Aim To compare inter-reader agreement for planar and SPECT/CT interpretation
of equivocal osseous lesions detected on bone scintigraphy in
cancer patients. Materials & Methods This prospective study recruited
patients known to have primary tumor referred for bone scintigraphy with
their planar images showing solitary or few equivocal osseous lesions.
Every patient underwent planar whole body scan followed by SPECT/CT
for the concerned region(s). The gold standard was based on clinical/
imaging follow-up for at least 6-12 months. Two experienced readers
(reader 1 has 8-year experience&reader 2 has 12-year experience) scored
each lesion on a subjective 5-point score for the possibility of being
malignant (1=benign, 2= probably benign, 3 = equivocal, 4=probably
malignant & 5 = malignant). True & false results were identified were
identified in relation to the gold standard. Kappa measure of agreement
was measured for both modalities. Results A total of 150 patients were
included in this study (110 females ,40 males) with median age 54 years
(range:7-84). On planar imaging, both readers successfully identified 48
true positive (TP) & 19 true negative (TN) patients. Both readers were
false positive (FP) in 31&false negative (FN) in 3 patients. They disagree
in 2 patients. Moderate agreement was noted (kappa = 0.55). While on
SPECT/CT imaging, both readers successfully identified 58 TP & 62 TN
patients. Both readers reported FP results in 11 & FN in 2 patients. They
disagree in a total of 9 patients. Perfect agreement was noted (kappa =
0.82). Conclusion SPECT/CT significantly improved inter-observer
agreement among readers for equivocal osseous lesions detected on planar
bone scintigraphy.