Development and Initial Validation of a Statistical Image Analysis Framework for Quantitative Bone SPECT/CT in Patients with Prostate Cancer.

Abstract

Background/Objectives: Quantitative bone single-photon emission computed tomography/computed tomography (SPECT/CT) enables standardized uptake value (SUV)-based assessment of skeletal tracer uptake; however, SUV does not directly indicate the degree of deviation from the physiological distribution at the corresponding anatomical location. This study aimed to develop a statistical image analysis software framework for quantitative bone SPECT/CT and perform its initial validation. Methods: Eighty-one male patients who underwent quantitative bone SPECT/CT for suspected bone metastases from prostate cancer were retrospectively analyzed, including 49 without and 32 with bone metastases. The developed software integrated CT-based anatomical standardization, construction of a bone-specific normal database (NDB), and voxel-wise Z-score mapping. Thirty successfully standardized subjects without bone metastases were used to construct the NDB. Preliminary clinical validation included 75 reference skeletal regions of interest (ROIs) in the lumbar vertebrae and iliac bones from three independent subjects and 74 metastatic lesions from 19 patients. Results: Anatomical standardization was successful in 33/49 subjects without bone metastases and 19/32 patients with bone metastases. NDBs comprising 8, 16, and 30 subjects showed no significant difference in the standard deviation-to-mean ratio (p = 0.37). The developed software successfully generated voxel-wise Z-score maps representing deviations from the reference distribution at corresponding skeletal locations. Z-score_max showed a strong positive correlation with SUV_max in metastatic lesions (ρ = 0.95, p < 0.0001), and all 74 metastatic lesions showed Z-score_max > 2. Conclusions: The developed software enables statistical analysis of quantitative bone SPECT/CT using anatomical standardization, a bone-specific NDB, and voxel-wise Z-score mapping. Z-score mapping provides quantitative information relative to the location-specific reference distribution and might complement conventional SUV-based assessment. Further validation using larger and independent cohorts is required to determine its clinical value.