Study Explores Prognostic Value of an Automated Bone Scan Index for Patients with Prostate Cancer
An international team of researchers investigates the automated bone scan index -- a quantitative assessment of bone scan data that uses artificial neural networks
A bone scan is the standard imaging test to determine whether a patient's prostate cancer has spread to his bones. A small amount of radioactive tracer material is injected into the bloodstream and taken up by metastatic lesions that show up as hot spots on a gamma camera.
Currently, the test relies on doctors' interpretations of the images, which is open to some degree of variability. As an alternative, researchers investigate the prognostic value of the automated Bone Scan Index (aBSI) -- a quantitative assessment of bone scan data that uses artificial neural networks to detect and classify metastatic hot spots. The study was published May 17 by JAMA Oncology.
The participants included 721 men with metastatic castration-resistant prostate cancer who received a whole-body bone scan. For aBSI analysis, each scan was segmented into different anatomic regions, and abnormalities detected by the computer were classified as either metastatic or benign lesions. The aBSI was then calculated from the total tumor burden as a fraction of total skeleton weight.
As expected, the index strongly correlated with the number of manually counted bone lesions. The index is also better at predicting the overall survival rate than that of manual lesion counting. These preliminary results support the prognostic ability of aBSI as an objective imaging biomarker in patients with prostate cancer.