AI4CC
Improving cervical cancer screening through the combined action of a PCR test and Artificial Intelligence
In brief
Cervical cancer is a humanitarian emergency, with more than 90% of deaths related to this disease occurring in LMICs. Early screening allows precancerous lesions to be eradicated in a rather simple way. The MSF Foundation, in collaboration with MSF and Epicentre, has joined an international consortium dedicated to improving cervical cancer screening. Led by the NCI in the United States, this consortium is conducting a wide-ranging study on more than 100,000 women to evaluate a new approach to cervical cancer screening: the PAVE (Human Papillomavirus and Automated Visual Evaluation) study.
This new approach combines identifying women infected with an oncogenic strain of Human Papilloma Virus (HPV) - at high risk of developing into cancer - and, among these women with high-risk HPV, using Artificial Intelligence (AI) to help identify those with precancerous lesions that can be treated on the spot
The MSF Foundation and its partners will conduct -upon ethical board validation of MSF and Malawian authorities- this clinical study in Malawi as part of MSF's program for care of women with cervical cancer at the Queen Elizabeth Hospital and the screening program in Blantyre health centers.
Status of the project
News
In detail
Cervical cancer: a humanitarian emergency
95% of cervical cancers are caused by persistent infection with certain so-called high-risk papilloma virus strains found all over the world. This infection will, in certain women, cause precancerous lesions that disappear without any specific treatment or develop into cancer after several years.
The solution to prevent the development of cervical cancer is well known. It involves:
- vaccination of young girls to protect them from infection with these high-risk strains
- regular screening and treatment of precancerous lesions.
There are many reasons for the terrible inequality in LMICs and vulnerable environments for women with cancer (access to prevention is limited, the disease is diagnosed too late and treatment of cancers detected at these late stages is difficult).
Over 4,000 Malawian women fall sick with cervical cancer every year and with 2,905 deaths due to cervical cancer in 2020, the country also had the second highest rate of mortality (https://www.wcrf.org. 2020). Malawi is probably among the countries with the highest number of cervical cancer cases.
In 2018, MSF launched a program to care for cervical cancer patients in the Blantyre district. Activities range from screening and treatment of precancerous and early cancerous lesions to specialized surgery and palliative care for patients with advanced cancer. The teams work at the Queen Elizabeth Central Hospital in Blantyre. They also support several health centers in the district and organize community awareness sessions.
-
The consortium and the PAVE study
In the past few years, research teams have been networking internationally to develop technologies to reduce the cost of screening and improve its performance.
This International Consortium led by the National Cancer Institute in the United States brings together a dozen institutions including the MSF Foundation and MSF to evaluate a new approach to cervical cancer screening.
The exploratory results are very positive and now need to be confirmed by a prospective, multicenter clinical study.
The consortium's study will make it possible to test and validate the combined performance of:
- a test to detect high-risk HPV infection
- an AI algorithm tool to help nurses to detect and characterize pre-cancerous lesions in the cervix during a VIA.
This consortium is a unique example of international collaboration to bring about genuine change in medical practices and create an open-source tool using artificial intelligence. This vast study is conducted on a scale rarely matched both in terms of numbers and representativeness of LMICs’ populations. It will evaluate this new protocol and, if validated will radically change access to care for thousands of women currently forgotten in the fight against this cancer.
The role of the MSF Foundation
The MSF Foundation’s role is to support MSF operation into improving standards of care, to evaluate what technology can contribute, in a fairly "simple" and direct way, to medical practices in vulnerable environments, and to bring back and adapt those to our fields. As such, we were screening for new ways to allow a better detection of women with pre-cancerous lesion. That’s how we got in contact with The NCI after they published a paper that highlighted the potential of artificial intelligence in photo analysis of VIA.
Since them, we are working daily to set up and coordinate a unique project with multiple partners, drawing on the best resources available both internally and externally. This means organizing the integration of MSF Malawi screening project within the study, collaborating to draw specific objectives relevant for MSF Malawi and other LMICs, relying on Epicentre’s expertise on study, Field staff motivation and knowledge on the context, MSF medical expert…
It is in Malawi, at the heart of the MSF program that we will launch the clinical study on nearly 10,000 women.
They will pay particular attention to evaluating, within the sub-population of women living with HIV (a major scourge in Malawi), the performance of the current treatment for precancerous lesions. This analysis will enable it to be confirmed that the proposed treatment can permanently eradicate these lesions and prevent their development into cancer.
Prospects and expected results
If the performance is as expected, this new protocol will considerably improve early detection of cervical cancer by:
- Making available systematic PCR screening of high-risk HPV strains : the most reliable method of identifying women at risk of developing cancer. Today HPV testing is only use marginally because of its cost. The new assay will halve costs and reduce the time required by a factor of 10 by focusing analysis on high-risk strains
- Improving screening for precancerous lesions during cervical examinations:This visual test is currently very widely practiced because it is easy to implement but is poorly accurate and operator-dependent.
- The study will validate the use of a new screening algorithm based on affordable PCR and open source algorithm. The cost efficiency of the method would allow a large scale up and massive impact on the disease burden