostate cancer. An analysis of PSA levels and b
one scan findings was carried out in a heterogeneous group of patients
with a view to determine whether PSA can accurately predict bone meta
stases in carcinoma prostate. Of the 48 patients studied, all 10 untre
ated cases had elevated PSA levels, eight having bone metastases. In 2
9 cases on follow-up after treatment of the primary, 10 out of 11 case
s with normal PSA had a negative bone scan. In the nine who received h
ormonal therapy, the PSA levels were generally lower than others in th
e study group. Two out of four with normal PSA had bone metastases. In
26 cases with positive bone scans, 23 had elevated PSA levels (mean 1
09.9 ng ml(-1)). Among 22 patients who had normal bone scans, all 10 w
ith high PSA were found to have soft tissue disease which could explai
n the elevated PSA. In those with normal PSA, 12 out of 15 patients ha
d negative scans. PSA has fairly high sensitivity (86.5%) and negative
predictive value (80%). But it suffers from low specificity (54.5%) a
nd low positive predictive value (69.7%) for bone metastases. In an un
treated patient with elevated PSA, a bone scan may be required to excl
ude bone metastases, whereas during follow-up after treatment, a norma
l PSA level may obviate a ''routine'' bone scan.