n/a
n/a
n/a
Advances in technology and scientific techniques created new horizons for improved understanding of the diseases’ processes at a molecular level. In the field of cervical pre-invasive and invasive disease, this allowed an in-depth exploration of the neoplastic mechanisms at a molecular level and led to the development of new test and biomarkers, many of which have become commercially available. With the explosion of new biomarkers targeting the viral DNA detection, the expression of oncoproteins and other cellular processes that promote carcinogenesis in the host, questions on how to best use these in different clinical settings are becoming increasing difficult to answer. With a continuously evolving evidence-base, the development a clinical decision support system is a current unmet need. This can assist clinicians to use these new technologies to promote prevention, personalise management and improve targeted management.
Since 2010 the Hellenic Cervical Pathology Academic (HeCPA) study group, is working on innovative approaches to exploit advanced mathematical and computing tools for the optimal use of ancillary tests that are available nowadays. More recently, we published a prospective multicentric study of a large patient cohort employed advanced neural networks and artificial intelligence techniques for the development of a Clinical Decision Support Scoring System (DSSS). The system developed had the ability to exploit all the biomarker information in order to accurately predict which women had clinically significant lesions with true oncogenic potential (CIN2 or worse) and give a quantified probability for different histological diagnoses. The results clearly and consistently demonstrated that this DSSS could achieve an optimal balance of increased sensitivity and specificity and minimize the rate of false negative and false positive results.
Improved accuracy and clinical decision support systems has important implications to patients, the health systems and policymakers. If these systems can predict with high accuracy women with or without the disease, they have the potential to significantly improve the management of these populations and the quantified probability for all possible histological diagnoses will be available for effective patient counseling at the primary care setting and in the colposcopy clinics.