Direct bisulphite conversion of cervical samples for DNA methylation analysis

DeepSF-4mC: A deep learning model for predicting DNA cytosine 4mC methylation sites leveraging sequence features

N4-methylcytosine (4mC) is a DNA modification involving the addition of a methyl group to the fourth nitrogen atom of the cytosine base. This modification may influence gene regulation, providing potential insights into gene control mechanisms. Traditional laboratory methods for detecting 4mC DNA methylation have limitations, but the rise of artificial intelligence has introduced efficient computational strategies for 4mC site prediction. Despite this progress, challenges persist in terms of model performance and interpretability. To tackle these challenges, we propose DeepSF-4mC, a deep learning model specifically designed for predicting DNA cytosine 4mC methylation sites by leveraging sequence features. Our approach incorporates multiple encoding techniques to enhance prediction accuracy, increase model stability, and reduce the computational resources needed. Leveraging transfer learning, we harness existing models to enhance performance through learned representations or fine-tuning. Ensemble learning techniques combine predictions from multiple models, boosting robustness and accuracy. This research contributes to DNA methylation analysis and lays the groundwork for understanding 4mC's multifaceted role in biological processes. The web server for DeepSF-4mC is accessible at: http://deepsf-4mc.top/and the original code can be found at: https://github.com/754131799/DeepSF-4mC.

Widening the offer of human papillomavirus self-sampling to all women eligible for cervical screening: Make haste slowly

Self-collection of samples for human papillomavirus (HPV) testing has the potential to increase the uptake of cervical screening among underscreened women and will likely form a crucial part of the WHO's strategy to eliminate cervical cancer by 2030. In high-income countries with long-standing, organised cervical screening programmes, self-collection is increasingly becoming available as a routine offer for women regardless of their screening histories, including under- and well-screened women. For these contexts, a validated microsimulation model determined that adding self-collection to clinician collection is likely to be cost-effective on the condition that it meets specific thresholds relating to (1) uptake and (2) sensitivity for the detection of high-grade cervical intraepithelial neoplasia (CIN2+). We used these thresholds to review the 'early-adopter' programme-level evidence with a mind to determine how well and how consistently they were being met. The available evidence suggested some risk to overall programme performance in the situation where low uptake among underscreened women was accompanied by a high rate of substituting clinician sampling with self-collection among well-screened women. Risk was further compounded in a situation where the slightly reduced sensitivity of self-sampling vs clinician sampling for the detection of CIN2+ was accompanied with lack of adherence to a follow-up triage test that required a clinician sample. To support real-world programmes on their pathways toward implementation and to avoid HPV self-collection being introduced as a screening measure in good faith but with counterproductive consequences, we conclude by identifying a range of mitigations and areas worthy of research prioritisation.

Evaluation of DNA methylation biomarkers ASCL1 and LHX8 on HPV-positive self-collected samples from primary HPV-based screening

BACKGROUND

Host-cell DNA methylation analysis can be used to triage women with high-risk human papillomavirus (HPV)-positive self-collected cervicovaginal samples, but current data are restricted to under-/never-screened women and referral populations. This study evaluated triage performance in women who were offered primary HPV self-sampling for cervical cancer screening.

METHODS

Self-collected samples from 593 HPV-positive women who participated in a primary HPV self-sampling trial (IMPROVE study; NTR5078), were tested for the DNA methylation markers ASCL1 and LHX8 using quantitative multiplex methylation-specific PCR (qMSP). The diagnostic performance for CIN3 and cervical cancer (CIN3 + ) was evaluated and compared with that of paired HPV-positive clinician-collected cervical samples.

RESULTS

Significantly higher methylation levels were found in HPV-positive self-collected samples of women with CIN3 + than control women with no evidence of disease (P values <0.0001). The marker panel ASCL1/LHX8 yielded a sensitivity for CIN3 + detection of 73.3% (63/86; 95% CI 63.9-82.6%), with a corresponding specificity of 61.1% (310/507; 95% CI 56.9-65.4%). The relative sensitivity for detecting CIN3+ was 0.95 (95% CI 0.82-1.10) for self-collection versus clinician-collection, and the relative specificity was 0.82 (95% CI 0.75-0.90).

CONCLUSIONS

The ASCL1/LHX8 methylation marker panel constitutes a feasible direct triage method for the detection of CIN3 + in HPV-positive women participating in routine screening by self-sampling.

Validation of ASCL1 and LHX8 Methylation Analysis as Primary Cervical Cancer Screening Strategy in South African Women with Human Immunodeficiency Virus

BACKGROUND

Compared with women who are human immunodeficiency virus (HIV) negative, women with human immunodeficiency virus (WWH) have a higher human papillomavirus (HPV) prevalence and increased cervical cancer risk, emphasizing the need for effective cervical cancer screening in this population. The present study aimed to validate methylation markers ASCL1 and LHX8 for primary screening in a South African cohort of WWH.

METHODS

In this post hoc analysis within the DIAgnosis in Vaccine And Cervical Cancer Screen (DiaVACCS) study, a South African observational multicenter cohort study, cervical scrape samples from 411 HIV-positive women were analyzed for hypermethylation of ASCL1 and LHX8 genes, HPV DNA, and cytology. Sensitivities, specificities, and positive and negative predictive values of primary methylation-based, HPV-based and cytology-based screening were calculated for the detection of cervical intraepithelial neoplasia of grade 3 or higher.

RESULTS

Single markers ASCL1 and LHX8 resulted in a good performance for the detection of cervical intraepithelial neoplasia of grade 3 or higher, with sensitivities of 85.9% (95% confidence interval [CI], 78.2%-93.6%) and 89.7% (83.0%-96.5%), respectively, and specificities of 72.9% (67.3%-78.5%) and 75.0% (69.5%-80.5%). Combining markers ASCL1 and LHX8 resulted in a lower sensitivity compared with HPV testing (84.6% vs 93.6%, respectively; ratio, 0.90 [95% CI, .82-.99]) and a higher specificity (86.7% vs 78.3%; ratio 1.11 [1.02-1.20]) and reduced the referral rate from 46.8% to 33.4%. ASCL1/LHX8 methylation had a significantly higher sensitivity than cytology (threshold, high-grade intraepithelial squamous lesion or worse), (84.6% vs 74.0%, respectively; ratio, 1.16 [95% CI, 1.01-1.32]) and similar specificity (86.7% vs 91.0%; ratio, 0.95 [.90-1.003]).

CONCLUSIONS

Our results validate the accuracy of ASCL1/LHX8 methylation analysis for primary screening in WWH, which offers a full-molecular alternative to cytology- or HPV-based screening, without the need for additional triage testing.

Posttreatment monitoring by ASCL1/LHX8 methylation analysis in women with HIV treated for cervical intraepithelial neoplasia grade 2/3

OBJECTIVE

Women with HIV (WWH) have an increased risk to develop recurrent cervical intraepithelial neoplasia grade 2/3 (rCIN2/3) after treatment compared with HIV-negative women. Therefore, appropriate posttreatment monitoring of WWH is important. This study evaluates the performance of ASCL1 and LHX8 methylation analysis as posttreatment monitoring test in WWH treated for CIN2/3, as alternative to cytology or human papillomavirus (HPV) as follow-up test.

DESIGN

Prospective observational cohort study.

METHODS

WWH treated for CIN2/3 by large loop excision of the transformation zone (LLETZ) (n  = 61) were invited for follow-up study visits at 1, 2.5 and 4 years after baseline. Baseline and follow-up cervical scrapes were tested for cytology, HPV and DNA methylation of ASCL1 and LHX8 genes. The performance of these strategies for the detection of rCIN2/3 was evaluated in the first follow-up cervical scrape.

RESULTS

Thirteen (21.3%) rCIN2/3 lesions were detected within 4 years of follow-up. In women without rCIN2/3 in follow-up, methylation levels of ASCL1 and LHX8 decreased significantly after LLETZ treatment (P  = 0.02 and 0.007, respectively). In women with rCIN2/3, methylation levels remained high after LLETZ treatment. The 4-year rCIN2/3 risk was 4.9% (95% CI: 0.6-16.5) for ASCL1/LHX8-negative women, 8.1% (95% CI: 1.7-21.9) for HPV-negative women and 7.7% (95% CI: 2.1-18.5) for cytology-negative women.

CONCLUSION

A negative ASCL1/LHX8 methylation test in follow-up is associated with a low rCIN2/3 risk and could serve as an objective test of cure and well tolerated alternative for HPV and/or cytology screening in the posttreatment monitoring of WWH.