Peripheral blood immune cell parameters in patients with high-grade squamous intraepithelial lesion (HSIL) and cervical cancer and their clinical value: a retrospective study

Study population

In this rigorous retrospective case-control analysis, we meticulously enrolled 39 subjects presenting with HSIL and a group of 42 patients diagnosed with cervical carcinoma, all of whom received treatment within our institution from July 2020 through September 2023. Stringent histopathological confirmation was obtained for each case of cervical neoplasia. Exclusion criteria were robustly applied to omit individuals with concurrent neoplastic entities, substantial organ pathologies, disturbances in endocrine metabolic homeostasis, or those who had partaken in immunosuppressive or antiviral pharmacotherapy in the proximity of the study commencement. Ensuring the integrity of the investigation, only patients with comprehensive clinical records and follow-up data were considered eligible. Parallelly, the study incorporated 31 healthy female participants, selected from those attending our hospital for routine health surveillance within the same temporal span. These participants were scrupulously screened to ensure the absence of any historical or current cervical neoplastic or precancerous conditions, chronic cervicitis, or ancillary gynecological reproductive afflictions.

The inclusion of subjects with HSIL in our study was motivated by the continuum of cervical neoplastic progression and the clinical need to explore the discriminatory potential of peripheral blood lymphocytic indices across the spectrum of cervical lesions. High-grade squamous intraepithelial lesions (HSIL) represent a precancerous stage in the transformation pathway of cervical neoplasms and are clinically significant due to their potential to progress to invasive cervical cancer. Given the close association between HSIL and the development of cervical cancer, investigating the peripheral blood neutrophil-to-lymphocyte ratio (NLR) and platelet-to-lymphocyte ratio (PLR) in patients afflicted with HSIL provides valuable insights into the hematologic alterations that may precede the onset of invasive malignancy. Moreover, the inclusion of HSIL patients allows for the examination of early hematologic changes that may serve as predictive biomarkers for disease progression, thereby offering a comprehensive assessment of the continuum from precancerous conditions to malignant tumors. By encompassing both HSIL and cervical carcinoma patients, our study seeks to elucidate the potential clinical value of peripheral blood NLR and PLR in discriminating the hematologic profiles associated with varying degrees of cervical neoplastic transformation.

All samples obtained in this study were approved by the ethics committee of the Shandong Provincial Maternal and Child Health Care Hospital and abided by the ethical guidelines of the Declaration of Helsinki, and ethics committee agreed to waive informed consent.

Methods

Clinical data were collected from the patients, including (1) general information: age, menopausal status, number of pregnancies, number of deliveries, hypertension, diabetes, etc.; (2) laboratory indices: routine blood examination data including platelet count, absolute neutrophil count, and lymphocyte count. The NLR and PLR were calculated, with NLR being the ratio of neutrophil count to lymphocyte count, and PLR being the ratio of platelet count to lymphocyte count; (3) pathological data: cervical tissue specimens were processed for slide preparation and staining by pathologists at our institution, with diagnoses made under the microscope. Immunohistochemical staining results were incorporated when necessary. The pathological staging of cervical cancer followed the clinical staging criteria established by the International Federation of Gynecology and Obstetrics (FIGO) in 2018 (Bhatla et al., 2018).

Statistical methods

The collated data from this investigative inquiry were subjected to rigorous statistical scrutiny utilizing the SPSS software, version 23.0. The graphical representations of the analyzed parameters were expertly generated via GraphPad Prism, version 8.0. The quantitative datasets were elucidated as means ± standard deviations. Continuous variables were first tested for normal distribution using the Shapiro–Wilk method. Dichotomous group comparisons were executed employing the Student’s t-test, while the analysis of variance (ANOVA) with Bonferroni correction was the statistical method of choice for assessing differences across multiple groups. Categorical variables were succinctly described as frequencies and percentages, and their interrelations were interrogated using the Chi-square (χ²) test. Bonferroni correction was used for eliminating potential inflation of type I error due to multiple subgroup analyses. The Pearson correlation coefficient was the selected metric for the evaluation of correlational dynamics. A P-value threshold of less than 0.05 was pre-established to denote the threshold of statistical significance. Area under the receiver operating characteristic (AUC-ROC) was used to assess diagnostic value of PLR and NLR (Mandrekar, 2010). Furthermore, logistic regression analysis was conducted to evaluate the impact of various clinical and hematologic parameters on the occurrence of HSIL and cervical cancer. The coefficients for age, menopause, number of pregnancies, number of parturitions, hypertension, diabetes, PLR, and NLR were calculated, providing key insights into their contributions to the likelihood of HSIL or cervical cancer. Subsequently, a nomogram was constructed to visually present the results of the logistic regression analysis, facilitating a comprehensive depiction of the relative impact of each parameter on the occurrence of HSIL and cervical cancer. This enables a more intuitive interpretation of the relationships between the analyzed variables and the likelihood of disease occurrence.

Comparison of clinical data between cervical cancer, HSIL patients, and the healthy control group

Upon scrutinizing the demographic and clinical parameters across the three groups—individuals diagnosed with cervical carcinoma, patients identified with High-Grade Squamous Intraepithelial Lesions (HSIL), and the healthy control group—our statistical evaluation revealed a homogeneity in age distribution, menopausal status, and gravidity (P >0.05), affirming comparability between groups as delineated in Table 1.

Comparison of peripheral blood PLR and NLR levels among cervical cancer, HSIL patients, and the healthy control group

In the comparative analysis of systemic inflammatory indices, specifically the PLR and NLR in peripheral blood, as presented in Table 2, a statistically significant disparity was observed among the three groups (P < 0.05) (Figs. 1–2). While the intergroup comparison between cervical cancer and HSIL patients yielded no significant variance in NLR (P = 0.061) and PLR (P = 0.759) levels, a remarkable elevation in these biomarkers was noted in the cervical cancer group compared to the healthy controls, manifesting statistically significant deviations (t = 5.094, 5.927; P < 0.001 for PLR and t = 4.206, 3.657; P < 0.001 for NLR).

Comparison of peripheral blood PLR levels in cervical cancer patients with different pathological features

In evaluating the PLR among cervical cancer patients with differing pathological features, our analysis revealed significant variations in PLR values associated with clinical stage and degree of muscle wall invasion (Table 3). Specifically, PLR values increased with advancing clinical stage, being 150.88 ± 49.53 for Stage I, 175.36 ± 33.52 for Stage II, and 193.15 ± 25.81 for Stage III, with a t value of 3.422, indicating a statistically significant difference (P = 0.043). Similarly, a significant association was identified between PLR and the degree of muscle wall invasion, with PLR values of 151.29 ± 39.48 for 2/3 to full-thickness invasion, 171.53 ± 30.57 for 1/2 to 2/3 thickness invasion, and 189.00 ± 26.83 for <1/2 thickness invasion, achieving an F value of 3.510 (P = 0.040). In contrast, comparisons based on tumor diameter (<four cm: 173.15 ± 31.63 vs. ≥4 cm: 179.89 ± 32.37), presence of vascular invasion (present: 176.15 ± 38.19 vs. absent: 174.69 ± 36.22), and nerve invasion (present: 179.32 ± 38.45 vs. absent: 177.07 ± 37.42) did not yield statistically significant differences, with P-values of 0.510, 0.903, and 0.863, respectively. These results highlight the significance of PLR variations in relation to clinical stage and the extent of muscle wall invasion in cervical cancer patients.

Comparison of peripheral blood NLR levels in cervical cancer patients with different pathological features

In the assessment of the NLR in patients with cervical cancer based on different pathological features, significant variations were observed across clinical stages (Table 4). Specifically, Stage I patients exhibited an NLR of 1.95 ± 0.44, which significantly increased in Stage II to 2.23 ± 0.38, and further to 2.94 ± 0.31 in Stage III, with an F value of 8.267 (P = 0.001), indicating a statistically significant difference in NLR values across these stages. Conversely, when comparing NLR based on tumor diameter (<four cm: 2.54 ± 0.57 vs. ≥4 cm: 2.77 ± 0.58), invasion depth (2/3 to full thickness: 1.96 ± 0.45 vs. 1/2 to 2/3 thickness: 2.30 ± 0.38 vs. <1/2 thickness: 2.93 ± 0.27), presence of vascular invasion (present: 2.94 ± 0.57 vs. absent: 2.83 ± 0.61), and presence of neural invasion (present: 2.92 ± 0.53 vs. absent: 2.79 ± 0.58), no statistically significant differences were found (P-values of 0.214 for tumor diameter, <0.001 for invasion depth, 0.570 for vascular invasion, and 0.518 for neural invasion). These findings underscore the significant association between NLR and the clinical stage of cervical cancer, while the correlations with tumor diameter, invasion depth, and presence of vascular or neural invasions did not exhibit statistical significance.

Correlation analysis of peripheral blood PLR and NLR with pathological features in cervical cancer patients

In the correlation analysis between peripheral blood markers and pathological features in patients with cervical cancer, the PLR and NLR were evaluated (Table 5). The results revealed a statistically significant positive correlation between PLR and clinical stage (r = 0.459, P = 0.013), as well as a similar significant correlation between NLR and clinical stage (r = 0.439, P = 0.020). Additionally, a significant correlation was observed between both PLR (r = 0.432, P = 0.029) and NLR (r = 0.417, P = 0.031) with invasion depth. Conversely, correlations between PLR and NLR with pathological type, tumor diameter, and the presence of vascular or neural invasion did not reach statistical significance. Specifically, PLR’s correlations with pathological type (r = 0.058, P= .616), tumor diameter (r = 0.244, P = 0.197), and vascular or neural invasion (r = 0.031, P = 0.823) exhibited P-values well above the threshold for significance. Similarly, NLR’s correlations with these parameters—pathological type (r = 0.218, P = 0.141), tumor diameter (r = 0.263, P = 0.165), and vascular or neural invasion (r = 0.044, P = 0.869)—also did not demonstrate statistical significance. These findings suggest that both PLR and NLR are significantly associated with the clinical stage and invasion depth of cervical cancer, highlighting their potential as indicators of disease progression.

ROC analysis of PLR and NLR and the occurrence of cervical cancer and HSIL

In the analysis of inflammatory markers’ discriminatory ability for cervical cancer diagnosis, the PLR (Fig. 3) exhibited a sensitivity of 67.9% and a specificity of 93.5%, while the NLR (Fig. 4) had a sensitivity of 65.4% and a specificity of 77.4%. The AUC for PLR was determined to be 0.851, indicating a superior diagnostic performance compared to the NLR, which presented an AUC of 0.734. Regarding the Youden index, a measure used to assess the effectiveness of a diagnostic test, PLR obtained a score of 0.614, thereby outperforming NLR, which had a Youden index of 0.428. Although both PLR and NLR demonstrated statistical significance in their diagnostic capacities, the results emphasize a notably higher specificity and overall diagnostic accuracy for PLR, suggesting its potential greater utility as a biomarker for cervical cancer compared to NLR.

Logistic regression analysis of various parameters and the occurrence of cervical cancer and HSIL

In the logistic regression analysis, the coefficients for age, menopause, number of pregnancies, number of parturitions, hypertension, and diabetes were 0.01, 0.136, 0.846, 0.479, 0.323, and 0.316, respectively (Table 6). The odds ratios for these variables were 1.011, 1.145, 2.331, 1.615, 1.381, and 1.372, suggesting the impact of each unit increase in the predictor variables on the likelihood of HSIL or cervical cancer. The 95% confidence intervals (CI) for the odds ratios provided additional insight into the precision of the estimates. Notably, the odds ratio for the number of pregnancies demonstrated a higher value, with an upper confidence interval greater than 1, indicating a potentially significant impact on the likelihood of HSIL or cervical cancer. Similarly, the odds ratios for hypertension, diabetes, and PLR were also noteworthy, emphasizing a potential association with the risk of HSIL or cervical cancer. Moreover, the analysis revealed a statistically significant association of neutrophil-to-lymphocyte ratio (NLR) with the presence of HSIL or cervical cancer, as evidenced by a P value of less than 0.001. The regression coefficient (B) values for PLR and NLR were 0.38 and 4.811, respectively, underlining the potential role of these hematologic parameters as discriminative biomarkers for the diagnosis of HSIL and cervical cancer. Finally, the results of logistic regression analysis were visualized by nomogram (Fig. 5).