If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Centre for Genomic and Experimental Medicine, Institute of Genetics and Cancer, University of Edinburgh, Crewe Road South, Edinburgh, EH4 2XU, UKNHS Lothian, Western General Hospital, Crewe Road South, Edinburgh, EH4 2XU, UK
Centre for Genomic and Experimental Medicine, Institute of Genetics and Cancer, University of Edinburgh, Crewe Road South, Edinburgh, EH4 2XU, UKNHS Lothian, Western General Hospital, Crewe Road South, Edinburgh, EH4 2XU, UK
NHS Lothian, Western General Hospital, Crewe Road South, Edinburgh, EH4 2XU, UKDepartment of Otolaryngology, University of Edinburgh, Lauriston Place, Edinburgh, EH3 9HX, UK
Corresponding author. Centre for Genomic and Experimental Medicine, Institute of Genetics and Cancer, University of Edinburgh, Crewe Road South, Edinburgh, EH4 2XU, UK.
Centre for Genomic and Experimental Medicine, Institute of Genetics and Cancer, University of Edinburgh, Crewe Road South, Edinburgh, EH4 2XU, UKNHS Lothian, Western General Hospital, Crewe Road South, Edinburgh, EH4 2XU, UK
Oropharyngeal squamous cell carcinoma (OPSCC) is increasing in global prevalence and is divided into two types dependent on association with human papillomavirus (HPV). Assay of HPV copy number in plasma cell-free DNA (cfDNA) provides a minimally invasive method for detecting and monitoring tumour-derived HPV, with potential for enhancing clinical care.
Materials and methods
In a prospectively recruited cohort of 104 OPSCC patients, we evaluate the utility of cfDNA droplet digital PCR (ddPCR) as a method for characterisation and longitudinal monitoring of patients with OPSCC.
Results
ddPCR assay of pre-treatment plasma cfDNA for five HPV types showed overall 95% concordance with p16 immunohistochemistry and PCR analysis of tumour tissue. Longitudinal sampling in 48 HPV+ve patients, with median follow-up of 20 months, was strongly associated with patient outcomes. Persistently elevated cfDNA-HPV post-treatment was associated with treatment failure (2/2 patients) and an increase of cfDNA-HPV in patients whose HPV levels were initially undetectable post-treatment was associated with disease recurrence (5/6 patients). No recurrence was observed in patients in whom cfDNA-HPV was undetectable in all post-treatment samples. In two patients, sequential HPV measurement could have avoided surgical intervention which did not confirm recurrence.
Conclusion
The high concordance of pre-treatment plasma cfDNA-HPV analysis with tissue-based assays, together with the clinical associations of sequentially measured post-treatment cfDNA-HPV copy number add to a growing body of evidence that suggest utility of cfDNA-HPV ddPCR in management of OPSCC. Standardised clinical trials based on these data are now needed to assess the impact of such testing on overall patient outcomes.
Oropharyngeal squamous cell carcinoma (OPSCC) is an increasingly prevalent cancer, linked in part to tobacco smoking, alcohol consumption and oncogenic types of the human papillomavirus (HPV) [
]. HPV-associated (HPV+ve) OPSCC has clinically significant differences from OPSCC without viral association (HPV-ve) including smaller primary tumours, more advanced regional disease and better prognosis [
]. The better prognosis of HPV+ve compared to HPV-ve disease has led to the interest in treatment deintensification for HPV+ve disease and highlights the importance of accurate tumour HPV status identification in OPSCC [
Deintensification of adjuvant treatment after transoral surgery in patients with human papillomavirus-positive oropharyngeal cancer: the conception of the PATHOS study and its development.
Radiotherapy plus cisplatin or cetuximab in low-risk human papillomavirus-positive oropharyngeal cancer (De-ESCALaTE HPV): an open-label randomised controlled phase 3 trial.
Pre-treatment assessment of OPSCC includes determining tumour HPV status by immunohistochemistry (IHC) against the protein p16 and/or HPV PCR testing of biopsy material [
]. Early-stage disease (AJCC TNM 7th edition stage I/II) is managed with radiotherapy (RT) or surgical resection while advanced disease (AJCC TNM 7th edition stage III/IV) is managed with combination therapy which may include surgery, radiotherapy and chemotherapy [
Treatment response is determined at 12-weeks following completion of therapy with clinical and radiological assessment. Current approaches to post-treatment assessment combine cross-sectional imaging, such as CT, with functional imaging, such as FDG-PET [
FDG-PET/CT for treatment response assessment in head and neck squamous cell carcinoma: a systematic review and meta-analysis of diagnostic performance.
]. For patients with a partial or indeterminate post-treatment locoregional response, management options include imaging surveillance, biopsy, or surgical intervention. Imaging surveillance to screen for recurrence is limited by the low positive predictive value of PET-CT [
Neck dissection after chemoradiotherapy for oropharyngeal and hypopharyngeal cancer: the correlation between cervical lymph node metastasis and prognosis.
]. A more accurate method of screening for treatment failure could allow identification of recurrence at an earlier stage with the potential to improve patient outcomes.
The analysis of circulating cfDNA, or “liquid biopsy”, represents a minimally invasive approach to diagnosis and management, with potential advantages for clinical care of patients with head and neck cancers [
]. Liquid biopsy has been shown to have applications across multiple cancer types, including HPV+ve OPSCC, for monitoring treatment response, predicting relapse and providing early detection of tumour recurrence [
]. The association of OPSCC with HPV infection offers an opportunity for diagnosis and disease tracking of HPV+ve cases through detection of HPV in plasma cfDNA, and a number of studies have taken this approach [
Rapid clearance profile of plasma circulating tumor HPV type 16 DNA during chemoradiotherapy correlates with disease control in HPV-associated oropharyngeal cancer.
Circulating human papillomavirus DNA detected using droplet digital PCR in the serum of patients diagnosed with early stage human papillomavirus-associated invasive carcinoma.
The primary aim of this study was to establish ddPCR assays of plasma cfDNA for multiple HPV strains and to compare these assays with conventional p16 IHC and qualitative PCR of solid tumour biopsies for definition of HPV status at presentation, in patients who were typed by tissue assays as either HPV+ve or HPV-ve. The secondary aim of this study was to carry out longitudinal blood sampling to test for association of plasma HPV cfDNA copy number with clinical course and to investigate its potential utility in clinical management.
2. Materials and Methods
Patients were screened and approached following discussion at the South East Scotland Multidisciplinary Head and Neck Cancer Meeting. Samples were collected under the Lothian NRS Human Annotated Bioresource (Reference: SR1171) and ethics approval by the East of Scotland Research Ethics Committee (Reference: 15/ES/0094). Patients with any-stage OPSCC and patients both with or without HPV association, managed with either curative or palliative intent by all treatment modalities were included (Table 1). Patients with non-squamous cell carcinoma, synchronous cancer or prior history of head and neck squamous cell carcinoma were excluded. Overall, 104 consecutive OPSCC patients gave informed written consent during their pre-treatment work-up. Patients underwent per-oral biopsy of primary and/or ultrasound-guided biopsy of nodal tissue with characterisation of tumour HPV status by p16 IHC as part of standard clinical care. A subset of tissue samples were later analysed with qualitative HPV-PCR from extracted tumour DNA.
Table 1Patient demographics, tumour stage and treatment.
Gender
-
Male
73
-
Female
31
Median Age (range)
61 (35–86)
Smoking status
-
Non-smoker
33
-
Ex-smoker
54
-
Current smoker
17
Alcohol consumption
-
Non/minimal
70
-
Occasional (≤10 units/wk)
2
-
Moderate (11–19 units/wk)
5
-
Excess (≥20 units/wk)
27
Subsite
-
Tonsil
58
-
Base of Tongue
37
-
Other
9
p16 IHC
-
Positive
88
-
Negative
16
TNM Stage
T
N
M
-
0
5
14
104
-
1
31
15
0
-
2
27
73
n/a
-
3
9
2
n/a
-
4
32
n/a
n/a
AJCC Stage (7th Edition)
-
1
3
-
2
5
-
3
12
-
4a
81
-
4b
3
Treatment
-
RT
13
-
CRT
73
-
Induction chemo + CRT
2
-
Induction chemo + RT
1
-
CRT + Immunotherapy
2
-
Surgery
3
-
Surgery + RT
4
-
Surgery + CRT
2
-
Palliative
4
Footnote: Two patients were treated with immunotherapy. One of these patients was HPV+ve and one was HPV-ve.
Pre-treatment blood samples were collected from all patients on recruitment to the study. Sequential post-treatment blood samples were also collected from each patient at follow-up clinic appointments, routinely scheduled as every 3 months for three years followed by every 6 months for two years for a total of 5 years, with additional appointments as clinically indicated. Some post-treatment blood samples were missed due to the cancellation of face-to-face appointments during the COVID-19 lockdown, however all clinical follow-up data was collected as planned. Blood collection, plasma separation and cfDNA extraction were carried out on a total of 268 blood samples. The overall study design is represented in Fig. 1 ddPCR assays of plasma cfDNA were carried out to detect the five most prevalent HPV genotypes in OPSCC, using primers designed and developed by Jeannot et al. (HPV16 & 18) and Chera et al. (HPV31, 33 & 35) [
Rapid clearance profile of plasma circulating tumor HPV type 16 DNA during chemoradiotherapy correlates with disease control in HPV-associated oropharyngeal cancer.
Circulating human papillomavirus DNA detected using droplet digital PCR in the serum of patients diagnosed with early stage human papillomavirus-associated invasive carcinoma.
]. Further details are given in Supplementary Material online.
Fig. 1Study design. All 104 patients with OPSCC underwent characterisation of HPV status by p16 IHC on tumour tissue and by assay of plasma cfDNA by HPV ddPCR. A subset of 68 patients were additionally characterised using HPV-PCR of tumour tissue. 48 patients who were HPV+ve by solid tissue characterisation and cfDNA ddPCR pre-treatment were followed longitudinally with sequential blood sampling for further HPV assay by cfDNA ddPCR. The “partial response” grouping includes patients with indeterminate imaging. OPSCC, oropharyngeal squamous cell carcinoma; HPV, human papilloma virus; cfDNA, cell-free DNA; CRT, chemoradiotherapy; Immuno, immunotherapy. RT, radiotherapy; EUA, examination under anaesthesia.
Clinical details of the 104 recruited patients are shown in Table 1, of whom 58 (56%) had primary tonsillar disease, 37 (36%) base of tongue and the remaining 9 (8%) had disease at other sub-sites (unknown primary, soft palate and lateral pharyngeal wall). This distribution is representative of the expected demographic in the UK [
3.2 Comparative analysis of pre-treatment HPV detection: plasma cfDNA-HPV ddPCR compared with solid tumour p16 IHC and HPV-PCR
We detected HPV by ddPCR of plasma cfDNA in 85 of the 104 patients pre-treatment, with a wide range of copy number (median 259 copies/ml; range 1–30,380 copies/ml; Fig. 2; Supplementary Fig. S1). Of these 85 patients, the overwhelming majority (n = 83) were positive for HPV16 (median 279 copies/ml). Of the remaining two patients, one was positive for HPV33 (5 copies/ml) and one for HPV35 (84 copies/ml). Of the HPV16 positive patients, three were dual positive for both HPV16 and HPV18 (median 154 copies/ml) (Fig. 2).
Fig. 2Comparison between plasma ddPCR assay and solid tumour assays. Bar chart and heatmap of plasma cfDNA-HPV copy number (plotted as Log10 copies/ml plasma) for the five most prevalent HPV genotypes in OPSCC. Each column denotes a patient. Bar chart displays the combined copies/ml for all genotypes in each patient. Light grey bars (middle panel) denote samples with no detectable copies beyond background levels. Bar chart above displays the combined copies/ml for all genotypes in each patient. Plots below describe results from matched solid tumour biopsy material for p16 immunohistochemistry and HPV-PCR. followed by T and N staging, using AJCC 7th edition. Note that T stages 4a and 4b have been combined into T4; N2 has been kept as 2a, 2b and 2c. The final plot highlights the 56 patients included in the longitudinal follow-up analysis.
All 104 patients had pre-treatment p16 IHC solid tissue data available. We found that 84/85 cases observed to be HPV+ve by ddPCR stained positively for p16 in matched tumour tissue (99% concordance), whilst 15/19 cases observed to be HPV-ve by ddPCR were negative for p16 IHC (79% concordance) giving an overall concordance of 95% between ddPCR and p16 IHC results (Fig. 2).
In 68/104 patients, we carried out qualitative PCR and genotyping on DNA extracted from solid tumour tissue collected pre-treatment. HPV-PCR results agreed with p16 IHC data in 64/68 (94%) of matched tumour samples. HPV-PCR results of solid tumour achieved similar concordance with cfDNA ddPCR results (63/68; 93%), although three of the samples showing discordance between cfDNA ddPCR and tumour PCR were concordant between cfDNA ddPCR and p16 IHC (Fig. 2). All patients in this subset who were observed to be HPV+ve by cfDNA ddPCR were also positive by either p16 IHC or HPV-PCR (52/52). Of the remaining 16 patients who were observed to be HPV-ve by cfDNA ddPCR, 88% (14/16) were also negative by either p16 IHC or HPV-PCR. The discordant cases included both early and late-stage disease, as calculated by the T and N staging (Fig. 2). Taken together, the data indicate that pre-treatment blood-based ddPCR results are strongly concordant with existing assessment methods employed on solid tumour tissue.
3.3 Longitudinal monitoring of plasma HPV cfDNA copy number
To test for correlation of treatment response with cfDNA HPV copy number, we collected sequential post-treatment blood samples in 48 of the 85 patients with detectable HPV cfDNA pre-treatment (range 2–105 weeks post-treatment) and in 8 of the 19 patients with undetectable HPV cfDNA pre-treatment (range 4–39 weeks post-treatment). All 8 patients with undetectable HPV cfDNA pre-treatment displayed negative results on all remaining samples reiterating the consistency of the ddPCR assays across time points and treatment stage.
All 48 patients with detectable HPV cfDNA pre-treatment in whom we collected sequential post-treatment blood samples underwent treatment with curative intent. Forty patients reached undetectable HPV levels post-treatment and remained undetectable for the remainder of the study (Fig. 3A). 39 of these achieved this by their first post-treatment blood draw (average 10 weeks, range 2–22 weeks) with the final patient (patient #04) reaching undetectable levels by 31-weeks post-treatment (Supplementary Fig. S2). Patient #04 was unique in that this was the only patient with detectable pre-treatment HPV cfDNA treated with combined CRT and adjuvant immunotherapy.
Fig. 3Treatment responses and patient outcomes in HPV+ve patients according to plasma cfDNA-HPV copy number. A. cfDNA-HPV copy number in patients for whom HPV levels became undetectable following treatment and remained undetectable for duration of follow-up (n = 40). B. cfDNA-HPV copy number in patients whose HPV levels never became undetectable following treatment or became undetectable but subsequently increased (n = 8). C. Kaplan-Meier plot and number at risk table of TTP for all patients with detectable HPV cfDNA pre-treatment with subsequent post-treatment samples (n = 48), stratified by presence or absence of detectable HPV in post-treatment cfDNA sampled at a median of 13 weeks post-treatment.
Eight patients demonstrated either a persistently elevated or a subsequent increase in their HPV copy number after initially becoming negative post treatment (Fig. 3B). Two of these patients never reached undetectable HPV levels post-treatment (patients #15 and #79). Early regional (patient #15) and distant (patient #79) disease progression was observed, suggesting treatment failure (Fig. 4). In six patients, HPV copy number fell to zero post-treatment and subsequently increased. In five of these patients (patients #11, #54, #55, #77, #84), the rise in HPV copy-number either pre-dated or correlated with detection of progressive regional or metastatic disease (Fig. 5A–E; Supplementary Figs. S3A–E). In the final patient (patient #26), a non-sustained rise was detected at 58 weeks post-treatment with no clinical or radiological evidence of disease progression (Fig. 5F).
Fig. 4Plasma cfDNA-HPV and CT/PET-CT scans in HPV+ve patients with consistently raised cfDNA-HPV. A-B. Pre- and post-treatment cfDNA-HPV copy number (left) in Patients #15 and #79 respectively with consistently raised cfDNA-HPV in whom disease progression was confirmed by clinical and imaging assessment. PR, partial response; CR, complete response. CT, MR and PET-CT scans from select time points (right). Yellow arrows demonstrate local disease, red arrows demonstrate regional disease and green arrows demonstrate distant disease.
Fig. 5Longitudinal cfDNA-HPV in HPV+ve patients with initial fall to undetectable and subsequent rise in HPV copy number. A-E. Patients #11, #54, #55, #77 and #84 showed a fall in HPV copy number with subsequent consistent rise associated with disease progression, as indicated. F. HPV copy number in Patient #26 fell to zero initially and showed a rise in one sample that was not sustained. PR, partial response; CR, complete response.
The utility of cfDNA HPV detection at ∼12-weeks post-treatment was assessed as a predictor for disease progression in our HPV+ve longitudinal cohort. Post-treatment imaging (CT ± PET-CT) at a similar time point classified 36 patients as complete response and 12 patients as partial or disease progression (Supplementary Materials and Fig. 1). In contrast, 44 patients were classified as having no detectable HPV levels by cfDNA ddPCR, suggesting a complete response in these patients. Given the apparently higher rates of response attained through analysis of cfDNA, we examined the outcomes of patients over the entire study period stratified by presence or absence of HPV cfDNA at the timepoint closest to their 12-week imaging (median 13 weeks, range 2-22-weeks). In all 48 patients, the 1-year time to progression (TTP) was 89.6%. When stratified by the presence or absence of detectable cfDNA-HPV on post-treatment blood sampling at a median of 13 weeks after cessation of treatment, the 1-year TTP was 50% vs. 93.2% respectively (p = 0.001; hazard ratio 10.0 (95% CI 2.1–47.1; Fig. 3C). Although this result was obtained using the wide range of sample timepoints closest to 12-week imaging, the result was robust to restricting the range of sample timepoints to 12 ±5 weeks (p = 0.001; hazard ratio 11.6 (95% CI 2.3–58.8)) and 12±3 weeks (p = 0.001; hazard ratio 11.8 (95% CI 2.1–66.7)).
3.4 Plasma cfDNA-HPV status and threshold for surgical intervention
In our practice, clinical assessment of HPV status in OPSCC is based on p16 IHC of solid tumour material, with HPV-PCR testing used only when requested.
In our cohort, one case (patient #32) was observed to be negative for HPV by p16 IHC but positive by ddPCR of cfDNA. Later HPV-PCR testing performed for this study on the pre-treatment solid tumour tissue (not available at the time of diagnostic decision making) was also positive. Longitudinal assessment of cfDNA-HPV copy number demonstrated a drop from 41 copies/ml pre-treatment to undetectable in all post-treatment samples (Fig. 6A). Clinically, the patient's tumour was managed as HPV-ve based on the p16 IHC result. Post-treatment imaging at ∼12-weeks demonstrated a partial response, and the clinical decision was taken to perform salvage surgery which resulted in a hypoglossal nerve injury. Pathological analysis of the resected tissue demonstrated no detectable tumour, consistent with the undetectable levels of HPV by ddPCR.
Fig. 6Post-treatment monitoring and influence on decision for surgical intervention. A-B. Case studies for Patient #32 and Patient #64 respectively who showed a partial response by imaging and underwent salvage surgery, with ddPCR results indicating a complete response to treatment. A. Patient #32 showed discordant pre-treatment HPV results (cfDNA-HPV+ve, tumour PCR HPV+ve, p16 -ve). B. All pre-treatment results in Patient #64 were HPV+ve.
Patient #64 demonstrated similar HPV kinetics (Fig. 6B), though pre-treatment p16 IHC, tissue HPV-PCR and cfDNA ddPCR results were all positive. Longitudinal assessment showed a drop from 18 copies/ml pre-treatment to undetectable levels at the patient's first post-treatment sample. Combined ∼12-week imaging suggested a partial response to treatment. This, in addition to the patient's young age (48 years), contributed to the decision to schedule salvage surgery at 21 weeks post-treatment, which was complicated by a severe wound infection. Pathological analysis of removed tissue showed squamous cells of questionable viability. All subsequent HPV ddPCR assays of plasma cfDNA have remained undetectable (Fig. 6B) and the patient to date has no clinical evidence of recurrence.
4. Discussion
In our cohort of 104 patients with OPSCC, we compared pre-treatment plasma cfDNA-HPV ddPCR with conventional p16 IHC and HPV-PCR from solid tumour tissue. We also carried out longitudinal blood sampling to test for association of sequentially measured plasma cfDNA-HPV copy number with clinical course and to investigate the potential role of cfDNA-HPV ddPCR in clinical management.
We found a high level of concordance (95%) between HPV detection via ddPCR of pre-treatment cfDNA and p16 IHC of tumour tissue, which was comparable to the concordance that we observed (94%) between p16 IHC and HPV-PCR of tumour tissue. This level of concordance is greater than that observed in other studies of PCR assays of cfDNA designed to detect HPV16 only (82% and 71% respectively) [
Rapid clearance profile of plasma circulating tumor HPV type 16 DNA during chemoradiotherapy correlates with disease control in HPV-associated oropharyngeal cancer.
HPV circulating tumoral DNA quantification by droplet-based digital PCR: a promising predictive and prognostic biomarker for HPV-associated oropharyngeal cancers.
Rapid clearance profile of plasma circulating tumor HPV type 16 DNA during chemoradiotherapy correlates with disease control in HPV-associated oropharyngeal cancer.
Circulating human papillomavirus DNA detected using droplet digital PCR in the serum of patients diagnosed with early stage human papillomavirus-associated invasive carcinoma.
]. In our study, in patients where all three assays (ddPCR of plasma cfDNA, p16 IHC and PCR of tumour tissue) were performed, 100% of samples that were positive by ddPCR were also positive by either p16 IHC or tissue PCR, whilst 88% of negative ddPCR tests were also negative by either p16 IHC or tissue PCR. Taken together with our data, these combined results demonstrate high sensitivity and specificity of ddPCR HPV assays for determination of HPV status at presentation, particularly those that assay for multiple HPV types [
Rapid clearance profile of plasma circulating tumor HPV type 16 DNA during chemoradiotherapy correlates with disease control in HPV-associated oropharyngeal cancer.
Circulating human papillomavirus DNA detected using droplet digital PCR in the serum of patients diagnosed with early stage human papillomavirus-associated invasive carcinoma.
We studied 48 HPV+ve patients longitudinally with a median follow-up of 20 months and observed seven cases of disease progression. In 40 of the 48 patients (83%), HPV copy number fell to zero after treatment and remained undetectable for the duration of the study. All 40 of these patients remained disease free. Previous studies of comparable size (22–67 cases studied longitudinally) have shown a similar percentage of patients (80–91%) who were cfDNA positive for HPV at presentation, who then became persistently cfDNA negative for HPV following treatment [
Rapid clearance profile of plasma circulating tumor HPV type 16 DNA during chemoradiotherapy correlates with disease control in HPV-associated oropharyngeal cancer.
]. Of note, each of these studies have different assay methods, sampling timepoints and study endpoints. For example, Chera et al. and Rutkowski et al. use ddPCR methodology to detect HPV16 from longitudinally collected cfDNA samples while Hanna et al. use multiple ddPCR assays to detect 5 types of oncogenic HPV. In contrast, Lee et al. employ next generation sequencing (NGS) for HPV16 detection. In assessing association with treatment outcomes, Chera et al. define a “favorable cfDNA clearance profile” based on pre-treatment HPV copy number and >95% clearance by week 4 of treatment [
Rapid clearance profile of plasma circulating tumor HPV type 16 DNA during chemoradiotherapy correlates with disease control in HPV-associated oropharyngeal cancer.
]. Rutkowski used the 12-week post-treatment timepoint to define sensitivity and specificity for detecting treatment failure up to 6 months post-treatment [
In 8 of the 48 patients that we followed longitudinally, two patients showed persistently raised HPV cfDNA and both were confirmed to have disease progression at ≤28-weeks post treatment. In the remaining six patients, HPV cfDNA was undetectable in the first post-treatment sample, which was followed by a rise in HPV cfDNA copy number in subsequent samples. Five of these six patients went on to develop either regional or distant disease progression. The rise in cfDNA HPV copy number was detected in two of these five patients before the clinical detection of disease recurrence. In a single case, there was an isolated increase in HPV cfDNA copy number in a single sample with subsequent return to undetectable and no evidence of disease progression. The result in this patient adds weight to the recommendation by Chera et al. to obtain two consecutive positive results before recall for clinical and radiological reappraisal [
Because of the routine 12-week radiological assessment timepoint, we tested for an association between TTP and detection of HPV cfDNA in the blood sample obtained closest to 12 weeks post-treatment. Detection of HPV cfDNA by ddPCR at a median of 13-weeks was associated with a lower TTP at one year in patients with HPV+ve OPSCC (50% compared to 93.2%, p = 0.001, hazard ratio = 10.0 (95% CI 2.1–47.1)). However, of the 7 patients that progressed, only 3 had detectable HPV cfDNA on the post-treatment blood sample that was closest to their 13-week timepoint. Progression in the remaining 4 patients was associated with a later rise in HPV cfDNA in all 4 patients, either preceding or at the time of clinical confirmation of disease progression. In a cohort of 66 HPV+ve patients, Rutkowski et al. reported a similar association between 12-week cell-free HPV16 DNA and patient outcomes after 6 months of follow up. In their study, five out of six patients who had detectable cell-free HPV16 DNA at 12 weeks had confirmed disease progression within 6 months while none of those who were HPV16 undetectable at 12 weeks had clinical evidence of disease progression [
Two cases of detectable pre-treatment cfDNA-HPV followed-by persistently undetectable post-treatment HPV cfDNA in our study demonstrate how pre-treatment and longitudinal post-treatment HPV cfDNA analysis can be of value in the clinical management of HPV+ve OPSCC. In Patient #32 and in Patient #64, indeterminate imaging at 12 weeks led to a clinical decision for surgical intervention which resulted respectively in a hypoglossal nerve injury and a wound infection. Tissue resection in neither case led to disease detection and the patients have remained in remission. In both cases, pre-treatment cfDNA-HPV was positive and was followed by a fall to consistently undetectable levels. Greater reliance on the ddPCR results in these two cases, had they been available for clinical decision making, could have avoided surgery. Similarly, Rutkowski et al. reported post-treatment biopsy or neck dissection in 7 HPV+ve OPSCC patients with indeterminate imaging but undetectable post-treatment HPV16 cfDNA. In all 7 patients, residual disease was not identified [
]. Lee et al. reported post-treatment neck dissection or biopsy in 6 patients with indeterminate imaging, all of whom had HPV16 cfDNA below the detection threshold, with no evidence of viable tumour found in the resected tissue [
]. These cases show the potential value of HPV cfDNA assays to assist in management decisions in OPSCC patients with suspected disease progression.
Our observation of undetectable pre-treatment cfDNA HPV by ddPCR in two patients who were positive by both p16 IHC and tissue HPV PCR indicates that ddPCR may have less than 100% sensitivity for detection of HPV-related OPSCC. Both tumours were N0 but T2 or T3, suggesting that limited tumour size or spread may have been a factor in the undetectable HPV result. However other factors may also contribute, because other tumours of similar stage in our cohort had positive HPV cfDNA results. These results may be viewed alongside the less than 100% sensitivity and specificity of the tissue-based assays, in which false positive results can occur with p16 IHC owing to activation of the p16 pathway by causes other than HPV [
Heterogeneity of p16 immunohistochemistry and increased sensitivity of RNA in situ hybridization in cytology specimens of HPV-related head and neck squamous cell carcinoma.
Evaluation of human papilloma virus diagnostic testing in oropharyngeal squamous cell carcinoma: sensitivity, specificity, and prognostic discrimination.
Our study has certain limitations. First, a small number of sampling timepoints were missed due to clinic closures during lockdown arising from the COVID-19 pandemic. This affected a minority of patients and replacement samples were sought where possible and all clinical data was collected as planned. Second, tumour HPV-PCR results were only available for a subset of patients, but allowed comparison of all three methods in over than 60% of the cohort. Third, longitudinal monitoring of HPV cfDNA was only available on 48 HPV+ve patients and 8 HPV-ve patients. This was restricted by patient availability for follow-up in an unbiased fashion and did not affect the overall significance of the results.
Taken together, these data show that pre-treatment assay of plasma cfDNA HPV is strongly concordant with solid tumour p16 IHC and qualitative HPV-PCR, that longitudinal measurement of cfDNA HPV can inform clinical decisions and that stratification of HPV+ve OPSCC patients by the presence or absence of detectable HPV cfDNA at a median of 13-weeks post-treatment is associated with 1-year TTP. Furthermore, regular post-treatment measurement of cfDNA HPV copy number can lead to the earlier detection of disease progression than standard clinical assessment and imaging alone. Together, our findings, contextualised with the current literature, provide a compelling case for evaluation of the cost effectiveness and clinical utility of cfDNA HPV ddPCR analysis in the management of patients with OPSCC. Given the multiple assays and timepoints on which the conclusions of studies to date have been based, testing of HPV cfDNA under standardised study design in prospective randomised trials will help to establish the long-term impact on patient outcomes. The results of this study may help in the successful design of such trials.
Ethics approval, consent to participate and consent for publication
All samples were collected under the Lothian NRS Human Annotated Bioresource (Reference: SR1171) with ethics approval by the East of Scotland Research Ethics Committee (Reference: 15/ES/0094). A total of 104 patients gave informed written consent, for participation and publication, during their pre-treatment work-up. Our study was performed in accordance with the Declaration of Helsinki.
Funding information
The work was supported by awards from Cancer Research UK (grant number C22524/A26254) and the Scottish Chief Scientist Office (grant numberTCS/20/11) to TJA, and from the Guthrie Trust of the Scottish Otolaryngological Society (Reference SEN10771984) to IJN.
Contribution Author(s)
Study concepts: TJA, IJN, SJW, MA, JPT. Study design: TJA, IJN, SJW, MA, JPT. Data acquisition: IJN, SJW, MA, JPT, RAW, LMC, HT, BC, KC, LQL. Quality control of data and algorithms: SJW, MA, JPT, IJN. Data analysis and interpretation: TJA, IJN, SJW, MA, JPT, RAW, CR. Statistical analysis: IJN, SJW, JPT, RAW, CR. Manuscript preparation: TJA, IJN, SJW, MA, JPT. Manuscript editing: TJA, IJN, SJW, MA, JPT, RAW, CR, KC. Manuscript review: TJA, IJN, SJW, MA, JPT, RAW, CR, KC
CRediT authorship contribution statement
Sophie J. Warlow: Formal analysis, Writing – original draft, Conceptualization, led on patient recruitment, generated and analysed data, Writing – review & editing. Martyna Adamowicz: Formal analysis, Writing – original draft, Conceptualization, led on patient recruitment, generated and analysed data, Writing – review & editing. John P. Thomson: Formal analysis, Writing – original draft, Conceptualization, generated and analysed data, Writing – review & editing. Robert A. Wescott: Formal analysis, Writing – original draft, generated and analysed data., Writing – review & editing. Christelle Robert: Formal analysis, Writing – original draft, generated and analysed data., wrote the manuscript. Lara M. Carey: Formal analysis, generated and analysed data. Helen Thain: led on patient recruitment. Kate Cuschieri: Formal analysis, generated and analysed data. Lucy Q. Li: Formal analysis, generated and analysed data. Brendan Conn: Formal analysis, generated and analysed data. Ashley Hay: led on patient recruitment. Iain J. Nixon: Formal analysis, Writing – original draft, Conceptualization, led on patient recruitment, generated and analysed data., wrote the manuscript. Timothy J. Aitman: Formal analysis, Writing – original draft, Conceptualization, generated and analysed data, Writing – review & editing.
Declaration of competing interest
The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:
TJA receives consultancy payments as Director of the company BioCaptiva. All remaining authors have declared no conflicts of interest.
Acknowledgements
We extend our thanks to the patients who contributed to this study. We thank the Lothian NRS Human Annotated Bioresource, NHS Lothian Department of Pathology, the Scottish HPV Reference Laboratory, Edinburgh Experimental Cancer Medicine Centre and NHS Lothian Clinical Genetics for clinical and laboratory support. We thank Peter G Morrice for laboratory assistance and Sophie Marion De Procé, Robert L Hollis and Simon C Herrington for comments on the manuscript.
Appendix A. Supplementary data
The following are the Supplementary data to this article:
Deintensification of adjuvant treatment after transoral surgery in patients with human papillomavirus-positive oropharyngeal cancer: the conception of the PATHOS study and its development.
Radiotherapy plus cisplatin or cetuximab in low-risk human papillomavirus-positive oropharyngeal cancer (De-ESCALaTE HPV): an open-label randomised controlled phase 3 trial.
FDG-PET/CT for treatment response assessment in head and neck squamous cell carcinoma: a systematic review and meta-analysis of diagnostic performance.
Neck dissection after chemoradiotherapy for oropharyngeal and hypopharyngeal cancer: the correlation between cervical lymph node metastasis and prognosis.
Rapid clearance profile of plasma circulating tumor HPV type 16 DNA during chemoradiotherapy correlates with disease control in HPV-associated oropharyngeal cancer.
Circulating human papillomavirus DNA detected using droplet digital PCR in the serum of patients diagnosed with early stage human papillomavirus-associated invasive carcinoma.
HPV circulating tumoral DNA quantification by droplet-based digital PCR: a promising predictive and prognostic biomarker for HPV-associated oropharyngeal cancers.
Heterogeneity of p16 immunohistochemistry and increased sensitivity of RNA in situ hybridization in cytology specimens of HPV-related head and neck squamous cell carcinoma.
Evaluation of human papilloma virus diagnostic testing in oropharyngeal squamous cell carcinoma: sensitivity, specificity, and prognostic discrimination.
00377-8&id=gr1.jpg){kind=link}
00377-8&id=gr2.jpg){kind=link}
00377-8&id=gr3.jpg){kind=link}
00377-8&id=gr4.jpg){kind=link}
00377-8&id=gr5.jpg){kind=link}
00377-8&id=gr6.jpg){kind=link}