Persistent high-risk human papillomavirus (hrHPV) infection is essential for the development of cervical cancer and its precursor lesions. High-risk HPV testing has a higher sensitivity than cytology does for detecting cervical epithelial lesions. However, a large study from a single institution showed 31% of patients with invasive cervical cancer had negative baseline hrHPV testing within 5 years preceding the diagnosis.
To investigate the limitation of hrHPV testing in detecting invasive cervical cancer.
Cases from 2012 with a histologic diagnosis of invasive cervical carcinoma were retrieved from multiple institutions. From those records, prior hrHPV testing and Papanicolaou test results in the 5 years before the cancer diagnosis were recorded.
Seventy patients with cervical carcinoma were included in the study. Negative HPV test result rates were 9% (5 of 53), 23% (6 of 26), and 25% (2 of 8) during the periods of less than 1 year, 1 to 3 years, and 3 to 5 years before the histologic diagnoses, respectively. Negative Papanicolaou testing results in the same time intervals were 3.4% (2 of 59), 33% (10 of 30), and 40% (6 of 15). Although the HPV− rate seemed to be different among different HPV test methods, no statistical significance was detected because of small sample size. Negative hrHPV rates in patients with adenocarcinoma were similar to those in patients with squamous cell carcinoma.
These data expose limitations for the potential use of primary HPV testing. In addition, current screening guidelines recommending cotesting at 5-year intervals should be evaluated further with additional historic data collection because there are women with negative results for both Papanicolaou tests and hrHPV testing within the period of 3 to 5 years before an invasive carcinoma diagnosis.
Cervical cancer is the third most common gynecologic cancer in the United States and remains the second most common type of cancer among all women in countries that do not have access to cervical cancer screening and prevention programs. Persistent high-risk human papillomavirus (hrHPV) infection is essential for the development of cervical cancer and its precursor lesions.1 High-risk HPV testing has been widely used in triage of patients with atypical squamous cells of undetermined significance (ASC-US) Papanicolaou (Pap) test results and also in conjunction with the Pap test in primary screening for women 30 years and older.
Although cytology screening with the Pap test has been very successful in lowering cervical cancer incidence and mortality worldwide, recent studies suggest hrHPV testing has a higher sensitivity than Pap tests and provides additional benefits, albeit with lower specificity and more false-positive findings. The addition of hrHPV testing to Pap tests increases detection rates of cervical intraepithelial neoplasia 3 and results in decreased cervical intraepithelial neoplasia 3 or invasive cancer detection rates in subsequent rounds of screening.2–4 A recent, large study from Kaiser Permanente Northern California (Oakland) reported a 5-year, cumulative incidence of cancer for all 315 061 women who were HPV− was extremely low, 3.8/100 000 women/y, which was about one-half of the cancer risk in women who were Pap-negative (7.5/100 000 women/y).5 In addition, cotesting increased the detection of endocervical adenocarcinoma in situ and adenocarcinoma5,6 and slightly decreased the risk of both squamous and endocervical cancers for women during a 10-year period.7 Based on those results, the 2012 cervical cancer screening guidelines recommended women aged 30 to 65 years be screened with cytology and HPV testing (cotesting) every 5 years.8,9
However, the Kaiser Permanente study also reported 31% (27 of 87) of patients with invasive cervical cancer had a negative baseline hrHPV test result within 5 years preceding the diagnosis of cervical cancer.5 Other studies found a similar range of negative detection rates for hrHPV in cervical cytologic specimens from patients with cervical cancers.10–13 Accurate estimates of the sensitivity of hrHPV testing for detection of invasive cervical cancer is important. Because some HPV infections are persistent for only a few years before invasive squamous cell carcinomas develop, a low sensitivity rate in hrHPV testing in detecting invasive cervical cancers would result in poor performance of screening programs that extend cotesting to 5 years. If there is a future move to hrHPV testing as the single, cervical cancer-screening tool, invasive cervical cancer might go undetected.
This retrospective study examined prior hrHPV test results in patients with invasive cervical cancer in a multicenter setting to the limitations of such testing in invasive cervical carcinoma diagnoses.
MATERIALS AND METHODS
A retrospective study was designed to find patients with a diagnosis of cervical carcinoma during 2012 from multiple institutions. The study was approved by University of Pittsburgh's institutional review board. Cases with a histologic diagnosis of invasive cervical carcinoma were retrieved from the pathology databases of the 18 laboratories of the authors. Only cases with prior hrHPV testing results during the 5 years before the cancer diagnosis were included in this study. From those records, prior hrHPV testing and Pap test results in the 5 years before the cancer diagnosis were recorded.
All Pap tests in this study were performed with the liquid-based cytology methods, either with ThinPrep Pap tests (Hologic, Marlborough, Massachusetts) or with SurePath Pap tests (Becton, Dickson and Company, Franklin Lakes, New Jersey). High-risk HPV testing was ordered by clinicians according to several ordering options as follows: reflex testing triggered by ASC-US Pap test results, cotesting with Pap tests in women 30 years and older, and testing regardless of age or Pap test results. High-risk HPV detection was performed using US Food and Drug Administration (FDA)–approved hrHPV testing methods, including Hybrid Capture 2 hrHPV DNA test (HC2) (Qiagen, Hilden, Germany), Cervista (Hologic), and cobas 4800 (Roche, Basel, Switzerland). Laboratories receiving SurePath Pap tests used internally validated HPV assays. The HC2 HPV test is an in vitro nucleic acid hybridization assay with signal amplification using microplate chemiluminescence for the qualitative detection of 13 high-risk types of HPV DNA (16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 68). The Cervista HPV test is an in vitro diagnostic test for the qualitative detection of hrHPV DNA from 14 hrHPV types (16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 66, 68) in cervical cytology specimens with the Invader chemistry platform (Hologic) and is able to genotype HPV 16 or 18. The cobas HPV test uses amplification of target DNA by polymerase chain reaction and subsequent nucleic acid hybridization for the detection of 14 hrHPV types in a single analysis. The test specifically identifies HPV types 16 and 18 while concurrently detecting the 12 remaining high-risk types (31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 66, 68).
The Pearson χ2 test was used for statistical analysis (Fisher exact test for small sample sizes) conducted on SAS 9.1 system software (SAS Institute Inc., Cary, NC, USA). A value of P < .05 was considered statically significant.
Seventy patients with cervical carcinoma were retrieved, including 55 squamous cell carcinomas (79%), 13 adenocarcinomas (19%), and 2 others (3%; 1 adenoid basal carcinoma and 1 carcinosarcoma). The average age of the patients was 49 years (range, 29 to 86 years). The number of patients from each laboratory varied from 0 to 16. High-risk HPV testing methods included HC2 in 58 patients (83%), Cervista in 9 patients (13%), and cobas 4800 in 3 patients (4%). Papanicolaou test methods included ThinPrep in 54 patients (77%) and SurePath in 16 patients (23%).
Prior HPV Testing Results
Fifty-three patients (76%) had 53 hrHPV tests within 1 year, 22 patients (31%) had 26 hrHPV tests from 1 to 3 years, and 7 patients (10%) had 8 tests 3 to 5 years before the diagnosis of cervical cancer (some patients had one or more HPV tests during different periods). Negative HPV test result rates were 9% (5 of 53), 23% (6 of 26), and 25% (2 of 8) during the periods of less than 1 year, 1 to 3 years, and 3 to 5 years before the histologic diagnoses of cervical cancer, respectively. Negative hrHPV test results were found in 12 of the 70 patients (17%) at some point within 5 years before the diagnosis of invasive cervical cancer. Eleven of these 12 patients (92%) had negative hrHPV test results only in all 5 years (one patient had 2 negative HPV tests), whereas 1 patient had both negative and positive HPV test results. The hrHPV− rate during the period of 1 to 3 years was significantly higher than the rate within 1 year. The lack of statistical significance for the rate during the period of 3 to 5 years compared to that within 1 year was probably due to small sample size in that group (Table 1).
Prior Pap Test Results
Five of 53 HPV tests (9%) were negative within 1 year before cancer diagnosis (Table 1). In contrast, 2 of 59 (3.4%) patients who had a Pap test within 1 year before their cancer diagnosis had only negative Pap test results. Papanicolaou test results varied from ASC-US to high-grade squamous intraepithelial lesion (HSIL+) or cancer with high-grade squamous intraepithelial lesion (58%; 34 of 59) as the most common abnormal Pap result (Table 2). The percentage of ASC-US cases in this study cohort was higher than in the general cervical cancer patient population because of selective bias created by including only cancer cases with both hrHPV testing and Pap tests.
Three of 59 cancer cases (5.1%) had at least a prior negative Pap test result within 1 year before a cervical cancer diagnosis; however, 37% (11 of 30) and 53% (8 of 15) of cancer cases showed a prior negative Pap test during the periods of 1 to 3 years and 3 to 5 years, respectively, and they were significantly higher than the number of cases within 1 year (Table 3).
Prior HPV Testing Results by Different Methods
All HPV testing results within 5 years before cancer diagnosis were stratified based on hrHPV detection methods. In this study, 58 patients (83%) had HC2 HPV testing (72 tests), 9 patients (13%) had Cervista HPV testing (12 tests), and 3 patients (4%) had cobas 4800 testing (3 tests). The HPV− rate was 15% (11 of 72) with HC2 tests, 8% (1 of 12) with Cervista, and 33% (1 of 3) with cobas 4800. Although the HPV− rate seemed to be lower in the Cervista group and higher in the cobas group when compared with the HC2 group, no statistical significance was detected, probably because of small sample size (Table 4).
Prior HPV Testing Results by Different Media
All HPV testing results within 5 years before cancer diagnosis were also stratified based on different preparation types. Although HPV− rates in the SurePath group (19%; 5 of 26) were higher than that in the ThinPrep group (13%; 8 of 61), the difference was not statistically significant (Table 5).
Prior HPV Testing Results in Different Histologic Cancer Types
Two of 13 patients (15%) with cervical adenocarcinoma had negative hrHPV testing results, which was similar to the HPV negative rate for those with squamous cell carcinoma (20%; 11 of 55) (Table 6).
In this multicenter retrospective study, we investigated prior hrHPV testing results in patients with invasive cervical carcinoma and found that the negative hrHPV testing rates were 9% (5 of 53), 23% (6 of 26), and 25% (2 of 8) during the periods of less than 1 year, 1 to 3 years, and 3 to 5 years before a cervical cancer diagnosis, respectively. Our findings are consistent with most previous studies showing similar negative HPV testing rates in invasive cervical cancer.5,10–12 In our study, HPV− rates in patients with invasive cervical cancer seemed to increase with increased time before cancer diagnosis, and it is reasonable to attribute that trend to the smaller size of the lesion or the lower viral titers during earlier periods.
The percentage of patients with cervical cancer showing hrHPV− testing in Pap test specimens varies widely. Data from Kaiser Permanente showed 37% (18 of 49) of cervical squamous cell carcinoma and 22% (322 of 1464) of cervical adenocarcinoma cases (31% [27 of 87] of all carcinomas) had a baseline negative HPV testing result 5 years before a histologic diagnosis of cervical cancer.5 A recent study13 also indicated that 4 of 29 patients (14%) had HPV− results more than 4 months to 3 years before a histopathologic diagnosis of cervical glandular neoplasia, including 3 cases of adenocarcinoma in situ and 1 case of invasive adenocarcinoma.
In this study, although the HPV− rate seemed to be lower in the Cervista group and higher in the cobas group when compared with the HC2 group, no statistical significance was detected. HC2 was the first FDA-approved HPV test, and studies from different countries, including the United Kingdom, Slovenia, China, and Korea, have demonstrated that HC2 is associated with negative results in cervical cytology specimens from patients with cervical cancer; HC2− rates have varied from 5.0% to 25%.14–18 The question is, whether those are true-negative or false-negative results. In a recent study, 3 of 31 patients10 (9.7%) diagnosed with invasive cervical squamous cell carcinoma and tested within the prior 12 months for hrHPV by HC2 had negative HC2 results. All three patients (100%) had HPV 18 detected by polymerase chain reaction in materials from paraffin sections of resected specimens, and 2 (66%) also had detectible HPV 16.10 Those data indicate that at least a certain percentage of HC2− cases are false-negative results, and the plausible explanations include (1) the detection assay does not cover the specific HPV type, (2) there is a low titer or low copy number of HPV,19 (3) there are inhibitors in the specimen, (4) the analytic sensitivity is limited, and (5) there is inadequate cellularity in the specimen.20
Besides these false-negative HPV cervical cancers, true HPV− cervical cancers also account for a small portion of HPV negative-testing cervical cancers. The true HPV negative cancers include unusual subtypes of cervical cancers, such as minimal deviation adenocarcinoma and gastric-type, intestinal-type, and mesonephric and clear cell carcinomas. Rapidly developing cervical carcinoma might also account for a very small portion of those HPV testing-negative cervical cancers because HPV testing might be performed outside of the HPV infection period.21
A significantly higher percentage of patients (25%; 15 of 59) in this study had an ASC-US Pap test within 1 year before their cancer diagnosis than in a previous study by Li et al10 (9 of 156; 5.8%). This might be caused by selection bias because only patients with prior hrHPV testing were included in our cohort.
Comparable to the Kaiser Permanente study,5 our study also demonstrated that approximately one-third of patients with invasive cervical carcinoma had a negative hrHPV testing result during the period of 3 to 5 years before their cancer diagnosis. To maximize the detection of cervical cancer, cotesting with both a Pap test and an FDA-approved hrHPV test is recommended in the newly updated screening guidelines8,9 for the prevention and early detection of cervical cancer. However, the current guidelines recommending cotesting at a 5-year screening interval need to be reevaluated because 20% (11 of 55) of cases with invasive cancer had negative cotesting between 1 year and 5 years before diagnosis, although a selective bias with enriched HPV− invasive cancers in highly screened populations could exist in our study.
Although this study was a multicenter study from 18 laboratories in the United States, only 70 cases with a 2012 histologic diagnosis of invasive cervical carcinomas and prior hrHPV testing were identified, indicating the rarity of these cases. Because of the small sample size in this study, the results of statistical analysis were limited. It was impossible to see whether there was a difference in negative hrHPV rates between the histologic type of cancer (squamous cell carcinoma versus adenocarcinoma) or among the detection methods (HC2, Cervista, and cobas 4800). Therefore, nationwide data collection with documentation of HPV test results in patients developing invasive cervical cancer is important to confirm these results in further studies.
The authors have no relevant financial interest in the products or companies described in this article.
Presented at the 61st Annual Scientific Meeting of the American Society of Cytopathology; November 10, 2013; Orlando, Florida.