Volume-based response evaluation with consensual lesion selection: a pilot study by using cloud solutions and...

Area of Science:

• Oncology
• Radiology
• Medical Imaging Analysis

Background:

• Response Evaluation Criteria in Solid Tumors (RECIST) is a standard for tumor measurement.
• Lesion volume-based response (VBR) offers a potentially more accurate and consistent alternative for tumor measurements.
• Accurate tumor measurements are crucial for early detection of temporal changes in cancer therapy.

Purpose of the Study:

• To evaluate the impact of consensual lesion selection on volume-based response (VBR) assessments.
• To compare the inter-reader agreement of VBR with RECIST.
• To assess the feasibility of cloud solutions for standardizing multicenter response evaluations.

Main Methods:

• Eleven patients with lung CT scans were analyzed by three geographically dispersed readers.
• Lesions were assessed using both RECIST 1.1 and VBR criteria.
• A consensus process using cloud solutions was implemented for lesion selection in VBR calculations.
• VBR thresholds were set at -35% and +55% based on measurement variability.
• Inter-reader agreement was measured using Fleiss kappa (κfleiss).

Main Results:

• Consensual lesion selection for VBR yielded a Fleiss kappa of 0.85 (SE=0.091).
• Adding consensus on new lesions further increased agreement to κfleiss=0.95 (SE=0.092).
• RECIST without consensus showed lower agreement at κfleiss=0.72 (SE=0.088).
• The observed differences in agreement were statistically significant.

Conclusions:

• Consensual lesion selection effectively reduces inter-reader variability in VBR assessments.
• Cloud-based solutions provide a feasible and effective strategy for standardizing response evaluations in multicenter trials.
• Standardized VBR assessments can enhance consistency and reduce variability in clinical trial results.