Cervical cancer prediction

Cervical cancer is one of the most common malignancies in women, developing from precursor lesions known as cervical intraepithelial neoplasia (CIN) and through persistent infection with high-risk human papillomavirus (HR-HPV). It is a well-known cause of death for women worldwide and continues to be a significant public health concern. Early detection and treatment are essential to prevent the progression of cervical cancer and reduce its impact on women's health (2)(5).

Size of the Problem

• In 2020, there were 604,000 new cases of cervical cancer reported, with 342,000 associated deaths (1).
• Approximately 30% of high-grade cervical intraepithelial neoplasia (CIN) lesions progress to invasive cancer within 30 years (2).
• HPV vaccine coverage is low, underscoring the ongoing importance of routine cervical cancer screening for women (4).
• Strategies for detecting and treating precancerous lesions in women are critical to prevent the progression of cervical cancer and minimize its impact on public health (5).

‍

It is essential to have accurate medical tests in the context of cervical screening and other aspects of healthcare. These tests play a fundamental role in detecting anomalies early, such as precancerous lesions, enabling timely treatment and preventing the development of invasive cancer. Additionally, by ensuring accuracy in diagnosis, these tests improve patient safety and reduce costs associated with diagnostic errors and unnecessary treatments. Therefore, improving the accuracy of medical tests not only directly impacts the quality of healthcare but can also have significant implications for reducing the global burden of cervical cancer (2)(3).

• Development of advanced algorithms for cervical image analysis: This advancement encompasses solutions such as CerviCARE AI, an artificial intelligence-based software that automates the interpretation of telecervicography images. Its purpose is to distinguish between low-grade and high-grade lesions, thus improving the detection of cervical precancerous lesions. A retrospective multicenter study demonstrated that CerviCARE AI achieved a sensitivity of 98% for high-risk groups, with a specificity of 95.5%. Although further research is needed to validate its clinical utility, this example illustrates how artificial intelligence can enhance diagnostic accuracy and reduce human interpretation errors in cervical image analysis, facilitating early detection and timely treatment of precancerous and cancerous lesions (2).

‍

• Implementation of automated triage systems based on artificial intelligence: This approach provides a detailed guide for designing and evaluating AI algorithms that complement automated triage systems. By following this systematic approach, algorithms can be developed to analyze and prioritize the results of HPV and other biomarker screening tests, allowing for rapid identification of high-risk cases. These systems optimize resources and improve the efficiency of the detection process, thereby facilitating early identification and timely treatment of precancerous lesions, significantly contributing to the prevention and effective management of cervical cancer (3).

‍

• Development of predictive tools based on artificial intelligence to assess risks: We highlight the study by the National Cancer Institute, which demonstrated that an artificial intelligence algorithm improved the accuracy and efficiency of cervical cancer detection compared to cytology, the current standard for monitoring women with positive results in primary HPV screening. This approach automates the assessment of dual stains, reducing human interpretation errors and expediting the detection and treatment of precancerous lesions. These advances promote greater medical efficiency and have clear implications for clinical care (4).

‍

• Advancement in artificial intelligence-assisted diagnostic systems: This perspective involves the creation of AI-assisted systems to improve the accuracy and efficiency of early detection and diagnosis of cervical cancer. These systems would use deep learning algorithms to analyze magnetic resonance imaging (MRI) images and clinical data, accurately identifying patterns associated with cervical cancer. AI would provide rapid and accurate evaluation, assisting medical professionals in detecting cervical precancerous lesions. These systems could be integrated into clinical settings, improving medical efficiency and enabling timely treatment of affected patients. However, further validation and close collaboration between AI and medical professionals would be required to ensure their accuracy and reliability in clinical practice (5).

‍

• Public cervical image databases, such as the Cervical Image Database from the National Cancer Institute (NCI).
• Medical image databases, like The Cancer Imaging Archive (TCIA) Medical Image Database.
• Clinical databases containing information on patients with HPV screening results and other biomarkers, such as electronic medical records and clinical trial databases.
• Population databases include demographic information, medical history, and screening test results, such as the National Cervical Cancer Screening (Papanicolaou) Program or national health survey databases.
• Medical image databases containing MRI images and clinical data from patients with and without cervical cancer, such as TCIA (The Cancer Imaging Archive) and the Human Brain MRI Database.