Ultrasensitive Liquid Biopsy in Patients with P16-Positive Oropharyngeal Cancer Using HPV-seq
E. Stutheit-Zhao,L. Penny,Z. Zhao,Y. Zheng,J. Zou,S. H. Huang,J. de Almeida,D. Goldstein,A. J. Hope,A. Hosni,J. Kim,F. F. Liu,G. Liu,E. Sanz Garcia,L. L. Siu,A. Spreafico,C. J. Tsai,J. Waldron,K. Han,S. V. Bratman
DOI: https://doi.org/10.1016/j.ijrobp.2024.01.021
2024-01-01
Abstract:Purpose/Objective(s)Plasma HPV DNA could detect molecular residual disease (MRD) after (chemo)radiotherapy (CRT/RT) for oropharyngeal cancer (OPC). Whole HPV genome sequencing (HPV-seq) could both genotype and quantify HPV DNA more sensitively than digital polymerase chain reaction (dPCR). We compared HPV-seq and dPCR pre- and post-RT in a large, prospective OPC cohort. We hypothesized that HPV-seq is more sensitive and more accurately predicts recurrence by MRD.Materials/MethodsWe recruited patients (pts) with non-metastatic p16-positive OPC (TNM8 stage I-III) undergoing curative-intent CRT/RT. Each pt had peripheral blood plasma prospectively collected at pre-RT and 3-months post-RT. The primary endpoint was recurrence-free survival (RFS) and secondary endpoint was overall survival (OS). Time-to-event analysis was performed using the Kaplan-Meier method with log-rank tests and multivariable Cox regression. HPV-seq was performed using a validated hybrid capture panel targeting 38 HPV genomes and 15 recurrently mutated genes in OPC. dPCR was performed using a validated multiplexed assay targeting HPV16 E6/E7. HPV DNA levels were expressed as copies/mL plasma and log-transformed for statistical tests.ResultsWe performed HPV-seq and dPCR on pre- and post-RT samples from 245 pts. Median follow-up was 52.3 (4.5-95.8) months, and 3-yr RFS was 87.9%. HPV-seq detected plasma HPV DNA in 237 (97%) pts pre-RT (HPV16-only: 161, non-HPV16: 16, mixed: 60) with 0.017-364867 copies/mL. Of pre-RT samples with HPV-16 reads detected on HPV-seq, HPV-16 dPCR was detectable in 202/221 (91%), with 0.32-37313 copies/mL on dPCR. Copies of HPV16 DNA per mL were strongly correlated between HPV-seq and dPCR (R=0.986). Post-RT, HPV-seq detected residual HPV DNA in 48 (20%) pts. Of these, dPCR detected 19 and did not detect 29, of which 6 had non-HPV16, and the other 23 had HPV16 with <1.37 copies/mL. Of pts with 3-year recurrence, post-RT detection of HPV DNA by HPV-seq predicted 16/23 (sensitivity=70%, NPV=95.5%, AUC=0.81) and by dPCR predicted 7/23 (sensitivity=30%, NPV=91.0%, AUC=0.63). Post-RT residual HPV DNA on HPV-seq was associated with poorer RFS (3-yr 58% vs. 96%, p<0.0001) and OS (3-yr 61.4% vs. 97.1%, p<0.0001). Of 192 pts with HPV16 undetectable by dPCR at post-RT, the 23 with residual HPV DNA detected by HPV-seq still had poorer RFS (3-yr 62.5% vs. 95.7%, p<0.0001) and OS (3-yr 75.0% vs. 96.9%, p<0.0001) than those with HPV-seq clearance. Adjusting for T and N staging, post-RT residual HPV16 DNA was independently prognostic of RFS and OS by either assay, with odds ratios for RFS of 14.91 [6.19, 35.94] for HPV-seq and 5.17 [2.09, 12.81] for dPCR.ConclusionWhole viral genome sequencing genotypes and quantifies plasma HPV DNA. Pts with residual HPV DNA had poorer RFS and OS, independent of stage. For HPV16, HPV-seq and dPCR are strongly correlated, but HPV-seq is more sensitive with higher negative predictive value and predicted 2.3x more 3-year recurrences in OPC pts.