**3. Results**

One hundred and seventy-two women scheduled to be conservatively treated with LEEP or laser conization or laser vaporization for CIN2+ were enrolled. The main characteristics of the study population at baseline are listed in Table 2.

Not all HPV tests were performed at baseline and follow-up due to lack of supply of reagents. Histological diagnosis on surgical specimens at baseline confirmed a CIN2+ lesion in 158 (91.9%) patients. Only histology confirmed samples were included in the final analysis.


**Table 2.** Patients' Characteristics at Baseline.

*<sup>a</sup>* Statistics are: Mean (SD) for Age, N (%) otherwise; SD = Standard Deviation; *<sup>b</sup>* min = 25, max = 61 years.

Overall, 150 patients were recruited at post-treatment follow-up, but only 118 in a time range between 3 and 6 months. Twenty-two patients (12.8%) were lost to follow-up and 32 (21.3%) were excluded due to incorrect timing of test-of-cure. Assuming HC2 as comparator, all HPV tests employed showed a good degree of comparison at both baseline (PABAK ranging from 0.81 to 0.90) and follow-up (PABAK ranging from 0.53 to 1). The concordance between different HPV tests and HC2 ranges from 91% to 95% at baseline and from 76% to 100% at follow-up, respectively, as shown in Tables 3 and 4.



*<sup>a</sup>* Column percent on non-missing counts; *<sup>b</sup>* McNemar test exact *p*-Values. PABAK = Prevalence and Bias adjusted Kappa; 95%CI = 95% Confidence Interval.


**Table 4.** HPV Test Results at Follow-Up *<sup>a</sup>* Compared to Hc2 (Reference). **HC2, N (col %)** *<sup>b</sup>*

**Table 4.** HPV Test Results at Follow‐Up *<sup>a</sup>* Compared to Hc2 (Reference).

*Diagnostics* **2022**, *12*, x FOR PEER REVIEW 6 of 9

*<sup>a</sup>* N = 118 patients with first visit at 3–6 months, median f.u. days = 108, (min = 90, max = 179) *<sup>b</sup>* Column percent on non-missing counts; *<sup>c</sup>* McNemar test; PABAK = Prevalence and Bias adjusted Kappa; 95%CI = 95% Confidence Interval. Sensitivity and specificity of all employed tests for CIN2+ at baseline and at follow‐

up, compared to HC2, are summarized in Figures 1 and 2. All HPV tests showed a very

Sensitivity and specificity of all employed tests for CIN2+ at baseline and at follow-up, compared to HC2, are summarized in Figures 1 and 2. All HPV tests showed a very good sensitivity in detecting CIN2+ at baseline, more than 92%, and a very good specificity at follow-up, more than 89%. good sensitivity in detecting CIN2+ at baseline, more than 92%, and a very good specific‐ ity at follow‐up, more than 89%.

**Figure 1.** Sensitivity for CIN2+ at baseline and at follow-up.

**Figure 1.** Sensitivity for CIN2+ at baseline and at follow‐up.

**Figure 2.** Specificity of all employed tests for CIN2+ at baseline and at follow-up.

#### **Figure 2.** Specificity of all employed tests for CIN2+ at baseline and at follow‐up. **4. Discussion**

**4. Discussion** The results of our study showed a very good concordance among different HPV tests performed in liquid‐based cervical samples from a group of women with high prevalence of preneoplastic cervical disease. The confidence intervals of these concordances overlap, further demonstrating the similarities of these HPV tests in performance. These data are in agreement with previous studies summarized in the 2020 list of human papillomavirus assays suitable for primary cervical cancer screening, published by Arbyn et al. (Arbyn et al., 2021). All HPV tests employed at baseline and follow‐up have been validated accord‐ ing to Meijer's guidelines [9–13]. As requested by validation guidelines (Meijer's guide‐ lines or Valgent protocol), the relative sensitivity and specificity must be high to be con‐ sidered "validated". Only LA is a test not fully validated according to Meijer's guidelines, due to the additional search for low‐risk (LR) HPV genotypes and the high sensitivity that does not correlate with CIN2+. However, data regarding positivity for LR‐HPV have not The results of our study showed a very good concordance among different HPV tests performed in liquid-based cervical samples from a group of women with high prevalence of preneoplastic cervical disease. The confidence intervals of these concordances overlap, further demonstrating the similarities of these HPV tests in performance. These data are in agreement with previous studies summarized in the 2020 list of human papillomavirus assays suitable for primary cervical cancer screening, published by Arbyn et al. (Arbyn et al., 2021). All HPV tests employed at baseline and follow-up have been validated according to Meijer's guidelines [9–13]. As requested by validation guidelines (Meijer's guidelines or Valgent protocol), the relative sensitivity and specificity must be high to be considered "validated". Only LA is a test not fully validated according to Meijer's guidelines, due to the additional search for low-risk (LR) HPV genotypes and the high sensitivity that does not correlate with CIN2+. However, data regarding positivity for LR-HPV have not been included in our analysis. Moreover, LA is a test previously validated according to Valgent protocol [8].

been included in our analysis. Moreover, LA is a test previously validated according to Valgent protocol [8]. Since validations have been usually performed in the screening setting, these data are only indicative for baseline. In the present analysis, we focused on comparing tests' performance not only at baseline, but also at the post‐treatment follow‐up. Interestingly, Since validations have been usually performed in the screening setting, these data are only indicative for baseline. In the present analysis, we focused on comparing tests' performance not only at baseline, but also at the post-treatment follow-up. Interestingly, our data showed comparable performance of the tests in terms of sensitivity and specificity at both baseline and test-of-cure.

our data showed comparable performance of the tests in terms of sensitivity and specific‐ ity at both baseline and test‐of‐cure. Due to the setting of samples, which show a high prevalence of positive at baseline and negative at follow‐up, respectively, sensitivities were found to be higher at baseline and lower at follow‐up. On the contrary, specificities are notably higher at follow‐up than at baseline. Due to the low prevalence of HPV after treatment, we chose the prevalence Due to the setting of samples, which show a high prevalence of positive at baseline and negative at follow-up, respectively, sensitivities were found to be higher at baseline and lower at follow-up. On the contrary, specificities are notably higher at follow-up than at baseline. Due to the low prevalence of HPV after treatment, we chose the prevalence and bias adjusted kappa, instead of either the simple or the weighted kappa, to estimate the HPV tests' agreement.

and bias adjusted kappa, instead of either the simple or the weighted kappa, to estimate the HPV tests' agreement. However, all tests perform similarly at baseline and follow‐up. Although Cobas However, all tests perform similarly at baseline and follow-up. Although Cobas seems to show better performance than other tests, these data might suffer from a bias related to the smaller number of samples that have been tested with the Cobas method.

seems to show better performance than other tests, these data might suffer from a bias related to the smaller number of samples that have been tested with the Cobas method. A negative HPV result at follow-up provides a good negative predictive value. Indeed, we found only a case of disease recurrence in the cohort of patients with a post-treatment negative HPV test result, for any validated HPV test. In this patient, cytology was HSIL

at follow-up and only LA revealed the presence of HPV18 and 73 at baseline, with the persistence of HPV 73 at relapse. Actually, LR HPV genotypes are not detected by validated HPV tests.

Furthermore, HPV tests that provide partial or extended genotyping showed comparable results.

The Aptima test, which detects HPV mRNA, showed no particular advantages in terms of sensitivity or specificity: the test performances are in line with other tests that detect HPV DNA.

Strengths of the present study include the type of population (women with only confirmed CIN2 + histology) and the timing of test-of-cure that was performed at 6 ± 3 months, as also suggested by the most recent guidelines from the Italian Group for Cervical Cancer Screening (GISCi) [15]. On the contrary, the main limit consists in the impossibility of performing all HPV tests in all samples.

In conclusion, our results demonstrate that validated HPV tests produce comparable results, and this cannot be extended to non-validated tests without proven evidence. Thus, only the use of validated HPV DNA or RNA tests is strongly recommended in both screening and test-of-cure settings. Moreover, HPV genotyping could be helpful in posttreatment management, by identifying women at higher risk of CIN2+ recurrence, due to the persistence of the same HPV genotype, and reassuring women who may present new HPV genotype infection after surgical treatment.

**Author Contributions:** F.B., A.D.I. and D.R. collected, analyzed data, and drafted the manuscript. E.P.P., M.P., D.F. and R.P. revised and edited the manuscript. M.T.S. and S.B. conceived the project. F.B., R.P. and A.D.I. participated in the coordination of the study and manuscript modification. All authors have read and agreed to the published version of the manuscript.

**Funding:** The funding for this study was provided by HPV test companies that supplied free kits for the assays. The Funder had the right to read and comment upon the manuscript, but without editorial rights, nor any role in the final interpretation of the data.

**Institutional Review Board Statement:** The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Institutional Review Board of the European Institute of Oncology, Milan, Italy (protocol IEO S544/210, date of approval: 26 May 2010).

**Informed Consent Statement:** Informed consent was obtained from all women at the entry of the study.

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author. The data are not publicly available due to patients' privacy restrictions. The data are safely stored in a private database of the European Institute of Oncology, Milan, Italy.

**Acknowledgments:** We are grateful to all HPV test manufacturing companies that supported the study, providing free kits to perform the tests. This work was partially supported by the Italian Ministry of Health with Ricerca Corrente and 5 × 1000 funds.

**Conflicts of Interest:** The authors declare no conflict of interest. None of the authors received financial support or funding for this work.
