Look away from Poonam Pandey's contraption: How technology can reduce cervical deaths | Health

Poonam Pandey Cervical cancer may not have killed you, but more than 70,000 Indian women die from it every year.

Mr. Pandey's tactics to draw attention to this disease are abhorrent and disrespectful to say the least, but the reality is that cervical cancer is the second leading cancer among women aged 15-44 in India. It is the cause. Not only is it highly preventable, it is also highly treatable if detected in a timely manner. However, cervical cancer, like other women's health issues, remains a low priority for technological disruption and funding. (Also read: Poonam Pandey's death stunt puts spotlight on cervical cancer)

There are more than 200 strains of human papillomavirus (HPV), and most infections caused by HPV are actually harmless and resolve on their own. However, about 12 strains cause the majority of cancers of the vulva, vagina, cervix, penis, anus, mouth, and throat. The sad reality is that HPV is actually the most prevalent sexually transmitted disease, and nearly everyone will be exposed to and infected with one of her HPV strains at some point in their life.

Although the statistics look scary, the good news is that there is hope. HPV is also easier to immunize and prevent because cervical cancer is caused by a virus, in contrast to other cancers that are caused by problems with internal cell replication. Merck & Co.'s Gardasil vaccine was until recently the only widely available vaccine on the market. However, the cost was prohibitive.

Well, the good news is that the Serum Institute of India has developed an HPV vaccine called Cervavac, which is now available at a tenth of the price. In fact, there is talk in the proposed union budget of making it part of the government's vaccination programme. This is a very welcome move. HPV is part of national immunization programs in many countries, including Bhutan (2010), Thailand (2017), Sri Lanka (2017), and Maldives (2019).

Another great news is that cervical cancer has excellent treatment and recovery outcomes when detected early. However, the good news ends here. Although great progress has been made in efforts to prevent cervical cancer, technology to predict risk, identify and treat HPV continues to lag.

First, the current process for identifying cervical cancer, the Pap smear, is not designed to be dilated. To obtain a Pap smear, a gynecologist scrapes some cells from a woman's cervix and sends it for analysis. Although this method is invasive and most women do not experience any pain, many avoid it due to discomfort or embarrassment.

Second, India has only 70,000 gynecologists for about 360 million women of reproductive age. Pap smears are usually performed at a gynecologist's office or hospital, so they are not available to most people.

Third, Pap smears rely on visual inspection, which is prone to human error and prevents detection in very early stages.

Many young, unmarried women are hesitant to ask their gynecologist about a pap smear because of the social stigma and taboo surrounding it. The World Health Organization is aware of these issues and recommends a more modern technology, her DNA testing for HPV. However, the cost of this test is prohibitive in India.

AI/ML-based predictive models have the potential to break through the barriers of invasiveness, over-reliance on human input, and cost. In addition to reported symptoms and medical history, there are many behavioral, socioeconomic, and demographic factors that influence the prediction of HPV. These include having multiple sex partners, being sexually active from a young age, belonging to a lower socio-economic class, smoking, and having herpes, HIV, or chlamydia. This includes having a medical history.

Symptomatically, irregular bleeding such as brownish discharge, spotting, and bloody discharge with an abnormal odor are factors that enable rapid triage and detection. This is a small sample set from a very extensive list. We've built predictive models that not only diagnose your current condition, but also predict issues like HPV and cervical cancer risk based on some of these variables.

Using AI/ML-based technology, symptoms and medical history can be diagnosed instantly, which can play a major role in not only detecting current problems but also predicting future problems such as cervical cancer. Such a model that uses AI to examine cervical cancer cells can greatly help improve early detection of cervical cancer and avoidance of unnecessary deaths.

Poonam Pandey's method of drawing attention to cervical cancer is untenable. But since we are talking about cervical cancer, let's focus not on the nature of cervical cancer, but on improving vaccine availability and technology for early detection, where AI/ML holds great promise.

(Divya Kamerkar is a biologist and global health researcher at Yale University and CEO of Pinky Promise, an AI/ML-based women's health platform. Dr. Reema Shah is an obstetrician-gynecologist and medical director of Pinky Promise, and Akanksha Vyas He is an engineer and mathematician. Over 15 years of AI/ML work and CTO of Pinky Promise)