Frontiers in Oncology (Apr 2024)

Human papillomavirus circulating tumor DNA: a diagnostic tool in squamous cell carcinoma of unknown primary—a pilot study

  • Amani Kais,
  • Stell Patadji Santiago,
  • Peng Cheng Han,
  • David A. Clump,
  • William A. Stokes,
  • Tanya Fancy,
  • Ruifeng Cui,
  • Elizabeth Martin,
  • Meghan T. Turner

DOI
https://doi.org/10.3389/fonc.2024.1376595
Journal volume & issue
Vol. 14

Abstract

Read online

IntroductionNeck mass is the most common presentation of human papillomavirus-related (HPV-related) oropharyngeal squamous cell carcinoma (OPSCC). Recently, circulating tumor HPV-DNA (ctHPVDNA) assays have been developed to detect active OPSCC. This pilot study investigates the diagnostic accuracy of ctHPVDNA in establishing HPV status for known vs. unknown OPSCC presenting as a neck mass.MethodsA single-institution pilot study was conducted on all patients with OPSCC presenting as a neck mass between 2021 and 2022. The diagnostic accuracy of ctHPVDNA was compared to that of standard diagnostic procedures used to obtain HPV status according to the American Society of Clinical Oncology (ASCO) guideline for squamous cell carcinoma of unknown primary (SCCUP). Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of ctHPVDNA were calculated.ResultsA total of 27 patients were included; 70.4% were current or former smokers, 48.1% (N = 13) had identifiable primaries, and 51.9% (N = 14) had SCCUP. Four patients with known primaries required operative direct laryngoscopy with biopsy (DLB) to establish HPV status. Two patients with SCCUP underwent diagnostic transoral robotic surgery (TORS) to establish HPV status and localize the primary. Twelve patients underwent therapeutic TORS and neck dissection. The gold standard for HPV status was based on final histopathologic p16 or HPV in situ hybridization (ISH) staining during workup/treatment. ctHPVDNA had 95.8% sensitivity, 100% specificity, 100% PPV, and 75% NPV in predicting HPV-positive OPSCC in the whole sample. Binary logistic regression model using ctHPVDNA results to predict HPV-positive OPSCC was significant (−2 log likelihood = 5.55, χ2 = 8.70, p <.01, Nagelkerke’s R squared = .67). Among patients with identifiable primaries, all patients had HPV-positive tumors on final pathology, and ctHPVDNA was positive in 100%. In the unknown primary patients, ctHPVDNA had 90.9% sensitivity, 100% specificity, 100% PPV, and 75% NPV.DiscussionctHPVDNA demonstrated good diagnostic accuracy for both known and unknown primaries. Incorporation of ctHPVDNA into the diagnostic algorithm for SCCUP may reduce the need for multiple procedures to establish HPV status.

Keywords