Clinical Meaningfulness of an Algorithm-Based Service for Analyzing Treatment Response in Patients with Metastatic...

Area of Science:

• Oncology
• Radiology
• Medical Imaging Analysis

Background:

• Assessing metastatic cancer treatment response is challenging for oncologists using text-only radiology reports.
• Current reports often lack comprehensive quantitative and spatial lesion data.

Purpose of the Study:

• To evaluate the clinical utility of an algorithm-based analysis for lesion quantification and spatial localization.
• To compare this algorithmic analysis against information present in standard US radiology reports.

Main Methods:

• Retrospective analysis of 228 metastatic cancer patients' FDG PET/CT radiology reports.
• Evaluation of qualitative and quantitative information in reports.
• Assessment of an algorithm providing quantitative data and spatial location for all lesion regions of interest (ROI) in 103 patients.
• Independent usefulness ratings by three oncologists for the algorithmic analysis.

Main Results:

• Standard reports provided quantitative size/uptake data for at least one lesion in 78%/95% of patients, but often lacked data across time points (52%/66%).
• Few reports (7%) quantified total lesions, and none quantified changes in all lesions for complex cases.
• The algorithm-based analysis was rated useful by 98% of oncologists for overall assessment.
• Oncologists found the algorithm highly useful for systemic therapy decisions (97%), spatial information (96%), and patient education (93%).

Conclusions:

• Algorithm-based analysis of all lesion regions of interest (ROI) enhances oncologists' understanding of treatment response in metastatic cancer.
• This quantitative and spatial data aids in optimizing patient therapy more precisely.