An intelligent clinical decision support system for patient-specific predictions to improve cervical intraepithelial neoplasia detection

Revolutionizing Women's Health: A Comprehensive Review of Artificial Intelligence Advancements in Gynecology

Artificial intelligence has yielded remarkably promising results in several medical fields, namely those with a strong imaging component. Gynecology relies heavily on imaging since it offers useful visual data on the female reproductive system, leading to a deeper understanding of pathophysiological concepts. The applicability of artificial intelligence technologies has not been as noticeable in gynecologic imaging as in other medical fields so far. However, due to growing interest in this area, some studies have been performed with exciting results. From urogynecology to oncology, artificial intelligence algorithms, particularly machine learning and deep learning, have shown huge potential to revolutionize the overall healthcare experience for women's reproductive health. In this review, we aim to establish the current status of AI in gynecology, the upcoming developments in this area, and discuss the challenges facing its clinical implementation, namely the technological and ethical concerns for technology development, implementation, and accountability.

Learning meaningful latent space representations for patient risk stratification: Model development and validation for dengue and other acute febrile illness

Background

Increased data availability has prompted the creation of clinical decision support systems. These systems utilise clinical information to enhance health care provision, both to predict the likelihood of specific clinical outcomes or evaluate the risk of further complications. However, their adoption remains low due to concerns regarding the quality of recommendations, and a lack of clarity on how results are best obtained and presented.

Methods

We used autoencoders capable of reducing the dimensionality of complex datasets in order to produce a 2D representation denoted as latent space to support understanding of complex clinical data. In this output, meaningful representations of individual patient profiles are spatially mapped in an unsupervised manner according to their input clinical parameters. This technique was then applied to a large real-world clinical dataset of over 12,000 patients with an illness compatible with dengue infection in Ho Chi Minh City, Vietnam between 1999 and 2021. Dengue is a systemic viral disease which exerts significant health and economic burden worldwide, and up to 5% of hospitalised patients develop life-threatening complications.

Results

The latent space produced by the selected autoencoder aligns with established clinical characteristics exhibited by patients with dengue infection, as well as features of disease progression. Similar clinical phenotypes are represented close to each other in the latent space and clustered according to outcomes broadly described by the World Health Organisation dengue guidelines. Balancing distance metrics and density metrics produced results covering most of the latent space, and improved visualisation whilst preserving utility, with similar patients grouped closer together. In this case, this balance is achieved by using the sigmoid activation function and one hidden layer with three neurons, in addition to the latent dimension layer, which produces the output (Pearson, 0.840; Spearman, 0.830; Procrustes, 0.301; GMM 0.321).

Conclusion

This study demonstrates that when adequately configured, autoencoders can produce two-dimensional representations of a complex dataset that conserve the distance relationship between points. The output visualisation groups patients with clinically relevant features closely together and inherently supports user interpretability. Work is underway to incorporate these findings into an electronic clinical decision support system to guide individual patient management.

The Influence of Sexual Behavior and Demographic Characteristics in the Expression of HPV-Related Biomarkers in a Colposcopy Population of Reproductive Age Greek Women

Simple Summary

This observational study aimed to investigate the possible influence of sexual behavior and demographic characteristics in the expression of HPV-related biomarkers in a colposcopy population. Lifestyle factors that impacted HPV DNA positivity in a statistically significant manner were younger age at first sexual intercourse, a history of more than four sexual partners as well as a partner change during the last year before HPV DNA assessment. Although ambitious, the development and validation of lifestyle scoring systems that combine molecular and epidemiological patient data to effectively predict underlying cervical pathology will represent a milestone in the quest of cervical cancer prevention.

Abstract

Despite the significant scientific evolution in primary and secondary cervical cancer prevention in the battle started by George Papanicolaou in the previous century, global cervical cancer mortality rates remain disappointing. The widespread implementation of HPV-related molecular markers has paved the way to tremendous developments in cervical cancer screening, with the transition from cytological approach to the more accurate and cost-effective HPV testing modalities. However, the academic audience and different health systems have not yet adopted a universal approach in screening strategies, and even artificial intelligence modalities have been utilized from the multidisciplinary scientific armamentarium. Combination algorithms, scoring systems as well as artificial intelligent models have been so far proposed for cervical screening and management. The impact of sexual lifestyle inherently possesses a key role in the prevalence of HPV-related biomarkers. This study aimed to investigate any possible influence of sexual behavior and demographic characteristics in the expression of HPV-related biomarkers in a colposcopy population from October 2016 to June 2017, and corroborated the determining role of age at fist intercourse; the older the age, the lower the probability for DNA positivity. Multivariate analysis illustrated additionally that a number of sexual partners exceeding 4.2 was crucial, with women with ≤5 partners being approximately four times less likely to harbor a positive HPV DNA test (p < 0.0001). Similarly, a reported partner change during the last year before HPV DNA assessment contributed to 2.5 times higher odds for DNA positivity (p = 0.0006). From this perspective, the further development and validation of scoring systems quantifying lifestyle factors that could reflect cervical precancer risk seems paramount.

Morbidity after local excision of the transformation zone for cervical intra-epithelial neoplasia and early cervical cancer

The awareness that cervical intra-epithelial neoplasia (CIN) treatment increases the risk of preterm birth has led to major changes in clinical practice. Women with CIN have a higher baseline risk of prematurity but local treatment further increases this risk. The risk further increases with increasing cone length and multiplies for repeat excisions; it is unclear whether small cones confer any additional risk to CIN alone. There is no evidence to suggest that fertility is affected by local treatment, although this increases the risk of mid-trimester loss. Caution should prevail when deciding to treat women with CIN of reproductive age. If treatment is offered, this should be conducted effectively to optimise the clearance of disease and minimise the risk of recurrence. Colposcopists should alert women undergoing treatment that this may increase the risk of preterm birth and that they may be offered interventions when pregnant. The cone length should be clearly documented and used as a risk stratifier.

Novel biomarkers with promising benefits for diagnosis of cervical neoplasia: a systematic review

BACKGROUND

Cervical cancer screening is slowly transitioning from Pappanicolaou cytologic screening to primary Visual Inspection with Acetic Acid (VIA) or HPV testing as an effort to enhance early detection and treatment. However, an effective triage tests needed to decide who among the VIA or HPV positive women should receive further diagnostic evaluation to avoid unnecessary colposcopy referrals is still lacking. Evidence from experimental studies have shown potential usefulness of Squamous Cell Carcinoma Antigen (SCC Ag), Macrophage Colony Stimulating Factor (M-CSF), Vascular Endothelial Growth Factor (VEGF), MicroRNA, p16INKa / ki-67, HPV E6/E7/mRNA, and DNA methylation biomarkers in detecting premalignant cervical neoplasia. Given the variation in performance, and scanty review studies in this field, this systematic review described the diagnostic performance of some selected assays to detect high-grade cervical intraepithelial neoplasia (CIN2+) with histology as gold standard.

METHODS

We systematically searched articles published in English between 2012 and 2020 using key words from PubMed/Medline and SCOPUS with two reviewers assessing study eligibility, and risk of bias. We performed a descriptive presentation of the performance of each of the selected assays for the detection of CIN2 + .

RESULTS

Out of 298 citations retrieved, 58 articles were included. Participants with cervical histology yielded CIN2+ proportion range of 13.7-88.4%. The diagnostic performance of the assays to detect CIN2+ was; 1) SCC-Ag: range sensitivity of 78.6-81.2%, specificity 74-100%. 2) M-CSF: sensitivity of 68-87.7%, specificity 64.7-94% 3) VEGF: sensitivity of 56-83.5%, specificity 74.6-96%. 4) MicroRNA: sensitivity of 52.9-67.3%, specificity 76.4-94.4%. 5) p16INKa / ki-67: sensitivity of 50-100%, specificity 39-90.4%. 6) HPV E6/E7/mRNA: sensitivity of 65-100%, specificity 42.7-90.2%, and 7) DNA methylation: sensitivity of 59.7-92.9%, specificity 67-98%.

CONCLUSION

Overall, the reported test performance and the receiving operating characteristics curves implies that implementation of p16ink4a/ki-67 assay as a triage for HPV positive women to be used at one visit with subsequent cryotherapy treatment is feasible. For the rest of assays, more robust clinical translation studies with larger consecutive cohorts of women participants is recommended.

The association between sexually transmitted infections, human papillomavirus, and cervical cytology abnormalities among women in Greece

OBJECTIVE

To investigate the diagnosis of sexually transmitted infections (STIs) with human papillomavirus (HPV) infection and the presence of cytological changes in the cervix in a cohort of sexually active women in Greece.

METHODS

Cervical cytology testing and the molecular typing of HPV and other STIs were performed for 345 sexually active women aged between 18 and 45 years (mean 33.2±7.2years) visiting a gynaecology clinic for routine cervical screening. The association of HPV and STI detection with cytological findings was investigated.

RESULTS

HPV was detected in 61 women (17.7%) and STIs in 82 (23.8%). Ureaplasma spp was the most frequently detected pathogen, which was found in 63 (18.2%) women, followed by Mycoplasma spp (21 women, 25.6%) and Chlamydia trachomatis (five women, 6.1%). HPV positivity only (with no co-presence of STI) was associated with an abnormal cytology (odds ratio 6.9, p<0.001), while women who were negative for both HPV and STIs had a higher probability of a normal cytology (odds ratio 0.36, p<0.01). Sixteen out of the 63 (25.4%) women who tested positive for Ureaplasma spp, harboured a high-risk HPV type (odds ratio 2.3, p=0.02).

CONCLUSIONS

In a population with a high prevalence of Ureaplasma spp, there was an association of this pathogen with high-risk HPV infection, a finding that needs further elucidation.

Immediate referral to colposcopy versus cytological surveillance for minor cervical cytological abnormalities in the absence of HPV test

BACKGROUND

A significant number of women are diagnosed with minor cytological abnormalities on cervical screening. Many authorities recommend surveillance as spontaneous regression might occur. However, attendance for cytological follow-up decreases with time and might put some women at risk of developing invasive disease.

OBJECTIVES

To assess the optimum management strategy for women with minor cervical cytological abnormalities (atypical squamous cells of undetermined significance - ASCUS or low-grade squamous intra-epithelial lesions - LSIL) at primary screening in the absence of HPV (human papillomavirus) DNA test.

SEARCH METHODS

We searched the following electronic databases: Cochrane Central Register of Controlled Trials (CENTRAL Issue 4, 2016), MEDLINE (1946 to April week 2 2016) and Embase (1980 to 2016 week 16).

SELECTION CRITERIA

We included randomised controlled trials (RCTs) comparing immediate colposcopy to cytological surveillance in women with atypical squamous cells of undetermined significance (ASCUS/borderline) or low-grade squamous intra-epithelial lesions (LSIL/mild dyskaryosis).

DATA COLLECTION AND ANALYSIS

The primary outcome measure studied was the occurrence of cervical intra-epithelial neoplasia (CIN). The secondary outcome measures studied included default rate, clinically significant anxiety and depression, and other self-reported adverse effects.We classified studies according to period of surveillance, at 6, 12, 24 or 36 months, as well as at 18 months, excluding a possible exit-examination. We calculated pooled risk ratios (RR) and 95% confidence intervals (CI) using a random-effects model with inverse variance weighting. Inter-study heterogeneity was assessed with I² statistics.

MAIN RESULTS

We identified five RCTs with 11,466 participants that fulfilled the inclusion criteria. There were 18 cases of invasive cervical cancer, seven in the immediate colposcopy and 11 in the cytological surveillance groups, respectively. Although immediate colposcopy detects CIN2+ and CIN3+ earlier than cytology, the differences were no longer observed at 24 months (CIN2+: 3 studies, 4331 women; 17.9% versus 18.3%, RR 1.14, CI 0.66 to 1.97; CIN3+: 3 studies, 4331 women; 10.3% versus 11.9%, RR 1.02, CI 0.53 to 1.97). The inter-study heterogeneity was considerable (I² greater than 90%). Furthermore, the inclusion of the results of the exit examinations at 24 months, which could inflate the CIN detection rate of cytological surveillance, may have led to study design-derived bias; we therefore considered the evidence to be of low quality.When we excluded the exit examination, the detection rate of high-grade lesions at the 18-month follow-up was higher after immediate colposcopy (CIN2+: 2 studies, 4028 women; 14.3% versus 10.1%, RR 1.50, CI 1.12 to 2.01; CIN3+: 2 studies, 4028 women, 7.8% versus 6.9%, RR 1.24, CI 0.77 to 1.98) both had substantial inter-study heterogeneity (I² greater than 60%) and we considered the evidence to be of moderate quality).The meta-analysis revealed that immediate referral to colposcopy significantly increased the detection of clinically insignificant cervical abnormalities, as opposed to repeat cytology after 24 months of surveillance (occurrence of koilocytosis: 2 studies, 656 women; 32% versus 21%, RR 1.49, 95% CI 1.17 to 1.90; moderate-quality evidence) incidence of any CIN: 2 studies, 656 women; 64% versus 32%, RR 2.02, 95% CI 1.33 to 3.08, low-quality evidence; incidence of CIN1: 2 studies, 656 women; 21% versus 8%, RR 2.58, 95% CI 1.69 to 3.94, moderate-quality evidence).Due to differences in trial designs and settings, there was large variation in default rates between the included studies. The risk for default was higher for the repeat cytology group, with a four-fold increase at 6 months, a six-fold at 12 and a 19-fold at 24 months (6 months: 3 studies, 5117 women; 6.3% versus 13.3%, RR 3.85, 95% CI 1.27 to 11.63, moderate-quality evidence; 12 months: 3 studies, 5115 women; 6.3% versus 14.8%, RR 6.39, 95% CI 1.49 to 29.29, moderate-quality evidence; 24 months: 3 studies, 4331 women; 0.9% versus 16.1%, RR 19.1, 95% CI 9.02 to 40.43, moderate-quality evidence).

AUTHORS' CONCLUSIONS

Based on low- or moderate-quality evidence using the GRADE approach and generally low risk of bias, the detection rate of CIN2+ or CIN3+ after two years does not appear to differ between immediate colposcopy and cytological surveillance in the absence of HPV testing, although women may default from follow-up. Immediate colposcopy probably leads to earlier detection of high-grade lesions, but also detects more clinically insignificant low-grade lesions. Colposcopy may therefore be the first choice when good compliance is not assured. These results emphasize the need for an accurate reflex HPV triage test to distinguish women who need diagnostic follow-up from those who can return safely to routine recall.

γH2Ax Expression as a Potential Biomarker Differentiating between Low and High Grade Cervical Squamous Intraepithelial Lesions (SIL) and High Risk HPV Related SIL

Background

γH2AX is a protein biomarker for double-stranded DNA breakage; its expression was studied in cervical squamous intraepithelial lesions and carcinomas.

Methods

Immunostaining for phospho-γH2AX was performed in sections from histologically confirmed cervical SIL and carcinomas, as well as from normal cervices used as controls. In total, 275 cases were included in the study: 112 low grade SIL (LGSIL), 99 high grade SIL (HGSIL), 24 squamous cell carcinoma (SCC), 12 adenocarcinoma and 28 cervical specimens with no essential lesions. Correlation of histological grading, high risk vs. low risk HPV virus presence, activated vs. non-activated status (by high risk HPV mRNA expression) and γH2AX expression in both basal and surface segments of the squamous epithelium was performed.

Results

Gradual increase of both basal and surface γH2AX expression was noted up from normal cervices to LGSIL harboring a low risk HPV type, to LGSIL harboring a high risk virus at a non-activated state (p<0.05). Thereafter, both basal and surface γH2AX expression dropped in LGSIL harboring a high risk virus at an activated state and in HGSIL.

Conclusions

γH2AX could serve as a potential biomarker discriminating between LGSIL and HGSIL, as well as between LGSIL harboring high risk HPV at an activated state.

Mapping Population Health Management Roadmap into Cervical Cancer Screening Programs

Population Health Management (PHM) aims to provide better health outcomes for preventing diseases, closing care gaps and providing more personalized care. Since the inception of the Pap test, cervical cancer (CxCa) decreased in countries applying cervical cancer programs, involving both prevention and treatment. In this article, the authors map the PHM roadmap to the design of cervical cancer screening programs and examine the effect on the supporting information technology systems. Notwithstanding screening programs have a tight relation to PHM; the mapping reveals numerous interventions involving additional data sources, and timeless reconfiguration.

HPV Detection Methods

Human papilloma viruses (HPVs) have been acknowledged to be the leading risk factor of cervical intra-epithelial lesion creation (CIN) and cervical cancer development (CxCa). Many different techniques have been created and utilized in HPV detection and monitoring with a vast amount of them being commercialized and few of them integrated in official screening strategies. A growing trend for DNA typing of the 14 most commonly accepted high risk HPV types has been introduced, supporting that in many cases molecular testing could replace classic morphologic diagnostic routines, even though DNA detection has lower specificity than other molecular and morphology tests. However, there have been limited attempts in combining data from all different techniques to provide efficient patient triaging schemes, since, apart from the obvious increase of patient cost, the amount of data and its interpretation in patient management has been impossible. Complex computer based clinical support decision systems, many of which are based on artificial intelligence may abolish these limitations.

Development of a clinical decision support system using genetic algorithms and Bayesian classification for improving the personalised management of women attending a colposcopy room

Cervical cancer (CxCa) is often the result of underestimated abnormalities in the test Papanicolaou (Pap test). The recent advances in the study of the human papillomavirus (HPV) infection (the necessary cause for CxCa development) have guided clinical practice to add HPV related tests alongside the Pap test. In this way, today, HPV DNA testing is well accepted as an ancillary test and it is used for the triage of women with abnormal findings in cytology. However, these tests are either highly sensitive or highly specific, and therefore none of them provides an optimal solution. In this Letter, a clinical decision support system based on a hybrid genetic algorithm - Bayesian classification framework is presented, which combines the results of the Pap test with those of the HPV DNA test in order to exploit the benefits of each method and produce more accurate outcomes. Compared with the medical tests and their combinations (co-testing), the proposed system produced the best receiver operating characteristic curve and the most balanced combination among sensitivity and specificity in detecting high-grade cervical intraepithelial neoplasia and CxCa (CIN2+). This system may support decision-making for the improved management of women who attend a colposcopy room following a positive test result.

A decision support system based on an ensemble of random forests for improving the management of women with abnormal findings at cervical cancer screening

In most cases, cervical cancer (CxCa) develops due to underestimated abnormalities in the Pap test. Today, there are ancillary molecular biology techniques available that provide important information related to CxCa and the Human Papillomavirus (HPV) natural history, including HPV DNA tests, HPV mRNA tests and immunocytochemistry techniques such as overexpression of p16. These techniques are either highly sensitive or highly specific, however not both at the same time, thus no perfect method is available today. In this paper we present a decision support system (DSS) based on an ensemble of Random Forests (RFs) for the intelligent combination of the results of classic and ancillary techniques that are available for CxCa detection, in order to exploit the benefits of each technique and produce more accurate results. The proposed system achieved both, high sensitivity (86.1%) and high specificity (93.3%), as well as high overall accuracy (91.8%), in detecting cervical intraepithelial neoplasia grade 2 or worse (CIN2+). The system's performance was better than any other single test involved in this study. Moreover, the proposed architecture of employing an ensemble of RFs proved to be better than the single classifier approach. The presented system can handle cases with missing tests and more importantly cases with inadequate cytological outcome, thus it can also produce accurate results in the case of stand-alone HPV-based screening, where Pap test is not applied. The proposed system may identify women at true risk of developing CxCa and guide personalised management and therapeutic interventions.

Computerized Clinical Decision Support: Contributions from 2014

OBJECTIVE

To summarize recent research and propose a selection of best papers published in 2014 in the field of computerized clinical decision support for the Decision Support section of the IMIA yearbook.

METHOD

A literature review was performed by searching two bibliographic databases for papers related to clinical decision support systems (CDSSs) and computerized provider order entry systems in order to select a list of candidate best papers to be then peer-reviewed by external reviewers. A consensus meeting between the two section editors and the editorial team was finally organized to conclude on the selection of best papers.

RESULTS

Among the 1,254 returned papers published in 2014, the full review process selected four best papers. The first one is an experimental contribution to a better understanding of unintended uses of CDSSs. The second paper describes the effective use of previously collected data to tailor and adapt a CDSS. The third paper presents an innovative application that uses pharmacogenomic information to support personalized medicine. The fourth paper reports on the long-term effect of the routine use of a CDSS for antibiotic therapy.

CONCLUSIONS

As health information technologies spread more and more meaningfully, CDSSs are improving to answer users' needs more accurately. The exploitation of previously collected data and the use of genomic data for decision support has started to materialize. However, more work is still needed to address issues related to the correct usage of such technologies, and to assess their effective impact in the long term.

The Application of Classification and Regression Trees for the Triage of Women for Referral to Colposcopy and the Estimation of Risk for Cervical Intraepithelial Neoplasia: A Study Based on 1625 Cases with Incomplete Data from Molecular Tests

OBJECTIVE

Nowadays numerous ancillary techniques detecting HPV DNA and mRNA compete with cytology; however no perfect test exists; in this study we evaluated classification and regression trees (CARTs) for the production of triage rules and estimate the risk for cervical intraepithelial neoplasia (CIN) in cases with ASCUS+ in cytology.

STUDY DESIGN

We used 1625 cases. In contrast to other approaches we used missing data to increase the data volume, obtain more accurate results, and simulate real conditions in the everyday practice of gynecologic clinics and laboratories. The proposed CART was based on the cytological result, HPV DNA typing, HPV mRNA detection based on NASBA and flow cytometry, p16 immunocytochemical expression, and finally age and parous status.

RESULTS

Algorithms useful for the triage of women were produced; gynecologists could apply these in conjunction with available examination results and conclude to an estimation of the risk for a woman to harbor CIN expressed as a probability.

CONCLUSIONS

The most important test was the cytological examination; however the CART handled cases with inadequate cytological outcome and increased the diagnostic accuracy by exploiting the results of ancillary techniques even if there were inadequate missing data. The CART performance was better than any other single test involved in this study.