Performance of Liquid Biopsy for Diagnosis and Surveillance of Human Papillomavirus–Associated Oropharyngeal Cancer

Abstract

There is growing interest in the use of circulating plasma tumor human papillomavirus (HPV) DNA for diagnosis and surveillance of patients with HPV-associated oropharyngeal squamous cell carcinoma (OPSCC). Recent advances in the assays, combining the identification of circulating HPV tumor DNA and tumor DNA fragment analysis (tumor tissue-modified viral [TTMV]-HPV DNA), have been shown to be highly accurate. However, use of these newer techniques has been limited to small cohort studies and clinical trials. To establish the clinical efficacy of plasma TTMV-HPV DNA testing in the diagnosis and surveillance of HPV-associated OPSCC in a contemporary clinical setting. This retrospective observational cohort study included patients with OPSCC who underwent TTMV-HPV DNA testing between April 2020 and September 2022 during the course of routine clinical care. For the diagnosis cohort, patients with at least 1 TTMV-HPV DNA measurement prior to initiation of primary therapy were included. Patients were included in the surveillance cohort if they had at least 1 TTMV-HPV DNA test performed after completion of definitive or salvage therapy. Per-test performance metrics, including sensitivity, specificity, positive predictive value, and negative predictive value, for TTMV-HPV DNA testing. Of 399 patients included in the analysis, 163 were in the diagnostic cohort (median [IQR] age, 63 [56-68.5] years; 142 [87.1%] male), and 290 were in the surveillance cohort (median [IQR] age, 63 [57-70] years; 237 [81.7%] male). Of the 163 patients in the diagnostic cohort, 152 (93.3%) had HPV-associated OPSCC while 11 (6.7%) had HPV-negative OPSCC. The TTMV-HPV DNA sensitivity in pretreatment diagnosis was 91.5% (95% CI, 85.8%-95.4% [139 of 152 tests]), and the specificity was 100% (95% CI, 71.5%-100% [11 of 11 tests]). In the surveillance cohort, 591 tests conducted in 290 patients were evaluated. A total of 23 patients had molecularly confirmed pathologic recurrences. The TTMV-HPV DNA test demonstrated sensitivity of 88.4% (95% CI, 74.9%-96.1% [38 of 43 tests]) and specificity of 100% (95% CI, 99.3%-100% [548 of 548 tests]) in detecting the recurrences. Positive predictive value was 100% (95% CI, 90.7%-100% [38 of 38 tests]), and negative predictive value was 99.1% (95% CI, 97.9%-99.7% [548 of 553 tests]). The median (range) lead time from positive TTMV-HPV DNA test to pathologic confirmation was 47 (0-507) days. This cohort study demonstrated that when evaluated in a clinical setting, the TTMV-HPV DNA assay demonstrated 100% specificity in both diagnosis and surveillance. However, the sensitivity was 91.5% for the diagnosis cohort and 88.4% for the surveillance cohort, signifying that nearly 1 in 10 negative tests among patients with HPV-associated OPSCC was a false negative. Additional research is required to validate the assay's performance and, if validated, then further research into the implementation of this assay into standard clinical practice guidelines will be required.