The main cause of invasive cervical cancer is the deregulated and persistent production of HPV E6 and E7 oncoproteins 
. Hence, HPV E6/E7 mRNA is a rational target for detecting HPV infections leading to cellular transformation. PreTect HPV-Proofer is a commercial assay for the detection of E6/E7 mRNA from the five main high-risk HPV types 16, 18, 31, 33, and 45. Due to the higher clinical specificity and PPV of this method compared to other HPV tests 
, this was the method of choice at our hospital when HPV testing was introduced in secondary screening in Norway.
Our data show that when including all age groups, the HPV mRNA test has a PPV for CIN2+ of 70%. A previous data analysis 
shows that the PPV of repeat cytology for CIN2+ is 29.8% (unpublished data; this number is overrated due to known HPV status). In a meta-analysis by Marc Arbyn and colleagues 
, a PPV of repeat cytology of 11.8% for ASC-US and 23.2% for LSIL is reported; the PPV of repeat cytology at cutoff ASC-US or worse ranged from 3.8% to 22.2%. Taken together, our results indicate that the PPV of the HPV mRNA test strongly exceeds that of repeat cytology, which is an important finding. The 30% with a positive HPV test but with no confirmatory high-grade lesion may be explained by the observation that colposcopic biopsy misses 26–42% of prevalent CIN2+ (depending on the severity of the lesion and the number of biopsies taken) (Marc Stoler, Eurogin 2010), with the main problem being the subjectivity in determining the location of the lesion 
. The question then arises whether the 30% of the HPV mRNA positive cases with no confirmatory high-grade histological diagnosis actually may represent false negatives due to a small lesion or a biopsy not representing the true lesion. Also, HPV 18 infections have been described as being specifically difficult to detect histologically and may therefore lead to a false negative histological diagnosis 
. Interestingly, in several cases we have demonstrated that women with an originally benign or low-grade histological diagnosis have underlying lesions when the biopsy is more thoroughly examined after a positive HPV mRNA test result (for example by making deeper sections of the biopsy or by p16 immunostaining) indicating that, in several cases, a false negative histological result was revealed.
The PPV of the diagnoses ASC-H and HSIL for CIN2+ are, according to Norwegian Cancer Registry data, 56.4% and 73.9% respectively 
. The respective numbers from our lab are 52.7% and 76.9% (unpublished data; first cytology 2005–2008, CIN2+ up to end 2008). The data in the present study show that by including HPV E6/E7 mRNA testing, the PPV of test positive women with ASC-H or HSIL is 79.4% and 94.2%, respectively. This strongly suggests that women with HSIL and a positive HPV mRNA test can be directly referred to treatment. For women with ASC-H, the HPV test will be a valuable tool in separating women requiring follow-up (or treatment if >40 years of age or not planning for children) from women only having a chronic inflammation (seen in more than 40% of women with ASC-H). Women with HSIL and a negative HPV mRNA test should be referred to colposcopy and histology in the traditional way.
Women with ASC-US or LSIL at repeat cytology and with a positive HPV mRNA test are referred directly to colposcopy and biopsy. When an HPV test is not used, women with ASC-US and LSIL are referred to colposcopy and biopsy after three times repetition of the diagnosis. This means that for women with an underlying high-grade lesion, the disclosure of their disease will be delayed with up to two years, which may be critical. The same applies to women with normal cytology at triage, for whom the PPV of HPV mRNA test positives for CIN2+ is 65.0%. Women with normal cytology (when the HPV test is not included) are referred back to the screening program, emphasizing the importance of an HPV test for these women. Of note is that the PPV for mRNA positive women with normal cytology is higher than the PPV for women with ASC-US (). This may be explained by small sample size and overlapping confidence intervals. It may also be explained by the fact that for women with normal cytology, the guidelines recommend biopsy only when a persistent infection is demonstrated (HPV positive with six months interval) and consequently the risk of having a CIN2+ is higher.
An HPV mRNA positive HPV 16 result has a PPV for CIN2+ of 80.4% when including all age groups, representing a higher PPV than the cytological HSIL diagnosis. For the other HPV types, the PPV is somewhat lower. One explanation for the lower PPV of HPV 18 and HPV 45 may be the lower degree of noticeable cell changes related to these infections and the fact that lesions caused by these HPV-types often are located further up in the cervical canal (glandular lesions) 
. Therefore, for these infections, if the original biopsy is benign or low-grade, diagnostic conization might be more important than for infections with other HPV types.
The motivation for looking at the age group >40 years is related to the lower probability of children planned among women in this group. Given the impact of cervical treatment (such as LEEP conization) on pregnancy outcomes 
, cervical treatment is more critical for women planning to have children. In addition, if a high-grade lesion is found in a woman above 40 years of age, there is a higher chance that the lesion has been present for a longer period of time, in turn making surgery more important. Accordingly, for women above 40 years of age, the PPV of 70.3% (83.7% for HPV16 and 84.6% for HPV 33) suggests that HPV mRNA test positive women can be referred directly to treatment without histological confirmation, contributing to a better clinical safety for this age group. For the 30 women above 40 years of age with no confirmed high-grade lesion, 14 had a negative HPV test at follow-up (out of a total of 15 women receiving a follow-up HPV test; one had a persistent HPV 45 infection). Fifteen women were followed up by cytology only, and of these, three had ASC-US, one had LSIL and 11 had at least one normal cytological smear and were consequently referred back to the screening program. However, these women may have lesions further up in the cervical canal that are difficult to detect morphologically, as often seen for HPV 18 or 45 infections. In our case, nine of these 15 women had either an 18 or 45 infection. We would therefore suggest that all women with a positive HPV result should be followed up with a repeated HPV test after six months. Also, among postmenopausal women, a higher degree of atrophic changes are present, often being difficult to distinguish from dysplasia. For these women, an HPV test will be crucial for the evaluation of the sample and subsequent follow-up of the woman.
and show that for 59 out of 406 women with a positive HPV mRNA test, no histological diagnosis is reported. There may be several reasons for this. First of all, for women with a normal cytological diagnosis at triage, referral to biopsy is recommended only if they are found to have a persistent HPV-infection or a HSIL at follow-up. A new HPV test is only performed for about half of these women, due to misunderstandings related to the new strategies for clinical management or to the fact that a conventional Pap-smear is taken, not allowing for HPV testing. Some times a new cytological smear is done, although a biopsy has been requested (for example after a persistent HPV result). If this smear is found to be normal, the woman may disappear from the system and is a typical example of a woman “lost in follow-up”. Another reason for lack of histological data may be that the woman is pregnant and a biopsy is not recommended for this group. Alternatively, the woman may have moved to another city and consequently the follow-up of this woman has been done elsewhere.
An added value of the HPV mRNA test in detecting women with high-grade lesions is presented in , among others showing that five women with an ASC-US or LSIL diagnosis at triage had underlying carcinoma revealed after a positive HPV test. Another added value of the HPV test is seen for women with a benign or low-grade diagnosis at first biopsy. If the HPV test is positive, a new Pap-smear and HPV test are recommended after six months. In our material, 23 women with a benign, uncertain or CIN1 diagnosis at first histology had an underlying high-grade lesion revealed during follow-up after a positive HPV result (). For these women, the first biopsy or histological slide may not have been representative for the underlying disease. Alternatively, a microlesion difficult to detect by colposcopy-directed biopsy or by histology may have been present. Nevertheless, this in turn emphasizes the need for additional methods in order to increase the overall sensitivity of biopsy and histology.
In total, 13 women (with ASC-US, LSIL or HSIL at triage) were diagnosed with cancer. Interestingly, the HPV test had an important role in detecting most of these women, explained by the knowledge that the cytological diagnosis is influenced by the HPV result and by the follow-up of women with normal or low-grade cytology due to a positive HPV result (). Consequently, the HPV mRNA testing led to an earlier diagnosis and better prognosis: All but one of these women had a final diagnosis of either microinvasive carcinoma or stage 1A carcinoma; one patient had stage 1B carcinoma. Nine women had hysterectomy and for four women, conization was sufficient to eliminate further disease.
To conclude, our data suggest a clinical “test and treat” application of the HPV mRNA test: Direct treatment may contribute to a better clinical safety for test positive women above 40 years of age and for women with a concurrent cytological HSIL diagnosis. Consequently, by decreasing the time between testing and treatment, thereby reducing the number of recalls, the patient management algorithm will be considerably improved and will in turn reduce follow-up costs and more importantly, unnecessary psychological stress among patients will be avoided.