Interrater Agreement and Reliability of PERCIST and Visual Assessment When Using 18F-FDG-PET/CT for Response Monitoring of Metastatic Breast Cancer.

Jonas S Sørensen,Jorun Holm,Jakob L Bülow,Poul-Erik Braad,Mie H Vilstrup,Lasse Ljungstrøm,Marianne Vogsen,Malene G Hildebrandt,Oke Gerke

doi:10.3390/diagnostics10121001

Jonas S Sørensen, Jorun Holm + Show 7 more

Open Access

https://doi.org/10.3390/diagnostics10121001

Copy DOI

Abstract

Response evaluation at regular intervals is indicated for treatment of metastatic breast cancer (MBC). FDG-PET/CT has the potential to monitor treatment response accurately. Our purpose was to: (a) compare the interrater agreement and reliability of the semi-quantitative PERCIST criteria to qualitative visual assessment in response evaluation of MBC and (b) investigate the intrarater agreement when comparing visual assessment of each rater to their respective PERCIST assessment. We performed a retrospective study on FDG-PET/CT in women who received treatment for MBC. Three specialists in nuclear medicine categorized response evaluation by qualitative assessment and standardized one-lesion PERCIST assessment. The scans were categorized into complete metabolic response, partial metabolic response, stable metabolic disease, and progressive metabolic disease. 37 patients with 179 scans were included. Visual assessment categorization yielded moderate agreement with an overall proportion of agreement (PoA) between raters of 0.52 (95% CI 0.44–0.66) and a Fleiss kappa estimate of 0.54 (95% CI 0.46–0.62). PERCIST response categorization yielded substantial agreement with an overall PoA of 0.65 (95% CI 0.57–0.73) and a Fleiss kappa estimate of 0.68 (95% CI 0.60–0.75). The difference in PoA between overall estimates for PERCIST and visual assessment was 0.13 (95% CI 0.06–0.21; p = 0.001), that of kappa was 0.14 (95% CI 0.06–0.21; p < 0.001). The overall intrarater PoA was 0.80 (95% CI 0.75–0.84) with substantial agreement by a Fleiss kappa of 0.74 (95% CI 0.69–0.79). Semi-quantitative PERCIST assessment achieved significantly higher level of overall agreement and reliability compared with qualitative assessment among three raters. The achieved high levels of intrarater agreement indicated no obvious conflicting elements between the two methods. PERCIST assessment may, therefore, give more consistent interpretations between raters when using FDG-PET/CT for response evaluation in MBC.

Highlights

Breast cancer (BC) is the most common cause of cancer-related deaths among women and the second most common cancer in the world [1]
In this retrospective clinical study of patients with metastatic breast cancer, we found that raters achieved a significantly higher proportion of agreement and reliability when using PET Response Criteria in Solid Tumors (PERCIST) evaluation compared to visual assessment of FDG-PET/CT for response monitoring
This indicated that a rater was more likely to assign the same response category to a scan as the other raters when using PERCIST compared with using visual assessment

Summary

Introduction

Breast cancer (BC) is the most common cause of cancer-related deaths among women and the second most common cancer in the world [1]. Approximate 10–15% of these women will develop metastatic breast cancer (MBC) [2]. Not all patients respond alike to the same treatment and when treatment fails, which may be caused by mutation in the BC, a change in treatment will be indicated. Accurate response monitoring tools are needed to direct patients to adequate treatment protocols. The Danish Breast Cancer Cooperative Group (DBCG) recommends the RECIST 1.1 [3] criteria for response monitoring in MBC. The European Society for Medical Oncology (ESMO) recommends imaging of chest, abdomen, and bone in staging of MBC, and imaging of target lesions every 2–4 month during treatment to evaluate response to treatment; they do not specify which imaging modality or response criteria should be used [5]

Objectives

Methods

Results

Conclusion