Purpose: Circulating cell-free (ccf) human papillomavirus (HPV) DNA may serve as a unique tumor marker for HPV-associated malignancies, including cervical cancer. We developed a method to genotype and quantify circulating HPV DNA in patients with HPV16- or HPV18-positive metastatic cervical cancer for potential disease monitoring and treatment-related decision making.

Experimental Design: In this retrospective study, HPV ccfDNA was measured in serum samples from 19 metastatic cervical cancer patients by duplex digital droplet PCR (ddPCR). Nine patients had received tumor-infiltrating lymphocyte (TIL) immunotherapy. ccfDNA data were aligned with the tumor HPV genotype, drug treatment, and clinical outcome.

Results: In blinded tests, HPV ccfDNA was detected in 19 of 19 (100%) patients with HPV-positive metastatic cervical cancer but not in any of the 45 healthy blood donors. The HPV genotype harbored in the patients' tumors was correctly identified in 87 of 87 (100%) sequential patient serum samples from 9 patients who received TIL immunotherapy. In three patients who experienced objective cancer regression after TIL treatment, a transient HPV ccfDNA peak was detected 2–3 days after TIL infusion. Furthermore, persistent clearance of HPV ccfDNA was only observed in two patients who experienced complete response (CR) after TIL immunotherapy.

Conclusions: HPV ccfDNA represents a promising tumor marker for noninvasive HPV genotyping and may be used in selecting patients for HPV type–specific T-cell-based immunotherapies. It may also have value in detecting antitumor activity of therapeutic agents and in the long-term follow-up of cervical cancer patients in remission. Clin Cancer Res; 23(22); 6856–62. ©2017 AACR.

Translational Relevance

Because HPV DNA is present in cervical cancer cells, we examined the feasibility of HPV ccfDNA detection in metastatic cervical cancer patients as a marker for the disease. Our data indicate that HPV ccfDNA can be effectively detected in serum samples from metastatic cervical cancer patients. We show that HPV ccfDNA is a reliable biomarker for HPV genotyping and may serve as a replacement for tumor biopsies in the selection of patients for HPV-targeted T-cell therapies. The detection of HPV ccfDNA may have additional value in long-term follow-up of cervical cancer patients in remission and as a pharmacodynamic biomarker for tumor cell death caused by anticancer therapies. In summary, HPV ccfDNA is a potential biomarker for detecting and monitoring metastatic cervical cancer and for genotyping patient tumors for HPV-targeted immunotherapies.

Despite major advances in early detection, including the Pap smear and co-human papillomavirus (HPV) testing (1), there are still an estimated 528,000 new cases of cervical cancer and 275,000 cervical cancer–related deaths worldwide, making it the fourth leading cause of cancer-related death in women (2). It is estimated that the 5-year survival rate is approximately 70% (3). With the routine Pap smear–based cytology test being largely insensitive for detecting recurrent and metastatic cervical cancer (4, 5), there is a need for a minimally invasive and specific test for disease monitoring, which could be beneficial for HPV-positive cancer patients with a risk of recurrent and metastatic disease. Furthermore, immunotherapies that target HPV oncoproteins are an attractive experimental strategy for the treatment of metastatic cervical cancer (6, 7). The oncoproteins of different HPV types have distinct sequences for the responses of T cells, and it is, therefore, important to correctly identify the tumor HPV genotype for patient enrollment in HPV-targeted T-cell therapy (8). Current genotyping methods require tissue specimens that may be difficult to acquire, thus presenting a need to genotype tumors with liquid biopsies for immunotherapies.

Circulating cell-free DNA (ccfDNA) has been widely evaluated as a liquid biopsy for detecting cancer, monitoring disease, characterizing drug targets, and uncovering resistance in various tumors (9–13). In colorectal cancer, ccfDNA was used to detect RAS pathway mutations prior to tumor progression by standard imaging (11, 14, 15) and to sense minimal residual disease as well as predict recurrence (16). In breast cancer, whole-genome sequencing revealed that ccfDNA could be detected in a significant percentage of advanced cancer patients and exhibited a greater dynamic range that appeared to be correlated with tumor burden (17). In prostate cancer, mutations in androgen receptors were found in plasma DNA and were found to be related to treatment sensitivity and resistance (18, 19). In lung cancer, the T790M mutation in EGFR was detected in the ccfDNA of many EGFR drug-resistant patients (20, 21). In addition, as the majority of ccfDNA is derived from apoptotic and necrotic tumor cells that have been shed into circulation (22), ccfDNA has a very short half-life (23). Another potential application of ccfDNA may be for detecting acute antitumor activity following the administration of investigational therapeutic agents in cancer patients.

HPV is pathogenic for cervical cancer, of which 70% are positive for HPV16 and HPV18 (24). Although HPV DNA detection has been routinely used for cervical cancer screening by Pap smear (25), an alternative source of samples will likely be necessary for the follow-up of patients with a potential for disease recurrence. There has been limited success in detecting plasma or serum HPV DNA for cervical cancer diagnosis due to a lack of sensitivity, as suggested in several studies (26–28). There is obviously a need to understand whether such a limitation is due to limited HPV DNA in blood circulation or whether new advances in DNA isolation and detection technologies would lead to improved detection. In this study, we established a method using an automated ccfDNA preparation system and duplex digital droplet PCR (ddPCR) assays for genotyping and quantifying HPV ccfDNA in sera from peripheral blood. With improved sensitivity and consistency, we assessed the potential of HPV ccfDNA as a noninvasive biomarker for patient selection and treatment follow-up in a T-cell immunotherapy clinical trial (6).

Study design

In this retrospective study, serum samples collected from HPV16- or HPV18-positive metastatic cervical cancer patients who were either screened for eligibility or enrolled in tumor-infiltrating lymphocytes (TIL) therapy at the U.S. National Cancer Institute were analyzed (6). Patient samples were de-identified, tested, and analyzed in an unbiased and blinded fashion.

Samples

All samples from the 21 cervical cancer patients were collected between July 2012 and November 2015 using an institutional review board-approved protocol at the U.S. NIH Clinical Center. Informed consent was obtained from all patients. Nineteen of the patients were confirmed to have HPV16- or HPV18-positive tumors. The 9 patients enrolled in the TIL clinical trial were treated with a 7-day nonmyeloablative chemotherapy followed by a single infusion of TIL (Fig. 2A; ref. 6). Tumor responses were determined according to the Response Evaluation Criteria In Solid Tumors (RECIST, version 1.0; ref. 6). A series of serum samples from each patient were collected at both pre- and posttreatment time points. Serum samples from 45 healthy donors were used as negative controls in this study; they were collected at the Blood Research Service of the NIH Clinical Center after providing signed informed consent (Supplementary Table S1). Healthy donors were not age- or gender-matched with the studied patient population.

Peripheral blood samples from patients or healthy donors were collected in red top vacutainer tubes (Becton Dickinson) and processed within the same day. First, blood was clotted at room temperature for 30 minutes. After centrifuging at 2,000 × g for 10 minutes in a refrigerated centrifuge, the serum samples were carefully transferred to new tubes without disturbing the lower layers of blood cells. Serum samples were stored in 1-mL aliquots in polypropylene tubes at –80°C. All samples from the cervical cancer patients were blindly tested for HPV16 and HPV18 in ccfDNA, and all results were included in the analysis without exclusion.

Procedure for HPV ccfDNA isolation and ddPCR analysis

The ccfDNA was extracted from 1 mL of sera with an automated Maxwell RSC Instrument (Promega) using a cellulose magnet resin–based ccfDNA purification process. Briefly, serum or plasma samples were thawed on ice and centrifuged at 2,000 × g for 10 minutes to remove any sediment formed during the freezing process before proceeding to DNA extraction following the manufacturer's instructions. The DNA purification process consists of binding the beads to DNA and several sequential washing steps. The isolated ccfDNA was finally eluted into 60 μL of elution buffer and stored in a −80°C freezer until analysis.

To detect single copies of HPV16 or HPV18 DNA, a ddPCR method was developed using the sequences of the HPV16 or HPV18 E7 genes that are common among the different subtypes (Supplementary Fig. S1; Supplementary Table S2). PCR primers and probes were designed for each HPV type to produce amplicons of different lengths (Supplementary Table S3). The probes for HPV16 and HPV18 were differentially fluorescence-tagged for the simultaneous detection and quantification of both types (Supplementary Fig. S2). The detection and quantification of HPV16 or HPV18 viral DNA in the ccfDNA samples were accomplished by dual-target ddPCR (Bio-Rad) using specific primers and primers for the E7 oncogene that includes the HPV16 and HPV18 variants (Supplementary Table S2). Primers and probes for HPV 16 and HPV18 were designed to avoid cross reactivity with 12 other common HPV types (HPV31, HPV33, HPV35, HPV39, HPV45, HPV51, HPV52, HPV56, HPV58, HPV59, HPV66, and HPV68) implicated in cervical cancer. Eight of the 60 μL elute was used for each ddPCR reaction. The ddPCR reactions were mixed by vortexing, equilibrated at room temperature for 3 minutes, and loaded into a Bio-Rad automated droplet generator. Upon completion of the droplet generation process, the plate was immediately sealed with a pierceable foil using an ABgene Combi Thermo-Sealer (Thermo Fisher Scientific). Immediately after completion of the PCR reaction, the droplets were detected using a Bio-Rad QX200 Droplet Reader. Data were analyzed to determine the HPV types and copy numbers using QuantaSoft software (Bio-Rad).

The ddPCR assay was capable of accurately quantifying target molecules. Although the assays have a maximum sensitivity of a single copy, the linear quantification range was between 10 and 100,000 copies for both HPV16 and HPV18 DNA, with a coefficient of variation (CV) less than 14% throughout the entire quantification range.

Analytic methods

Limit of detection.

Analytic validation was performed with HPV16 and HPV18 ccfDNA assays. The results showed that the limit of blank (LoB) for both HPV16 and HPV18 was zero for all healthy donor samples tested (n = 24; Fig. 1A), and the limit of detection (LoD) was 50 copies/mL (cp/mL). The results showed that the spiked HPV DNA were detected in 97% of HPV16 spikes (n = 34) and 100% of the HPV18 spikes (n = 36), with average recovery efficiencies of 68% and 77%, respectively. Adjusted for DNA recovery and the fraction of sample tested, there was an expected 4–5 cp of HPV DNA per ddPCR reaction, which gives a greater than 95% detection rate for spiked plasma samples. Thus, the sensitivity or LoD for HPV16/18 ccfDNA assay was ≤50 cp/mL.

Figure 1.

Detection and genotyping HPV ccfDNA in metastatic cervical cancer patients. A, LoD for HPV ccfDNA test with spiked blood samples. B, Detecting and typing HPV16 and HPV18 ccfDNA with blood samples from metastatic cervical cancer patients and healthy donors. Tumor HPV analysis was performed using tumor specimens from the cancer patients. C, HPV16 or HPV18 ccfDNA copies per mL of blood serum from cervical cancer patients, as quantified using ddPCR. Data are shown as mean ± SD. D, ccfDNA HPV16 (n = 32) or HPV18 (n = 55) detected in multiple sequential blood samples (n = 87) of cervical cancer patients showed 100% correlation with that detected in the tumors (n = 19).

Figure 1.

Detection and genotyping HPV ccfDNA in metastatic cervical cancer patients. A, LoD for HPV ccfDNA test with spiked blood samples. B, Detecting and typing HPV16 and HPV18 ccfDNA with blood samples from metastatic cervical cancer patients and healthy donors. Tumor HPV analysis was performed using tumor specimens from the cancer patients. C, HPV16 or HPV18 ccfDNA copies per mL of blood serum from cervical cancer patients, as quantified using ddPCR. Data are shown as mean ± SD. D, ccfDNA HPV16 (n = 32) or HPV18 (n = 55) detected in multiple sequential blood samples (n = 87) of cervical cancer patients showed 100% correlation with that detected in the tumors (n = 19).

Close modal

Recovery.

To determine the spike recovery and linearity of the tests, HPV16 or HPV18 DNA was serially diluted and spiked into plasma samples independently at 100,000, 10,000, 1,000, and 100 cp/mL, followed by ccfDNA isolation and testing. The results showed linear relationship with R2 for Goodness of Fit of 1.000 for both DNA (Supplementary Fig. S3). The average CVs of the assays are 14.0% and 11.1%. Thus, the results indicated a linear range for accurate quantification of both HPV DNA molecules was between 100 and 100,000 cp/mL.

Precision.

To determine the precision of the test, plasma samples were spiked with HPV16 DNA at two different levels, 670 and 10,000 cp/mL. The spiked samples were tested in six runs each, in triplicates, and the average daily numbers were used for the analysis. The results showed overall recoveries of about 70%, and inter-day CVs of 8.1% and 5.6%.

Interference studies.

To evaluate the potential interference of endogenous and exogenous interferents, bilirubin, hemoglobin, triglycerides, carboplatin, and paclitaxel were added to the HPV16 DNA-spiked blood samples, followed by ccfDNA isolation and testing. When normalized to the spiked blood samples without any interferent, no tested interferent had any inhibitory effect on HPV detection and quantification (Supplementary Table S4). Therefore, no interference with the performance of the HPV ccfDNA assays was observed with the compounds tested.

Statistical analysis

Statistical analysis was performed using GraphPad Prism 7. Unpaired t test was used for determining the differences in baseline HPV16 and HPV18 ccfDNA copy numbers, and the differences in pretreatment HPV ccfDNA levels between response and nonresponse groups. Pearson correlation statistics were used to determine the association between expected DNA copy number and the actual detected DNA copy value of HPV16 and HPV18 in analytic validation studies for linear dynamic range. Fisher exact test was performed to determine the correlation between ccfDNA and tumor cDNA–based genotyping of HPV16 and HPV18. Receiver-operating characteristic (ROC) analysis was performed to determine the sensitivity and specificity of HPV ccfDNA test for metastatic cervical cancer.

Analytic studies

Analytic validation was performed with HPV16 and HPV18 ccfDNA assays. Results showed that the LoB for both HPV16 and HPV18 was zero for all 24 healthy donor samples that were tested (Fig. 1A). The LoD was 50 copies/mL; at this level, HPV DNA was detected in 97% of the HPV16 spikes (n = 34) and 100% of the HPV18 spikes (n = 36). After adjusting for DNA recovery and the fraction of sample tested, there was an expected 4–5 copies of HPV DNA per ddPCR reaction. Using spiked DNA, the linear range of the tests was determined to be 100 to 10,000 cp/mL, with an R2 Goodness of Fit of 1.000 for both DNA molecules and an average CV of 14% and 11% for HPV16 and HPV18, respectively (Supplementary Fig. S3). The average recovery rates for spiked HPV16 and HPV18 were 77.6% and 63.7% when compared with the prespiked DNA. To determine the precision of the test, plasma samples were spiked with HPV16 DNA at 670 cp/mL. The spiked samples were tested for 6 days in triplicate. The results showed that the overall recovery was approximately 70%, and the interday CV was 16%.

To evaluate the potential interference of endogenous and exogenous interferents, bilirubin, hemoglobin, triglycerides, carboplatin, and paclitaxel were added to the HPV16 DNA-spiked blood samples, and then, ccfDNA isolation and testing were performed. When normalized to the samples without any interferents, no inhibitory effect from the tested compounds on HPV detection and quantification was found (Supplementary Table S3).

Effectiveness of the test in detecting and genotyping HPV ccfDNA in metastatic cervical cancer patients

We evaluated the effectiveness of the HPV16 and HPV18 ccfDNA test by analyzing serum samples from 21 metastatic cervical cancer patients. The corresponding tumors of nineteen patients were found to be positive for HPV16 or HPV18. Control sera were collected from 45 healthy blood donors. The results showed that HPV16 or HPV18 was specifically detected in the ccfDNA of all 19 HPV16- or HPV18-positive cervical cancer patients, with each identification matching that in the tumor tissues (Fig. 1B). Two patients negative for HPV16 or HPV18 ccfDNA were also found negative for HPV16 and HPV18 in their tumor specimens. HPV ccfDNA quantification further revealed a large variation among cancer patients. The median level of HPV ccfDNA was 21,600 copies/mL for HPV16, which was significantly different than 1,360 copies/mL for HPV18 (P = 0.040) (Fig. 1C). Thus, the results showed 100% concordance between the results of HPV16/18 ccfDNA testing and tumor specimen evaluation. None of the 45 healthy donors had detectable HPV DNA in their sera (Fig. 1B). ROC analysis revealed that the area under the ROC curve was 1, with a P < 0.0001. Thus, HPV ccfDNA detection appears to be effective in detecting tumor DNA in metastatic cervical cancer patients who are positive for HPV16 or HPV18.

We further examined the utility of ccfDNA for HPV genotyping with multiple sequential serum samples from the patients enrolled in the immunotherapy study. Among all 87 samples that tested positive for HPV ccfDNA, 32 were HPV16 positive and 55 were HPV18 positive. There was a 100% match between the HPV types determined by ccfDNA and those determined by tumor cDNA analyses for all samples (Fig. 1D). Fisher exact test revealed that the correlation between tumor- and ccfDNA-based HPV typing results was highly significant (P < 0.0001). Thus, the HPV ccfDNA test is accurate for genotyping HPV in metastatic cervical cancer patients.

Robust clearance of HPV ccfDNA in cervical cancer patients with complete responses

We explored the feasibility of using HPV ccfDNA as a circulating biomarker in the follow-up of cervical cancer patients who had undergone an investigational TIL therapy (6). Of the 9 studied patients, three experienced an objective response to TIL therapy: 2 complete responses (CR) and 1 partial response (PR); 6 patients had progressive disease (PD; Table 1; ref. 6). HPV ccfDNA was present in every sample (n = 37) from the 6 patients with PD (Table 1). The levels of baseline HPV ccfDNA were not related to the treatment response (Fig. 2A). One cervical cancer patient (Patient 2) exhibited a very short PR after TIL therapy before experiencing progression of the disease after 3 months posttreatment. Her ccfDNA levels briefly fell below detection in four samples around the first month posttreatment; subsequently became detectable again at day 70 and day 98 (Fig. 2B), by which time she exhibited signs of progression by CT. To examine the potential of HPV ccfDNA for long-term follow-up of the treatment responses, we tested sequential serum samples from patient 3 and patient 6 who had experienced CR at 4.2 and 10.6 months posttreatment, respectively (6). The data show the clearance of HPV ccfDNA from days 32 and 8 for patient 3 and patient 6, respectively (Fig. 2C and D). All subsequent blood samples tested negative for HPV ccfDNA until days 404 and 318, respectively, with both patients still exhibiting CRs. Thus, persistent HPV ccfDNA clearance was only achieved in two metastatic cervical patients with long-lasting CRs.

Table 1.

Summary of HPV ccfDNA results for metastatic cervical patients on HPV-TIL study

ccfDNA HPV Detection in blinded test
PatientAgeTumor HPV typePrior RTPrior chemoTherapyResponseDuration TTP (months)No. of samplesNo. of positive testsHPV typeBaseline copies (mL)No. of negative testsDuration of negative tests (months)ccfDNA recurrence
30 18 Yes Yes HPV-TIL PD 18 4,347 NA NA 
53 18 Yes Yes HPV-TIL PR 14 10 18 2,002 Yes 
36 16 Yes Yes HPV-TIL CR 22+ 25 17 16 86 12+ No 
55 16 Yes Yes HPV-TIL PD 16 3,305 NA NA 
44 18 Yes Yes HPV-TIL PD 18 453 NA NA 
36 18 Yes Yes HPV-TIL CR 15+ 23 13 18 1,193 10 10+ No 
59 18 Yes Yes HPV-TIL PD 18 3,113 NA NA 
31 18 No Yes HPV-TIL PD 18 50 NA NA 
37 18 Yes Yes HPV-TIL PD 18 1,522 NA NA 
ccfDNA HPV Detection in blinded test
PatientAgeTumor HPV typePrior RTPrior chemoTherapyResponseDuration TTP (months)No. of samplesNo. of positive testsHPV typeBaseline copies (mL)No. of negative testsDuration of negative tests (months)ccfDNA recurrence
30 18 Yes Yes HPV-TIL PD 18 4,347 NA NA 
53 18 Yes Yes HPV-TIL PR 14 10 18 2,002 Yes 
36 16 Yes Yes HPV-TIL CR 22+ 25 17 16 86 12+ No 
55 16 Yes Yes HPV-TIL PD 16 3,305 NA NA 
44 18 Yes Yes HPV-TIL PD 18 453 NA NA 
36 18 Yes Yes HPV-TIL CR 15+ 23 13 18 1,193 10 10+ No 
59 18 Yes Yes HPV-TIL PD 18 3,113 NA NA 
31 18 No Yes HPV-TIL PD 18 50 NA NA 
37 18 Yes Yes HPV-TIL PD 18 1,522 NA NA 

Abbreviations: ccfDNA, circulating cell-free DNA; CR, complete response; HPV, human papillomavirus; PD, progressive disease; PR, partial response; RT, radiotherapy; TIL, tumor-infiltrating T cells; TTP, time to progression.

Figure 2.

Robust clearance of HPV ccfDNA associated with complete responses. A, Comparisons of baseline HPV ccfDNA copy numbers (mean ± SD) in patients with or without clinical response to TIL therapy (Table 1). B, Transient clearance of HPV ccfDNA for Patient 2 with partial response (PR). C, Clearance of HPV ccfDNA in cervical cancer Patient 3 with complete response (CR), and durable clearance for 370 days. D, Clearance of HPV ccfDNA in cancer Patient 6 with CR, and durable clearance for 310 days. Insert panels in C and D show the transient induction of ccfDNA following administration of TIL therapy.

Figure 2.

Robust clearance of HPV ccfDNA associated with complete responses. A, Comparisons of baseline HPV ccfDNA copy numbers (mean ± SD) in patients with or without clinical response to TIL therapy (Table 1). B, Transient clearance of HPV ccfDNA for Patient 2 with partial response (PR). C, Clearance of HPV ccfDNA in cervical cancer Patient 3 with complete response (CR), and durable clearance for 370 days. D, Clearance of HPV ccfDNA in cancer Patient 6 with CR, and durable clearance for 310 days. Insert panels in C and D show the transient induction of ccfDNA following administration of TIL therapy.

Close modal

The ability to quantitatively detect changes in ccfDNA also allowed for the examination of the kinetic details of the antitumor activities of TIL therapy. In patient 3, the administration of TIL at day 0 led to a 20-fold increase in HPV ccfDNA that peaked at day 2 (Fig. 2C, insert). In patient 6, the administration of TIL led to an 8-fold increase that peaked at day 3 (Fig. 2D, insert). In both cases, the HPV ccfDNA quickly declined and returned to baseline within a week. Thus, the pharmacodynamic analysis showed a rapid and transient induction of HPV ccfDNA associated with the infusion of reactive T cells.

This study reveals a method that is effective for detecting HPV16 and HPV18 circulating DNA in patients with metastatic cervical cancer. In addition, the assay was reliable in genotyping HPV16 and HPV18 in ccfDNA, as the results from 87 serum samples showed a 100% match with that of the patients' tumors in blinded tests. Furthermore, in patients with an objective response after TIL immunotherapy, a transient induction of HPV ccfDNA peaked after 2–3 days of T-cell infusion, which may suggest therapy-induced antitumor effects. Finally, long-term clearance of HPV ccfDNA was associated only with two patients who experienced CR after TIL immunotherapy.

Potential of HPV ccfDNA detection technology

The HPV ccfDNA assay used in the current study was analytically validated. We showed that the ccfDNA assay had a LoD ≤50 cp/mL with 1 mL of serum or plasma samples and at least a 95% positive rate among the HPV16/18 reference samples (n = 70) and a 0% positive rate among the healthy donor controls (n = 24). Taking into consideration the purification efficiency and the fraction of DNA eluent (8 of 60 μL) used for testing, the assay has a minimum detection limit of 4–5 HPV DNA molecules per reaction for a ≥95% success rate in detection. Additional improvements in the LoD of up to 10-fold might be achieved by using a larger volume of blood and a larger fraction of the purified DNA for ddPCR analysis. Furthermore, our preliminary analysis indicates the feasibility of an expanded HPV ccfDNA assay to include 12 other HPV types implicated in cervical cancer without a loss of specificity (Supplementary Fig. S4). The expanded HPV genotype coverage of HPV ccfDNA assay may thus cover about 95% of HPV types implicated in cervical cancer.

Effective detection of HPV ccfDNA in metastatic cervical cancer patients

HPV ccfDNA was detected in all metastatic cervical cancer patients who were positive for HPV16 or 18 in tumors (n = 19), with a complete concordance. The sensitivity of this assay was further confirmed in 6 patients with PD, in whom HPV ccfDNA was detected at every time point that was tested (n = 37). Like other blood-based tests, it would still have practical limitation in detecting circulating HPV DNA in patients with low disease burden, as showed in the negative HPV ccfDNA data in one patient with PR. However, the assays could be among the most sensitive ones for circulating DNA detection due to multiple integration sites of HPV16 or HPV18 in cancer cells and the focal amplification of the HPV E7 oncogene used in the detection (29). The median levels were 21,000 cp/mL for HPV16 and 1,300 cp/mL for HPV18 in the current study and were significantly above the LoD of ≤50 cp/mL using 1 mL sera. Our findings agree with a recent publication that revealed the higher sensitivity of ddPCR detection (30) and are consistent with reports showing that high HPV16/18 ccfDNA positive rate in plasma samples from patients with head and neck squamous cell carcinoma (31–33). Our study further presents examples of long-term clearance of HPV ccfDNA in metastatic cervical cancer patients with CRs. Because the Pap smear–based cytology test is ineffective at detecting recurrent and metastatic cervical cancer after initial therapies (4, 5), the detection of HPV ccfDNA may potentially be beneficial for the routine follow-up of cervical cancer patients after achieving remission.

Genotyping HPV ccfDNA for investigational T-cell–based immunotherapies

The results suggest that HPV ccfDNA assay can be a reliable method for genotyping tumor HPV in patients with metastatic cancer. An immediate application of the test would be in selecting patients for clinical trials involving novel T-cell receptor (TCR)-engineered T cells directed against the E6/E7 proteins of HPV16 (8). Two clinical trials (NCT02280811; NCT02858310) with the TCR-engineered T cells are in progress and require the selection of cancer patients with HPV16. As it is often logistically difficult, time consuming, and potentially invasive to obtain tumor samples from recurrent cervical cancer patients, a blood-based test would be a far better alternative. Another strength of blood-based genotyping is that in the case of doubtful results, a confirmatory test could be requested and readily performed with a second blood sample from the patient. Our results suggest that for metastatic cervical cancer patients, it may be feasible to use blood samples as liquid biopsies for genotyping tumor HPV, thus allowing for the selection of patients for virus-specific tumor immunotherapies.

A potential biomarker for antitumor activities of investigational therapies

The ability to noninvasively obtain multiple blood samples for the quantification of ccfDNA molecules allow for the detection of the cytotoxic activities of investigational agents against tumors within days after the initiation of therapy. In the time-course analysis of HPV ccfDNA from two patients with a complete and long-term response to TIL therapy, a rapid tumor ccfDNA peak at 2–3 days after initiation of TIL treatment was observed, which may be interpreted as the tumor DNA being released from dying cervical cancer cells (12). Such a rapid induction of ccfDNA was also observed in melanoma patients who were treated with TIL therapy (34). Thus, ccfDNA may serve as a new type of pharmacodynamic biomarker for cancer patients undergoing experimental therapies and may have potential value in the early-stage clinical development of investigational cancer therapeutics.

In conclusion, the findings of this study suggest that ccfDNA may serve as a promising biomarker for detecting recurrent metastatic cervical cancer and for tumor HPV genotyping. Moving forward, the blood-based tumor HPV genotyping method may be incorporated as a biomarker for patient selection in prospective cell immunotherapy trials. Limitations of the study include retrospective design, the limited number of metastatic cervical cancer patients available for this investigation, and the identification of only two metastatic cervical cancer patients with CRs for a long-term follow-up study. Because of the rarity of CR for metastatic cervical cancer patients, the immediate utility of the assays for current disease management may still be limited.

No potential conflicts of interest were disclosed.

Conception and design: Z. Kang, S. Stevanović, C.S. Hinrichs, L. Cao

Development of methodology: Z. Kang, C.S. Hinrichs, L. Cao

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): Z. Kang, S. Stevanović,

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): Z. Kang, S. Stevanović, C.S. Hinrichs, L. Cao

Writing, review, and/or revision of the manuscript: Z. Kang, S. Stevanović, C.S. Hinrichs, L. Cao

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): Z. Kang, S. Stevanović

Study supervision: L. Cao

We thank all patients and blood donors for participating in the clinical studies at NCI. The authors thank Cindy Clark, NIH Library, for manuscript editing assistance.

This project was supported by the Intramural Research Program of the NIH, NCI, and CCR. This project was also funded in part with federal funds from NCI, under contract HHSN261200800001E. The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the U.S. Government.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Schiffman
M
,
Doorbar
J
,
Wentzensen
N
,
de Sanjose
S
,
Fakhry
C
,
Monk
BJ
, et al
Carcinogenic human papillomavirus infection
.
Nat Rev Dis Primers
2016
;
2
:
16086
.
2.
Ferlay
J
,
Shin
HR
,
Bray
F
,
Forman
D
,
Mathers
C
,
Parkin
DM
. 
Estimates of worldwide burden of cancer in 2008: GLOBOCAN 2008
.
Int J Cancer
2010
;
127
:
2893
917
.
3.
Schwartz
SM
,
Daling
JR
,
Shera
KA
,
Madeleine
MM
,
McKnight
B
,
Galloway
DA
, et al
Human papillomavirus and prognosis of invasive cervical cancer: a population-based study
.
J Clin Oncol
2001
;
19
:
1906
15
.
4.
Morice
P
,
Deyrolle
C
,
Rey
A
,
Atallah
D
,
Pautier
P
,
Camatte
S
, et al
Value of routine follow-up procedures for patients with stage I/II cervical cancer treated with combined surgery-radiation therapy
.
Ann Oncol
2004
;
15
:
218
23
.
5.
Koh
WJ
,
Greer
BE
,
Abu-Rustum
NR
,
Apte
SM
,
Campos
SM
,
Chan
J
, et al
Cervical cancer
.
J Natl Compr Canc Netw
2013
;
11
:
320
43
.
6.
Stevanovic
S
,
Draper
LM
,
Langhan
MM
,
Campbell
TE
,
Kwong
ML
,
Wunderlich
JR
, et al
Complete regression of metastatic cervical cancer after treatment with human papillomavirus-targeted tumor-infiltrating T cells
.
J Clin Oncol
2015
;
33
:
1543
50
.
7.
Zsiros
E
,
Tsuji
T
,
Odunsi
K
. 
Adoptive T-cell therapy is a promising salvage approach for advanced or recurrent metastatic cervical cancer
.
J Clin Oncol
2015
;
33
:
1521
2
.
8.
Draper
LM
,
Kwong
ML
,
Gros
A
,
Stevanovic
S
,
Tran
E
,
Kerkar
S
, et al
Targeting of HPV-16+ epithelial cancer cells by TCR gene engineered T cells directed against E6
.
Clin Cancer Res
2015
;
21
:
4431
9
.
9.
Alix-Panabieres
C
,
Pantel
K
. 
Clinical applications of circulating tumor cells and circulating tumor DNA as liquid biopsy
.
Cancer Discov
2016
;
6
:
479
91
.
10.
Dawson
SJ
,
Rosenfeld
N
,
Caldas
C
. 
Circulating tumor DNA to monitor metastatic breast cancer
.
N Engl J Med
2013
;
369
:
93
4
.
11.
Bettegowda
C
,
Sausen
M
,
Leary
RJ
,
Kinde
I
,
Wang
Y
,
Agrawal
N
, et al
Detection of circulating tumor DNA in early- and late-stage human malignancies
.
Sci Transl Med
2014
;
6
:
224ra24
.
12.
Diaz
LA
 Jr
,
Bardelli
A
. 
Liquid biopsies: genotyping circulating tumor DNA
.
J Clin Oncol
2014
;
32
:
579
86
.
13.
Jahr
S
,
Hentze
H
,
Englisch
S
,
Hardt
D
,
Fackelmayer
FO
,
Hesch
RD
, et al
DNA fragments in the blood plasma of cancer patients: quantitations and evidence for their origin from apoptotic and necrotic cells
.
Cancer Res
2001
;
61
:
1659
65
.
14.
Diaz
LA
 Jr.
,
Williams
RT
,
Wu
J
,
Kinde
I
,
Hecht
JR
,
Berlin
J
, et al
The molecular evolution of acquired resistance to targeted EGFR blockade in colorectal cancers
.
Nature
2012
;
486
:
537
40
.
15.
Misale
S
,
Yaeger
R
,
Hobor
S
,
Scala
E
,
Janakiraman
M
,
Liska
D
, et al
Emergence of KRAS mutations and acquired resistance to anti-EGFR therapy in colorectal cancer
.
Nature
2012
;
486
:
532
6
.
16.
Tie
J
,
Wang
Y
,
Tomasetti
C
,
Li
L
,
Springer
S
,
Kinde
I
, et al
Circulating tumor DNA analysis detects minimal residual disease and predicts recurrence in patients with stage II colon cancer
.
Sci Transl Med
2016
;
8
:
346ra92
.
17.
Dawson
SJ
,
Tsui
DW
,
Murtaza
M
,
Biggs
H
,
Rueda
OM
,
Chin
SF
, et al
Analysis of circulating tumor DNA to monitor metastatic breast cancer
.
N Engl J Med
2013
;
368
:
1199
209
.
18.
Joseph
JD
,
Lu
N
,
Qian
J
,
Sensintaffar
J
,
Shao
G
,
Brigham
D
, et al
A clinically relevant androgen receptor mutation confers resistance to second-generation antiandrogens enzalutamide and ARN-509
.
Cancer Discov
2013
;
3
:
1020
9
.
19.
Romanel
A
,
Gasi Tandefelt
D
,
Conteduca
V
,
Jayaram
A
,
Casiraghi
N
,
Wetterskog
D
, et al
Plasma AR and abiraterone-resistant prostate cancer
.
Sci Transl Med
2015
;
7
:
312re10
.
20.
Piotrowska
Z
,
Niederst
MJ
,
Karlovich
CA
,
Wakelee
HA
,
Neal
JW
,
Mino-Kenudson
M
, et al
Heterogeneity underlies the emergence of EGFRT790 wild-type clones following treatment of T790M-positive cancers with a third-generation EGFR inhibitor
.
Cancer Discov
2015
;
5
:
713
22
.
21.
Sundaresan
TK
,
Sequist
LV
,
Heymach
JV
,
Riely
GJ
,
Janne
PA
,
Koch
WH
, et al
Detection of T790M, the acquired resistance EGFR mutation, by tumor biopsy versus noninvasive blood-based analyses
.
Clin Cancer Res
2016
;
22
:
1103
10
.
22.
Alix-Panabieres
C
,
Schwarzenbach
H
,
Pantel
K
. 
Circulating tumor cells and circulating tumor DNA
.
Annu Rev Med
2012
;
63
:
199
215
.
23.
Lo
YM
,
Zhang
J
,
Leung
TN
,
Lau
TK
,
Chang
AM
,
Hjelm
NM
. 
Rapid clearance of fetal DNA from maternal plasma
.
Am J Hum Genet
1999
;
64
:
218
24
.
24.
Khan
MJ
,
Castle
PE
,
Lorincz
AT
,
Wacholder
S
,
Sherman
M
,
Scott
DR
, et al
The elevated 10-year risk of cervical precancer and cancer in women with human papillomavirus (HPV) type 16 or 18 and the possible utility of type-specific HPV testing in clinical practice
.
J Natl Cancer Inst
2005
;
97
:
1072
9
.
25.
Schiffman
M
,
Wentzensen
N
,
Wacholder
S
,
Kinney
W
,
Gage
JC
,
Castle
PE
. 
Human papillomavirus testing in the prevention of cervical cancer
.
J Natl Cancer Inst
2011
;
103
:
368
83
.
26.
Dong
SM
,
Pai
SI
,
Rha
SH
,
Hildesheim
A
,
Kurman
RJ
,
Schwartz
PE
, et al
Detection and quantitation of human papillomavirus DNA in the plasma of patients with cervical carcinoma
.
Cancer Epidemiol Biomarkers Prev
2002
;
11
:
3
6
.
27.
Liu
VW
,
Tsang
P
,
Yip
A
,
Ng
TY
,
Wong
LC
,
Ngan
HY
. 
Low incidence of HPV DNA in sera of pretreatment cervical cancer patients
.
Gynecol Oncol
2001
;
82
:
269
72
.
28.
Shimada
T
,
Yamaguchi
N
,
Nishida
N
,
Yamasaki
K
,
Miura
K
,
Katamine
S
, et al
Human papillomavirus DNA in plasma of patients with HPV16 DNA-positive uterine cervical cancer
.
Jpn J Clin Oncol
2010
;
40
:
420
4
.
29.
Akagi
K
,
Li
J
,
Broutian
TR
,
Padilla-Nash
H
,
Xiao
W
,
Jiang
B
, et al
Genome-wide analysis of HPV integration in human cancers reveals recurrent, focal genomic instability
.
Genome Res
2014
;
24
:
185
99
.
30.
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
:
201
9
.
31.
Wang
Y
,
Springer
S
,
Mulvey
CL
,
Silliman
N
,
Schaefer
J
,
Sausen
M
, et al
Detection of somatic mutations and HPV in the saliva and plasma of patients with head and neck squamous cell carcinomas
.
Sci Transl Med
2015
;
7
:
293ra104
.
32.
Ahn
SM
,
Chan
JY
,
Zhang
Z
,
Wang
H
,
Khan
Z
,
Bishop
JA
, et al
Saliva and plasma quantitative polymerase chain reaction-based detection and surveillance of human papillomavirus-related head and neck cancer
.
JAMA Otolaryngol Head Neck Surg
2014
;
140
:
846
54
.
33.
Cao
H
,
Banh
A
,
Kwok
S
,
Shi
X
,
Wu
S
,
Krakow
T
, et al
Quantitation of human papillomavirus DNA in plasma of oropharyngeal carcinoma patients
.
Int J Radiat Oncol Biol Phys
2012
;
82
:
e351
8
.
34.
Xi
L
,
Pham
TH
,
Payabyab
EC
,
Sherry
RM
,
Rosenberg
SA
,
Raffeld
M
. 
Circulating tumor DNA as an early indicator of response to T-cell transfer immunotherapy in metastatic melanoma
.
Clin Cancer Res
2016
;
22
:
5480
86
.