False Positive Rate from Prospective Studies of PET-CT in Cutaneous Malignant Melanoma: A Systematic Review and Meta-Analysis

Abstract

Background

Cutaneous malignant melanoma (CMM) is increasing in prevalence and possesses the highest mortality rate of any skin cancer. Positron Emission Tomography and Computed Tomography (PET-CT) may be utilised in either radiological staging or surveillance, primarily in stage III-IV disease. False positive (FP) results lead to patient distress, increased costs, and unnecessary follow-up. The FP rate in CMM literature varies widely, altering calculations of positive predictive value and has not undergone pooled meta-analytic.

Materials and Methods

A systematic review and meta-analysis of FP results in prospective studies of PET-CT in CMM was performed in accordance with PRISMA guidelines.

Results

The systematic review produced 14 trials for inclusion. Patient-based reporting had the lowest pooled proportion of FP results with 5.8 % (95 % CI = 3.3 % to 8.8 %), lesion-based was highest with 9.1 % (95 % CI = 3.4 % to 17.2 %) and combined was 6.1 % (95 % CI = 4.3 % to 8.1 %). Bias was low to unclear other than for FP reporting. Heterogeneity (I2) was variable across all analyses. FP findings were mainly lymphatic, dermatological, respiratory, or skeletal. Diagnostic information was not provided.

Conclusions

This study was the first attempt to quantify the pooled proportion of FP results from PET-CT in CMM. A small number of studies (n = 14) were available due to the predominance of retrospective methodology. Due to inconsistent reporting the true proportion of FP results is unclear. Systemic distribution was expected but limited diagnostic information was provided. Repeat meta-analysis using retrospective work should be performed. Future work should be prospective with clearly documented FP proportion, distribution, diagnosis, and follow-up.