Unresolved issues in the management of human papillomavirus-associated mucosal high-grade pre-cancers

This article reviews human papillomavirus-associated mucosal high-grade pre-cancers and their management. It examines pre-cancer classification systems, the natural history of HPV-associated pre-cancers, the various types of management and treatment for HPV pre-cancers, the various mucosal site-specific considerations, and then some of the unresolved issues. Different conclusions are reached for each of the relevant mucosal sites, which are cervix, vagina, vulva, anus, penis and oro-pharynx, and indeed there are differing volumes of evidence relating to each of these sites, and thus differing degrees of certainty/uncertainty in the recommendations.

Accurate genome-wide genotyping from archival tissue to explore the contribution of common genetic variants to pre-cancer outcomes

PURPOSE

The contribution of common genetic variants to pre-cancer progression is understudied due to long follow-up time, rarity of poor outcomes and lack of available germline DNA collection. Alternatively, DNA from diagnostic archival tissue is available, but its somatic nature, limited quantity and suboptimal quality would require an accurate cost-effective genome-wide germline genotyping methodology.

EXPERIMENTAL DESIGN

Blood and tissue DNA from 10 individuals were used to benchmark the accuracy of Single Nucleotide Polymorphisms (SNP) genotypes, Polygenic Risk Scores (PRS) or HLA haplotypes using low-coverage whole-genome sequencing (lc-WGS) and genotype imputation. Tissue-derived PRS were further evaluated for 36 breast cancer patients (11.7 years median follow-up time) diagnosed with DCIS and used to model the risk of Breast Cancer Subsequent Events (BCSE).

RESULTS

Tissue-derived germline DNA profiling resulted in accurate genotypes at common SNPs (blood correlation r2 > 0.94) and across 22 disease-related polygenic risk scores (PRS, mean correlation r = 0.93). Imputed Class I and II HLA haplotypes were 96.7% and 82.5% concordant with clinical-grade blood HLA haplotypes, respectively. In DCIS patients, tissue-derived PRS was significantly associated with BCSE (HR = 2, 95% CI 1.2-3.8). The top and bottom decile patients had an estimated 28% and 5% chance of BCSE at 10 years, respectively.

CONCLUSIONS

Archival tissue DNA germline profiling using lc-WGS and imputation, represents a cost and resource-effective alternative in the retrospective design of long-term disease genetic studies. Initial results in breast cancer suggest that common risk variants contribute to pre-cancer progression.

Use of artificial intelligence in diagnosis of head and neck precancerous and cancerous lesions: A systematic review

This systematic review analyses and describes the application and diagnostic accuracy of Artificial Intelligence (AI) methods used for detection and grading of potentially malignant (pre-cancerous) and cancerous head and neck lesions using whole slide images (WSI) of human tissue slides. Electronic databases MEDLINE via OVID, Scopus and Web of Science were searched between October 2009 - April 2020. Tailored search-strings were developed using database-specific terms. Studies were selected using a strict inclusion criterion following PRISMA Guidelines. Risk of bias assessment was conducted using a tailored QUADAS-2 tool. Out of 315 records, 11 fulfilled the inclusion criteria. AI-based methods were employed for analysis of specific histological features for oral epithelial dysplasia (n = 1), oral submucous fibrosis (n = 5), oral squamous cell carcinoma (n = 4) and oropharyngeal squamous cell carcinoma (n = 1). A combination of heuristics, supervised and unsupervised learning methods were employed, including more than 10 different classification and segmentation techniques. Most studies used uni-centric datasets (range 40-270 images) comprising small sub-images within WSI with accuracy between 79 and 100%. This review provides early evidence to support the potential application of supervised machine learning methods as a diagnostic aid for some oral potentially malignant and malignant lesions; however, there is a paucity of evidence using AI for diagnosis of other head and neck pathologies. Overall, the quality of evidence is low, with most studies showing a high risk of bias which is likely to have overestimated accuracy rates. This review highlights the need for development of state-of-the-art deep learning techniques in future head and neck research.

FAM19A4/miR124-2 methylation analysis as a triage test for HPV-positive women: cross-sectional and longitudinal data from a Dutch screening cohort

OBJECTIVES

The aim was to evaluate the cross-sectional and long-term triage performance of FAM19A4/miR124-2 methylation analysis in human papillomavirus (HPV)-based cervical screening.

METHODS

We conducted a post hoc analysis within a Dutch population-based HPV-positive study cohort of women aged 30-60 years (n = 979). Cross-sectional cervical intraepithelial neoplasia (CIN) 3+ sensitivity, specificity, positive predictive value and negative predictive value as well as cumulative CIN3+ or cervical cancer risks after 9 and 14 years were compared for three baseline triage strategies: (1) cytology, (2) FAM19A4/miR124-2 methylation analysis and (3) combined FAM19A4/miR124-2 methylation with cytology.

RESULTS

CIN3+ sensitivity of FAM19A4/miR124-2 methylation analysis was similar to that of cytology (71.3% vs 76.0%, ratio 0.94, 95% CI 0.84 to 1.05), at a lower specificity (78.3% vs 87.0%, ratio 0.90, 95% CI 0.86 to 0.94). Combining FAM19A4/miR124-2 methylation analysis with cytology resulted in a CIN3+ sensitivity of 84.6% (95% CI 78.3 to 90.8) at a specificity of 69.6% (95% CI 66.5 to 72.7). Similar 9- and 14-year CIN3+ risks for baseline cytology-negative women and baseline FAM19A4/miR124-2 methylation-negative women were observed, with risk differences of -0.42% (95% CI -2.1 to 1.4) and -0.07% (95% CI -1.9 to 1.9), respectively. The 14-year cumulative cervical cancer incidence was significantly lower for methylation-negative women compared to cytology-negative women (risk difference 0.98%, 95% CI 0.26 to 2.0).

DISCUSSION

FAM19A4/miR124-2 methylation analysis has a good triage performance on baseline screening samples, with a cross-sectional CIN3+ sensitivity and long-term triage-negative CIN3+ risk equalling cytology triage. Therefore, FAM19A4/miR124-2 methylation analysis appears to be a good and objective alternative to cytology in triage scenarios in HPV-based cervical screening.

Status of thiocyanate levels in the serum and saliva of non-smokers, ex-smokers and smokers

BACKGROUND

Use of tobacco is often implicated in the development of oral diseases. Questionable accuracy of the traditional questionnaires to assess cigarette exposure necessitates the use of biomarkers like thiocyanate which provide a definitive quantitative measure.

OBJECTIVE

To assess the rise in the level of thiocyanate for measurement of smoking behaviour in adults.

MATERIALS AND METHODS

Serum and salivary thiocyanate levels were estimated in 20 non-smokers, 20 ex-smokers and 40 smokers. Smokers were divided into two groups based on the presence or absence of oral mucosal lesions.

RESULTS

The mean serum and salivary thiocyanate levels were increased significantly in smokers when compared to non-smokers and ex-smokers. The levels were not significantly different between ex-smokers and non-smokers and between smokers with tobacco related oral mucosal lesions and those without. Statistically significant correlation was seen between the serum and salivary levels of thiocyanate.

CONCLUSION

This study highlights the high level of thiocyanate in the serum and saliva of smokers when compared to non-smokers and ex-smokers. Significant increase in thiocyanate level was also seen in saliva. Hence it can be stated that saliva can be used as a reliable, non-invasive tool to assess smoking behaviour in the population and its changes over time.

The genomic road to invasion-examining the similarities and differences in the genomes of associated oral pre-cancer and cancer samples

BACKGROUND

It is frequently assumed that pre-invasive lesions are simpler precursors of cancer and will contain a limited subset of the genomic changes seen in their associated invasive disease. Driver mutations are thought to occur early, but it is not known how many of these are present in pre-invasive lesions. These assumptions need to be tested with the increasing focus on both personalised cancer treatments and early detection methodologies.

METHODS

We examined genomic copy number changes in 256 pre-invasive and invasive samples from 69 oral cancer patients. Forty-eight samples from 16 patients were further examined using exome sequencing.

RESULTS

Evidence of a shared ancestor of both dysplasia and carcinoma was seen in all but one patient. One-third of dysplasias showed independent copy number events. The remainder had a copy number pattern that was similar to or simpler than that of the carcinoma. All dysplasias examined contained somatic mutations absent in the related carcinoma. Previously observed copy number changes and TP53 mutations were very frequently observed, and almost always shared between dysplasia and carcinoma. Other gene changes were more sporadic. Pathway analysis confirmed that each patient's disease developed in a different way. Examining the numbers of shared mutations and the rate of accumulation of mutations showed evidence that all samples contain a population of sub-clones, with little evidence of selective advantage of a subset of these.

CONCLUSIONS

These findings suggest that most of the genomic changes driving oral cancer occur in the pre-cancerous state by way of gradual random accumulation rather than a dramatic single event.