Introduction

Cervical Cancer is caused by the human papillomavirus (HPV) and is one of the deadliest cancer types for women in low to middle income countries (LMIC), where the proportion of women vaccinated against HPV is low, and precise, cost-intensive screening programs are not available. Cervical cancer is the second leading cause of cancer deaths in women between the ages of 15 and 44 years in many African countries [1] as well as LMICs outside Africa, like Cambodia [2]. Many of these cases are preventable when diagnosed and treated early.

HPV infects the basal cells of the cervical epithelium, resulting in abnormal cells that proliferate further and, if left untreated, can break through the basement membrane and invade surrounding tissues. As many early-stage infections regress on their own without treatment, it is important to identify high-grade cases where an immediate treatment can save the woman from developing invasive cervical cancer. To detect these severe cases, a non-invasive screening system is used and if there are abnormal findings during the examination, a tissue sample is taken and analyzed under the microscope to give the final diagnosis. The currently used screening method in LMICs is the Visual Inspection of the Cervix After Acetic Acid Application (VIA), in which acetic acid is applied onto the cervix and the diagnosis is done with the naked eye. Although cheap, VIA is highly subjective with a poor sensitivity and an even lower specificity, which leads to a high rate of overtreatment.

Current Screening Methods

An improvement over VIA is the use of a device called the colposcope, which magnifies the cervical area to facilitate the detection of advanced pre-cancerous lesions. Other screening methods include:

• The Papanicolaou (Pap) test: A cytology test in which cells are examined under the microscope for abnormalities.
• HPV DNA test: Detects the presence of high-risk HPV types (like 16 and 18) that are associated with cervical cancer.
• Dual Stain (p16/Ki-67): Detects co-expression of p16 and Ki-67 proteins, which strongly indicates transforming HPV infections.

These methods can be combined for an even more accurate diagnostic result. Unfortunately, most of these methods are expensive and require trained personnel and well-equipped laboratories.

Our Solution at PAICON

At PAICON, we are developing an AI-based diagnostic assistant that can be seamlessly integrated into digital colposcopes as well as smartphone devices. This tool offers the following core functionalities:

• Detection of HPV-induced lesions: Automatically segments colposcopic images to highlight regions affected by HPV-related lesions.
• Grading of lesion severity: Identifies and classifies precancerous lesions by severity, distinguishing high-grade lesions from low-grade ones to support treatment decisions.
• Proposing localization of biopsy: Detects the point of highest grading inside the segmented lesions and gives advice for the optimal biopsy spot.

This AI-powered solution provides an accurate, easy-to-use, and cost-effective alternative to VIA in LMICs.

Conclusion

Cervical cancer remains a major public health challenge in LMICs, where access to reliable and affordable screening methods is limited. The high mortality rate associated with cervical cancer in these regions is largely due to the lack of early detection and timely treatment, despite the fact that most cases are preventable. Current screening techniques like VIA are low-cost but suffer from poor accuracy, while more advanced diagnostic methods, though effective, are often inaccessible due to infrastructure and cost constraints.

Our AI-based diagnostic assistant at PAICON addresses this critical gap by offering a reliable, non-invasive, and affordable tool for early detection and risk stratification of HPV-induced cervical lesions. By automating lesion detection, grading severity, and guiding biopsy localization, our solution empowers healthcare providers to make timely, informed decisions—even in resource-limited settings.

Ultimately, our technology has the potential to reduce unnecessary treatments and increase the overall accuracy of cervical cancer screening programs. With wide implementation, this innovation could play a transformative role in reducing the global burden of cervical cancer and saving the lives of women in LMICs.

References

[1] Bray, F., Parkin, D. M., Gnangnon, F., Tshisimogo, G., Peko, J.-F., Adoubi, I., Assefa, M., Bojang, L., Awuah, B., Koulibaly, M., et al. Cancer in Sub-Saharan Africa in 2020: a review of current estimates of the national burden, data gaps, and future needs. The Lancet Oncology 23, 6 (2022), 719–728.

[2] Bruni, L., Albero, G., Serrano, B., Mena, M., Gómez, D., Muñoz, J., Bosch, F., and De Sanjosé, S. Human Papillomavirus and Related Diseases Report. ICO/IARC Information Centre on HPV and Cancer (HPV Information Centre). Human Papillomavirus and Related Diseases in Cambodia. Summary Report 10 March 2023 (2023).