From planned neck dissection to PET response to circulating tumour DNA: changes in post‐treatment assessment of HPV‐driven oropharyngeal cancer

Abstract

Introduction

Circulating tumour human papillomavirus DNA (ctHPVDNA) is an emerging tool to assess post‐treatment response in patients with HPV+ oropharyngeal squamous cell carcinoma (OPSCC). Its use is not a standard practice, however, with interval F‐18 FDG PET/CT and fiberoptic examination preferred. Post‐treatment PET/CT at 3 months has a low positive predictive value (PPV), especially in patients with HPV+ OPSCC treated with (chemo)radiation therapy (CRT). We aimed to compare 3–6 month post‐treatment PET/CT and ctHPVDNA test results to determine the most effective option for post‐treatment response assessment.

Methods

Patients with HPV+ OPSCC that underwent commercially available ctHPVDNA blood testing after curative intent treatment were identified. Demographic, clinical, treatment, surveillance and oncologic outcome information were collected for each patient. Specificity and false positive rate were calculated for post‐treatment PET/CT and ctHPVDNA.

Results

80% of patients had Stage I disease. 52% of the population was treated with definitive chemoradiation (43%) or accelerated radiation (9%), with the remaining patients treated with transoral robotic surgery (TORS) +/− risk‐adapted adjuvant therapy. In total, 25 patients underwent ctHPVDNA testing and PET/CT at 3–6 months after finishing treatment. At 3–6 months post‐treatment, specificity of ctHPVDNA and PET/CT was 96% and 56%, correlating to false positive rates of 4% and 44%, respectively.

Conclusions

ctHPVDNA is more reliable than PET/CT following treatment in patients with HPV+ OPSCC, and its incorporation in post‐treatment response assessment will decrease the rate of anxiety over persistent disease and lead to a decrease in unnecessary medical procedures, including completion of neck dissection.

Circulating tumour human papillomavirus DNA (ctHPVDNA) is an emerging tool to assess post‐treatment response in patients with HPV+ oropharyngeal squamous cell carcinoma (OPSCC). This study highlights the benefits of this test in post‐treatment response assessment and offers a strategy for its incorporation into clinical practice.

Introduction

Patients with human papillomavirus (HPV)‐positive oropharyngeal squamous cell carcinoma (OPSCC) have superior outcomes compared to HPV‐negative patients, ¹ and as a result, effective and reliable post‐treatment response assessment is of paramount importance. In lieu of completion of neck dissection after definitive (chemo)radiation, post‐treatment surveillance consists of fiberoptic examination and F‐18 fluorodeoxyglucose (FDG) PET/CT. ² The National Comprehensive Cancer Network (NCCN) recommends delaying PET/CT until after 12 weeks given its poor positive predictive value (PPV). ³ , ⁴

Two retrospective studies have focused on the question of PET/CT utilisation at 3 months after definitive radiation therapy with or without chemotherapy in HPV‐positive oropharyngeal cancers. Both studies reported poor positive predictive value of F‐18 FDG PET/CT at 3 months, ⁵ , ⁶ which is especially the case for HPV+ OPSCC (compared to HPV‐ OPSCC). ⁵ Conversely, plasma circulating tumour HPV DNA (ctHPVDNA) is a sensitive and specific biomarker of HPV‐positive OPSCC, with reported PPV and negative predictive value (NPV) of 94% and 100%, respectively. ⁷

Current NCCN guidelines recommend post‐treatment surveillance of F‐18 FDG PET/CT at 3–6 months only, with additional imaging reserved for patients with ‘worrisome or equivocal signs/symptoms’. If clinical suspicion for recurrent or residual disease is high, pathologic confirmation is necessary. ³ With a current focus on treatment de‐intensification in HPV‐positive OPSCC, it is important to characterise the optimal response assessment strategy, so as to avoid the morbidity of undergoing an unnecessary post (chemo)RT neck dissection.

We decided to evaluate the clinical utility of ctHPVDNA after definitive treatment of HPV‐positive OPSCC and compare it to standard‐of‐care PET/CT evaluation.

Materials and Methods

Our clinic at Oregon Health & Science University completed an Institutional Review Board (IRB)‐approved retrospective review to highlight our experience of ctHPVDNA detection in patients diagnosed with HPV‐positive OPSCC through the utilisation of a commercially available ctHPVDNA detection kit to identify an optimal treatment response assessment strategy. Given the retrospective nature of this study, and after meeting all criteria mandated by the IRB, informed consent was waived. Patients selected for analysis included HPV+ OPSCC patients treated definitively with chemoradiotherapy, accelerated radiotherapy, or surgery with adjuvant (chemo)radiation as needed for adverse pathologic features. All patients must have completed at least one commercially available ctHPVDNA test after treatment.

Collected information included patient age, sex, tumour site/stage, pre‐treatment ctHPVDNA results (if collected), treatment specifics (date of surgery, radiation dosing/fractionation etc.) and dates/results of post‐treatment surveillance studies, including F‐18 FDG PET/CT and ctHPVDNA.

Our institutional standard for treatment response assessment consisted of PET/CT and physical examination (including fiberoptic examination) at 3 months. At that time, if the PET/CT signified a complete response to treatment, no further imaging was pursued unless clinical signs, symptoms and history justified additional imaging. In addition to standard post‐treatment evaluation, we have also begun to collect ctHPVDNA at 12 weeks after treatment completion, with subsequent collections for disease surveillance. ctHPVDNA collection was first implemented by our department in 2020, and given positive experiences reported by patients, coupled with the diagnostic utility of this test as reported in the literature, it later became widely adopted as a part of post‐treatment surveillance.

Regarding equivocal responses noted on PET/CT, the test was ideally repeated in 1–3 months. Each PET/CT was interpreted by a board‐certified diagnostic radiologist with extensive nuclear medicine experience, with impressions documenting complete response, equivocal findings indeterminate for residual disease versus treatment‐related inflammation, or evidence of residual disease/disease progression. Equivocal findings were discussed with otolaryngology colleagues to determine the best subsequent treatment, which included completion of neck dissection, biopsy, or surveillance. For the purposes of quantifying the sensitivity, specificity, PPV, NPV and false positive rate of each test, equivocal results were treated as positive results (as they would require additional work‐up/completion neck dissection, or have the potential to increase patient concerns over residual/recurrent disease), unless subsequent testing during the follow‐up period proved contrary.

Results

Patients (N = 65; Table 1) were included in the analysis (Stratum 1; Fig. 1), with a median follow‐up of 29.2 (13.0–112.3) months. The majority of patients were male with stage I OPSCC treated with chemoradiotherapy or surgery alone. Ninety‐one per cent of patients (n = 59) obtained post‐treatment imaging with F‐18 FDG PET/CT, typically completed 3–6 months following treatment (median 14.7 weeks, range 10.1–25.0). Given the retrospective nature of this study, this surveillance schedule was not strictly adhered to. Though 57.6% of patients demonstrated a complete response, 24 patients (40.7%) had PET/CT scans that yielded equivocal results, which led to further work‐up for 11 patients. Two of the 11 patients underwent biopsy of the PET‐equivocal area, which revealed no evidence of disease. Ten patients underwent repeat PET/CT (6 equivocal results, 4 complete responses), with a median time to subsequent PET/CT of 9.6 weeks (range 4.6–15.8). Twenty‐nine patients (44.6%) had pre‐treatment ctHPVDNA collection, which identified detectable ctHPVDNA levels in 96.6% of cases. Twenty‐eight of 65 patients had ctHPVDNA completed within 3–6 months after treatment with 96.4% of tests resulting as negative (1 indeterminate result—repeat of test resulted negative).

Table 1.

Clinical characteristics of the patient cohort.

Characteristic	HPV+ OPX T0‐3N0‐3M0 patients (N = 65)
Age (years)	59.3
Sex
Male	60 (92.3%)
Female	5 (7.7%)
Primary site
Oropharynx	58 (89.2%)
Recurrent disease	5 (8.6%)
Unknown	7 (10.8%)
Stage group
Stage I (T1–2 N0‐1)	52 (80.0%)
Stage II (T3 or N2)	12 (18.5%)
Stage III (T4 or N3)	1 (1.5%)
Treatment
TORS alone	15 (23.1%)
TORS + Adjuvant RT	11 (16.9%)
66 Gy/33 fx	2 (18.2%)
60 Gy/30 fx †	8 (72.7%)
62.7 Gy/33 fx	1 (9.1%)
TORS + Adjuvant CRT	5 (7.7%)
70 Gy/35 fx + cisplatin	2 (40%)
66 Gy/30–33 fx + cisplatin	2 (40%)
63 Gy/35 fx + cetuximab	1 (20%)
Definitive RT	6 (9.2%)
70 Gy/35 fx (5–6 fx per week)	3 (50%)
69.96 Gy/33 fx	1 (16.7%)
66 Gy/30 fx	1 (16.7%)
40 Gy/5 fx	1 (16.7%)
Definitive CRT	28 (43.1%)
70 Gy/35 fx + cisplatin ‡	22 (78.6%)
70 Gy/35 fx + cetuximab	3 (10.7%)
70 Gy/35 fx + carbo/taxol	1 (3.6%)
69.96 Gy/33 fx + cisplatin	2 (7.1%)

Abbrevaitions: carbo/taxol, carboplatin/paclitaxel; CRT, chemoradiation therapy; fx, fraction; Gy, Gray; HPV, human papilloma virus; OPX, oropharynx; RT, radiation therapy; TORS, transoral robotic surgery.

^† 1 with treatment discontinuation at 18 Gy per patient preference.

^‡ 1 with treatment discontinuation at 46 Gy after traumatic brain injury.

Figure 1.

Flow diagram of the patient strata analysed and relevant diagnostic test results. ctHPVDNA, circulating tumor human papillomavirus DNA; HPV; human papillomavirus; OPSCC; oropharyngeal squamous cell carcinoma; PET/CT, positron emission tomography.

Twenty‐five of 65 patients had both ctHPVDNA and PET/CT results 3–6 months after treatment (Stratum 2; Fig. 1). The median follow‐up period for this sub‐group was 21.2 (14.9–29.3) months. Seventy‐two per cent of patients had pre‐treatment ctHPVDNA testing completed. We observed a complete response by PET/CT in 56% and by ctHPVDNA in 96% of cases. Eighty‐eight per cent of patients (N = 22) underwent additional post‐treatment ctHPVDNA testing (usually every 3 months), with a median of four additional tests and median surveillance time of 17.7 months (7.3–25.0). Eleven patients had equivocal PET/CT results, six of which had a subsequent PET/CT 1–3 months later (median time to subsequent PET/CT 8.3 weeks, range 4.6–13.0) which revealed equivocal findings in 5 of 6 cases. None of the patients who had equivocal results underwent additional surgical evaluation (including neck dissection). Two patients (8%) developed recurrent disease (both in the lungs) roughly 1.5 years after treatment, and both patients had positive ctHPVDNA testing that preceded pathologic confirmation by more than 3 months.

The specificity of PET/CT and ctHPVDNA testing at 3–6 months post‐treatment was 56% and 96%, which complements with a false positive rate of 44% and 4%, respectively. No local or regional recurrences were observed in the first year of follow‐up, and as such, sensitivity, PPV and NPV calculations were not necessary.

Discussion

Our false positive rate (44%) for F‐18 FDG PET/CT at 3–6 months after definitive treatment of HPV‐positive OPSCC is in line with other studies. A UK analysis revealed that PPVs of 12‐week PET/CT in HPV‐positive and negative disease were 30% and 81.8%, respectively (P < 0.01), while the negative predictive value (NPV) of 12‐week PET/CT in HPV‐positive and negative disease was 92.9% and 55.6% respectively (P < 0.01). A striking 67% of HPV‐positive patients with equivocal responses on 12‐week PET/CT had a complete response (CR) by 24 weeks. Notably, patients with initial equivocal results had similar survival to patients with a complete response on PET/CT. ⁵ A retrospective analysis conducted in Australia of patients with HPV‐positive oropharyngeal cancer who achieved less than a complete response by PET/CT at 12 weeks and had repeat PET/CT at 16 weeks revealed a 71% conversion rate from incomplete response to CR. All patients with a persistent incomplete response on 16‐week PET/CT had no evidence of regional recurrence, despite proceeding to neck dissection. NPV of PET/CT at 12 and 16 weeks remained high at 98% and 97% respectively, whereas positive predictive value rose from 12% to 33%. ⁶

ctHPVDNA testing following definitive chemoradiation demonstrates high NPV and PPV for detecting disease recurrences, which is similar to our reported experience. A prospective biomarker clinical trial was conducted among patients with nonmetastatic HPV‐associated (p16‐positive) OPSCC, and detection of ctHPVDNA in two consecutive plasma samples during post‐treatment surveillance had high PPV (94%) and NPV (100%) for identifying disease recurrence. False‐positive rate was 1%. The median lead time between ctHPVDNA positivity and biopsy‐proven recurrence was 4 months, thus potentially facilitating earlier initiation of salvage therapy. ⁷ Furthermore, Kowalchuk et al. demonstrated ctHPVDNA testing to be the most cost‐effective strategy, even in the case of equivocal results through cost‐effective analysis. ⁸

Application of these previously mentioned studies, coupled with our own institutional experience has led us to modify our clinical algorithm for post‐treatment assessment of patients with HPV‐positive OPSCC, as detailed in Figure 2.

Figure 2.

Treatment response assessment algorithm in our clinic for patients with HPV‐positive OPSCC after definitive (chemo)RT, incorporating ctHPVDNA as an adjunct to standard of care PET/CT response‐based assessment. ctHPVDNA, circulating tumor human papillomavirus DNA; NED, no evidence of disease; PET/CT, positron emission tomography.

Our work is limited by patient number, median follow‐up time and inter‐observer variability by diagnostic radiologists. This work also has limitations that are inherent to retrospective studies, notably the timing differences in the completion of the first post‐treatment PET/CT. No patients underwent completion neck dissection, thus treating equivocal PET/CT scans as false positive results without pathologic confirmation creates some challenges. Even so, tumour recurrence would coincide with positive ctHPVDNA testing, which typically precedes abnormal imaging by more than 3 months, ⁷ as was the case for the 2 recurrences in our cohort.

We believe this study adds to the growing body of evidence supporting the use of ctHPVDNA testing as an effective treatment response assessment strategy in patients with treated HPV‐positive OPSCC and confirms the fallibility of PET/CT in response assessment. The incorporation of ctHPVDNA de‐intensifies post‐treatment response assessment and surveillance, reduce financial toxicity ⁸ and eases patient concerns secondary to equivocal results, justifying it as a component to either replace imaging intensive strategies or as an adjunct to better interpret equivocal imaging results.

Conflicts of Interest

The authors declare no conflict of interest.

Funding

Naveris, Inc.—responsible for open‐access journal publication fees. No involvement in drafting, editing or revision of the manuscript.

Data Availability Statement

Research data are stored in an institutional repository and will be shared upon request with the corresponding author.