Cervical cancer screening comes to your mobile phone

June 10, 2019 | Disease

Melanie senior

Effective screening programs have helped reduce cervical cancer-linked mortality by over 50% in the last forty years. But thousands of women still die each year from the disease, including many in developing countries with little or no access to screening and early diagnosis.

New technologies and artificial intelligence-based software are now making cervical cancer screening and detection easier and cheaper, and opening it up to non-specialists – including, potentially, the patients themselves.

Scientists at Duke University in Durham, North Carolina, have developed a pocket-sized colposcope – a small camera – which can be used to generate clear images of the cervix without the use of a speculum. The device, which is far smaller and cheaper than a conventional colposcope, can be inserted as easily as a tampon, by the woman herself or a non-specialist health provider. It has FDA clearance and has been tested in eight countries and over 1000 women, across the US, Peru and some African nations.

If it gains traction, the device would enable community clinics and relatively unskilled providers to detect and potentially treat cervical cancer much earlier – at the screening stage, when the traditional pap smear happens. (Colposcopy is the more involved procedure that follows a positive smear test.)

The Duke researchers are also working on algorithms that could help non-specialists detect abnormalities in the images generated by the device, and even on ways to treat early stage, pre-invasive cancer. These include an ethanol-based gel that could target and destroy cancerous lesions, scalpel-free.

Israel-based Mobile ODT is pursuing a similar goal. Its ‘smart’ colposcope, which generates advanced images of the cervix on a mobile phone, is already used in 29 countries including the US. And it has recently been enhanced with an artificial-intelligence based algorithm trained to detect abnormalities in the images. A study among 212 patients in South Korea suggests that the algorithm may be better at detecting pre-cancerous or cancerous lesions than trained humans using pap cytology. And it can do so much faster – in one minute, while the patient is still present, rather than in days.

That study was only a pilot. But the potential of artificial intelligence in medical imaging more generally is already proven, making it likely that cervical screening and early detection will continue to become more widely available.

More treatments are on the way for metastatic or recurrent cervical cancer, which remains the biggest unmet need. But there are bumps in the road: FDA in January 2019 placed a partial hold on Advaxis’ Phase III study of therapeutic cancer vaccine Axal, following a trial re-design. Two earlier trials, including combination studies, were forcibly halted due to patient deaths. Axal is a strain of listeria bacteria programmed to generate cancer-targeting T cells.

Regeneron/Sanofi’s checkpoint inhibitor Libtayo (cemiplimab), meanwhile, has shown early promise both as monotherapy and in combination with radiotherapy in early clinical trials of patients with advanced or recurrent cancer; a larger Phase III is underway. Checkpoint inhibitors like Libtayo have thus far not proven very effective in first-line cervical cancer, where platinum-based therapies and Avastin dominate.