ccr-23-3941-html.html

Title: Cell-free human papillomavirus (HPV) DNA is a sensitive biomarker for prognosis and for

early detection of relapse in locally advanced cervical cancer

Authors: Lars Sivars*

1,2

, Cecilia Jylhä

1,3

, Ylva Crona Guterstam

4,5

, Mark Zupancic

6,7

, Britta

Lindqvist

, Magnus Nordenskjöld

1,3

, Emma Tham

1,3

and Kristina Hellman

6,8

*corresponding author

Department of Molecular Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden.

Department of Immunology, Genetics and Pathology, Uppsala University, Uppsala, Sweden

Dept of

Clinical Genetics and Genomics, Karolinska University Hospital, Stockholm,

Sweden.

Department of Clinical Science, Intervention and Technology, Karolinska Institutet,

Stockholm, Sweden

Department of Gynaecology and Reproductive medicine, Karolinska University Hospital,

Huddinge, Sweden

Dept. of Oncology-Pathology, Karolinska Institutet, Stockholm, Sweden,

Medical Unit Head, Neck, Lung, and Skin Cancer, Theme Cancer, Karolinska University

Hospital, Stockholm, Sweden

Department of Gynaecologic Cancer, Theme Cancer, Karolinska University Hospital,

Stockholm, Sweden

Corresponding author: Dr Lars Sivars, Department of Molecular Medicine and Surgery,

Karolinska Institutet,

K1 MMK Klinisk genetik, 171 76 Stockholm

, Sweden,

lars.sivars@ki.se

The authors declare no potential conflicts of interest.

Running title: Cell-free HPV DNA as a biomarker in cervical cancer

Downloaded from http://aacrjournals.org/clincancerres/article-pdf/doi/10.1158/1078-0432.CCR-23-3941/3448290/ccr-23-3941.pdf by CDL - University of California - San Diego user on 01 May 2024

Statement on translational relevance

Our results indicate that detection of cell-free HPV DNA (ctHPV DNA) in plasma 1-4 months after

finished treatment of locally advanced cervical cancer could in the future be used as a biomarker for

poor treatment response and an increased risk for relapse. Patients with a positive ctHPV DNA-post-

treatment plasma could thus potentially be offered additional treatment or an intensified follow-up

program.

Furthermore, during surveillance, detection of ctHPV DNA appears to be an early indicator of

relapse. In this study,

ctHPV DNA was found in plasma before relapse was diagnosed on radiology in

all patients (n=10) who experienced relapse after complete response to treatment, with a median

315 days lead time. In the future, p

atients with a ctHPV DNA-positive plasma during follow-up could

thus be offered e.g. an extra radiologic examination, potentially leading to detection of relapse

earlier than by todays follow-up regimens. This in a simple blood test.

Downloaded from http://aacrjournals.org/clincancerres/article-pdf/doi/10.1158/1078-0432.CCR-23-3941/3448290/ccr-23-3941.pdf by CDL - University of California - San Diego user on 01 May 2024

Abstract

Background: Human papillomavirus (HPV) is the cause of the majority of cervical cancer cases and

has been showed to be released as cell-free tumour DNA (ctHPV DNA) into the circulation. We here

analyse if ctHPV DNA could be used as a prognostic biomarker and/or to detect relapse earlier than

traditional methods in locally advanced cervical cancer (LACC).

Patients and methods: 74 patients with LACC were included, 66/74 were positive for 13 high-risk

HPV-types using a bead-based assay on tumour biopsies. HPV-type-specific droplet digital PCR

(ddPCR) assays were developed. Longitudinal plasma samples were then analysed for the biopsy-

verified HPV-type for each patient. 418 plasma samples were analysed. Patients were followed for a

median of 37 months. Results were correlated to tumour- and clinical characteristics.

Results: 92.4% of pre-treatment plasma samples were positive for ctHPV DNA. Persistent ctHPV DNA

in end-of-treatment, early follow-up (1-2 months after end-of-treatment) or tumour evaluation (3-4

months after end-of-treatment) plasma was correlated with worse progression-free survival (p <

0.001) compared to if ctHPV DNA was not found. The positive predictive value of ctHPV-status at

early follow-up for predicting disease progression was 87.5% and the negative predictive value was

89.3%. ctHPV DNA was found in plasma before relapse was diagnosed on radiology in all patients

(n=10) who experienced relapse after complete clinical response to treatment with a median 315

days lead time.

Conclusion: ctHPV DNA in follow-up plasma is a promising prognostic biomarker in patients with

LACC, useful for analysis of response to therapy and for early detection of relapse.

Keywords:

cell-free DNA, human papillomavirus, liquid biopsy, cervical cancer, cell-free tumor DNA,

biomarker

Downloaded from http://aacrjournals.org/clincancerres/article-pdf/doi/10.1158/1078-0432.CCR-23-3941/3448290/ccr-23-3941.pdf by CDL - University of California - San Diego user on 01 May 2024

Introduction

Cells release small amounts of their genomic material into the circulation, so called cell-free DNA

(cfDNA), either when the cells die or by active transport (1). cfDNA from cancer cells, so called cell-

free tumour DNA (ctDNA) can thus, with ultra-sensitive techniques, be analysed in a blood sample,

sometimes denoted as liquid biopsy (1). In order to be sensitive enough, these techniques are

usually targeted such as droplet digital PCR (ddPCR) or next generation sequencing (NGS) with error

suppression (1). Cervical cancer (CC) is one of the most common malignancies among women

worldwide and the 12 high-risk human papillomavirus types (HR-HPV) are the cause of the majority

of cases, while a few more putative high-risk types have been suggested (2). HR-HPV can exist as

episomes or can be integrated into the host genome and is often present in multiple copies per

cancer cell (3). Cell-free tumour HPV DNA (ctHPV DNA) is thus a possible biomarker in cervical

cancer. Indeed, in a pilot study on HPV16/18/45-positive CC, we showed that ctHPV16/18/45 DNA

was present in 94.4% of pre-treatment plasma from 18 patients with locally advanced cervical

cancer (LACC) and in 26.7% from 15 patients with early-stage CC, but not in any plasma from 21

patients with pre-malignant lesions of CC (4). Furthermore, patients with LACC and persistent ctHPV

DNA at end of treatment had worse progression-free survival (PFS) compared to patients without

detectable ctHPV DNA (4). In addition, the normal cfDNA level was studied, using cfAlbumin DNA

(cfALB) as a surrogate marker and high normal cfDNA load in pre- or end-of-treatment plasma was

correlated to poor PFS (4).

Similar results to ours regarding ctHPV DNA have been shown in a few studies in both cervical cancer

and other HPV-related malignancies (5-13) and ctHPV DNA is thus a very promising prognostic

biomarker. However, larger studies also including follow-up samples are lacking, and needed, before

ctHPV DNA can be used clinically. Of note, the possibility of using ctHPV DNA for surveillance during

follow-up in order to detect relapse early, is intriguing, but has not been studied extensively.

Therefore, in this study we expand the cohort of LACC, to also include follow-up samples, and

analyze plasma samples for 13 HR-HPV-types using ddPCR. Levels of ctHPV DNA and normal cfDNA

were then correlated to patient and tumour characteristics as well as outcome. Patients with pre-

malignant lesions of CC are used as a control group.

Patients, Material & Methods

Patients and samples

Consecutive patients diagnosed with locally advanced cervical cancer at the Karolinska University

Hospital, Stockholm, Sweden between 2019 and 2021 were included in the study. Of 112 patients

accessed for eligibility 66 were included in the analyses, see Figure 1. Treatment was individualized

for each patient but the standard treatment of the patients with LACC consisted of

chemoradiotherapy (CRT) including 6 weeks of external-beam radiotherapy (RT), 6 weekly cisplatin

cycles and 3-4 fractions of brachytherapy. 51 patients received CRT while 8 patients received

radiotherapy only due to co-morbidities, 4 patients had primary surgery followed by CRT, two

patients had primary chemotherapy (CT) followed by RT and one patient had primary CT followed by

CRT. Treatment was evaluated 3 months after end of treatment with an FDG-PET/CT for most of the

patients. The standard for detection of recurrences today also include clinical controls every 3-6

months and radiology once a year or if the patient present with any symptoms or clinical findings.

Malignant lesions are confirmed by cytology or histopathology in most cases.

Downloaded from http://aacrjournals.org/clincancerres/article-pdf/doi/10.1158/1078-0432.CCR-23-3941/3448290/ccr-23-3941.pdf by CDL - University of California - San Diego user on 01 May 2024

Tumour biopsies were collected at diagnosis and were used for HPV-typing. 418 plasma samples

were collected before, during, and at end-of treatment as well as during follow-up for up to 3 years.

Patients were followed until May 16th 2023 for a median of 37 months (range 17 – 50 months).

Furthermore, 15 patients treated with cone biopsy at the Karolinska University Hospital, Stockholm,

Sweden in 2022, for HPV16 and/or HPV18 verified high-grade squamous intra-epithelial lesion (HSIL),

low-grade squamous intra-epithelial lesion (LSIL) or adenocarcinoma in situ (AIS) were used as a

control group. Plasma was collected before the procedure and tested for both HPV-types if double-

positive. For 2 patients the histopathology analysis on the cone showed invasive carcinoma, see

Table 1. The study was approved by the Ethical Review Board at Stockholm County and written

informed consent was obtained from all patients prior to inclusion. The study was

conducted in

accordance with the Declaration of Helsinki.

HPV-typing

DNA had previously been extracted from tumour biopsies using the EZ1 DNA Tissue Kit (Qiagen,

Hilden, Germany) and stored in −20 °C. The biopsies were analysed using an established multiplex

bead-based assay (Luminex) for 27 HPV-types, including all HR-HPV-types, with

β-globin as an

internal control for presence of amplifiable DNA

as previously described (14-16).

Development of ddPCR-assays

Type-specific ddPCR-assays for the E7 gene of HPV16, 18 and 45, respectively, had previously been

developed (4). Additional type-specific ddPCR-assays were now developed for HPV 31, 33, 35, 39, 51,

52, 56, 58, 59 and 66, as previously described (4), see also Supplementary Methods for details on

assay design. HPV-assays were individually combined with an assay for cell-free albumin DNA (cfALB)

on a separate fluorescent channel, as a surrogate marker for normal cell-free DNA and as an internal

control for presence of amplifiable cfDNA (4). See Supplementary Table S1 for primers and probes

used (Integrated DNA Technologies, Coralville IA, USA).

The assays were first tested on serial

dilutions of plasmids with cloned reference DNA-genomes (obtained from the The International

HPV-Reference Center, Stockholm, Sweden), see supplementary and

Supplementary

Table S2 for

details) as well as normal genomic DNA and non-template controls (NTCs) using

ddPCR Multiplex

Supermix on a QX200 ddPCR System (all BioRad, Hercules, CA, USA).

The ddPCR QX Manager analysis

system (BioRad) uses Poisson distribution for absolute quantification providing a concentration and

95% confidence intervals.

The Limit of Blank (LoB) and limit of detection (LoD) of our assays was

calculated as previously described (17)

The LoB was 0 for all HPV-assays except HPV31 (LoB 1.58)

and HPV58 (LoB 1.94) based on average 12 reactions per HPV-type. Thus, the limit of detection (LoD)

was 3 for all HPV types except for

HPV31 (LoD 5.4 and HPV58 (LoD 6.0). Using 3 molecules as a cut-

off has also previously been used to determine ctHPV DNA-positivity (5, 9 ,11). All assays performed

well in the dilution series of the above-mentioned plasmids and showed linearity down to at least 5-

10 copies HPV DNA (Supplementary Figure S1).

Analysis of cell-free DNA

Blood samples were collected in Cell-Free DNA BCT tubes (Streck). Plasma was isolated using two

centrifugation steps and then stored at -80°C. Cell-free DNA was isolated from 3ml plasma per

sample using the QIAamp Circulating Nucleic Acid Kit (Qiagen) according to manufacturers’

instructions and eluted in 40ul AVE buffer.

Samples were then analysed using the above-mentioned ddPCR-assays and ddPCR-system, according

to manufacturers’ instructions.

For each patient, only the HPV-type found in the patient’s tumour

Downloaded from http://aacrjournals.org/clincancerres/article-pdf/doi/10.1158/1078-0432.CCR-23-3941/3448290/ccr-23-3941.pdf by CDL - University of California - San Diego user on 01 May 2024

biopsy was analysed. Samples were run in triplicates, 11ul of eluted ctDNA, 5ul of ddPCR Multiplex

Supermix, 900nM primers and 250nM probes per well, with dH20 to make it a 20ul reaction. The

total number of droplets in the three wells were combined for the Poisson calculation of

concentration and quantification of ctHPV DNA. If one of the triplicates failed, the average from the

other two were used (n= 2). If more than one of the triplicates failed the sample was re-run (n = 3).

The following 20ul PCR protocol was used: 95°C for 10 minutes followed by 40 cycles of 30 seconds

at 94°C, 30 seconds at 57.3°C and 30 seconds at 60°C and then finally 10 minutes at 98°C.

Fluorescence intensity thresholds were set manually for each HPV-type using plasmids with cloned

HPV-reference genomes of the respective HPV-type as a reference and as a positive control. Plasma

from anonymized blood donors, extracted from peripheral blood using QIAamp DNA Blood Maxi kit

(Qiagen, Germany) and then processed in the same manner as above, was used as negative controls

in addition to non-template controls and nuclease-free water as blanks. Samples were deemed

ctHPV DNA-positive if at least 3 molecules per sample were ctHPV DNA-positive for all HPV types

except HPV31 (5.4 molecules) and HPV58 (6.0 molecules) and deemed negative if less than 3

molecules per sample were ctHPV DNA-positive and normal cfDNA levels were >333 copies/ml. If

less <333 copies normal cfDNA, samples were excluded (n = 2).

Statistical analysis

SPSS (IBM) (RRID:SCR_002865) was used for statistical analysis. The Kaplan-Meier method was used

to calculate survival curves and the Log-Rank test was used to calculate the difference between

groups. Hazard Ratios (HR) were calculated using Cox regression. For comparing cfDNA between

groups, non-parametric tests, i.e. Mann-Whitney U or Kruskal-Wallis were used. All p-values

reported were 2-sided and p < 0.05 was considered as statistically significant. Progression-free

survival (PFS) was defined as time from start of treatment to verified progression in disease or death

by any cause. Disease-free survival (DFS) was defined as time from end of treatment to verified

progression in disease or death by any cause, with patients not having complete response to

treatment at tumour evaluation excluded from analysis. Early- mid- and late treatment timepoints

were defined as thirds of the normal treatment time for patients treated with CRT or RT. For other

treatment regimes, early-treatment was defined as during the first treatment modality, mid-

treatment was defined as after the first and before the second treatment modality, and late-

treatment defined as during the second modality. End-of-treatment was defined as -4 to +14 days

after last date of treatment. Early follow-up was defined as 1-2 months after the last date of

treatment and tumour evaluation was defined as 3 - 4 months after last date of treatment. Late

follow-up was defined as more than 4 months after last date of treatment. Complete remission (CR)

was defined as no sign of tumour at tumour evaluation. Disease progression was defined as either

progression of a verified tumour/metastasis on CT or occurrence of a new metastasis on radiology

(verified by cytology in most cases and, if that was not possible (n=1), when treatment for recurrent

disease was decided). Positive predictive value (PPV) was defined as the number of patients during a

certain time period with at least one ctHPV DNA-positive sample experiencing disease progression

divided by the total number of patients with at least one ctHPV DNA-positive sample during the

same period. Negative predictive value (NPV) was the defined as the number of patients with only

ctHPV DNA-negative samples during a time period not experiencing disease progression divided by

the total number of patients with only ctHPV DNA-negative samples during the same time period.

Data availability statement

Downloaded from http://aacrjournals.org/clincancerres/article-pdf/doi/10.1158/1078-0432.CCR-23-3941/3448290/ccr-23-3941.pdf by CDL - University of California - San Diego user on 01 May 2024

The data generated in this study are not publicly available due to information that could

compromise patient privacy or consent but are available upon reasonable request from the

corresponding author.

Results

Patients and HPV-type distribution

In total, 74 patients were included in the study. 66 patients (89%) had HPV-positive biopsies and

could be further studied. Not all 66 patients had plasma samples available at all time points. See

Figure 1 for a flow-chart of inclusion and exclusion and number of patients with available samples at

different time points, and Table 1 for a summary of patient characteristics. The majority had

squamous cell carcinoma (SCC) (Table 1). Tumour biopsies from 8 patients (11%) were negative for

all 27 HPV types: 2 SCC, 2 adenocarcinomas, 2 adenocarcinomas of gastric type, one neuroendocrine

tumour and one adenocarcinoma of mesonephric type. HPV16 was found in 29 of 66 (43.9%) HPV-

positive biopsies and HPV18 was found in 11 (16.7%), HPV 33 and 45 in 5 each (7.6%), HPV31 and 39

in 4 each (6.1%), HPV52, 56 and 58 in 2 each (3.0%) and HPV35 and 66 in 1 (1.5%) biopsy each.

biopsies were positive for more than one HPV-type.

ctHPV DNA detection in plasma

Pre-treatment

Pre-treatment plasma was available from 53/66 (80.3%) patients and each sample was analysed for

the HPV-type detected in the paired tumour biopsies, Figure 1. ctHPV DNA was detected in 49/53

(92.4%) cases, median 58 copies/ml, range 0 – 37008 copies/ml (Figure 2). Higher ctHPV DNA copy

number was significantly correlated to more advanced FIGO 2018 stage (Table 2, p = 0.020), but not

to tumour size (Supplementary Table S3, p = 0.103), age (Supplementary Table S3, p= 0.409), HPV-

type (p = 0.126) or histopathological type (p = 0.128) when comparing SCC to adenocarcinomas.

During treatment

90 samples from 59 patients collected during treatment were available, Figure 1. 23, 35 and 30

patients had at least one sample available from early-, mid- and late treatment, respectively. The

percentages of patients with detectable ctHPV DNA in at least one sample declined from 91.3% early

in treatment (median of the ctHPV DNA-positive samples 58.9 copies/ml, range 1.4 – 10 295), to

57.3% mid-treatment (median 3.93 copies/ml, range 1.4 – 36 155), to 46.7% late in treatment

(median 2.5 copies/ml, range 1.3 – 16 238) (Figure 2). 9/17 (52.9%) patients with available samples

from both pre- and early treatment had an increase in ctHPV DNA level early in treatment compared

to pre-treatment. 7/9 (77.8%) then decreased during treatment while 2/9 (22.2%) continued to

increase.

End of treatment and follow-up

End-of-treatment plasma was available from 52 patients, Figure 1. 14/52 (26.9%, median 3.7

copies/ml, range 1.2 – 32 000) had detectable ctHPV DNA (Figure 2). Early follow-up plasma was

available from 36 patients (37 samples), Figure 1; 8/36 (22.2%) had detectable ctHPV DNA (median

6.8, range 1.7 – 38 607 copies/ml). Tumour evaluation plasma was available from 49 patients (56

samples), Figure 1; 12/49 (24.5%) patients had at least one sample with detectable ctHPV DNA

(median 4.2, range 1.7 – 204 106 copies/ml). If combining early follow-up and tumour evaluation

periods, 61 patients had at-least one sample available (in total 93 samples) and 17/61 (27.9%)

patients had at least one ctHPV DNA-positive sample. 130 late follow up samples were available

Downloaded from http://aacrjournals.org/clincancerres/article-pdf/doi/10.1158/1078-0432.CCR-23-3941/3448290/ccr-23-3941.pdf by CDL - University of California - San Diego user on 01 May 2024

from 46 patients, 17/46 (37.0%) patients had at least one ctHPV DNA-positive sample during follow-

up (Figure 2). In total, 27 late follow up samples were ctHPV DNA-positive (median 15.9, range 1.33 –

896 copies/ml).

ctHPV DNA as a prognostic marker

ctHPV DNA detection or not in pre-treatment samples did not correlate to PFS (p = 0.504,

Supplementary Figure S2a) or DFS (p = 0.958). When dividing patients with ctHPV DNA-positive pre-

treatment samples into thirds depending on the level of ctHPV DNA, the results did not correlate to

PFS (p = 0.349, Supplementary Figure S2b) or DFS (p = 0.880). Detection or not of ctHPV DNA during

treatment did not correlate to PFS (p = 0.808) or DFS (p = 0.351). See Figure 2 for positive predictive

values (PPV) and negative predictive values (NPV) for ctHPV DNA-status to predict disease

progression, as well as percentage of ctHPV DNA-positive patients and the number of patients and

samples per time point.

Patients with persistent ctHPV DNA at end of treatment had significantly worse PFS (p = 0.018,

Figure 3a) than patients without detectable ctHPV DNA, but the difference did not reach statistical

significance for DFS (p = 0.054, Supplementary Figure S3a) with a median follow-up time of 37

months (range 17 – 50 months). The Hazard Ratio (HR) was 2.921 for PFS (95% CI: 1.149 – 7.428).

PPV and NPV for prediction of disease progression was 57.1% (95% CI: 29.6%-81.2%) and 73.7% (95%

CI: 56.6% - 86.0%), respectively, Figure 2.

Patients with at least one ctHPV DNA-positive sample during early follow-up (n=36) had statistically

significantly worse PFS (p < 0.001, Figure 3b) and DFS (p < 0.001, Supplementary Figure S3b)

compared to patients with only ctHPV DNA-negative samples. HR for PFS was 18.078 (95% CI: 4.451-

73.427) and for DFS 17.843 (95% CI: 2.877-110.681). The PPV was 87.5% (95% CI: 46.7%-99.3%) and

NPV 89.3% (95% CI: 70.6%-97.2%). Patients with at least one ctHPV DNA-positive sample during

tumour evaluation (n=49) had statistically significantly worse PFS (p < 0.001, Figure 3c) and DFS (p <

0.001, Supplementary Figure S3c) compared to patients with only ctHPV DNA-negative samples. HR

for PFS was 10.361 (95% CI: 3.740-28.701) and for DFS 10.187 (95% CI: 2.656-39.074). The PPV was

85.7% (95% CI: 56.1%-97.5%) and NPV 82.9% (95% CI: 65.7%-92.8%). If the early follow-up and

tumour evaluation groups are combined (n=61), the results are very similar with p < 0.001 for both

PFS and DFS, see Supplementary Figures S4a and S4b. See also Supplementary Figure S5 for a

comparison with a Kaplan-Meier plot of PFS in patients stratified using CR and non-CR at tumour

evaluation.

For patients with at least one ctHPV DNA-positive sample during late follow-up the PPV for

predicting relapse during the study period was 93.3% (95% CI: 66.0% – 99.7%) and the NPV for

patients with only ctHPV DNA-negative samples was 96.7% (95% CI: 80.9% – 99.8%), Figure 2.

Temporal detection of ctHPV DNA in plasma in relation to detection of disease progression

23/66 (34.8%) patients experienced disease progression, while one patient died from a co-morbidity.

10/23 (43.5%) patients who experienced disease progression were deemed to have had clinical

complete response to treatment and later experienced relapse (6 distant, 4 locoregional). In all 10

cases, ctHPV DNA was found in plasma before diagnosis of relapse (range 122-578 days, median 315

days, mean 280 days), at end-of-treatment and/or during follow-up. 4/10 (40%) patients with clinical

complete response, were ctHPV-positive at all time points measured, while 5/10 (50%) patients had

cleared the ctHPV DNA by end-of-treatment and one patient, with a stage IIIC1r adenocarcinoma,

Downloaded from http://aacrjournals.org/clincancerres/article-pdf/doi/10.1158/1078-0432.CCR-23-3941/3448290/ccr-23-3941.pdf by CDL - University of California - San Diego user on 01 May 2024

had been ctHPV DNA-negative in all 10 previous samples, including the pre-treatment one, see

Figure 4.

13/23 (56.5%) patients who experienced disease progression were not deemed to have had

complete clinical response to treatment. Three patients were diagnosed with disease progression

during or just at the end of treatment and all three had high, and increasing, levels of ctHPV DNA

during treatment, Supplementary Figure S6. Nine patients were diagnosed with disease progression

(2 distant, 7 locoregional) at early follow-up/tumour evaluation or shortly thereafter and one

relapsed after 13 months (Supplementary Figure S7). In 7/10 (70%) patients ctHPV DNA was found in

plasma before diagnosis of relapse: at end-of-treatment or during early follow-up (28, 35, 73, 79, 90,

106 and 127 days before, respectively. In 2/10 (20%) patients ctHPV DNA was not found at end-of-

treatment and no plasma samples were collected before relapse. HPV ctDNA was then found after

diagnosis of relapse (10 and 122 days respectively, Supplementary Figure S7). In one case of a

singular pelvic lymph node metastasis, no ctHPV DNA was found, either at end-of-treatment or at

follow-up (Supplementary Figure S7).

ctHPV DNA-positive follow-up samples, but no diagnosed relapse

In total, 4 patients had ctHPV DNA-positive early- or late follow-up samples (three patients had one,

and one patient had two positive samples), but have not been diagnosed with relapse,

Supplementary Figure S8. ¾ (75%) patients had however, a suspected viable local tumour or

metastasis on radiology near the time of the positive ctHPV DNA sample, that spontaneously

regressed. The ctHPV DNA was then cleared in 2/3 (33.3%) cases, while the third patient has not

contributed any more samples. The remaining patient with at least one ctHPV DNA-positive follow-

up sample had not had any radiologic examination within 6 months of the positive sample and

subsequently cleared the ctHPV DNA.

In addition, one patient who has not been diagnosed with disease progression, had a ctHPV DNA-

positive sample at end-of-treatment, but has not contributed any more samples. PET-CT at tumour

evaluation showed partial response, Supplementary Figure S8, light blue line.

ctHPV DNA-negative follow-up samples and no diagnosed relapse

The remaining 38 patients had only ctHPV DNA-negative samples at end-of-treatment and at early-

and late follow-up and have not developed disease progression, median follow-up time 37.5 months,

range 18.2 – 50.0 months (Supplementary Figure S9).

ctHPV DNA detection in the control group

ctHPV DNA was not found in any pre-treatment plasma from the 15 patients in the control group

with pre-malignant cervical lesions, including the 2 patients where the histopathology analysis on

the removed cone showed invasive cancer.

Normal cell-free DNA detection in plasma and its correlation to prognosis

Pre-treatment levels of normal cfDNA (using cfALB as a surrogate marker) were higher in the cancer

patient group compared to the control group (median 1263 and 648 copies/ml, respectively, p <

0.001). However, in cancer patients there were no statistical correlation of normal cfDNA to stage,

HPV-type, histopathology or outcome (PFS or DFS), see Supplementary Results and Supplementary

Table S4.

Discussion

Downloaded from http://aacrjournals.org/clincancerres/article-pdf/doi/10.1158/1078-0432.CCR-23-3941/3448290/ccr-23-3941.pdf by CDL - University of California - San Diego user on 01 May 2024

We here demonstrate that detectable ctHPV DNA in end-of-treatment and early-follow-up plasma is

a promising prognostic biomarker in locally advanced cervical cancer and that ctHPV DNA shows

great promise for relapse surveillance during follow-up. This has been noted in other HPV-related

malignancies (e.g. head and neck- and anal cancer (6,7)) but this is one of the first large studies

investigating ctHPV DNA in long term follow-up samples in LACC.

We found ctHPV DNA in 92.4% of pre-treatment plasma which is higher than in some other recent

studies using ddPCR (5,8-10) or NGS (13), potentially due to a larger starting volume plasma (3 ml)

than some of the studies and thus a greater chance of finding the sometimes miniscule amount of

ctHPV DNA. This is especially important during follow-up when the ctHPV DNA levels in the positive

samples were very low. Pre-treatment ctHPV DNA levels were correlated to stage, as also noted in

another recent study (13), but not to prognosis, which also has been noted before (5,10). ctHPV DNA

levels decreased during treatment in most patients. However, some patients had an initial increase

in ctHPV DNA level early in the treatment, something previously noted in head- and neck cancer (6)

which might be due to the increased cell death during the initial stages of radiotherapy (18). Thus, a

wash-out phase might exist before true response to treatment can be analysed using ctHPV DNA. By

mid-treatment all but one, and by late-treatment all, patients responding to treatment had

decreased levels of, or had cleared, the ctHPV DNA.

Strikingly however, is that patients with ctHPV DNA-positive plasma at end of treatment, during

early follow-up or at tumour evaluation had significantly worse PFS and DFS compared to patients

with negative ctHPV DNA-status. The PPV for predicting a relapse increased from 57% at end of

treatment to 86% and 88% during early follow-up and tumour evaluation respectively, suggesting

that waiting a few weeks after finished treatment before analysing ctHPV DNA might be beneficial.

However, it does not seem to matter if the sample is taken after 1-2 or 3-4 months. Analysing ctHPV

DNA during this time period might thus be a useful adjunct to standard follow-up and could

potentially aid in determining response to treatment and identifying those in need of additional

treatment. Furthermore, analysis of ctHPV DNA in a blood sample might help identify patients at

high risk for developing relapse who need more intensive monitoring or, on the contrary, identify

patients with low risk who could perhaps be offered less frequent radiological surveillance.

Another potential use for ctHPV DNA analysis is for surveillance during follow-up. In this study, 10

patients deemed to have had complete response to treatment later developed a relapse. Strikingly,

in all 10 cases, ctHPV DNA was found in plasma, on average 9 months, before the relapse was

diagnosed clinically. Similar results to ours have been showed in a recent study in cervical cancer (5).

Being able to diagnose relapse and start treatment months earlier might be desirable and could

potentially confer a survival benefit for these patients. Furthermore, a simple blood test would be

easy and affordable to implement in clinical practice. Still, it is too early to conclude that finding

recurrences earlier would lead to a difference in outcome, and if so be the case, how much earlier

would be clinically relevant, However, with recent advances in treatment of recurrences such as

Stereotactic Body Radiation Therapy (SBRT), more advanced surgery and new drugs, e.g.

immunotherapy, this is something that should be investigated further.

However, we also found four patients with positive ctHPV DNA plasma during follow-up who have

not been diagnosed with a relapse. It is noteworthy that 3/4 of these patients had a suspected

malignant lesion on CT that spontaneously regressed. Thus, ctHPV DNA alone will likely not be

sufficient to diagnose a relapse, but a ctHPV DNA-positive plasma during follow-up should in the

future warrant an additional CT or PET-CT and intensified follow-up

Downloaded from http://aacrjournals.org/clincancerres/article-pdf/doi/10.1158/1078-0432.CCR-23-3941/3448290/ccr-23-3941.pdf by CDL - University of California - San Diego user on 01 May 2024

Another possible use for ctHPV DNA is when a suspected malignant lesion is found on CT but other

conditions or co-morbidities do not allow for a biopsy to be taken. ctHPV DNA status could then

potentially be used to guide clinical decisions. See Figure 4 (dark blue line) for an example where this

approach could have been used. This patient had a lung lesion on CT 4 months after completion of

treatment, but a biopsy could not be taken. Instead, the lesion was followed using CT. After 14

months the lesion started to grow, the relapse could therefore be confirmed and the patient

received treatment. However, plasma samples were positive for ctHPV DNA 16 months before the

lesion was found on CT. Thus, ctHPV DNA might be of great help in differentiating radiological

changes such as post-radiotherapy effects or reactive lesions from true relapse/progression and

could perhaps be used as a complement to or even replace repeated imaging in some cases (19).

ctHPV DNA was not found in any of the pre-treatment samples from patients initially diagnosed with

pre-malignant lesions of cervical cancer, including the 2 patients later diagnosed with early-stage

invasive carcinoma. This is in line with our pilot study where we investigated another 21 patients

using the same method as in this paper, as well as with all other recent studies using ddPCR on pre-

malignant lesions (4,8,11).

A limitation with the study was that samples were prospectively collected during the Covid-19

pandemic. This occasionally made sample collection difficult during follow-up where physical visits

were sometimes replaced by telephone appointments, leading to some patients not contributing

samples at all time points. Despite this, a total of 418 samples were collected, making this the ctHPV

DNA-study with the largest amount of follow-up samples in cervical cancer. Furthermore, ctHPV DNA

analysis is only possible in patients with HPV-positive tumours. Although the vast majority of cervical

cancers are HPV-positive (89% of our cohort), other biomarkers will be needed for surveillance of

the negative patients.

Conclusion

In summary, by using sensitive ddPCR assays for 13 HR-HPV-types on one of the largest reported

longitudinal follow-up cohorts of locally advanced cervical cancer (LACC), we show that ctHPV DNA is

a sensitive biomarker for cervical carcinoma. Furthermore, we can confirm that detectable ctHPV

DNA at end-of-treatment or during follow-up is a negative prognostic marker in LACC. Finally, we

show that detection of ctHPV DNA after clinical complete response to treatment is indicative of later

relapse with a median lead time of 10 months.

Acknowledgements

This work was supported by Karolinska Institutet (L. Sivars), Magnus Bergvall Foundation (L. Sivars),

Region Stockholm (E. Tham) and Cancerfonden (E. Tham). The authors would like to thank Stefan

Holzhauser and Torbjörn Ramqvist for help with HPV-typing of cervical biopsies and The

International HPV Reference Center (Karolinska Institutet, Stockholm, Sweden) for providing

plasmids with HPV reference genomes for method development. Recombinant clones of HPV-types

16, 18, 45, 51 and 52 were originally isolated by Prof. Dr. Ethel-Michele De Villiers, Deutsches

Krebsforschungszentrum, Heidelberg, Germany. HPV58 and HPV59 were originally isolated by T.

Matsukura at National Institute of Health, Japan, HPV31, 35 and 56 were originally isolated by Lorna

Sammoury at Qiagen and HPV 33 and 39 were originally isolated by Michel Favre at Institut Pasteur,

Paris, France.

References:

Downloaded from http://aacrjournals.org/clincancerres/article-pdf/doi/10.1158/1078-0432.CCR-23-3941/3448290/ccr-23-3941.pdf by CDL - University of California - San Diego user on 01 May 2024

Barbany G, Arthur C, Lieden A, Nordenskjöld M, Rosenquist R, Tesi B,

et al

. Cell-free

tumour DNA testing for early detection of cancer - a potential future tool. J Intern Med.

2019; 286(2): 118-136.

Monographs on the Evaluation of Carcinogenic Risks to Humans, No. 100B.

IARC Working Group on the Evaluation of Carcinogenic Risks to Humans. Biological

agents. IARC Monogr Eval Carcinog Risks Hum. 2012;100(Pt B):1-441.

Bodelon C, Untereiner ME, Machiela MJ, Vinokurova S, Wentzensen N. Genomic

characterization of viral integration sites in HPV-related cancers. Int J Cancer. 2016 Nov

1;139(9):2001-11.

Sivars L, Hellman K, Crona Guterstam Y, Holzhauser S, Nordenskjöld M, Falconer H,

et al

Circulating cell-free tumor human papillomavirus DNA is a promising biomarker in

cervical cancer. Gynecol Oncol. 2022 Oct;167(1):107-114.

Jeannot E, Latouche A, Bonneau C, Calmejane MA, Beaufort C, Ruigrok-Ritstier K,

et al

Circulating HPV DNA as a Marker for Early Detection of Relapse in Patients with Cervical

Cancer. Clin Cancer Res. 2021; 27(21): 5869-5877.

Chera BS, Kumar S, Beaty BT, Marron D, Jefferys S, Green R,

et al

. Rapid Clearance Profile

of Plasma Circulating Tumor HPV Type 16 DNA during Chemoradiotherapy Correlates

with Disease Control in HPV-Associated Oropharyngeal Cancer. Clin Cancer Res. 2019;

25(15): 4682-4690.

Bernard-Tessier A, Jeannot E, Guenat D, Debernadi A, Michel M, Proudhon C,

et al

Clinical Validity of HPV Circulating Tumor DNA in Advanced Anal Carcinoma: An Ancillary

Study to the Epitopes-HPV02 Trial. Clin Cancer Res. 2019; 25(7): 2109-2115

Bonlokke S, Stougaard M, Sorensen BS, Booth BB, Hogdall E, Nyvang GB,

et al

. The

diagnostic value of circulating cell-free HPV DNA in plasma from cervical cancer

patients.

Cell

. 2022;

(14):2170.

Cabel L, Bonneau C, Bernard-Tessier A, Hequet D, Tran-Perennou C, Bataillon G,

et al

HPV ctDNA detection of high-risk HPV types during chemoradiotherapy for locally

advanced cervical cancer. ESMO Open. 2021 Jun;6(3):100154

10.

Cheung TH, Yim SF, Yu MY, Worley Jr MJ, Fiascone SJ, Chiu RWK,

et al

. Liquid biopsy of

HPV DNA in cervical cancer.

J Clin Virol

. 2019;

114

: 32-36.

11.

Jeannot E, Becette V, Campitelli M, Calmejane, MA, Lappartient E, Ruff E,

et al

Circulating human papillomavirus DNA detected using droplet digital PCR in the serum of

patients diagnosed with early stage human papillomavirus-associated invasive

carcinoma. J Pathol Clin Res. 2016; 2(4): 201-209.

12.

Sivars L, Palsdottir K, Crona Guterstam Y, Falconer H, Hellman K, Tham E. The current

status of cell-free human papillomavirus DNA as a biomarker in cervical cancer and other

HPV-associated tumors: A review. Int J Cancer. 2023 Jun 1;152(11):2232-2242.

13.

Bønløkke S, Steiniche T, Sorensen BS, Nyvang GB, Lindegaard JC, Blaakaer J,

et al

Circulating cell-free HPV DNA is a strong marker for disease severity in cervical cancer.

Mol Oncol. 2023 Oct 19. doi: 10.1002/1878-0261.13538. Epub ahead of print. PMID:

37853962.

Downloaded from http://aacrjournals.org/clincancerres/article-pdf/doi/10.1158/1078-0432.CCR-23-3941/3448290/ccr-23-3941.pdf by CDL - University of California - San Diego user on 01 May 2024

14.

Schmitt M, Bravo IG, Snijders PJ, Gissman L, Pawlita M, Waterboer T. Bead-based

multiplex genotyping of human papillomaviruses. J Clin Microbiol. 2006 Feb;44(2):504-

15.

Schmitt M, Dondog B, Waterboer T, Pawlita M. Homogeneous amplification of genital

human alpha papillomaviruses by PCR using novel broad-spectrum GP5+ and GP6+

primers. J Clin Microbiol. 2008 Mar;46(3):1050-9

16.

Sivars L, Landin D, Rizzo M, Haeggblom L, Bersani C, Muck-Wikland E,

et al

. Human

papillomavirus (HPV) is absent in branchial cleft cysts of the neck distinguishing them

from HPV positive cystic metastasis. Acta Otolaryngol. 2018 Sep;138(9):855-858.

17.

Milbury CA, Zhong Q, Lin J, Williams M, Olson J, Link DR,

et al

. Determining lower limits

of detection of digital PCR assays for cancer-related gene mutations. Biomol Detect

Quantif. 2014 Aug 20;1(1):8-22.

18.

Lo YM., Leung SF., Chan LY, Chan AT, Lo KW, Johnson PJ,

et al

. Kinetics of plasma

Epstein-Barr virus DNA during radiation therapy for nasopharyngeal carcinoma. Cancer

Res. 60, 2351–2355 (2000)

19.

Sundstrom H, Blomqvist L, Hellman K. Uterine cervical carcinoma treated with

chemoradiotherapy: impact of three-month MRI follow-up on clinical management and

outcome. J Cancer Metastasis Treat 2021;7:65 DOI: 10.20517/2394-4722.2021.139

Downloaded from http://aacrjournals.org/clincancerres/article-pdf/doi/10.1158/1078-0432.CCR-23-3941/3448290/ccr-23-3941.pdf by CDL - University of California - San Diego user on 01 May 2024

Patients (

= 66) Control group

(

= 15)

Age, median
(range)

52 (29-84)

33 (25-47)

Stage, n (%)

Pre-malignant

13 (86.7)

5 (7.6)

2 (13.1)*

14 (21.2)

III

43 (65.1)

4 (6.1)

Type, n (%)

Squamous

57 (86.4)

Adenocarcinoma

6 (9.1)

2 (13.3)*

Clear cell

1 (1.5)

Sarcomatoid

1 (1.5)

Adenosquamous

1 (1.5)

HSIL

9 (60.0)

AIS

1 (6.7)

HSIL + AIS

3 (20.0)

HPV-type, n (%)

29 (43.9)

6 (40.0)

11 (16.7)

5 (33.3)

Non 16/18

26 (39.4)

16 & 18

4 (26.7)

Treatment, n (%)

CRT

51 (77.3)

8 (12.1)

Surgery + CRT

4 (6.1)

CT + CRT

1 (1.5)

CT + RT

2 (3.0)

Surgery

2 (13.3)*

Conization

13 (86.7)

Table 1. Patient- and lesion characteristics. CRT – chemoradiotherapy. RT – radiotherapy, CT –

chemotherapy. HSIL - high-grade squamous intra-epithelial lesion. LSIL - low-grade squamous intra-

epithelial lesion. AIS - adenocarcinoma in situ. *Originally diagnosed with a pre-malignant lesion

where conization subsequently showed minimally invasive carcinoma

Downloaded from http://aacrjournals.org/clincancerres/article-pdf/doi/10.1158/1078-0432.CCR-23-3941/3448290/ccr-23-3941.pdf by CDL - University of California - San Diego user on 01 May 2024

ctHPVDNA copies/ml

Median

Mean

FIGO 2018 Stage

I & II

III

697

3494

11004

p = 0.020

HPV-type

HPV16

2055

HPV18

1845

HPV non 16/18

217

p = 0.126

Histopathology

Squamous

1276

Adenocarcinoma

Sarcomatoid

Clear cell

12585

Adenosquamous

868

p = 0.128*

Total

1322

Table 2. Cell-free HPV DNA levels in pre-treatment plasma in relation to FIGO 2018 stage, HPV-type

and histopathology. * squamous vs adenocarcinoma

Downloaded from http://aacrjournals.org/clincancerres/article-pdf/doi/10.1158/1078-0432.CCR-23-3941/3448290/ccr-23-3941.pdf by CDL - University of California - San Diego user on 01 May 2024

Figure legends:

Figure 1. Flow chart of inclusion and exclusion and number of samples collected at each timepoint

for those included in the study (n=66). Early follow-up = 1-2 months after end-of-treatment. Tumour

evaluation = 3-4 months after end-of-treatment. Follow-up = 4+ months after end-of-treatment.

Figure 2. Percentage of patients with at least one ctHPV DNA-positive sample as well as the PPV

(Positive predictive value) and NPV (Negative predictive value) of ctHPV DNA-status as a predictor of

disease progression at different time-points in relation to treatment.

Not all patients had samples

available at all time points. Dotted lines indicate trends between different time points. Before –

before treatment start. Early- mid- and late – during treatment; these timepoints were defined as

thirds of the normal treatment time. End – end-of-treatment; defined as -4 to +14 days after last

date of treatment. Follow-up: more than 4 months after last date of treatment.

Figure 3. Kaplan-Meier graphs of comparisons between patients with ctHPV DNA-positive (dotted,

bottom lines) and ctHPV DNA-negative (solid, top lines) plasma at different time points. A:

Progression-free survival (PFS) in end-of-treatment plasma. B: PFS in early-follow-up plasma (1-2

months after finished treatment). C: PFS in tumour evaluation plasma (3-4 months after finished

treatment). For additional Kaplan-Meier curves on disease-free survival, see Supplementary Figure

S8.

Figure 4. Number of copies ctHPV DNA at different time periods for all patients (n = 10) that were

deemed to have complete response to treatment (i.e. with no tumour detected on FDG-PET

performed 3-4 months after end of treatment, time point designated with dotted line) and later

were diagnosed with relapse. In all 10 cases, ctHPV DNA was found in plasma before the diagnosis,

either around tumour evaluation at 3-4 months, or later (median 315, mean 280 days before

diagnosis of relapse). Arrows indicate approximate time of diagnosis of relapse. The standard for

detection of recurrences today include a FDG-PET/CT 3 months after finished treatment, clinical

controls every 3-6 months and radiology once a year or if the patient presents with any symptoms or

clinical findings.

Downloaded from http://aacrjournals.org/clincancerres/article-pdf/doi/10.1158/1078-0432.CCR-23-3941/3448290/ccr-23-3941.pdf by CDL - University of California - San Diego user on 01 May 2024