3639 - Comparative Performance of ctHPVDNA and Imaging Surveillance for Oropharyngeal Cancer Following Radiotherapy

Y. Ding¹, J. E. Bates¹, S. Rudra¹, J. S. Remick¹, M. R. Patel², A. S. Kaka², M. W. McDonald³, and W. A. Stokes¹; ¹Department of Radiation Oncology, Winship Cancer Institute of Emory University, Atlanta, GA, ²Department of Otolaryngology, Winship Cancer Institute of Emory University, Atlanta, GA, ³Emory Proton Therapy Center, Atlanta, GA

Purpose/Objective(s): Imaging is routinely used in post-treatment surveillance of HPV-related oropharyngeal cancer (HPVOPC) but frequently identifies indeterminant areas of concern that prompt further imaging to ensure stability or resolution. Serum ctHPVDNA has emerged as an adjunctive surveillance strategy with high sensitivity for recurrence. To the best of our knowledge, there are no studies directly comparing ctHPVDNA with imaging in surveillance. Hypothesizing that ctHPVDNA would lead to fewer false positive results than imaging, we compared the performance of these modalities in early surveillance. Materials/

Methods: We retrospectively reviewed patients evaluated by our multidisciplinary clinic from 2022 onward with a new diagnosis of HPVOPC who received radiotherapy with planned dose >=50Gy in either the definitive or adjuvant setting. Eligible patients had 1) detectable pretreatment ctHPVDNA, 2) surveillance imaging and ctHPVDNA at 3 months posttreatment, and 3) 6-month posttreatment assessment. ctHPVDNA was quantified via liquid biopsy TTMV score (Naveris). Minimum thresholds for a “positive” test were results that would prompt short-interval repeat testing: an indeterminate or detectable ctHPVDNA and imaging NI-RADS 2+ OR at least mild avidity on PET. Recurrences were ascertained up to the 6-month assessment.
Results: 41 patients met eligibility criteria, with subsites divided evenly between tonsil (51%) and base of tongue (49%). Patients were clinically T0-4, N0-2. Fifteen (37%) received surgery, and 31 (76%) received concurrent systemic therapy. Delivered radiation doses ranged from 46-70 Gy(E), with 28 (62%) receiving proton therapy. Median pretreatment ctHPVDNA was 471 (range 8-174,881).

At 3 months follow up, 40 (98%) patients had undetectable ctHPVDNA and 14 (34%) had nonsuspicious imaging; 24 (59%) patients had NI-RADS 2+ imaging. By 6 month follow-up, there was 1 true recurrence that was detected at 3 months by both imaging and ctHPVDNA. Imaging yielded 26 false positives (63%) versus none for ctHPVDNA. Comparing ctHPVDNA versus imaging, sensitivity was identical at 100%, specificity was 100% vs 35%, positive predictive value (PPV) was 100% vs 4%, and negative predictive value (NPV) was identical at 100% (Table).
Conclusion: When setting the threshold for a positive surveillance test at a level that would prompt short-interval repeat testing, ctHPVDNA outperforms imaging as a 3-month surveillance strategy, with superior PPV and specificity. Relying on ctHPVDNA rather than imaging at 3-month posttreatment assessment may enable a significant portion of HPVOPC survivors to forgo unnecessary short-term follow-up imaging. Abstract 3639 – Table 1

		ctHPVDNA		Imaging
		-	+	-	+
Recurrence	-	40	0	14	26	100%/ 35%	Specificity (ctHPVDNA/Imaging)
	+	0	1	0	1	100%/100%	Sensitivity (ctHPVDNA/Imaging)
		100%	100%	100%	4%
		NPV	PPV	NPV	PPV