The author is responsible for the MSF Innovation project in Malawi
Cervical cancer is the second deadliest cancer in low- and middle-income countries. In Malawi, East Africa, more than 4,000 women get this type of cancer every year (37% of new cancers affecting women). With 2,905 deaths related to this cancer in 2020, the country also had the second highest death rate.
Such a high mortality from a disease which, in high-income countries, is easily preventable and generally less fatal, is linked to limited access to prevention and screening and to diagnoses that are not always reliable. Clara Nordon, MSF project manager, has just returned from Malawi where the humanitarian medical organization has been engaged in the fight against this cancer since 2018 with the aim of reducing the incidence and mortality among women in the districts of Blantyre and Chiradzulu .
A new protocol
The project started three years ago now. The idea was to work on a new protocol to improve cervical cancer screening using artificial intelligence, to detect cancer before it reaches an advanced stage and consequently treat it more easily. Because early detection allows you to start treatment immediately in a health center relatively painlessly and very quickly, in just a few minutes.
We are therefore working on the combined action of a highly accurate, widely accessible PCR test and improved visual assessment through the use of artificial intelligence.
There is a lack of access to necessary care
Living with cervical cancer in Malawi means not having access to the necessary treatments but also suffering caused by the disease sometimes combined with the stigma, because it is the entire family unit that is affected in a country where almost 10% of adults are HIV positive.
Unfortunately, access to the HPV vaccine remains limited, even though it is one of the few vaccines that can prevent this type of cancer. Furthermore, the current screening method, based on a visual assessment of the cervix, is unreliable. While it is inexpensive and has the advantage that it can be done immediately in a health center, it leaves too much room for interpretation and thus for error. For this reason, many women return home after screening, risking returning a few years later with advanced stage cancer that can no longer be treated.
The quality of the screening
MSF, like the other actors involved in the prevention and treatment of this cancer, has focused on improving the quality of screening. At the beginning of 2020, we became aware of a study published by the Journal of the National Cancer Institute (NCi) which highlighted the potential of artificial intelligence for better identification of precancerous lesions.
We contacted the authors, including Mark Schiffmanan expert in molecular epidemiology who has been studying HPV for over 35 years within the NCI and we discussed it with some MSF experts.
As a first step, we agreed that everyone rework the algorithm and then compare the results. To do this, we decided to use the database of images collected through MSF screenings and create a medical annotation tool. We have partnered with KTH (Royal Institute of Technology in Sweden) and have been working on this preliminary AI tool for more than a year.
A new screening in the world
When we made it, NCI informed us that their algorithm was already complete with results that were conclusive enough to start the validation study. This is also part of the innovation! Work for months and not necessarily use the result obtained. The good news was that we would be able to move forward faster and potentially be useful in the field quicker.
We therefore focused all of our efforts on partnering and collaborating on the large-scale study (the PAVE study) initiated by NCI, involving 100,000 women worldwide to evaluate a new approach to cervical cancer screening.
With the help of MSF teams in Malawi we aim to include 10,000 women in the country, a particularly interesting cohort in this study, because given the high prevalence of HIV positivity, we could potentially have findings that will also help us improve the fight to this other disease.
Find out who has high risk HPV
This study is fundamentally based on the introduction of two major innovations. A systematic screening using a PCR test, which will triage women and identify high-risk HPV carriers (with a high risk of developing cancer), thanks to an innovative, faster and less expensive test technology for the Ministry of Health; a better visual evaluation thanks to the use of artificial intelligence.
The study will start in May and run for 12 months to provide data to the PAVE study. The results of the PAVE study on 100,000 women will be known by the end of 2024 and if the results are conclusive, the NCI will recommend that this new protocol be widely promoted and used in all low- and middle-income countries with significant impact globally .