Zoe Guan

Purpose: To evaluate long-term outcomes and determine whether longitudinal plasma profiling provides biological insight into disease progression after proton hypofractionated ablative radiation therapy (HART) for hepatocellular carcinoma (HCC) and intrahepatic cholangiocarcinoma (ICC). Patients and Methods: In this single-arm, multi-institutional phase II trial, 92 patients with localized unresectable HCC or ICC with Child-Pugh class A/B liver function received HART (up to 67.5 GyE in 15 fractions). Overall survival (OS) and progression-free survival (PFS) were estimated using the Kaplan-Meier method. Plasma samples were collected at baseline, days 1, 8, 15, and 1-month after treatment for exploratory correlative analyses. Results: Among evaluable patients (44 HCC, 39 ICC), median OS was 29.2 months (HCC) and 20.8 months (ICC), with 5-year OS rates of 25.9% and 12.3%, respectively. Median PFS was 13.8 months (HCC) and 8.4 months (ICC). In-field local failure occurred in 2 HCC (4.5%) and 7 ICC (17.9%) patients, whereas out-of-field intrahepatic/distant progression remained the predominant pattern of failure. No grade >3 treatment-related toxicities were observed. Exploratory plasma profiling identified distinct biological signatures associated with progression: HCC was characterized by markers of hypoxia, inflammation, vascular remodeling, stromal activation, and hepatic regeneration, whereas associations in ICC predominantly involved signatures of hypoxia and vascular remodeling. Conclusions: HART provides durable long-term in-field local control with minimal toxicity, but disease progression remains predominantly out-of-field intrahepatic/distant. Exploratory correlative analyses suggest that biological signatures reflecting vascular remodeling, hypoxia, inflammation, stromal activation, and hepatic reserve are associated with progression despite successful local therapy, informing future translational studies and combination strategies.

Surgery is a standard treatment for early-stage human papillomavirus–associated head and neck cancer (HPV+HNC), yet adjuvant therapy decisions rely on clinicopathologic features with limited prognostic accuracy. Circulating tumor HPV DNA is a promising biomarker, but current assays lack the sensitivity needed for detecting minimal residual disease after surgery. This study evaluated the prognostic value of an HPV whole-genome sequencing assay, HPV-DeepSeek, and compared this with existing methods and clinical standards. One hundred three patients with stage I to IV HPV+HNC treated with surgery were prospectively enrolled (NCT06730412). Blood samples were collected before surgery, postoperatively, and during surveillance and were analyzed using HPV-DeepSeek and droplet digital polymerase chain reaction (PCR) assays. After a median follow-up of 27 months, patients with circulating tumor HPV DNA detected after surgery had worse 2-year disease-free and overall survival than those without detection. Detection after completion of all treatment was also associated with poorer outcomes. Minimal residual disease status was a stronger predictor of recurrence than standard clinicopathologic criteria. Molecular recurrence was identified up to 17.5 months earlier than clinical recurrence and nearly twice as early as with droplet digital PCR. These findings demonstrate that sensitive HPV whole-genome sequencing enables accurate detection of minimal residual disease, early identification of recurrence, and improved postoperative risk stratification, supporting its potential as a tool for guiding personalized adjuvant therapy in HPV+HNC.

Background: Early detection of HPV-associated oropharyngeal squamous cell carcinoma, the most common HPV cancer in the United States, could reduce disease-related morbidity and mortality, yet currently, there are no early detection tests. HPV circulating tumor DNA (ctDNA) is a sensitive and specific biomarker for HPV–associated oropharyngeal squamous cell carcinoma at diagnosis. It is unknown if ctDNA HPV is detectable prior to diagnosis, and thus its potential as an early detection test is also unknown. Methods: Plasma samples from the Mass General Brigham Biobank collected 1.3-10.8 years prior to diagnosis from HPV–associated oropharyngeal squamous cell carcinoma patients (n = 28) and age- and sex-matched controls (n = 28) were blinded and run on a newly developed and validated multifeature HPV whole genome sequencing liquid biopsy assay and a validated HPV antibody assay. Results: HPV ctDNA results were positive in 22 of 28 prediagnostic samples from HPV–associated oropharyngeal squamous cell carcinoma cases (sensitivity 79%) with a maximum lead time of 7.8 years. HPV ctDNA results were negative in all controls (0 of 28, 100% specificity). Diagnostic accuracy was highest within 4 years of cancer diagnosis and was higher than HPV Ab detection within the same timeframe (P = .004). Application of a machine-learning model trained and tested on an independent cohort of 306 cases and controls increased the sensitivity of detection to 27 of 28 cases (overall sensitivity 96%) and the maximum lead time to 10.3 years. Conclusions: HPV ctDNA can be detected in the blood years prior to diagnosis with HPV–associated oropharyngeal squamous cell carcinoma, with high specificity, in a case-control cohort of 56 participants. HPV ctDNA detection alone, or in combination with previously identified serological biomarkers, may be a feasible approach to early detection of HPV–associated oropharyngeal squamous cell carcinoma.

Purpose:The incidence of human papillomavirus (HPV)–associated head and neck squamous cell carcinoma (HPV + HNSCC) is increasing in the United States. Currently, there are no early detection approaches for HPV + HNSCC. Two blood-based analytes for early detection and diagnosis of HPV + HNSCC, circulating tumor HPV DNA (ctHPVDNA) and HPV early protein antibodies (HPV Ab), show promise, yet current approaches lack adequate diagnostic accuracy for broad clinical utility. Further, performance metrics across various assays for detecting these analytes alone or in combination have not been compared head-to-head. To address these limitations and knowledge gaps, we developed a multifeature HPV whole-genome sequencing (WGS) liquid biopsy for improved low-level ctHPVDNA detection. We defined the performance characteristics of this WGS-based approach and compared it head-to-head with existing blood-based HPV detection approaches to determine the optimal single or combinatorial biomarker strategy for a future prospective study of HPV + HNSCC early detection.Experimental Design:We tested blood samples from 304 participants: 152 patients with untreated incident HPV + HNSCC (77% stage I) and 152 general population control patients. We compared WGS-based ctHPVDNA detection, single-plex Droplet Digital PCR (ddPCR)–based ctHPVDNA detection, multiplex ddPCR-based ctHPVDNA detection, multiplex HPV Ab detection, and clinical standard-of-care tissue biopsy, benchmarked to gold-standard HPV + HNSCC tissue diagnosis. We then modeled the operational feasibility of these approaches as screening biomarkers for HPV + HNSCC.Results:HPV WGS sensitivity and specificity were 98.7% and 98.7%, respectively. Single-plex ddPCR sensitivity and specificity were 94.2% and 98.6%, respectively. Multiplex ddPCR sensitivity and specificity were 90.6% and 96.3%, respectively. HPV Ab sensitivity and specificity were 86.4% and 96.3%, respectively. A combinatorial approach using both HPV WGS and HPV Ab yielded a sensitivity and specificity of 87.4% and 98.8%, respectively. In a head-to-head comparison, HPV WGS demonstrated significantly improved diagnostic accuracy compared with ddPCR (Youden index for HPV WGS, 0.99 vs. ddPCR, 0.90; P < 0.001), HPV Ab (HPV WGS, 0.99 vs. HPV Ab, 0.83; P < 0.001), and clinical workup (HPV WGS, 0.99 vs. clinical workup, 0.82; P < 0.001), which was maintained when evaluating only early-stage disease cases. For men ages 55 to 74, HPV WGS yielded the lowest number needed to screen (2,903 men) and the highest positive predictive value (2.6).Conclusions:HPV WGS–based ctHPVDNA detection demonstrated the highest sensitivity, specificity, and diagnostic accuracy and thus the lowest number needed to screen and highest positive predictive value compared with ddPCR-based ctHPVDNA detection, HPV Ab–based detection, and combinatorial approaches. These results highlight the promise of HPV WGS liquid biopsy for screening and early detection of HPV + HNSCC and the need for modeling and cost-effectiveness studies to evaluate and guide screening implementation.See related commentary by Haring et al., p. 3359See related article by Sim et al., p. 3494

Purpose: The incidence of human papillomavirus (HPV)–associated head and neck squamous cell carcinoma (HPV + HNSCC) is increasing in the United States. Currently, there are no early detection approaches for HPV + HNSCC. Two blood-based analytes for early detection and diagnosis of HPV + HNSCC, circulating tumor HPV DNA (ctHPVDNA) and HPV early protein antibodies (HPV Ab), show promise, yet current approaches lack adequate diagnostic accuracy for broad clinical utility. Further, performance metrics across various assays for detecting these analytes alone or in combination have not been compared head-to-head. To address these limitations and knowledge gaps, we developed a multifeature HPV whole-genome sequencing (WGS) liquid biopsy for improved low-level ctHPVDNA detection. We defined the performance characteristics of this WGS-based approach and compared it head-to-head with existing blood-based HPV detection approaches to determine the optimal single or combinatorial biomarker strategy for a future prospective study of HPV + HNSCC early detection. Experimental Design: We tested blood samples from 304 participants: 152 patients with untreated incident HPV + HNSCC (77% stage I) and 152 general population control patients. We compared WGS-based ctHPVDNA detection, single-plex Droplet Digital PCR (ddPCR)–based ctHPVDNA detection, multiplex ddPCR-based ctHPVDNA detection, multiplex HPV Ab detection, and clinical standard-of-care tissue biopsy, benchmarked to gold-standard HPV + HNSCC tissue diagnosis. We then modeled the operational feasibility of these approaches as screening biomarkers for HPV + HNSCC. Results: HPV WGS sensitivity and specificity were 98.7% and 98.7%, respectively. Single-plex ddPCR sensitivity and specificity were 94.2% and 98.6%, respectively. Multiplex ddPCR sensitivity and specificity were 90.6% and 96.3%, respectively. HPV Ab sensitivity and specificity were 86.4% and 96.3%, respectively. A combinatorial approach using both HPV WGS and HPV Ab yielded a sensitivity and specificity of 87.4% and 98.8%, respectively. In a head-to-head comparison, HPV WGS demonstrated significantly improved diagnostic accuracy compared with ddPCR (Youden index for HPV WGS, 0.99 vs. ddPCR, 0.90; P < 0.001), HPV Ab (HPV WGS, 0.99 vs. HPV Ab, 0.83; P < 0.001), and clinical workup (HPV WGS, 0.99 vs. clinical workup, 0.82; P < 0.001), which was maintained when evaluating only early-stage disease cases. For men ages 55 to 74, HPV WGS yielded the lowest number needed to screen (2,903 men) and the highest positive predictive value (2.6). Conclusions: HPV WGS–based ctHPVDNA detection demonstrated the highest sensitivity, specificity, and diagnostic accuracy and thus the lowest number needed to screen and highest positive predictive value compared with ddPCR-based ctHPVDNA detection, HPV Ab–based detection, and combinatorial approaches. These results highlight the promise of HPV WGS liquid biopsy for screening and early detection of HPV + HNSCC and the need for modeling and cost-effectiveness studies to evaluate and guide screening implementation.

Inferring the cancer-type specificities of ultra-rare, genome-wide somatic mutations is an open problem. Traditional statistical methods cannot handle such data due to their ultra-high dimensionality and extreme data sparsity. To harness information in rare mutations, we have recently proposed a formal multilevel multilogistic “hidden genome” model. Through its hierarchical layers, the model condenses information in ultra-rare mutations through meta-features embodying mutation contexts to characterize cancer types. Consistent, scalable point estimation of the model can incorporate 10s of millions of variants across thousands of tumors and permit impressive prediction and attribution. However, principled statistical inference is infeasible due to the volume, correlation, and noninterpretability of mutation contexts. In this paper, we propose a novel framework that leverages topic models from computational linguistics to effectuate dimension reduction of mutation contexts producing interpretable, decorrelated meta-feature topics. We propose an efficient MCMC algorithm for implementation that permits rigorous full Bayesian inference at a scale that is orders of magnitude beyond the capability of existing out-of-the-box inferential high-dimensional multi-class regression methods and software. Applying our model to the Pan Cancer Analysis of Whole Genomes dataset reveals interesting biological insights including somatic mutational topics associated with UV exposure in skin cancer, aging in colorectal cancer, and strong influence of epigenome organization in liver cancer. Under cross-validation, our model demonstrates highly competitive predictive performance against blackbox methods of random forest and deep learning.

Table S3: AUCs of benchmark WES models using loss-of-function and missense variants.

Table S1: Known risk genes (from Rahman 2014) and SNPs (from NHGRI-EBI GWAS Catalog) in WES models.

Table S1: Known risk genes (from Rahman 2014) and SNPs (from NHGRI-EBI GWAS Catalog) in WES models.

Table S3: AUCs of benchmark WES models using loss-of-function and missense variants.