Although HR-HPV16 was the most prevalent genotype among Korean women, it exhibited a similar malignant potential to HR-HPV52 and HR-HPV58 in the uterine cervix. Additionally, the risk of requiring medical intervention did not significantly differ among the three single infection groups.
The results of this study can help physicians explain the malignant potential of HR-HPV genotypes 16, 52, and 58 to their patients. HR-HPV16, the most prevalent genotype in Korea, did not exhibit a significantly higher risk of carcinogenesis in the uterine cervix than HR-HPV52 and HR-HPV58, which are the second and third most prevalent genotypes, respectively. Physicians should explain that the malignant potential of HPV52 or HPV58 is not inferior to that of HPV16. Furthermore, as the quadrivalent vaccine does not cover HPV52 or HPV58, our results support the need for a nine-valent vaccine against HR-HPVs, covering genotypes 6, 11, 16, 18, 31, 33, 45, 52, and 58.
To evaluate the malignant potential while controlling for age, subgroup analysis was performed. The entire patient cohort was divided into two subgroups on the basis of age, one with individuals aged 50 years or younger and the other with individuals older than 50 years. In the subgroup of individuals aged 50 years or younger, the number of HPV16, HPV52, and HPV58 single infection groups were 79 (53.0%), 38 (25.5%), and 32 (21.5%) cases, respectively. The prevalence of CIS + was observed in 45 (57.0%), 19 (50.0%), and 14 (43.8%) cases, respectively. In this subgroup, the CIS + ratios were not significantly different (p = 0.426). Logistic regression did not show significant ORs between any of the three single infection groups. Specifically, for HPV16 versus HPV52, HPV16 versus HPV58, and HPV52 versus HPV58, the ORs were 0.756 (95% CI = 0.348–1.642, p = 0.479), 0.588 (95% CI = 0.257–1.345, p = 0.208), and 0.778 (95% CI = 0.302–2.000, p = 0.602), respectively. In the subgroup of individuals older than 50 years, 40 (49.4%), 22 (27.2%), and 19 (23.5%) cases of HPV16, HPV52, and HPV58, respectively, were noted. The prevalence of CIS + was observed in 22 (55.0%), 9 (40.9%), and 12 (63.2%) cases, respectively. Similar to the subgroup of individuals aged 50 years or younger, no significant differences were noted in the CIS + ratio, and the risk of CIS + was not significantly different in the logistic regression test (HPV16 vs. HPV52, OR = 0.566 [95% CI = 0.197–1.625, p = 0.290]; HPV16 vs. HPV58, OR = 1.403 [95% CI = 0.457–4.304, p = 0.554]; HPV52 vs. HPV58, OR = 2.476 [95% CI = 0.701–8.742, p = 0.159]) (Table 2).
In Tables 1 and 2, we identified a discrepancy between the ratios of HSIL and CIS in the initial and final diagnoses within all three single infection groups. As mentioned in the Methods section, the initial diagnosis was determined on the basis of cervical swab cytology or punch biopsy in the office. Therefore, we assessed the diagnostic accuracy of cervical swab cytology and punch biopsy under colposcopy with acetic acid application to detect HSIL, CIS, or invasive carcinoma. All patients, regardless of their HPV infection status or whether they had undergone an HPV test within 1 year before conization, were included in this analysis, except for patients without cervical swab cytology or biopsy under colposcopy results. The pathologic diagnoses of both screening tests and cervical conization were categorized into HSIL − and HSIL+, with HSIL − encompassing normal, atypical cells, and LSIL, whereas HSIL + included HSIL, CIS, and invasive carcinoma.
In the cervical swab cytology for detecting HSIL+ (n = 738), the sensitivity (SS), specificity (SP), positive predictive value (PPV), and negative predictive value (NPV) was 32.8% (150/458), 86.1% (241/280), 79.4% (150/189), and 43.9% (241/549), respectively. The overall diagnostic accuracy was 53.0% (391/738). In the cervical punch biopsy under colposcopy with acetic acid application for detecting HSIL+ (n = 597), the SS, SP, PPV, and NPV was 89.5% (365/408), 42.3% (80/189), 77.0% (365/474), and 65.0% (80/123), respectively. The overall diagnostic accuracy was 74.5% (445/597).
Based on some recent studies, the Pap smear appeared to have a sensitivity of 47.2–55.5% and a specificity of 64.8–75.0%, whereas cervical biopsy under colposcopy showed a sensitivity and specificity of 64.7% and 52.74%, respectively23. Considering our data, cervical swab cytology and colposcopy examination appear to offer the benefits of high specificity and high sensitivity, respectively. However, as these examinations can exhibit low reproducibility, further studies are needed to establish optimal guidelines, especially in limited or resource-constrained condition.
The count and ratio of each HR-HPV genotype including all genotypes not only HPV16, 52, and 58 among the patients enrolled in this study are presented in Fig. 2. All subgroups represent the single infection group. The 10 most prevalent genotypes, in descending order, were HPV16, 52, 58, 18, 33, 31, 51, 53, 56, and 66.
Recent studies have analyzed the HPV genotype-specific risk for carcinogenesis in the uterine cervix10, 11. Park E et al. evaluated the risk on the basis of cervical biopsy. In this study, we identified HPV16, HPV52, and HPV58 as the three most common HR-HPV genotypes in Korean women, which is consistent with the reports of those authors. However, unlike the results of their study, our results were based on the pathologic results of uterine cervical conization. Moreover, they did not elucidate whether the patients infected with multiple HR-HPV genotypes were excluded. Thus, although they demonstrated a significantly higher carcinogenic risk for several HR-HPV genotypes including HPV16, HPV52, and HPV58 and a much higher OR of HPV16 than other genotypes, multiple infections with HR-HPV genotypes may have influenced these results. In this study, the HR-HPV genotype 16-, 52-, and 58-specific risk for carcinogenesis was clearly shown owing to the exclusion of concurrent multiple HR-HPV infections. On the basis of age, the authors stratified the patients into the following five subgroups: ≤34, 35–44, 45–54, 55–64, and ≥ 65 years. This enabled them to evaluate age- and HR-HPV genotype-specific risks11.
Another similar study conducted by So KA et al. was based on cervical biopsy results. They did not describe the inclusion or exclusion criteria for age; based on the result, the authors seemed to have included patients of all ages. In their study, HR-HPV genotypes 16, 52, and 58 were the most prevalent, which is consistent with our results. Their results showed that some HR-HPV genotypes such as 16, 31, 33, 52, and 58 were significantly more in CIN2, CIN3, and cervical cancer than those in the normal or CIN1 group. However, as patients with concurrent multiple infections were included in their study, the genotype-specific risk was unclear10.
The strength of this study was the homogeneous cohort with a single infection of HPV16, HPV52, or HPV58. This enabled the study to evaluate the malignant potential of each genotype with less bias compared with previous studies.
This study had some limitations. First, this was a retrospective study with a small sample size from a single center. Owing to the small sample size, the cohort could not be further stratified on the basis of age. Therefore, bias may arise from the different sexual activities according to age. A selection bias might also have affected the results because some patients, postmenopausal or not wanting more pregnancies, could have chosen to receive hysterectomy due to HSIL, CIS, or microinvasive carcinoma rather than cervical conization. Second, several patients in the cohort did not undergo follow-up HPV test before surgery. Thus, we could not prove or control the influence of persistent HR-HPV infection despite it being known as a critical factor in cervical cancer4, 5. Third, we could not review the HPV vaccination history. The Korean government included the quadrivalent and bivalent HPV vaccines (against HR-HPV6, 11, 16, and 18; against HR-HPV16 and 18) in the National Program 2016 and provided these vaccines for girls aged 12–17 years. According to statistics from the Korean government, approximately 70% of Korean girls have received the vaccines24. Finally, the HPV test in this study included two different methods, including DNA microarray and RT-PCR. Heterogeneity might have arisen from this aspect, raising concerns about the reliability of detecting HPV genotypes. When the titration of a certain HPV DNA was very high, other HPV genotypes’ DNA could not have been detected on the HPV test (references?? Detection accuracy??). Some patients could have had concurrent multiple infection of HR-HPVs, even though they had been shown to have a single HR-HPV infection on the HPV test. Furthermore, the HPV test was not centrally reviewed.