Search

Purpose To determine whether plasma human papillomavirus (HPV) DNA predates clinical recurrence and compare its accuracy with 3-month fluorodeoxyglucose positron emission tomography (FDG-PET) in locally advanced cervical cancer. Methods This prospective multicenter study accrued 23 women with stage IB to IVA cervical cancer planned for definitive chemoradiation therapy (CRT). Plasma HPV DNA was measured serially by digital polymerase chain reaction, and FDG-PET was performed at 3 months post-CRT. Results Of the 19 women with HPV+ cervical cancer included in this analysis, 32% were stage IB, 58% IIB, and 10% IIIB/IVA. Median follow-up was 24 months (range, 18 to 30 months). All patients had detectable plasma HPV DNA before treatment. Six patients had detectable plasma HPV DNA at the end of CRT, and three of them developed metastases at 3 months. Of the 13 patients with undetectable plasma HPV DNA at end of CRT, to date, only one has developed recurrence. Six of those 13 patients had a positive 3-month FDG-PET with no definite residual disease on subsequent imaging or clinical examination to date, and four of these six had undetectable plasma HPV DNA at 3 months. Patients with undetectable plasma HPV DNA at end of CRT had significantly higher 18-month progression-free survival than those with detectable plasma HPV DNA (92% v 50%; P = .02). The area under the receiver operating characteristic curve (accuracy) of 3-month plasma HPV DNA and 3-month FDG-PET imaging for predicting recurrence at 18 months were 77% and 60%, respectively ( P = .008). Conclusion Detectable plasma HPV DNA at end of CRT predates the clinical diagnosis of metastases and is associated with inferior progression-free survival. Moreover, 3-month plasma HPV DNA level is more accurate than 3-month FDG-PET imaging in detecting residual disease. The clinical utility of plasma HPV DNA detection for guiding adjuvant/salvage therapy should be evaluated in future studies.