Practical approach for comparative analysis of multilesion molecular imaging using a semiautomated program for PET/CT

Area of Science:

• Oncology
• Nuclear Medicine
• Radiology

Background:

• Quantitative molecular imaging (MI) in cancer patients with multiple lesions presents challenges in standardization.
• Accurate assessment of tumor burden and treatment response requires reliable imaging metrics.

Purpose of the Study:

• To propose and validate a standardized 5-step approach for quantitative molecular imaging in cancer patients with multiple lesions.
• To enhance the reproducibility and accuracy of PET/CT scan interpretation for treatment response assessment.

Main Methods:

• Twenty patients with castration-resistant prostate cancer underwent (18)F-FDG and (18)F-FDHT PET/CT scans.
• A 5-point confidence scale, site lexicon, and semiautomated PET software were used for lesion interpretation.
• Background correction factors were calculated using receiver-operating-characteristic analysis, and a novel Larson-Fox-Gonen plot was developed for response assessment.

Main Results:

• High agreement (>99%) was achieved for negative sites, with 83-85% agreement for positive sites.
• Consensus-lesion maximum standardized uptake value (SUV(max)) demonstrated high reproducibility (concordance correlation coefficient > 0.98).
• The approach yielded 4 correction factors and a novel scatterplot for depicting tumor burden and changes in SUV(max).

Conclusions:

• A 5-step approach optimizes multilesion molecular imaging in cancer.
• This method incorporates a confidence scale, site lexicon, semiautomated PET software, background correction, and Larson-Fox-Gonen graphing.
• The standardized approach enhances quantitative accuracy and reproducibility for cancer patient management.