[18F]Fluoro-Deoxy-Glucose positron emission tomography to evaluate lymph node involvement in patients with muscle-invasive bladder cancer receiving neoadjuvant pembrolizumab.
Academic Article
Overview
abstract
BACKGROUND: Data regarding the role of positron emission tomography/computed tomography (PET/CT) to stage lymph nodes in patients receiving neoadjuvant immunotherapy before radical cystectomy are lacking. The aim of this study is to evaluate the role of PET/CT to predict the pathologic lymph node involvement (LNI) in patients with MIBC receiving neoadjuvant pembrolizumab within the PURE-01 trial (NCT02736266). MATERIAL AND METHODS: Three courses of pembrolizumab were administered before radical cystectomy and extended pelvic lymph node dissection in clinical T2-4aN0M0 MIBC based on contrast-enhanced CT scan. LNI was also assessed with PET/CT before and after treatment. PET/CT results were compared with histopathological findings. The ability of baseline and post-therapy PET/CT to evaluate LNI was assessed, and univariate logistic regression analyses were performed. RESULTS: From February 2017 to August 2019, a total of 108 patients and 105 patients had evaluable baseline and post-pembrolizumab scans, respectively. The sensitivity to detect LNI was 27% and 37.5% for pre- and post-pembrolizumab PET/CT, and specificity was 97% and 98%, respectively. In total, 4 of 7 patients (57%) showing baseline FDG-uptake had LNI vs. 11 of 101 (11%) with no baseline uptake. All but 1 of the 7 patients did not respond to pembrolizumab. Both pre- and post-pembrolizumab PET/CT significantly predicted LNI (P = 0.004 and P < 0.001) at univariate analyses. Our results warrant further validation in larger datasets. CONCLUSIONS: PET/CT performance does not justify its use in routine practice for cN0 MIBC. However, our preliminary data revealed opportunities for the use of baseline PET/CT, within clinical trials, to optimally select patients with MIBC who are best suited for neoadjuvant immunotherapy strategies. Validation in larger datasets, as well as a cost analysis, are needed.
