Introduction: Cervical Cancer burden in Kenya

As per the latest stats from Globocan, Kenya has a high burden of cervical cancer. In 2020, at least 5600 new cases of the disease were diagnosed, a conservative figure by all means. 

Despite cervical cancer being preventable and curable when detected early, Kenya faces a big challenge in rolling out screening and treatment of early cervical lesions. In one study, only 14% of eligible women had ever had cervical cancer screening. 

There are many contributing factors to the low screening rates. The 2021 WHO guidelines on cervical cancer screening detail 3 modalities: HPV testing, VIA/VILI and cytology. Each of these has its unique challenges, while some like inadequate infrastructure, low funding and lack of adequate trained personnel are cross-cutting.

Kenya, with a population of 56 million people, faces significant challenges in providing efficient and accurate cervical cancer screening services. For this post, we will focus on cytology and the potential for emerging technologies such as machine learning adn AI.

Healthcare workforce

With only 100 pathologists and less than 30 cytopathologists available, the demand for screening far exceeds the capacity of the existing workforce. However, the emergence of AI-enabled cytology slide image analysis offers a potential solution to this problem. By leveraging this technology, Kenya can benefit from improved diagnostic accuracy, increased screening efficiency, and enhanced access to cervical cancer screening services.

Photo:   Credit 

Benefits of AI-Enabled Cytology Slide Image Analysis in Kenya

1. Improved diagnostic accuracy: 

AI algorithms can be trained to detect and classify abnormal cells with a high degree of accuracy, reducing the risk of false negatives and false positives in cervical cancer screening (Liu et al., 2021). This can help ensure that women receive timely and appropriate follow-up care based on their screening results.

1. Increased screening efficiency:

 AI-enabled systems can prescreen normal slides, allowing pathologists and cytotechnologists to focus on reviewing and interpreting abnormal slides (Liu et al., 2021). 

This can help reduce the workload of the existing workforce and potentially increase the number of slides that can be screened in a given time.

2. Enhanced access to screening services: 

Point-of-care digital diagnostic systems, such as portable slide scanners and cloud platforms for deep learning analysis, can be deployed in remote areas with limited access to pathology services (Liu et al., 2021). This can help bridge the gap in cervical cancer screening between urban and rural areas in Kenya.

3. Cost-effective screening solutions: A

I-enabled technology can potentially offer cost-effective screening solutions, reducing the need for expensive equipment and infrastructure. This can make cervical cancer screening more affordable and accessible to a larger portion of the Kenyan population.

But..does it work?

Short answer – Yes. Many studies have been conducted, including in Kenya^3,  have demonstrated the viability of this technology. We believe going forward we’ll see the application of digital cervical cancer screening across public and private healthcare spaces.

Conclusion: A Time to Transition to Digital Cytology.

As Kenya continues to address the challenges of cervical cancer screening, AI-enabled cytology slide image analysis offers a promising solution to improve diagnostic accuracy, increase screening efficiency, and enhance access to screening services. 

By embracing this technology, Kenya can take significant strides towards reducing the burden of cervical cancer and improving women’s health outcomes across the country.

If you liked this article, please like and share (and comment as well)

Also, follow me on X and check out more of my content on my blog.

References:

1. https://gco.iarc.fr/today/data/factsheets/populations/404-kenya-fact-sheets.pdf
2. Liu, Y., Li, X., Li, Y., Zhang, Y., & Zhang, X. (2021). Artificial intelligence in cervical cancer screening: Progress and prospects. Journal of Cancer Research and Therapeutics, 17(3), 527-532. doi: 10.4103/jcrt.JCRT_1677_20
3. https://pubmed.ncbi.nlm.nih.gov/33729503/