Determinants of Viral Oncogene E6-E7 mRNA Overexpression in a Population-Based Large Sample of Women Infected by High-Risk Human Papillomavirus Types

ABSTRACT

Cervical cancer screening by human papillomavirus (HPV) DNA testing with cytology triage is more effective than cytology testing. Compared to cytology, the HPV DNA test's higher sensitivity, which allows better protection with longer intervals, makes it necessary to triage the women with a positive result to compensate its lower specificity. We are conducting a large randomized clinical trial (New Technologies for Cervical Cancer 2 [NTCC2]) within organized population-based screening programs in Italy using HPV DNA as the primary screening test to evaluate, by the Aptima HPV assay (Hologic), the use of HPV E6-E7 mRNA in a triage test in comparison to cytology. By the end of June 2016, data were available for 35,877 of 38,535 enrolled women, 2,651 (7.4%) of whom were HPV DNA positive. Among the samples obtained, 2,453 samples were tested also by Aptima, and 1,649 (67.2%) gave a positive result. The proportion of mRNA positivity was slightly higher among samples tested for HPV DNA by the Cobas 4800 HPV assay (Roche) than by the Hybrid Capture 2 (HC2) assay (Qiagen). In our setting, the observed E6-E7 mRNA positivity rate, if used as a triage test, would bring a rate of immediate referral to colposcopy of about 4 to 5%. This value is higher than that observed with cytology triage for both immediate and delayed referrals to colposcopy. By showing only a very high sensitivity and thus allowing a longer interval for HPV DNA-positive/HPV mRNA-negative women, a triage by this test might be more efficient than by cytology.

INTRODUCTION

Screening with the human papillomavirus DNA (HPV DNA) test proved to be more effective in reducing cervical cancer incidence than Pap test screening (1). The prevalence of HPV infections is quite high, and the specificity of the HPV DNA test for cervical intraepithelial neoplasia grade 2 or more severe (CIN2+) is quite low. Therefore, a triage test is necessary in order to reduce colposcopy referrals. To date, guidelines recommend cytology as a triage test for HPV-positive women (2–5). To increase test specificity, a new generation of HPV tests targeting E6-E7 oncogene expression instead of simply the presence of viral DNA has been developed (6). The activation of these two genes correlates with viral DNA integration and has been proven to cause several changes in the cell related to cell cycle deregulation.

The tests targeting HPV E6-E7 mRNA have been suggested both as a putative primary screening test (7, 8) and as a triage test instead of cytology (9).

A large randomized trial, denominated New Technologies in Cervical Cancer 2 (NTCC2), is recruiting women into organized population-based cervical cancer screening programs based on HPV DNA as the primary screening test in 5 Italian regions (NCT01837693), with the aim of evaluating different biomarkers (HPV E6-E7 mRNA; p16INK4a/Ki67 expression) as putative triage tests for HPV DNA-positive women. Preliminary data on HPV E6-E7 mRNA test results in HPV DNA-positive samples are now available.

The aim of this paper is to describe the association between age, cytology results, viral genotype, HPV DNA load, and HPV mRNA positivity in a large population-based sample of HPV DNA-positive women.

RESULTS

Preplanned interim analyses include 35,877 women for whom the HPV DNA test results were available as of June 2016; of 38,535 recruited women within the NTCC2 study (Fig. 1), 12,657 women were tested by the Hybrid Capture 2 (HC2) assay (Qiagen) and 23,220 by the Cobas assay (Roche) (Table 1). Overall, 2,651 (7.4%) were HPV DNA positive and the HPV mRNA test results were available for 2,453 women. Four samples yielded invalid results, i.e., 2 samples showed a negative internal control, while the other 2 samples dried during transportation. For the 2,449 samples with a valid result, the positivity for HPV E6-E7mRNA was 67.3%. The distribution of HPV E6-E7 mRNA results, expressed as signal strength, showed a peak at zero for negatives and a second peak for positives at 11 signal-to-cutoff (S/CO) with a low frequency of intermediate values (Fig. 2). Among the negative results, 91.9% (735/800) showed a value of zero whereas 65 samples (8.1%) showed a nonzero value below the 0.5 S/CO threshold.

FIG 1.

Flow chart of the study. Pos, positive; Neg, negative.

TABLE 1.

Women tested for HPV DNA and HPV E6-E7 mRNA and results by center and HPV DNA assay

Center	HPV DNA testing						HPV E6-E7 mRNA testing
	HPV DNA assay	Age range (yr)	No. of samples tested	No. of samples positive	% of positive to tested	95% CI	No. of samples tested	No. of samples	% of HPV DNA- positive samples	95% CI	% of total screened	95% CI
Umbria	Cobas	35–64	14,950	885	5.9	5.6–6.3	807	617	76.5	73.4–79.3	4.5	4.2–4.9
Veneto	Cobas	25–59	7,469	484	6.5	5.9–7.1	407	287	70.5	65.8–74.9	4.6	4.1–5.1
Veneto	HC2	25–59	2,604	160	6.1	5.3–7.1	160	112	70.0	62.3–77.0	4.3	3.6–5.2
Florence	Cobas	34–59	801	73	9.1	7.2–11.3	52	38	73.1	59.0–84.4	6.7	4.8–9.1
Florence	HC2	34–59	7,133	711	10.0	9.3–10.7	710	408	57.5	53.7–61.1	5.7	5.8–6.3
Turin	HC2	30–59	2,920	338	11.6	10.4–12.8	313	187	59.7	54.1–65.2	6.9	6.0–7.9
All centers	Cobas		23,220	1,442	6.2	5.9–4.5	1,266	942	74.4	72.0–76.8	4.6	4.3–4.9
All centers	HC2		12,657	1,209	9.5	9.0–10.1	1,183	707	59.8	57.0–62.3	5.7	5.3–6.1
Total			35,877	2,651	7.4	7.1–7.7	2,449	1,649	67.3	65.4–69.2	5.0	4.7–5.2

FIG 2.

Distribution of HPV E6-E7 mRNA test results reported as signal/cutoff (S/CO) values. All 2,449 HPV DNA-positive samples with a valid mRNA result are included. White bar, values below the positivity threshold of 0.5 S/CO.

Among HPV DNA-positive samples, HPV mRNA positivity did not vary with age (Table 2). Conversely, it varied among centers, with higher figures in the centers of Veneto and Umbria, where HPV DNA positivity in the screened population was lower than in Florence and Turin (Tables 1 and 2). Thus, the rate of HPV mRNA-positive women for the total recruited population was similar in all centers.

TABLE 2.

HPV E6-E7 mRNA positivity in women HPV-DNA positive, by age and center^a

Age range of women tested (yr)	Umbria			Veneto			Florence			Turin			Total
	No. tested	No. (%) mRNA+	95% CI	No. tested	No. (%) mRNA+	95% CI	No. tested	No. (%) mRNA+	95% CI	No. tested	No. (%) mRNA+	95% CI	No. tested	No. (%) mRNA+	95% CI
<30				52	41 (78.8)	65.3–88.9							52	41 (78.8)	65.3–88.9
30–34				97	60 (61.9)	51.4–71.5	48	31 (64.6)	49.5–77.8	49	33 (67.3)	52.5–80.1	194	124 (63.9)	56.4–70.3
35–39	203	155 (76.4)	69.9–82.0	97	71 (73.2)	63.2–81.7	144	80 (55.6)	47.1–63.8	99	58 (58.6)	48.2–68.4	543	364 (67.0)	56.7–70.7
40–44	190	142 (74.7)	67.9–80.7	103	71 (68.9)	59.1–77.7	181	105 (58.0)	50.5–65.3	55	33 (60.0)	45.9–73.0	529	351 (66.4)	62.1–70.4
45–49	147	119 (81.0)	73.7–87.0	80	53 (66.3)	54.8–76.4	163	102 (62.6)	54.7–70.0	49	24 (49.0)	34.4–63.7	439	298 (67.9)	63.3–72.2
50–54	143	108 (75.5)	67.6–82.3	81	62 (76.5)	65.8–85.2	153	83 (54.2)	46.0–62.3	35	21 (60.0)	42.1–76.1	412	274 (66.5)	61.7–71.1
≥55	124	93 (75.0)	66.4–82.3	57	41 (71.9)	58.5–83.0	73	45 (61.6)	49.5–72.8	26	18 (69.2)	48.2–85.7	280	197 (70.4)	64.6–75.6
Total	807	617 (76.5)	73.4–79.3	567	399 (70.4)	66.4–74.1	762	446 (58.5)	54.9–62.1	313	187 (59.7)	54.1–65.2	2,449	1,649 (67.3)	65.4–69.2

^amRNA⁺, positive for mRNA.

Finally, after adjusting for center, we did not find any difference in the positivity rate (P = 0.735) or in the S/CO values (P = 0.163) between the samples processed immediately after the sampling and those stored for more than 1 year at −80°C.

HPV E6-E7 mRNA positivity by HPV DNA type.

For 1,266 samples tested with Cobas 4800 with valid mRNA results, we analyzed the HPV mRNA results by virus type identified.

The proportion of HPV E6-E7 mRNA positivity was slightly higher in HPV16-positive (81.8%) women and, less markedly, in HPV18-positive (76.7%) women than in women with other high-risk (HR) genotypes (72.2%) (Table 3). The differences were significant when taking into account center and age.

TABLE 3.

HPV E6-E7 mRNA positivity by cytology, type of HPV DNA, and level of HPV DNA positivity expressed in sample/control RLU ratio

Test category	No. of samples tested	No. of mRNA+ samplesi	No. mRNA+/no. tested (%)	95% CI
Cytology
Negativea	1,747	1,062	60.8	58.5–63.1
Low gradeb	479	396	82.7	79.0–86.0
High gradec	153	143	93.5	88.3–96.8
Total	2,379	1,601	67.3	65.4–69.2
HPV type
HPV16d	247	202	81.8	76.4–86.4
HPV18e	86	66	76.7	66.4–85.2
Other HRh HPV types	933	674	72.2	69.2–75.1
Totalf	1,266	942	74.4	71.9–76.8
HC2 RLU ratio
1–1.99	227	24	10.6	6.9–15.3
2–3.99	131	36	27.5	20.0–35.0
4–9.99	124	51	41.1	32.4–50.3
≥10	701	596	85.0	82.2–87.6
Totalg	1,183	707	59.8	56.9–62.6

^aIncludes 14 inadequate samples.

^bIncludes 476 low-grade squamous intraepithelial lesions (L-SIL) and 3 atypical glandular cells (AGC). No slides were classified as ASC-US.

^cIncludes 66 atypical squamous cells suggesting high grade lesion (ASC-H), 86 high grade squamous intraepithelial lesions (H-SIL), and 1 cancer.

^dIncludes samples positive also for HPV18 or other high-risk HPV types.

^eIncludes samples positive also for other high-risk HPV types.

^fOnly samples tested by Cobas.

^gOnly samples tested by Hybrid Capture 2 (HC2).

^hHR, other high-risk HPV types included in the Cobas HPV DNA test: HPV31, -33, -35, -39, -45, -51, -52, -56, -58, -59, -66, -68.

ⁱmRNA⁺, positive for mRNA.

Among HPV mRNA-positive samples, linear regression showed that the ln(HPV mRNA S/CO) was stronger in samples positive for HPV18 than in those positive for HPV16 (P = 0.005), which in turn were higher than in those positive for other high-risk types (P < 0.0005). The HPV type explained only 4.4% of the HPV-mRNA S/CO variability (Fig. 3a).

FIG 3.

Box plot reporting the distribution of the HPV E6-E7 mRNA test results expressed as S/CO values, by HPV type, with 942 cases positive for both HPV DNA and mRNA included (a), and by cytology, with 1,601 cases with HPV DNA and mRNA positive and available cytology results included (b). The line within the gray box represents the median, the limits of the box include 50% of the results, and the lines include 95% of the observations. HPV16 includes samples positive also for HPV18 or other high-risk HPV types. HPV18 includes samples positive also for other high-risk HPV types. Other high-risk HPV types included in the Cobas 4800 HPV DNA test are HPV31, -33, -35, -39, -45, -51, -52, -56, -58, -59, -66, and -68. Negative cytology includes also 14 inadequate samples; L-SIL includes 476 low-grade squamous intraepithelial lesions and 3 atypical glandular cells (AGC); no slides were classified as ASC-US; H-SIL includes 66 atypical squamous cells suggesting high-grade lesion (ASC-H), 86 high-grade squamous intraepithelial lesions (H-SIL), and 1 cancer.

HPV E6-E7 mRNA positivity by cytology results.

A clear trend in mRNA positivity with increasing cytology severity was observed (P for trend < 0.0005) (Table 3); among the high-grade cytology results, 7 of 66 atypical squamous cells suggesting high-grade lesion (ASC-H) (11%) and 3 of 86 high-grade squamous intraepithelial lesions (H-SIL) (3%) showed a negative mRNA result, while the only sample with squamous cervical cancer cells was mRNA positive. The linear regression among mRNA-positive women showed a moderate association with the natural logarithm of HPV mRNA S/CO (P for trend < 0.0005). Cytology explained only 5.5% of the mRNA S/CO variability (Fig. 3b).

Association between HC2 RLU ratios and HPV mRNA results.

For 1,183 samples positive for HPV DNA by the HC2 test and with a valid HPV mRNA result, the correlation between the semiquantitative values of the viral DNA and mRNA results was investigated. Positivity to HPV mRNA strongly increased with higher HPV DNA relative light unit (RLU) ratio (Table 3). The mean DNA RLU ratio of mRNA-negative samples was 17.7 (standard deviation, 101.4), while the mean RLU ratio of mRNA positives was 501.4 (standard deviation, 763.7). Nevertheless, among mRNA-positive samples, ln(HPV mRNA S/CO) increased only weakly with increasing ln(DNA RLU ratio), i.e., beta of 0.09 (95% confidence interval [CI], 0.07 to 0.11), explaining only 10% of the HPV mRNA S/CO variability (Fig. 4).

FIG 4.

Plot of the natural logarithms of the HPV DNA and HPV E6-E7 mRNA results reported as sample/control RLU ratios and S/CO values, respectively. All 1,183 Hybrid Capture 2 HPV DNA- and HPV mRNA-positive samples are included. The regression slope is 0.09 (95% CI, 0.07 to 0.11).

DISCUSSION

In this large population-based study, we observed a 67% positivity rate for HPV E6-E7 mRNA among HPV DNA-positive women, with 2 centers over 70% and 2 centers below 60%. When we applied this value to the whole screened population, and with the assumption that HPV mRNA-positive HPV DNA-negative women are a small proportion of the screened population, ranging from 0.6% to 1.5% (7, 9–11), the positivity rate would be about 5.5 to 6.5%. Our data are consistent with those reported in other studies, ranging from 57% to 85% in HPV DNA-positive women (7, 9–12) and from 4.7% to 10% in the general population (8–13).

In our study, HPV E6-E7 mRNA positivity among HPV DNA-positive women varied between centers and in relation to the HPV DNA assay used, with lower values in centers with higher HPV DNA positivity and vice versa. This phenomenon can be observed also within other studies (11) and when comparing different studies (e.g., references 11 and 12) and could result from lack of HPV mRNA positivity in the presence of HPV DNA test results that are either analytical false positives or true positives for very low viral loads (see discussion below). Indeed, variations in HPV DNA test reproducibility and specificity with different preanalytical procedures have been observed (13, 14, 31).

Interestingly, we did not observe any effect of age on the HPV E6-E7 mRNA positivity rate in the range that we explored, suggesting that despite a higher probability of CIN2 and CIN3 regression at a younger age (15), the probability of an infection to initiate the transforming process is not age dependent. This is consistent with the observation that the positive predictive value of HPV infection for CIN2+ is also not age dependent (16, 17).

As expected, HPV mRNA positivity increased with cytology severity. Almost all the samples with high-grade cytology were HPV mRNA positive (93.5%), with the majority of HPV mRNA-negative samples observed among the ASC-H samples (7 of 10).

We observed higher rates of HPV mRNA positivity in cases of HPV16 and, to a lesser extent, of HPV18 infections, than for other high-risk types. This is consistent with the higher transforming potential of HPV16 and HPV18, but it could also result from differential sensitivity of the test for the mRNA of different HPV types. Surprisingly, among positive samples, HPV E6-E7 mRNA S/CO values were higher for HPV18 than for HPV16, which usually is more frankly related with high-grade squamous lesions. Although this observation can be related to analytical biases, it could also reflect the true transforming ability of HPV18. In fact, this type is strongly associated with cancer, in particular with adenocarcinoma. Maybe, given that HPV mRNA expression is related to squamous cells as well as glandular cells transformation, this test could overcome the problem of low sensitivity of the Pap test for glandular lesions.

Finally, HPV mRNA positivity was strongly associated with the HC2 RLU ratio. This suggests that increasing the HC2 threshold, as suggested by many authors (15), could give an HC2 accuracy similar to that of the HPV E6-E7 mRNA, only slightly affecting sensitivity but considerably increasing specificity. Indeed, samples weakly positive by HC2 have been shown to be more frequently associated with false-positive results, i.e., no HPV types detected by other assays or detection of nontargeted types (18, 19), with infection clearance (20), and with low risk of high-grade lesions (15).

Implication for future research and practice.

Due to the very high positivity rate that we observed, the use of the HPV E6-E7 mRNA test as a triage test for HPV DNA-positive women will entail a greater number of colposcopy referrals than would occur with cytology triage. The baseline colposcopy referral with the HPV mRNA test would be 4.9%, higher than the 3.6% overall recall rate observed in routine Italian activity, in which cytology is the triage test (10), even when including the referrals due to 1-year retesting of HPV DNA-positive/cytology-negative women.

A strategy adopting HPV E6-E7 mRNA as the triage test can be competitive with the cytology triage only if the sensitivity of the HPV mRNA test is very high. This would allow a much longer interval for HPV DNA-positive/mRNA-negative women (21). Published data suggest very high cross-sectional sensitivity of HPV E6-E7 mRNA testing (22, 23), although also triage cytology has shown high sensitivity (24), and 18- or 24-month intervals have been suggested also for it (24–26).

The use of HPV E6-E7 mRNA has also been proposed as a primary screening test. A crucial element is represented by its prospective sensitivity and the resulting safe screening intervals. With the HPV DNA test, a 5-year interval has been shown to be about twice more protective than a 3-year interval with the Pap test (1). To date, longitudinal long-term data for mRNA are based on quite small studies (9).

On the other hand, our data show an important reduction of test positivity in women screened by mRNA, which in the worst hypothesis would be lower by about one percent absolute point than with HPV DNA. This clearly suggests better specificity, as observed by previous studies (9–11, 22, 27, 28).

Furthermore, the test was confirmed to be very robust, with a good separation of S/CO values between negative and positive samples and a negligible proportion of nonvalid results, as observed in previous studies (6, 29, 30). However, it must be considered that the Aptima internal control is only for analytic efficiency and not for sample adequacy, in contrast to the β-globin amplification in the Cobas HPV DNA test.

Limitations.

Data from histological endpoints are not yet available and will be available at the end of 2017. This will allow to calculate cross-sectional sensitivity of all the biomarkers under study. Moreover, given the randomized design for management of women that are HPV DNA positive/cytology negative, we might establish the ability of HPV E6-E7 mRNA test to distinguish between progressive and regressive lesions.

Another main limitation is that the data comparing the results of HPV DNA positivity by the Cobas 4800 and HC2 assays came from only two centers, and this comparison was not based on randomized populations. Since we observed relevant differences between centers, it is difficult to disentangle the effect of the population characteristics in different geographic areas from the effect of a different first-level screening test.

Conclusions.

Due to the observed HPV E6-E7 mRNA positivity rate, the use of this test in the triage of HPV DNA-positive women would bring a 4 to 5% immediate referral to colposcopy. This value is higher than that observed with cytology triage, including both immediate and delayed referrals. Only a very high sensitivity, allowing a long interval before retesting HPV DNA-positive/mRNA-negative women, can make the triage by this test efficient in comparison with cytology.

MATERIALS AND METHODS

Setting.

The NTCC2 trial is still recruitng women in 5 Italian centers. The preliminary results presented here come from 4 centers: Turin, Florence, Umbria, and Veneto. Recruitment started in April 2014 in Umbria, in September 2014 in Veneto, in June 2015 in Florence, and in February 2016 in Turin. Data from the fifth center, Trento, which started recruitment in June 2016, were not available at the moment of preliminary analyses. Here we report data of preplanned preliminary analyses for women recruited up to May 2016.

Women were recruited within HPV DNA-based organized screening programs for cervical cancer with a well-established call and recall invitation system and a fixed protocol observing Italian recommendations (3). All second-level and follow-up visits are managed with fixed appointments within the screening program facilities. Software for the women's management and the registration of individual level data is available in each center.

Study design.

The trial aimed at determining long-term (>5 years) cumulative incidence of CIN2+ in HPV DNA-positive women with cytology and study-targeted selected biomarkers (HPV E6-E7 mRNA and p16INK4a/ki67) as triage tests. Cytology-positive women are directly referred to colposcopy, as in routine practice, while HPV DNA-positive/cytology-negative women are randomly assigned to immediate colposcopy or to repetition of the HPV DNA test after 1 year (the latter being routine practice in Italian HPV-based screening programs) (3). Here we present only data on HPV E6-E7 mRNA testing in HPV DNA-positive women at recruitment.

Women aged 25 to 59 years showing up for a new screening episode were eligible for the study. In Umbria, also women aged 60 to 64 were recruited to get cross-sectional information at all the screening target ages. Exclusion criteria were pregnancy, treatment for a CIN2+ lesion in the previous 5 years, not a new screening episode (i.e., repetition after inadequate or postcolposcopy), and not resident in the program area.

All women matching the inclusion criteria were asked to participate in the trial, and those who accepted signed an informed consent. Nonconsenting women were screened by HPV DNA as for routine practice. The study was approved by the ethics committee of the S. Giovanni Battista University Hospital, Turin, Italy, on 20 June 2012 (N. CEI513) and then by the local ethics commitees of all the recruiting centers. Amendments were approved on 30 September 2015.

Cervical samples were collected in PreservCyt solution (Thin Prep; Hologic).

The HPV DNA test was performed locally, with different tests according to the period of collection and the center: Cobas4800 (Roche Diagnostic) in Umbria, Hybrid Capture 2 (HC2; Qiagen) in Turin; initially HC2 and then Cobas 4800 in Veneto and Florence.

From samples that gave positive results to HPV DNA testing, one milliliter of the remaining PreservCyt liquid cytology specimen was transferred into an Aptima specimen transfer tube (Hologic), according to the manufacturer's instructions. Simultaneously, another two 1-ml aliquots were stored at −80°C for biobanking. Aptima samples were sent to two reference laboratories (Centro Unico Cervico-Vaginale, Turin, for samples from Turin, Veneto, and Trento; ISPO, Florence, for samples from Florence and Umbria) for the Aptima mRNA test on the Panther system (Hologic). PreservCyt samples collected before June 2015 were stored at −80°C and then mailed frozen to the reference laboratories, where they were transferred into Aptima tubes.

Cytology.

Liquid-based slides (ThinPrep; Hologic) were prepared, and cytology was interpreted in the local screening pathology laboratory. Cytology results were classified using the Bethesda 2001 system. Data were grouped as negative for intraepithelial lesions, low grade (including L-SIL, ASC-US, and AGC), and high grade (including ASC-H, H-SIL, adenocarcinoma in situ [AIS], and cancer). It is worth noting that in the triage of HPV DNA-positive women, the ASC-US category is no longer used in Italy (10).

Laboratory methods. (i) DNA Cobas 4800 HPV test.

The Cobas 4800 HPV test (Roche Diagnostics) is a qualitative in vitro test that utilizes amplification of target DNA by real-time PCR for the detection of 14 high-risk (HR) HPV types in a single analysis allowing a concurrent partial genotyping for HPV16 and HPV18 and for the other 12 high-risk HPV types (HPV31, -33, -35, -39, -45, -51, -52, -56, -58, -59, -66, and -68) as a pooled result. Beta-globin from cellular input is used as the internal control for specimen quality. The test is performed on the completely automated Cobas 4800 System (Roche Diagnostics) according to the manufacturer's instructions.

(ii) HC2 HR HPV DNA test.

The Hybrid Capture 2 assay (HC2, Qiagen, Hilden, Germany) is a nucleic acid hybridization assay with signal amplification for qualitative detection of 13 HR HPV types (HPV16, -18, -31, -33, -35, -39, -45, -51, -52, -56, -58, -59, and -68) in cervical specimens. DNA was extracted using the QIAsymphony DSP HPV Media kit (Qiagen) and hybridized using the Rapid Capture System (RCS; Qiagen) according to the manufacturer's instructions.

Semiquantitative HC2 results were expressed as the ratio of the specimen's light emission (relative light units [RLU]) to the mean of three concurrently tested positive controls (CO). Samples were considered positive when the ratio (RLU/CO) was ≥1.

(iii) HPV mRNA Aptima test.

The Aptima HPV assay (Hologic) detects E6-E7 viral mRNA from 14 HR HPV types (HPV16, -18, -31, -33, -35, -39, -45, -51, -52, -56, -58, -59, -66, and -68) detected as a pool. The assay has been performed on the Panther System according to the manufacturer's instructions.

The assay incorporates an internal control (IC) to monitor nucleic acid capture, amplification, and detection, as well as operator or instrument error. Results are reported as positive or negative according to light emitted from the labeled RNA-DNA hybrids: a signal-to-cutoff (S/CO) value is calculated, and samples with an S/CO of ≥0.5 are considered positive.

An interlaboratory quality control on 20 samples (8 positive and 12 negative samples) showed a perfect concordance for positivity and 0.95 correlation coefficient for the S/CO values.

Statistical analyses.

Prevalence of HPV E6-E7 mRNA positives among HPV DNA-positive women by age, center, HPV type, and level of HC2 RLU ratio are given with their relative 95% confidence intervals (95% CI). To take into account the effect of age and center, the associations of mRNA results with HPV type, HPV DNA load (semiquantitative value), or cytology were evaluated with a logistic regression model for dichotomous results of positivity or with a linear regression model for HPV E6-E7 mRNA semiquantitative S/CO values.

In HPV mRNA-positive samples, the natural logarithm of mRNA S/CO was studied as a function of ln(HPV DNA RLU ratio) by linear regression. R² is interpreted as the proportion of the variance explained by the independent variable.