**4. Discussion**

Persistent infection caused by HR HPV is the leading risk factor for developing cervical intraepithelial lesions and cervical cancer (reviewed in [20]). Many countries have introduced screening programs based on the cytomorphological examination of cervical samples using the PAP test during the last 60 years to reduce the morbidity and mortality caused by cervical cancer. However, this procedure has shown less than optimal sensitivity (50%) and high inter- and intra-individual variability [21,22]. HPV DNA detection and genotyping provide an efficient screening method and enable risk stratification. However, just proving the presence of the HPV genome in a cervical epithelium does not provide insight into the type of infection. It does not answer whether there is a transient or persistent infection in which the virus is actively multiplying and in which there is a high risk of cancer developing (reviewed in [23]). It was established that cervical carcinogenesis is strongly associated with the HPV-caused infection in which the transcription of E6 and E7 HR HPV oncogenes occurs, with the consequent increase of their mRNA and protein levels. For this reason, the detection of E6 and E7 mRNA of HR HPVs can serve as a promising biomarker of their persistence and oncogenic activity (reviewed in [23]), which could enable a better assessment of the progression to high-grade cervical intraepithelial lesions and, in this regard, significantly influence the algorithm for monitoring patients (reviewed in [18,23,24]).

In our study, 365 female patient samples from Serbia were tested for the twelve HR HPV genotypes. It should be emphasized that the study has limitations. Used HPV tests are non-validated according to European Guidelines for use in primary screening (Meijer HPV test criteria) or according to the VALidation of HPV GENoyping Tests (VALGENT) protocols [25]. The commercial molecular HPV test was chosen for research according to Serbia's general requirements for molecular diagnostics and the limited research expenses. From the total number of tested samples, 246 (67.4%) samples were positive for at least one of the 12 tested HR genotypes. These results agree with previous studies in the same area and European countries, where a high prevalence of HPV was registered. Studies in Serbia have reported that the presence of HR HPV infections range from 50% to 79% of women [26,27]. In the countries of our region, such as Croatia, a similar prevalence was shown (59%) [28] and in Bulgaria (61%) [29], while a slightly lower prevalence was registered in Italy (53%) [30]. Contrarily, a low prevalence of HPV infection was registered

in Western and North European countries. Some of them are Great Britain (20.6%), Sweden (9.7%), and the Netherlands (3.8%) [7].

HPV genotype 16 (38%) emerged as the most frequently detected genotype in the study group, in concordance with previous reports [26,27]. The presence of HPV 16 is more often present (76%) in high-grade lesions compared to lower-grade lesions of the cervical epithelium. Studies on the HPV 16 genotype's prevalence concerning cytological findings confirm these results (reviewed in [6]). The following frequencies also represented the Alpha-9 genotype, HPV 31 (17%) and HPV 33 (9%). A meta-analysis study on five continents shows that HPV 31 is especially frequent in Europe [31]. The results of this research indicate that the frequency of the HPV 31 genotype statistically significantly decreases with the progression of the intraepithelial lesion. The decreasing trend of the same genotype's presence according to the degree of severity of the cytological lesion, more precisely, a lower prevalence in cancers compared to precancerous lesions, is observed in research from other studies [32–34]. Like HPV 31, a prevalence decrease of HPV 33 in HSIL compared to normal cytology was also registered in this research. Globally, HPV 33 ranks fourth in frequency and is responsible for 4.2% of all registered cervical cancers [6]. The surprising fact is a markedly high prevalence (9%) of the Alpha-5 genotype HPV 51 in our research. In the context of the causative agents of cervical cancer, HPV 51 is not included in the top ten most frequent HPV genotypes registered worldwide [6]. However, the detection of this genotype within this research, as well as previous studies from our region [26] and certain European countries [35–41], places it among the first four most prevalent genotypes detected in precancerous lesions of the cervix. Since HPV vaccines do not protect against all oncogenic HPVs, such as HPV 51, a complete understanding of its oncogenic activity is particularly significant [42]. The remaining tested genotypes (HR HPV 52, 56, 45, 18, 59, 58, 39, and 35) were present in a low percentage which is in concordance with previous reports [26]. Although HPV 18 is considered to be responsible for 15% of invasive cervical cancers, it is essential to note that its prevalence is similar to some studies from neighboring countries [29,43,44]; we found that in our area, HPV 18 was present at a lower percentage. Its frequency (3.6%) was 10-fold lower than that of HPV 16 (Figure 1). The HPV vaccine was introduced in over half of the WHO member countries in 2020 [5]. There is scientific data from numerous countries that have implemented HPV vaccines in their routine immunization programs on decreasing the burden of cervical HPV infections and precancers [45]. According to our data, using the nine-valent vaccine could prevent more than 80% of the cervical precancerous lesions identified in this study.

The presence of HR HPVs determined further examination of their oncogenic potential. In our study, the expression of E6/E7 mRNA HR HPV was identified in 67% of the HPVpositive samples (Table 6). The percentage of expression of the E6 and E7 HPV-examined HR HPVs is proportional to the degree of severity of the cervical lesion (Table 7). It can be observed that approximately every tenth HPV-infected woman (11%) with normal cytological findings is infected with an oncogenically expressed HPV. At the same time, in HSIL, this relationship is the opposite. Namely, the absence of indicators of HPV oncogenic activity is detected in approximately every tenth woman with HSIL status (11.1%). An undoubted trend in E6/E7 mRNA HR HPV positivity with increasing cytology severity has been observed in the data of studies conducted in different regions of the world [46–50]. In agreement with the report of Argyri et al. (2013) and according to the mRNA test, E6/E7 expression was prevalent in 9.1% of women with normal cytology, similar to our study [51]. However, other previously published data indicated the prevalence of those transcripts in a lower proportion of women with normal cytology findings than our study (0%) [52].

In our study, HPV 16 constituted 29.7%, followed by genotype 31 (19.2%), genotype 33 (9.3%), and genotype 51 (8.7%). The results showed that approximately every second HPV 16 genotype is oncogenically expressed (48.6%). Our data agree with the results reported in other studies. Rossi et al. (2017) observed a positivity rate for E6/E7 mRNA HR HPV, ranging from 58% to 77% in HPV-DNA-positive women [50]. A study by Tüney et al. (2017) in Turkey registered 55.6% E6/E7 mRNA HR HPVs, where HPV genotype 16 constituted

57.8% [24]; similar to that, Bruno et al. (2018) found that the HPV 16 genotype was the most oncogenetically active [46]. As confirmed in this research, it is also stated by several other authors that the transcription product HPV 16 is most often detected, which indicates that the tendency of expression of the essential oncogenes of this genotype is significantly higher compared to the other examined genotypes [24,53,54], which gives it the status of the HPV with the most carcinogenic potential [55]. The oncogenic activity of HPV 31 was detected in approximately every fifth (19.2%) HPV-DNA-positive sample. Contrary to HPV 16, the opposite trend is observed with the degree of cervical lesion severity. It can be assumed that the negative transcriptional status of E6 and E7 oncogenes is more prevalent due to the presence of an episomal form of the virus or an established transcriptional control that enables the spontaneous elimination of infection. Within this research, the results of E6/E7 mRNA HPV 51 detection are approximately the same as those of HPV 33. The distribution of oncogenic activity of these two genotypes is approximately the same across cytological groups and remains at a low level (2.2–11.4%) (Table 7). Other studies have registered different percentage ranges of transcriptional detection of oncogenes of a particular genotype (HPV 33), from its absence [52] to complete expression of 100% [55]. The reason for those differences can be explained by the variations in their number in the total sample [55]. Furthermore, the total oncogenic activity increases with the degree of cervical lesion severity (60–100%). The oncogenic activity of individual genotypes (20–86%) is higher than that of multiple genotypes (0–40%) in the same type of lesion and increases with the degree of its severity (Table 8). These data are supported by the literature. Each of the detected genotypes is considered to have an independent mechanism of action in oncogenesis [56–59], which supports the hypothesis that different cells being infected with different viral genotypes rather than their intracellular coexistence is possible [60].

To analyze the oncogenic activity of the four most frequently diagnosed HR HPVs related to the age categories, a statistically significantly higher prevalence of positive E6/E7 mRNA HPV 16 findings was observed in the older age groups (Table 9). In agreement with this result, studies have shown that E6 and E7 mRNA HR HPV detection is significantly higher after 30 years [48,61]. The same result is supported by population-based cohort studies, which state that the majority of young women have an asymptomatic HPV infection which, thanks to the immune response, acquires transitory status [49,62].

The growing interest in molecular diagnostics methods has led various authors to compare the characteristics of the HPV DNA test with the mRNA test, evaluating the diagnostic accuracy in identifying high-grade cervical lesions. Our data showed that the specificity (89%) and PPV (70–79%) of the mRNA test are statistically significantly higher, while the same test has a statistically significantly lower sensitivity than the HPV DNA test. These results agree with the statements of different authors, who emphasized slightly lower values of the sensitivity of the mRNA test compared to the sensitivity of the HPV DNA test. At the same time, the specificity is expressed at a higher level [48,49] (reviewed in [63]). In contrast, others suggest that the sensitivity is similar and that its application would significantly improve the assay's specificity characteristics [49]. Studies evaluating the reliability of the mRNA assay showed heterogeneous findings (reviewed in [32]). The meta-analysis results indicated that the obtained sensitivity value ranged from 41% to 95%, while the registered specificity was 42–97% (reviewed in [64]). The observed difference is explained by the heterogeneous participation of cervical pathology [64] and methodological quality in different studies, which highlights the limitations in the general interpretation of these test characteristics [32]. Although the results are presented broadly, they maintain a unique trend and suggest that the mRNA test over the HPV DNA test improves specificity [64]. By comparing the clinical characteristics of the test, it can be concluded that the detection of E6 and E7 mRNA HR HPV compared to HPV DNA represents a much better marker for more accurate screening of high-grade cellular atypia of the cervix (reviewed in [32]), which make it an appropriate tool for the secondary screening of cervical cancer [47,64].

In our study, the results of the ROC curve analysis indicated that the probability of detecting HPV infection with the mRNA test (81%, AUC = 0.812) in patients with highgrade lesions is higher than the possibility of diagnosing them with the HPV DNA test (74%, AUC = 0.740). The obtained results are in line with the other studies emphasizing the potential usefulness of this test. Sun et al. (2021) also found higher AUC values for the mRNA assay compared to the DNA HPV (0.929 vs. 0.833) [65], while according to Yao et al. (2017), the clinically relevant portion of the AUC of mRNA was 0.721 [66].

According to the previously established degree of influence on diagnosing HSIL, relevant factors were selected in our research. Factors that represent an increased risk for more severe cervical changes were gradually examined (the HPV 16 genotype, the total oncogenic activity of all of the tested HPV genotypes, the HPV 16 oncogenic activity, and the age category of the patient).

Firstly, the HPV-DNA-16-positive results have a statistically significant predictive value for diagnosing HSIL (Table S1). In the same context, the study of Bruno et al. (2018) stated that the presence of the HPV 16 genotype was associated with a five-fold higher risk of developing a high-grade lesion compared to women with the presence of another HPV genotype [46]. Similarly, the data from a recent study indicate that type-specific HPV persistence predicted high-grade lesions, with HPV 16 being the most common type [67].

Secondly, the oncogenic activity of all of the tested HR HPVs has a statistically significant predictive value for diagnosing HSIL (Table S2). The obtained results support the statements related to the research on the importance of oncogenic activity, which show that the detection of E6 and E7 mRNA HR HPV could have a prognostic value in monitoring the development of carcinogenesis [68,69]. The results of the study by Fontecha et al. (2016) [55] showed that in the highest percentage of E6 and E7 mRNA HPV-positive women, the progression of the lesion is diagnosed over time (53%), followed by the persistence of abnormal cytological findings (42%), while regression was recorded in 10-fold lower cases (4%).

Furthermore, in patients with positive results of indicators of the oncogenic activity of HPV 16, a statistically significantly higher probability of diagnosing a high-grade lesion was determined in all types of cytological groups (Table S3). According to the previous insights into the published statements, one of the few prospective follow-up studies that dealt with the predictive value of E6 and E7 mRNA HPV 16 indicated that through the detection of this biomarker, it is possible to identify 87.5% of the HPV infections that progressed. In this case, the risk of progression of negative cytology and low-grade cervical lesions was 10-fold higher than that detected in mRNA-negative women during a follow-up period of 35 months [70]. This is supported by the results of Johansson et al. (2015), which indicated that the absence of E6/E7 mRNA HPV demonstrated a high negative predictive value for the future development of high-grade lesions of the cervix among HR-HPV-DNA-positive women with ASCUS and LSIL [68].

Within the results of this research, a statistically significant final model with all of the predictors was constructed, which shows the strength of the potential of the examined factors, that is, biomarkers for diagnosing precancerous lesions in women with HR HPV infection. Looking at each cytological group individually, two independent variables made a statistically significant contribution to the model and were thus named as the strongest predictors. These are the oncogenic activities of HR HPV 16 and the age category (≥45 years). It is known that the presence of HPV infection is confirmed in all age categories. However, belonging to a particular age group is a determinant significantly associated with the risk of acquiring this infection, depicting the peak in prevalence, which generally takes place around 20–25 years of age. For lesions to progress to more severe forms of cervical disease, a period of 5–14 years is necessary, during which the infection persists and the process of oncogenesis takes place [71]; the results of this research confirm the stated findings. In the examined women, the age category (≥45 years) is a statistically significant prognostic factor for diagnosing HSIL in all of the cytological groups (Tables 12 and S4). Loopik et al. (2020) [58] indicated that the risk of progression of an existing high-grade

lesion increases with age, i.e., in women over 50, the risk of developing cervical cancer increases by seven fold. In addition, the risk of developing cervical abnormalities and the need to use an mRNA test in the diagnostic protocol of HPV-DNA-positive postmenopausal women with normal cytology is emphasized by Asciutto et al. (2020) [72].
