The importance of HPV genotypes varies substantially depending on their carcinogenicity and geographic distribution. To advance cervical cancer prevention and control strategies in Ningbo, we determined the predominant HR-HPV genotypes among local women who were diagnosed with high-grade cervical precancers and cancers. The findings from our study can be summarized as follows: (1) HPV16 was the most prevalent genotype with the highest attribution in both cervical precancers and cancers; (2) HPV18 and 45 were similarly important in adenocarcinoma (ADC), together accounting for 46.8% of ADC; (3) HPV31/33/52/58 each ranked second to fourth in terms of attributable proportions in cervical precancers, collectively accounting for approximately 40% of precancers and over 10% of invasive cervical cancer (ICC); and (4) HPV35/39/51/56/59/66/68 individually accounted for less than 2% of precancers or cancers and collectively accounted for less than 5% of ICC. These findings could inform the prioritization of specific genotype to be targeted in Ningbo’s future genotype-based interventions.

A meta-analysis summarizing the HPV genotype distribution among 8786 Chinese CIN women [22] showed that in CIN2/3, the overall HPV positive rate was 87%, and the top five HPV genotypes in descending order were HPV16 (45.7%), 58 (15.5%), 52 (11.7%), 33 (9.4%), and 31 (4.3%). Compared to this, our study had a higher HPV positive rate (98%) among CIN2/3 population, but the genotype distribution was completely consistent: HPV16 (51.7%), 58 (15.7%), 52 (15.4%), 33 (9.1%), and 31 (6.3%). Numerical difference in positive rates should mainly be attributed to different HPV detection methods and sample types. Among the nine studies included in the meta-analysis, eight used exfoliated cell samples instead of the FFPE used in our study, and the HPV detection methods were primarily PCR-based with longer primers, such as MY09/11 and GP5/6.

We found four studies reported the HPV genotype distribution among women with CIN and cervical cancer in Zhejiang province, where Ningbo is located, during the periods of 2004–2006 [23], 2007–2008 [24], 2008–2013 [25], and 2012–2014 [26]. The samples tested in these studies were all exfoliated cells, and three of them used the HPV GenoArray Diagnostic Kit (HybriMax, Chaozhou Hybribio Limited Corp., Chaozhou, China), while the other study used an MY09/11 primers-based PCR method. Similarly, due to the differences in sample types and the use of more sensitive SPF10-DEIA-LiPA25 detection system [27], our study had the highest overall HPV positive rates both in ICC and in CIN2/3 compared to these studies. Our study describes the genotype distribution among women with CIN2+ diagnosed in 2017–2020. Although our study and the above four studies covered different time periods, HPV16, 18, 31, 33, 52, and 58 remained the common five genotypes found in women with ICC and CIN2/3. Among ICC, positive rate of these five genotypes across different studies were: 56–73.6% for HPV16, 7.2–12.2% for HPV18, 2.4–4.4% for HPV31, 1.7–7.2% for HPV33, 2.2–8.6% for HPV52, and 4.8–11.2% for HPV58; Among CIN2/3, the positive rates were respectively: 42.3–51.7% for HPV16, 3.2–5.2% for HPV18, 4.6–7.0% for HPV31, 8.4–12.8% for HPV33, 2.3–15.4% for HPV52, and 15.7–20.7% for HPV58. Except for HPV16, which was the most dominant genotype in ICC and CIN2/3 across all studies, the distribution of other genotypes varied slightly depending on the study. The positive rates of these genotypes generally did not fluctuate much over the years. The possible reasons for the numerical change, like the positive rate of HPV52 among CIN2/3, could be related to the sample size and the study population besides the detection methods and sample types.

Due to the shared mode of sexual transmission of all alpha HPV genotypes, the concurrent presence of multiple genotypes is common, especially among women younger than 25 years old and older than 65 years old. [28] It is challenging to determine which genotype is truly carcinogenic and warrants greater concern based solely on its positive rate in the population. From a virologic perspective, to be confirmed as a definite human carcinogen, one HPV genotype must be transcriptionally active in a tumor [29]. Previous studies have suggested that only a single genotype is transcriptionally active and pathogenic in lesions that tested positive for multiple genotypes [21, 30]. Therefore, we adopted the fractional allocation method reported by Insinga et al., which assigns weights to the contribution of each genotype in muti-type infected lesions based on their presence as single-type infections. We then ranked the importance of individual HR-HPV genotypes by their estimated attributable proportions to cervical precancers and cancers to indicate which genotype should be of concern to health policymakers.

In Ningbo, HPV16 (alpha 9 species) could attribute to 77.3% of SCC during the period from 2017 to 2020, which is consistent with a nationwide study in 2009 reporting that HPV16 accounted for 76.7% of SCC [11]. The International Agency for Research on Cancer (IARC) also lists HPV16 as the foremost carcinogen among all carcinogenic HPV genotypes and reports that 60% of SCC could be singularly attributed to this genotype worldwide [31]. The above evidence highlights the importance of HPV16 and suggests that controlling HPV16 infection should remain a top priority in future large-scale vaccination and HPV-based cervical screening in Ningbo.

HPV18 and 45 were rare in women with cervical precancers but accounted for nearly 50% of ADC in our study. A retrospective cross-sectional worldwide study has suggested that ICC cases with HPV18 or HPV45 infection tended to present at an earlier age, indicating a shorter time for progression to invasive cancer, sometimes even without transformation through the preinvasive stages [32]. These two genotypes are also categorized with the second highest attributable risk of cancer by IARC [31]. Screening has been less effective in preventing adenocarcinomas than in preventing squamous cancers [33]. Given the sizable contribution of HPV18 and HPV45 to ADC, these two genotypes should not be overlooked in future genotype-based interventions.

Five non-HPV16 alpha 9 HR-HPV (HPV31/33/35/52/58) genotypes are categorized as the third highest attributable risk group by the IARC [31] and are at medium risk of developing cervical cancer [34]. A study analyzing HPV detection data from one million cervical samples in Belgium from 2006 to 2014 indicated that testing for HPV16/18 combined with HPV31, 33, 45, and 52 at certain viral load thresholds could predict 86.5% of cervical cancers occurring within a year after testing [35]. This study found comparable effectiveness by testing for all 14 HR-HPV genotypes, which predicted 89.4% of cervical cancers [35]. Additionally, the specificity also increased considerably in the former testing algorithm. However, the Belgium study did not find HPV58 to be predictive of any cervical cancer. This discrepancy could be attributed to regional differences in the distribution of HPV58, which has a much greater prevalence in high-grade cervical lesions in East Asia than in Europe [36].

Above non-HPV16 alpha 9 genotypes, excluding HPV35, are already covered by nonavalent vaccines. These genotypes (HPV31/33/52/58) combined accounted for a significant proportion of cervical precancers and ICCs in our study. However, HPV35 had a minimal contribution (1.4% in CIN2, 0.9% in CIN3, and zero in ICC) to cervical lesions. A population-based study in Ningbo similarly reported that only 0.5% of women who attended clinics from 2019 to 2021 were infected with HPV35 [37]. These findings suggest that controlling HPV31/33/52/58 infections might provide considerable additional protection for women in Ningbo, while HPV35 appears to be less important in both the general population and women with cervical lesions.

In 2022, the IARC reported that for women with cervical cancer, HPV39/51/56/59/66/68 have negligible attributable risk, each of which is responsible for less than 1% of cancers [31]. Notably, the estimated attributable risk for HPV66 was zero. These findings are consistent with our study of women with cervical cancer. However, the population-based study in Ningbo [37] revealed that the prevalence of these genotypes among all HPV-positive individuals reached as high as 25.9%. Current HPV testing assays mainly distinguish HPV16/18 separately but detect the universal gene region of the remaining 12 HR-HPV genotypes. Thus, in the general population, using HPV-based testing assays covering HPV39/51/56/59/66/68 will yield a high proportion of positive test results with a low probability of progression to cancer. We recommend that future HPV-based testing assays distinguish HPV16/18/31/33/45/52/58 and combine or exclude HPV39/51/56/59/66/68 for general population screening.

To the best of our knowledge, this study is the first to rank the importance of HR-HPV genotypes using biopsy-confirmed CIN2+ blocks and attributable proportion estimation in Ningbo. The data from our study could provide direct insights for upgrading current prevention and control measures. Second, HPV genotyping based on tissue specimens can more accurately indicate the pathogenic risk associated with HR-HPV genotypes. This is because exfoliated cell samples, which are commonly used for genotyping, represent comprehensive infections in the vagina-cervix areas, but most of these infections do not persist or progress to disease. However, the fixation process and long-term preservation of formalin-fixed paraffin-embedded blocks can cause nucleic acid degradation and fragmentation. This can potentially fail the detection of HPV DNA using PCR assays, resulting in reduced sensitivity of current PCR-based HPV genotyping assays in tissues. To minimize the impact of DNA degradation and fragmentation on the genotyping results, we applied the SPF10-DEIA-LiPA25 detection system. This system targets the smallest amplicon compared to any available HPV DNA genotyping system and is particularly suitable for detection in tissues [27]. Meanwhile, we employed a sandwich sectioning technique and strict pathology confirmation process to ensure the sample’s qualification for testing and reduce the possibility of sample loss.

One limitation of our study is that we provided HR-HPV genotyping data for only one city. China is a vast country with disparities in economic development and the allocation of medical resources among cities and regions. Due to regional variations in HR-HPV genotype-specific distribution, it is necessary to analyze region-specific HPV infection patterns to tailor cervical cancer prevention and control policies accordingly. Another limitation is that we reported cross-sectional information for HPV genotyping. When future large-scale vaccine programs are implemented, ongoing assessments will be needed to monitor any changes in HR-HPV genotype distribution and attribution over time to ensure efficient resource utilization.

In conclusion, it is important to prioritize HPV16/18 control efforts, and the focus could be broadened to HPV31/33/45/52/58 in resource-rich settings, such as Ningbo, for maximum health benefits. However, HPV39/51/56/59/66/68 should be considered with caution for HPV-based testing assays and vaccines due to their lower carcinogenic risks.