Cervical cancer is a prevalent gynecological malignancy, ranking fourth in both the incidence and mortality of female malignant tumors worldwide. According to the statistics in 2020, there were 604,000 newly diagnosed cases of cervical cancer globally, with 342,000 deaths (1). This population is predominantly located in developing and less developed countries. Cervical squamous intraepithelial lesion (SIL) encompasses a precancerous lesion, including low-grade cervical squamous intraepithelial lesion (LSIL) and high-grade cervical squamous intraepithelial lesion (HSIL). Approximate 80% of LSILs is characterized by the spontaneous regression, while some of them have capacity to progress to HSILs or cervical cancer (2). Thus, accurate diagnosis of female with cervical squamous intraepithelial lesions (SIL) is crucial to prevention of cervical cancer.

Persistent high-risk human papillomavirus (HR-HPV) infections are closely related to the development of precancerous lesions and cervical cancer (3). HPV screening having higher sensitivity and lower specificity compared to cytology screening, Thinprep cytology test (TCT) combined with HPV-DNA detection is commonly used for screening cervical cancer (4). Abnormal screening results require further evaluation of cervical biopsy via colposcopy (5). Colposcopy-directed biopsy (CDB) is a pivotal tool for early detection of cervical SIL and is widely regarded as the gold standard for diagnosing cervical intraepithelial neoplasia (6). Nonetheless, the diagnostic process has inherent limitations owing to the minimal tissue sample collected during biopsy. Furthermore, CDB primarily relies on the personal judgment of the colposcopy specialists, and biopsy sites may not be the most severe lesions for analysis. Some studies have reported that the accuracy of CDB in detecting SIL is suboptimal, limiting its use as a confirmatory diagnostic tool (7, 8). The cervical transformation zone (TZ) is also known as the squamocolumnar junction, as it is situated at the juncture of the squamous epithelium and the columnar epithelium of the cervix. The cervical transformation zone is classified into three types based on the visibility and location of the squamous-columnar junction under colposcopy: (1) TZ1: found at the external orifice of the cervix or outside the cervix and is entirely visible; (2) TZ2: located in the cervical canal but can be fully visualized after exposure; (3) TZ3: situated in the cervical canal and only partially visible (9). Some studies have shown that colposcopy biopsy carries the risk of histologic upgrade in SIL patients with incomplete visible cervical transformation zone (10).

It is currently considered that patients with LSIL may be closely monitored through regular follow-up visits, whereas those with HSIL require surgical intervention. Surgical approaches include cold knife conization (CKC) and loop electrosurgical excision procedure (LEEP). Conization not only serves therapeutic objectives but also allows for a definitive diagnosis of colposcopy-directed biopsy’s accuracy.

Accurate diagnosis of cervical SIL is critical for implementing appropriate management strategies, which range from active surveillance for LSIL to surgical excision for HSIL. The choice of management directly impacts patient quality of life, balancing the risks of overtreatment against those of disease progression. Therefore, evaluating and understanding the limitations of our primary diagnostic tool, CDB, by comparing it with the more comprehensive pathological assessment from excisional procedures like LEEP, is of paramount importance. Such comparisons not only enhance diagnostic accuracy but also pave the way for more targeted and individualized patient management. In present work, we compared with the differences between CDB and LEEP pathological results from 1,496 patients who were diagnosed with SIL by CDB and subsequently underwent LEEP surgery. We aimed to identify factors associated with histologic upgrade—defined as a more severe diagnosis on the LEEP specimen compared to the preceding CDB—recognizing that this discrepancy may stem from CDB sampling limitations or, less likely, interval disease progression.

Our study underscores the importance of correlating different diagnostic steps—CDB and LEEP—in the management pathway of cervical SIL. By identifying specific risk factors for CDB histologic upgrade, such as HPV16/18 infection and non-fully visible transformation zones, our findings provide actionable insights for clinicians. Our study showed that 93.38% (1,397/1,496) of cervical SIL were diagnosed by CDB successfully, highlighting its significant role in early detection of cervical lesions. Nonetheless, the limited sampling obtained during CDB only allows for superficial tissue collection, resulting in potential histologic upgrade. This study aimed to investigate the influence of risk factors on biopsy histologic upgrade in patients with cervical SIL. Additionally, we utilized a substantial cohort to uncover the association between HPV genotype and the likelihood of biopsy histologic upgrade. The majority of cervical SIL typically do not present any noticeable clinical symptoms or signs. If cervical SIL progresses to invasive cancer, the prognosis is often unfavorable. Therefore, accurate identification of cervical HSIL and early cervical cancer is significantly crucial. Colposcopy-directed cervical biopsy serves as a crucial method for diagnosing cervical lesions. Research has indicated that biopsies performed in four quadrants at 3, 6, 9, and 12 o’clock on the cervix yield a higher diagnostic rate than those conducted solely at the site of suspected lesions (12). In our study, we adopted a combined approach: besides obtaining biopsies from suspicious sites, we also collected samples from designated locations within the four quadrants.

In our study, the prevalence of HR-HPV infection among the patients was 88.30%, which aligns with previous research findings (13). The three most commonly detected HR-HPV genotypes were HPV16 (35.63%), HPV52 (17.25%), and HPV58 (16.38%), consistent with a retrospective study (14). HPV18, with an infection rate of 7.75%, ranked as the sixth most prevalent genotype. Previous studies have established that HPV infection is essential factors contributing to the development of cervical squamous intraepithelial lesions and cervical cancer. Notably, HPV16 and HPV18 are responsible for approximately 70% of cervical squamous cell carcinomas and about 90% of cervical adenocarcinomas (3, 15). Additionally, it’s worth noting that the pathological grading after LEEP surgery was lower than CDB in 348 patients (23.26%). This could result from complete lesion excision during cervical biopsy or endocervical curettage. Consequently, to ensure accuracy, the final diagnosis was determined based on the higher pathological assessment achieved either through CDB or LEEP.

We conducted an explicit analysis to examine the relationship between HPV genotypes and biopsy histologic upgrade. Our analysis identified a strong association between HPV16/18 infection and histologic upgrade. While this could be partly attributable to CDB sampling limitations (e.g., deeper or more localized lesions), it is crucial to consider that HPV16/18 are biologically more oncogenic and are associated with a higher prevalence and rapid progression of high-grade lesions. Therefore, the observed upgrade may not solely reflect a CDB-specific sampling error but could also be influenced by the intrinsic aggressiveness of these viral genotypes, potentially leading to faster disease progression even within our defined 3-month window or the development of small, focal high-grade lesions that are easily missed by biopsy. Our sensitivity analysis in the ≤6-week subgroup supports sampling error as a dominant, but likely not exclusive, factor. Our findings regarding HPV16/18 differ from those of Stoler et al. (16), who reported that these genotypes did not increase the risk of biopsy histologic upgrade. This discrepancy may be explained by evolving clinical contexts. Their study was conducted in the placebo arm of vaccine trials (2001–2003) with a different patient population and colposcopy referral patterns. Crucially, as the authors note, genotyping for non-16/18 HR-HPV was not available at that time, preventing a direct comparison of genotype-specific risks as performed in our contemporary cohort. Furthermore, differences in biopsy protocols and the definition of the reference standard may contribute to the contrasting results.

We concluded that type 2 transformation zone (TZ2) and type 3 transformation zone (TZ3) were more susceptible to histologic upgrade than type 1 transformation zone (TZ1) in cases of LSIL diagnosed by CDB. In addition, all the 19 upgraded cervical cancer patients had TZ2 and TZ3, and TZ3 was more likely to be underestimated than TZ2. For lesions within the cervical canal, it is challenging to fully expose the affected sites during CDB, making observation and sampling difficult. Furthermore, the detection rate of endocervical curettage is low, which increases the likelihood of misdiagnosis. Gustafson et al. (17) analyzed 102 patients with TZ3 and found that the HSIL detection rate was significantly higher in LEEP specimens than in biopsies, resulting in more than half of HSIL cases being underestimated in biopsies. Age has also been reported as a risk factor for histologic upgrade (18). Our results also indicate that in HSIL group, the age of the underestimated patients was higher than that of the patients who violated the histologic upgrade. This may be because estrogen levels gradually decrease with age, especially in perimenopausal women, leading to cervical atrophy and upward migration of the transformation zone. This makes it difficult to fully expose the lesions for examination, increasing the rate of histologic upgrade. A meta-analysis revealed that oral estradiol, intravaginal estradiol, and intravaginal misoprostol can be used to convert TZ3 to TZ1 or TZ2. Intravaginal misoprostol is well-tolerated and more feasible due to its availability and shorter treatment schedule when compared to oral or intravaginal estradiol (19).

In this study, we found that the type of transformation zones and HPV16/18 infection were significant risk factors associated with biopsy histologic upgrade. Given these results, special attention should be accorded to patients with the aforementioned high-risk factors during the process of a cervical biopsy. Previous research has demonstrated that the number of biopsy sites affects the accuracy of cervical biopsy diagnosis (16, 20). A decrease in the number of biopsy sites correlates with an increased likelihood of histologic upgrade. Therefore, we can aim to increase either the number of biopsies or the depth of the biopsy sites to enhance the precision of colposcopy, particularly in cases with the aforementioned high-risk factors. Alternatively, in a case with a CDB diagnosis of LSIL for more than 1 year, we may consider cervical conization if it has HPV16/18 infection and type 2 and type 3 cervical transformation zones. It has been reported that implementing conization before carrying out radical hysterectomy yields improved survival rates and decreased recurrence rates for patients suffering from early-stage cervical cancer, especially those opting for minimally invasive surgery (21). Consequently, patients suspected of having cervical cancer based on colposcopy findings may proceed directly to cervical conization. Studies have further indicated that cervical lesions infected with HPV16 and 18 are susceptible to recurrence post cervical conization (22). Therefore, patients with HPV16/18 infection should be subject to close postoperative follow-up. For patients with a type 3 transformation zone, placement of vaginal misoprostol can help better visualize the cervical transformation zone, thereby enhancing the accuracy of biopsy.

Our study has limitations inherent to its retrospective design. We were unable to account for several potential technical confounders in our regression models, such as the precise number of biopsy fragments taken or individual colposcopist experience, as these data were not consistently recorded in the medical records. Furthermore, the analysis did not stratify the risk of upgrade based on the severity of the antecedent cytology result. These factors could influence the likelihood of histologic discrepancy and represent important avenues for future prospective studies with standardized data collection protocols. In addition, while we included only patients who underwent LEEP within 3 months of CDB to minimize the risk of disease progression, the lack of exact procedural dates precluded a sensitivity analysis using a standardized shorter interval (e.g., ≤6 weeks). Future prospective studies should record exact dates to allow for more granular time-to-event analyses, which would help further distinguish sampling error from rapid progression.

In conclusion, our study comprehensively analyzed the cervical biopsy pathology and postoperative LEEP pathology of 1,496 patients with cervical squamous intraepithelial lesion. Our findings indicate that HPV16/18, as well as cervical transformation zones, especially type 3 transformation zone, represent significant contributing factors to histologic upgrade resulting from colposcopy biopsy procedures. The comparative analysis between CDB and LEEP pathology in this large cohort allows us to move beyond a one-size-fits-all approach. The identified risk factors for histologic upgrade create a framework for personalized medicine in cervical precancer management. Integrating these factors into clinical decision-making can help tailor the diagnostic and therapeutic intensity to the individual patient’s risk profile. This targeted strategy, aiming for precise diagnosis and appropriate management, is a fundamental step towards improving healthcare efficiency and patient quality of life in the context of cervical cancer prevention.