Cervical cancer (CC) is the second most common tumor found in women of reproductive age [1]. Human papillomavirus (HPV) was proven be a mandatory prerequisite for most malignant neoplasms (MN) of the cervix [2, 3]. In recent years, most countries have changed the paradigm of national CC screening programs and chose HPV testing over cytological examination as a more sensitive method. Up to 90% of HPV infections are transient: the virus self-eliminates within 1–2 years [3, 4]. However, in 10% of cases, it persists, and sometimes cause development of a tumor [5, 6]. The global prevalence of HPV varies from 43% to 72.8% [7, 8]. In the USA alone, more than 14 million new cases of HPV infection are registered every year [9]. It should be noted that the development CC is preceded by squamous intraepithelial lesions of the cervix. Among HPV-positive women, low-grade squamous intraepithelial lesions of the cervix (LSIL) occur in 11.5% of cases, while 10.3% of them suffer from high-grade squamous intraepithelial lesions (HSIL) [10].

Currently, the detection of DNA of HPV directly in the affected cervical tissue is perceived as an important marker of viral activity that is fundamentally important for the prognosis of the likelihood of development of precancerous and cancerous lesions of the cervix [11–13]. The detection of HPV in a histological sample is a reliable sign of the level of contamination because the said sample originates from the affected part of the tissue, where the registered virus of a certain genotype and in a certain amount acts as the etiological cause of the pathological process. However, there is still no accurate understanding of the difference between finding DNA of HPV in a cervical smear and in a cervical biopsy.

This study aimed to investigate the specifics and patterns peculiar to the detection of DNA of HPV in the paired cervical canal smears and cervix histological samples.

METHODS

For this cross-sectional study, we used the biological material from 99 HPV-positive patients who sought cervical pathology diagnosing at the Scientific Outpatient Department of the V. I. Kulakov National Medical Research Center for Obstetrics, Gynecology and Perinatology from January to December 2023. In the context of the comparative analysis aimed at learning the diagnostic value of detecting HPV DNA, the genotype of the virus was identified in a pair of samples taken from each participant, one being the secretions of the cervical canal, another — cervix biopsy (colposcopy-guided). All patients were divided into three groups depending on the histological verification of the diagnosis:

1. Group 1 — diagnosed with LSIL (n = 34);
2. Group 2 — diagnosed with HSIL (n = 31);
3. Group 3 — diagnosed with chronic cervicitis (n = 34) (control group).

The inclusion criteria were: age 18–65 years; HPV-positive status confirmed by the cervical canal secretion test; LSIL, HSIL, chronic cervicitis confirmed by the histopathological examination of the cervix biopsy.

The exclusion criteria were: pregnancy, lactation; malignant neoplasms of the cervix; non-specific inflammatory diseases at the decompensation stage.

In the context of this single-stage cross-sectional study, all patients underwent the following:

1. medical history taking, clinical examination (general and bimanual examinations);
2. real-time PCR testing for DNA of HPV using a diagnostic panel with 21 HPV genotypes (6, 11, 16, 18, 26, 31, 33, 35, 39, 44, 45, 51, 52, 53, 56, 58, 59, 66, 68, 73, 82);
3. extended colposcopy to visualize a potentially pathological area of the cervical epithelium for subsequent targeted cervical biopsy;
4. cytological examination (liquid-based cytology);
5. colposcopy-guided cervical biopsy followed by histological verification of the diagnosis;
6. identification of DNA of HPV, its genotyping and assessment of the viral load by real-time PCR (in histological samples).

For the real-time PCR test for DNA of HPV, we used an array designed to detect 21 genotypes of HPV. The materials for the test were cervical canal secretion samples and cervix bioptates harvested under CT guidance. The bioptates were fixed in neutral buffered formalin (pH = 7.0) for 24 hours. After fixation, the samples were put into Histo-Tek VP1 (Sakura, Japan) for tissue processing, and semi-automatic paraffin embedding station (Leica HistoCore Arcadia, Germany) was used to form paraffin blocks. Paraffin sections 5 µm thick were put into 1.5 ml dry tubes and transported for HPV DNA testing, same as the tubes with the cervical canal secretion samples. To prevent compromising of the results of comparison of HPV DNA identification in smears and bioptates, all samples were harvested within a single menstrual cycle. Before DNA isolation from the bioptates, paraffin was removed and samples treated with proteinase K using the Proba-PK kit (DNA Technology, Russia), then we isolated the DNA with the help of the ProbaNK kit that precipitates nucleic acids in alcohol. The viral load in the range from 0 to 3.0 lg was regarded as low, in the range from 3.1 to 5 lg — moderate, and values from 5.1 lg and up were considered to represent a high viral load [14].

Statistical data processing methods

For statistical data analysis, we used the IBM SPSS Statistics Version 20 software package (IBM, USA). The normality of distribution of the variables (under the Gaussian law) was checked with the help of the Kolmogorov–Smirnov test. At this step, it was established that most variables do not distribute normally, therefore, as is customary for data violating the Gaussian law, we used the median, which is less susceptible to extreme variations. The dispersion measures were represented by the upper and lower quartiles (Me (Q₁; Q₃)). Categorical data were presented as a percentage with a 95% confidence interval (95% CI) calculated using the Wilson method. Nonparametric Mantel-Haenszel chi-square value tests were used to compare variables depending on their properties (quantitative or categorical). The level of significance was set at p < 0.05.

RESULTS

Among the 99 pairs of samples, we registered DNA of HPV in 175 cervical canal smears and 111 biopsy specimens. In the smears, the distribution was as follows: chronic cervicitis — 53/175 (30.3%), LSI — 60/175 (34.3%), HSIL — 62/175 (35.4%). In the biopsy specimens, this characteristic was less uniform: chronic cervicitis — 22/111 (19.8%), LSIL — 40/111 (36.0%), HSIL — 49/111 (44.1%). A noteworthy fact is that in the biopsy samples from patients diagnosed with chronic cervicitis and HSIL, we registered HPV DNA significantly less frequently than in the secretion samples paired with the respective biopsy specimens (p = 0.048 and p = 0.034, respectively). We have also considered the distribution of HPV genotypes through the lens of the method of sampling (fig. 1, fig. 2). For 25/111 (22.5%) cases of HPV DNA detection in the biopsy specimens, we failed to find identify HPV genotypes in the paired smears. In 8/25 (32%) of such cases, the registered genotype was HPV-16, in 7/25 (28%) cases — HPV-33, and in 6/25 (24%) cases — HPV-51. It is important to note that HPV-16 was most frequently identified in the biopsy specimens of the patients from Group 1 (6/8, 75%).

The comparative analysis has revealed that in 49.1% of cases HPV DNA was detected in both the cervical canal smears and the paired cervical biopsy specimens. It is also important to take into account the intra-group overlap of the HPV testing data. In Group 3 (chronic cervicitis), the level of coincidence of detection of HPV DNA in both samples of the pair was 28.3%, in Group 1 *(LSIL) — 45%, in Group 2 (HSIL) — 67.7% (fig. 3).

One of the important components of this study was the assessment of the prevalence of simultaneous detection of two or more HPV genotypes. Several HPV genotypes were detected in 46/99 (46.5%) of the cervical canal smears. However, the respective tests made on the biopsy specimens returned only 13/99 (13.1%) of such cases. This difference is statistically significant (p < 0.001).

Analyzing the correlations between HPV test results, we discovered a noteworthy phenomenon. HPV-44, which is one of the three most prevalent genotypes (14/175. 8%) and is more associated with HSIL than with LSIL or chronic cervicitis (50% vs 28.6% vs 21.4%, respectively), was not detected in of the cervical biopsy specimens. At the same time, for other most common types, the matches within the pairs were frequent: HPV-16 — 30/32 (93.8%), HPV-33 — 13/16 (81.3%), HPV-31 — 11/16 (68.8%) (fig. 4).

The aspects of the study that are of particular interest are investigation of the effect of HPV viral load on the various indicators, such as the incidence, frequency of simultaneous detection of several genotypes, and the degree of cervical lesion. In particular, we analyzed how the amounts of HPV affect the frequency of matching results in sample pairs. Among cervical canal smears, 62 exhibited high HPV viral load, 65 — moderate, and 48 low viral load. Among the biopsy specimens, the distribution was as follows: high viral load — 52 samples, moderate viral load — 37 samples, low viral load — 23 samples. It should be borne in mind that 22.5% of HPV DNA detections in the biopsy specimens were first-time, which means that these cases were not factored into the calculation of correlations between quantitative HPV test results. The analysis has shown that when the viral load is low, the results of the HPV DNA detection tests matched in 27.1% of sample pairs, when the load was moderate — in 35.4% of the pairs, and when it was high, 82.3% of sample pairs exhibited matching results (fig. 5). It is important to note that the differences in the amounts of matching results in high vs. moderate viral load, and in high vs. low viral load cases was significant (p < 0.001 for both comparisons). Cramér's V showed a relatively strong correlation between the level of viral load and the probability of matching HPV test results (table). Comparison of the frequency of matching HPV DNA detections in moderate and low viral load cases revealed no significant differences (p = 0.456).

In addition, we calculated the correlation of HPV test results depending on each viral load range, taking 1 lg as the increment. It was established that an increase of 1 lg raises the probability of confirmation of infection with the HPV genotype detected in the cervical canal smear (in an initially determined quantitative ratio) by 9.3%.

DISCUSSION

There are very few studies dedicated to this subject. For our study, we took as a reference an experiment that paired 74 paired cytological and histological samples of cervical cancer and tested them using a 51 HPV genotype diagnostic array [13]. In that experiment, the HPV test results matched in 93% of cases, but when the viral load was low, the results differed in 78% of cases. In addition, two or more types of HPV were detected significantly less often in bioptates than in smears (14% vs 47%; p < 0.001). It should be noted that the results of our study confirm and largely complement these findings. In Group 2 (HSIL), the proportion of matching results of HPV DNA tests was 67.7%, in Group 1 (LSIL) — 45%, and in Group 3 (chronic cervicitis) is was 28.3%. The values reported by the authors of the mentioned experiment support the upward trend reflecting the dependence of the amount of matching results on the degree of damage to the cervix, which we registered. Our analysis shows that when the HPV viral load is low, the detection results in the sample pairs match in 27.1% of cases, when the load is moderate, this value goes up to 35.4%, and in high viral load circumstances, it increases to 82.3%. This is also consistent with what was reported in the foreign study. Another noteworthy aspect in which our findings coincide with those reported in the referenced experiment concerns registration of several HPV genotypes in a single sample: we encountered such situations significantly less often in cervix bioptates than in the cervical canal smears (13.1% vs 46.5%; p < 0.001). It is also important to note the specific distribution of HPV-44 we have identified: this genotype was the most common in cervical canal smears, but it was not found in any of the cervix bioptates. This probably indicates that HPV-44 has a tropism for columnar epithelial cells. Given the results of our retrospective study, which demonstrate the association between this genotype and precancerous condition of the cervix, HPV-44 should probably be studied further.

CONCLUSIONS

Thus, increasing HPV viral load translates into a growing probability of detection of HPV genotypes in the diseased cervical tissue. In addition, the amount of matching results of HPV tests in paired samples depends on the pathomorphological conclusion: the more severe the cervical lesion, the more likely HPV is to be detected in the bioptates.