Related Experiment Video
Updated: Nov 8, 2025
Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model
Published on: August 16, 2020
Response evaluation of bone metastasis in prostate cancer: Preliminary comparison of computerized bone scan index
Lars J Petersen1, Michael Gade, Randi F Fonager
1Department of Nuclear Medicine, Aalborg University Hospital, Hobrovej 18-22,DK-9000 Aalborg,Denmark. lajp@rn.dk.
Objective:
The correlation between the computer-assisted bone scan index (BSI) responses versus clinical response classification if bone metastases in prostate cancer patients are not clear. We compared changes in BSI with Prostate Cancer Working Group-3 (PCWG3) and MD Anderson (MDA) criteria.
Materials And Methods:
Fifty-six consecutive patients with at least two bone scans (BS) within 12 months were included, who had BS before and after treatment with the same anticancer agent.
Results:
Progressive disease (PD) by PCWG3 criteria was seen in 28% of the cases (median BSI increased by 1.69 units) versus non-PD in 72% (BSI change -0.13). MDAnderson showed PD in 34% (BSI increase 0.49), 45% stable disease (BSI change 0.00), and 20% partial responses (BSI decrease 1.44). Absolute BSI changes differed significantly among response categories by PCWG3 and MDA criteria (both P<0.0001). Response classification using dichotomized BSI data (>0/≤0 and >0.3/≤0.3 BSI units) showed a significant correlation with PCWG3 and MDA criteria (all P<0.001). Absolute BSI changes and dichotomized BSI correlated to prostate-specific antigen responses (both P<0.001) but not to clinical responses.
Conclusion:
Absolute changes in BSI and BSI response classification correlated significantly with standardized clinical response criteria for the assessment of treatment responses of skeletal metastases in prostate cancer.
