Machine Learning Algorithm Predicts Short-Term Cancer Mortality in Outpatients

A study published in Biophysics Reviews explored how nanotechnology and computer learning could help develop new methods of screening for human papillomavirus (HPV), the cause of most cases of cervical cancer.

The current standard for HPV detection is the Pap smear, which detects around 80% of cases of developing cervical cancer but requires repeated screenings, trained clinicians, and high-quality laboratories.

“The Pap smear has done wonders in terms of reducing mortality from a cancer that is very treatable when caught early and almost invariably fatal when it is caught late,” said co-author Cesar Castro, MD, oncologist at Massachusetts General Hospital and associate professor at Harvard Medical School, via press release. “Part of its imperfection is that there is subjectivity to it. The trained eye is the limiting step in the process. The untrained eye, or relatively untrained eye, can miss cancers.”

Some of the technologies that show promise include DNA and mRNA testing, which show a 39% improvement in sensitivity compared to Pap testing. Existing tests highlighted in this review include the Cobas HPV test, CareHPV test, Xpert HPV test, Aptima HPV Assay, and the OncoE6 Cervical Test, as well as some promising next generation molecular testing technologies. The review also explored novel multiplexed fluorescence screening platform that detect HPV antibodies in patient plasma, enzyme-assisted nanocomplexes that allow for visual and modular detection of HPV nucleic acids, artificial intelligence monitoring, and other promising technologies.

With these new technologies comes the hope that, with increased sensitivity and less barriers to screening, cervical cancer diagnosis can improve in low- and middle-income nations.

“Of the approximately 311,000 annual deaths from cervical cancer, more than 85% of these occur in low- and middle-income countries,” wrote the authors.

“Similar to COVID-19 testing, we have great technology in places like the United States that does not work well enough in other countries,” said author Hyungsoon Im, PhD, biomedical engineer at Mass General and assistant professor at Harvard Medical School. “This is why there is great motivation to find next-generation, affordable technology to address this problem.”

Credit: Original article published here.