Logistic regression analysis of mRNA expression changes and prognosis after cervical surgery

Many studies have confirmed that the occurrence and development of cervical cancer is a process from quantitative change to qualitative change, usually cervical intraepithelial neoplasia (CIN)—an early invasive cancer. According to statistics, it takes about ten years.¹ From this, we can see that it takes a long time for cervical cancer to occur. We can intervene through cervical screening and early detection to block further development of the disease. Cervical intraepithelial neoplasia (CIN) is a precancerous lesion of cervical cancer.² Therefore, the purpose of cervical cancer screening is to detect and treat cervical precancerous lesions as soon as possible, block their further development and deterioration, and thus prevent the occurrence of cervical cancer and reduce cervical cancer mortality.³ 

With an in-depth study of cervical cancer and precancerous lesions, cervical cancer is one of the few malignant tumors with a clear etiology, and HPV infection is the cause of cervical cancer. Cervical cytology and HPV testing as the main means of cervical screening are beyond reproach. In recent years, research on new screening methods and indicators has emerged one after another.⁴ In 2003, Heselmeyer used FISH technology to detect the amplification of the TERC gene in cervical cells for the first time and compared it with the results of a pathological biopsy. He found that the positive rate of TERC gene amplification was only 7.14% in low-grade precancerous lesions and in CIN 2 and CIN 3, the positive rates of TERC gene amplification were 62.5% and 76.4%, respectively, in low-grade and high-grade precancerous lesions.⁵ FISH technology was used to detect the hTERC gene in cervical exfoliated cells, and it was found that the hTERC gene was amplified in CIN and SCC, and the detection rate of the hTERC gene increased with an increase in lesion degree. The expression rate of HPV LI protein decreased with an increase in cervical lesions, which can be used as a biological marker to predict the progression of cervical precancerous lesions.⁶ A study was conducted on patients with ZASCUS and high-risk HPV infection at the diagnostic level of cervical cytology, and it was found that expression rates of LI protein in CIN 1, CIN 2, CIN 3 cervical cancers were 83.5%, 41.8%, and 3.1%, respectively. The expression rate of L1 protein decreased with an increase in lesion grade (P < 0.01).⁷ This shows that the telomerase RNA gene (hTERC) and HPVL1 protein are biological markers for predicting the recovery of HPV infection. All the above are solid foundations for the prevention or early intervention of precancerous lesions and cervical cancer.

Although HPV-DNA detection has high sensitivity, its specificity is low and it is easy to cause false positives, leading to excessive examination and treatment. Obviously, it is very important to seek a simple, accurate, and reliable screening technique. Persistent infection with high-risk HPV is the primary factor causing cervical lesions, and the expression of oncogenes E6 and E7 is a necessary condition for tissue malignancy. Therefore, the presence of HPV E6 and E7 mRNA transcripts can predict the risk of CIN residual or recurrence in patients.⁸ By comparing the diagnostic effects of HPV DNA detection, the authors explored the diagnostic value of HPV E6/E7 mRNA detection for the residual or recurrence of CIN lesions after cervical conization.

This study contributes significantly to cervical intraepithelial neoplasia (CIN) research by investigating the diagnostic value of HPV E6/E7 mRNA combined with TCT (thin-layer liquid-based cytology test) detection in predicting the risk of CIN residual or recurrence after cervical conization. It enrolled patients diagnosed with high-grade CIN and employed a comprehensive follow-up.

Table I provides an overview of important studies in the field of cervical intraepithelial neoplasia (CIN) research. It highlights the major results of these studies and their significance to the present investigation, which focuses on the use of HPV E6/E7 mRNA paired with TCT detection to predict the persistence or recurrence of CIN after surgery.

A total of 154 patients with CIN 2 and CIN 3 who underwent loop electrosurgical excision procedure (LEEP) cervical conization in the obstetrics and gynecology department of our hospital from January 2019 to January 2021 and were diagnosed by histopathology were retrospectively selected. The mean age was 42.4 (26–66) years. The initial conization specimens had a pathological diagnosis of CIN 2 in 84 cases (54.5%) and CIN 3 in 70 cases (45.5%, 70/154). In 49 cases (31.8%), the lesions were in glands. Before initial conization, 135 cases (87.7%, 135/154) had abnormal TCT, 143 cases (92.9%, 143/154) had HPV-DNA (+), and 69 cases had HPV E6/E7 mRNA (+) (44.8%, 69/154; Table II).

The inclusion criteria were as follows:

1. Married non-pregnant women.

2. Aged 21–70 years.

3. No medical immune system disease or history of using related drugs.

4. No history of total hysterectomy; no history of chemotherapy or pelvic radiotherapy.

5. The initial conical resection margin was negative, and postoperative pathological results were not upgraded.

6. Complete clinical follow-up data.

Three days after menstruation was clean, exfoliated cells of the cervical canal and external os of the cervix were collected with a special brush for TCT, and the TCT system produced by Guangzhou Anbiping Company was used for automatic staining, according to the classification criteria of the Bethesda System (TBS) (2001).

Samples were taken after menstruation was clean, and sex was prohibited within 3 days before the collection of specimens. Using a special HPV sampler to collect cervical secretions, the PCR-reverse dot blot technique was used to detect 25 types of HPV. Patients were instructed to strictly follow the instructions of the human papillomavirus gene detection kit (produced by Shenzhen Yaneng Biotechnology Co., Ltd.).

Use a special brush to rotate anticlockwise at the external cervical orifice for three weeks, stay for 10 s, and place the obtained specimen in a specimen preservation tube. We used the QuantiVirusTM diagnostic kit and QuantiVirusT luminometer, which were made by the Sino-US joint venture Cotia Biotechnology Co., Ltd. We followed the kit’s instructions to the latter and used branched DNA (bDNA) technology to do so. Copy number <1 copy/ml is negative, and that >1 copy/m is positive.

Full-time colposcopy physicians use electronic colposcopy to observe abnormalities in the acetowhite and iodine staining tests. In areas showing abnormalities, they perform selective multi-point biopsies. Additionally, routine biopsies are taken at 3, 6, 9, and 12 points across the cervix. Neck scraping is also conducted if deemed necessary. The histopathological diagnosis of cervical conization was performed for those who were positive for TCT and HPV-DNA but negative for cervical biopsy. Pathology reporting method: normal or inflamed cervix; mild atypical hyperplasia of squamous epithelium (CIN 1); moderate atypical hyperplasia of squamous epithelium (CIN 2); severe dysplasia of squamous epithelium; and squamous cell carcinoma in situ (CIN 2).

Those who still have CIN lesions within 3–6 months after operation are regarded as residual lesions; those who have no CIN lesions within 6 months after operation but lesions are found again after 6 months are judged to be CIN lesion recurrence; and those who have postoperative HPV subtypes the same as that before operation are regarded as residual or recurrence and different from that before operation are regarded as HPV reinfection.

During follow-up examination at 3–6 months after operation, 9 cases were found to have residual lesions, and at 12 months after operation, 22 cases of lesions were found to recur, and there were 31 cases of residual or recurrence of lesions, and the residual or recurrence rate was 20.1% (31/154). 13 patients (21.7%, 13/60) aged <40 years had residual or recurrence; 18 (19.1%, 18/94) patients aged >40 years had residual or recurrence (Table II), including residual or recurrence of lesions.

There were 13 (15.5%, 13/84) patients with residual or recurrence in initial conization specimens diagnosed as CIN 2 and 18 (25.7%, 18/70) patients with residual or recurrence in pathological diagnosis of CIN 3 (Table III). There were 17 cases (34.7%, 17/49) of lesions involving glands in initial conization and 14 cases (13.3%, 14/105) of lesions that did not involve glands (Table III). When lesions involved glands, the residual or recurrence rate of lesions increased. Twenty-six patients (37.7%, 26/69) with HPV E6/E7 mRNA (+) had residual or recurrent lesions, and only five patients (5.9%, 5/85) with HPV E6/E7 mRNA (−) had residual lesions or recurrence (Table II).

The cumulative follow-up results up to the 12th month after conization showed 41 cases (26.6%, 41/154) with abnormal TCT, 57 cases (37.0%, 57/154) with HPV-DNA (+), and 26 cases with HPV E6/E7 mRNA (+) (16.9%, 26/154). Among all residual/recurrent cases, 27 (87.1%, 27/31) had abnormal TCT, 30 (96.8%, 30/31) had HPV DNA (+), and 24 (77.4%, 24/31) had HPV E6/E7 mRNA (+) (Table III and Fig. 1).

The Se, Sp, PPV, NPV, and diagnostic accuracy rates of HPV-DNA detection were 96.8%, 78.0%, 52.6%, 99.0%, and 81.8%, respectively, and the Se, Sp, PPV, NPV, and diagnostic accuracy of HPV E6/E7 mRNA detection were 77.4%, 98.4%, 92.3%, 94.5%, and 94.2%, respectively (Table IV and Fig. 2).

Univariate analysis was performed by the R × C contingency table chi-square test to understand the relationship between each factor and the residual or recurrence of lesions: age at discovery (residual/recurrence rate of <40 years old was 4.2%; residual/recurrence rate of >40 year old was 10.6%, X = 10.471, P = 0.001), preoperative HPV burden (residual/recurrence rate of <100 pg/ml was 2.6%; 100–1000 pg/ml recurrence/residual rate was 5.5%; and >1000 pg/ml residual/recurrence rate was 13.9%, X = 17.654, P = 0.000), cervicovaginal pathological diagnosis grade (<CIN 2 residual/recurrence rate was 1.5%; CIN 2–CIN 3 residual/recurrence rate was 5.8%; and CIN 3 residual/recurrence rate was 13.9%, X = 19.503, P = 0.001), cervical canal pathological diagnosis grade (<CIN 2 residual/recurrence rate was 1.6%; CIN 2–CIN 3 residual/recurrence rate was 17.4%; and CIN 3 residual/recurrence rate was 42.9%, X = 109.372, P = 0.001), external cervicovaginal margins (negative residual/recurrence rate was 4.6%; positive residual/recurrence rate was 43.5%, X = 61.335, P = 0.001), cervicovaginal internal margin (negative residual/recurrence rate was 3.1%; positive recurrence/residual rate was 17.9%, X = 43.571, P = 0.001), cervical incision margin (residual/recurrence rate of negative was 3.8%; residual/recurrence rate of positive was 34.0%, X = 77.585, P = 0.001), lesion involvement quadrant (residual/recurrence rate of 0 was 1.1%; residual/recurrence rate of 1 quadrant was 2.5%; residual/recurrence rate of 2–3 quadrants was 8.9%; and residual/recurrence rate of 4 quadrants was 22.6%, X² = 44.472, P = 0.001), and glands involved in disease (residual/recurrence rate of no glands involved was 2.6%; the residual/recurrence rate with gland involvement was 8.0%, X² = 9.406, P = 0.002) (see Table V and Figs. 3 and 4).

Note that the patients were assessed according to their preoperative HR-HPV-DNA load. There were 230 patients with ≤100 pg/ml (quadrant 4), of which 6 patients were found to have residual/recurrence, and the residual/recurrence rate was 2.6%. There were 274 patients with 100–100 pg/ml [quadrant 2–3], of which 15 were found to have residual/recurrence. There were 115 patients with >1000 pg/ml (quadrant 4), and the residual/recurrence rate was 5.5%; 16 of them were found to have residual/recurrence of lesions, and the residual/recurrence rate was 13.9%. After R × C contingency table analysis, the value is 17.654, P = 0.000.

Note that the patients were grouped according to the quadrant of lesions; there were 186 patients in 0 (the pathology before operation was CIN 2–CIN 3, and the pathology after operation was <CIN 2), of which 2 patients were found to have residual/recurrence, and the residual/recurrence rate was 1.1%; a total of 246 patients had lesions involving 1 quadrant, of which 6 patients were found to have residual/recurring lesions, and the residual/recurrence rate was 2.5%; and a total of 257 patients had lesions involving 2–3 quadrants, of which 23 patients were found to have residual/recurring lesions, and the residual/recurrence rate was 8.9%. A total of 53 patients had lesions involving 4 quadrants, of which 12 patients were found to have residual or recurring lesions, and the residual or recurrence rate was 22.6%. After R × C contingency table analysis, the value was 44.472, P = 0.000.

Univariate analysis considered factors that may affect the residual or recurrence of lesions and performed logistic regression analysis: pathological diagnosis grade of cervical canal (CIN 2–CIN 3: OR = 14.690, 95% CI: 5.150–41.899, P = 0.000; CIN 3: OR = 30.597, 95% CI: 7.065–132.517, P = 0.000); external cervicovaginal margin status (OR = 5.312, 95% CI: 1.620–17.413, P = 0.006); endocervical margin status (OR = 2.574, 95% CI: 1.071-6.187, P = 0.035); and LEEP. There was a correlation between residual or recurrence of later lesions (Tables VI and VII and Figs. 5 and 6).

1. All factors are compared with first in this factor.

2. Logistic analysis method: backward (LR).

To further compare the predictive value of each factor, we calculated the above risk factors considered by multiple regression analysis and factors considered in current clinical practice to predict the residual or recurrence of lesions, including specificity and sensitivity; see Table VIII and Fig. 7. The cervicovaginal pathological diagnosis grades were divided into <CIN 2, CIN 2–CIN 3, and CIN 3. When taking <CIN 2 as a negative population and >CIN 2 as a positive population, sensitivity and specificity were 93.02% and 28.06%, respectively; when <CIN 3 was regarded as a negative population and >CIN 3 was regarded as a positive population, sensitivity and specificity were 34.88% and 86.62%, respectively; see Table IX.

The findings of the study have significant clinical relevance in understanding the prognosis after cervical surgery. By demonstrating the effectiveness of combining HPV E6/E7 mRNA detection with TCT in predicting the risk of CIN residual or recurrence, the study offers a more accurate and efficient method for post-surgical monitoring. This can lead to better patient management as it allows for more targeted follow-up treatments and reduces unnecessary interventions. These insights are crucial for improving the overall care and treatment outcomes of patients undergoing cervical surgery for CIN.

The occurrence and development of cervical cancer is a multi-factor and multi-step biological process that requires a long period of precancerous lesions.⁹ In this process, persistent infection with high-risk HPV is a key factor. Cervical cancer screening is of great significance in helping people detect and treat cervical precancerous lesions early and prevent cervical cancer. Due to the high reversal rate of CIN 1, most scholars tend to advocate continuous monitoring and follow-up for such patients to reduce the risk of over-treatment.¹⁰ Therefore, most guidelines include CIN 2 and CIN 3 in the category of precancerous lesions. Due to the malignant potential of CIN 2–CIN 3, destructive treatment is generally not selected in clinical practice and conization is the treatment method for most doctors. As an effective method for the diagnosis and treatment of CIN 2–CIN 3, LEEP surgery is easy to operate, is less harmful to patients, and can be performed in an outpatient operating room under local anesthesia. Compared with cold knife conization and laser conization, LEEP surgery has great advantages.¹¹ It has been widely used in clinical practice and is an important means of treating cervical CIN 2–CIN 3. However, due to differences in surgical methods and operators, residual or recurrence rates after LEEP reported in the literature vary greatly, ranging from 4.2% to 51.8% (13.4%). The residual/recurrence rate of lesions in this study was 5.83%, which was slightly lower than most reported in the literature.¹² The higher pathological diagnosis grade of cervical squamous intraepithelial neoplasia implies deeper involvement of the squamous epithelium. Some studies have shown a higher lesion grade and more obvious biological characteristics of skipping and multicenter growth. However, in the current research on pathological diagnosis and residual or recurrence of lesions, the grade of the lesion tends not to be a risk factor for residual or recurrence.¹³ A retrospective study was conducted on 449 patients. In a univariate study, it was found that the residual/recurrence rate of CIN 2–CIN 3 was 20% (61/298), and the residual/recurrence rate of CIN 3 was 2.0% (3/151; P < 0.001), but after multivariate analysis, there was no correlation between grade of lesion and residual and recurrence of lesion (OR = 0.288, 95% CI: 0.062–1.336).¹⁴ In this study, a total of 236 CIN 1–CIN 3 women were included, and it was found that 4% (1/26) of CIN 1 patients had residual or recurrence cases, 43 CIN 2 patients had no residual or recurrent cases, and 5 CIN 3 patients were found to have residual or recurrence. In our study, preoperative pathology, cervicovaginal pathology, and cervical canal pathological grades were statistically analyzed, and in a univariate study, it was found that there was no correlation between preoperative pathological grades and residual or recurrence of lesions. The incidence of residual disease/recurrence in patients diagnosed with CIN 2–CIN 3 was 5.8% (25/432), and the residual/recurrence rate in patients with CIN 3 was 13.9% (15/108; P = 0.000); the pathological diagnosis of cervical canal was CIN 2. The incidence of residual disease/recurrence in 3 patients was 17.4% (26/149), and the residual/recurrence rate of CIN 3 patients was 42.9% (9/21; P = 0.000). The patients with pathological diagnosis of CIN 2–CIN 3 and CIN 3 were compared with cases below CIN 2, and it was found that there was no correlation between the grade of cervicovaginal lesions and residual or recurrence of lesions (CIN 2–CIN 3: OR = 2.151, 95% CI: 0.211–21.974, P = 0.518; CIN 3: OR = 2.306, 95% CI: 0.188–28.328, P = 0.514); the grade of cervical canal lesions was associated with residual or recurrence of lesions (CIN 2–CIN 3: OR = 14.690, 95% CI: 5.150–41.899, P = 0.000; CIN 3: OR = 30.597, 95% CI: 7.065–132.517, P = 0.000). This conclusion is different from some of the current research opinions.¹⁵ We believe that the reason may be that, although there are internationally uniform pathological diagnostic criteria for the pathological diagnosis of cervical intraepithelial neoplasia, the repeatability of the diagnosis is not satisfactory.¹⁶ There are differences in the interpretation of lesions by pathologists. Most of the research seen so far uses the “one knife method” to excise cervical lesions, and the “two knife method” to remove cervicovaginal and cervical canals, respectively, is rare.¹⁷ 

Surgical margins refer to the free margins of excisional surgery. Lesions involving surgical margins indicate that there are still lesions at the resection site and the risk of residual or recurrence of lesions is high. In the pathological evaluation of LEEP surgical specimens, some scholars believed that due to its thermal damage effect, LEEP may destroy the cellular structure of resection margin tissue, thereby affecting pathologists’ assessments of resection margin status.¹⁸ With the adjustment of electric power, LEEP can basically maintain the integrity of the resected specimen.¹⁹ Currently, it is hypothesized that Loop Electrosurgical Excision Procedure (LEEP) may induce minor thermal injury to tissues during the electroexcision phase. This is further exacerbated by the application of electrocautery on the wound bed and the extensive utilization of electrocoagulation, which potentially amplifies tissue damage at the resection margins. Such a damage, while initially appearing detrimental, may paradoxically exert a therapeutic effect. Consequently, patients presenting with positive surgical margins—indicative of residual malignant cells at the excision boundary—do not invariably experience persistence or recurrence of lesions, suggesting a potential therapeutic benefit from the thermal and electrocautery interventions inherent in LEEP.²⁰ In a retrospective study of 120 women, a total of 41 women with involvement of the external resection margin had residual or recurrence and 4 of the 43 women with involvement of the internal resection margin had residual or recurrence. By factor analysis, it was found that there was a correlation between the involvement of the external resection margin and the residual and recurrence of the lesion (P = 0.002).²¹ In another study, 449 women with CIN 3 received LEEP and it was shown that internal resection margin was a risk factor for residual or recurrence of the lesion through multiple regression analysis (OR = 2.972, 95% CI: 1.401–6.281). The surgical methods of LEEP surgery are diverse. Different operations have different names for incision margins.²² The traditional “one-shot method” forms two incision margins, named “internal incision margin” and “external incision margin,” which is also expressed in most papers.²³ We adopted the “two-knife method” and named surgical margins “cervicovaginal internal margin,” “cervicovaginal external margin,” and “cervicovaginal margin.”²⁴ Among them, 17.9% (24/134) of patients with positive cervicovaginal margins had residual disease or recurrence and 3.1% (19/604) of patients with negative internal margins had residual disease or recurrence, P = 0.000; 43.5% (10/23) of patients with positive margins had residual disease or recurrence, and 4.6% (33/715) of patients with negative external margins had residual disease or recurrence, P = 0.000; and 34.0% (17/50) had residual disease or recurrence, and 3.8% (26/688) of patients with negative cervical resection margin had residual or recurrence, P = 0.000. Logistic regression analysis found that external cervicovaginal margin (OR = 5.312, 95% CI: 1.620–17.413, P = 0.006) and endocervical margin (OR = 2.574, 95% CI: 1.071–6.187, P = 0.035) were both associated with residual disease or recurrence.^25–27 

Our study’s findings suggest that the HPV E6/E7 mRNA combined with TCT detection could be a valuable tool in the follow-up examinations to timely and effectively predict the risk of CIN residual or recurrence, which may help in reducing excessive examinations and treatments. This approach is relatively novel as it combines HPV E6/E7 mRNA detection with TCT, offering a more nuanced understanding and prediction capability compared to traditional methods.

The study was supported by “Cangzhou Key Research and Development Projects (Grant No. 213106062).”

The authors have no conflicts to disclose.

The project complies with the principles of the Helsinki Declaration and regulations of China and agrees to accept the supervision and inspection of the Ethics Committee; all patients gave informed consent. Issue No. CZH-2018046.

The study was approved by the review committee of our ethics committee. All patients gave informed consent. The study adheres to the principles outlined in the Declaration of Helsinki.

Yinghui Song: Supervision (equal); Writing – original draft (equal); Writing – review & editing (equal). Min Zhang: Conceptualization (equal); Data curation (equal). Cui Zhang: Investigation (equal); Methodology (equal). Shiyu Du: Resources (equal); Software (equal). Furui Zhai: Visualization (equal).

The data that support the findings of this study are available from the corresponding author upon reasonable request.