Published: 28th January 2019
Could AI and bots replace trained doctors for better diagnosis? Only time will tell
Artificial intelligence (AI) may be poised to wipe out cervical cancer after a study showed that computer algorithms can detect precancerous lesions far better than trained experts
As any doctor can tell you, the most crucial step toward healing is having the right diagnosis. If the disease is precisely identified, a good resolution is far more likely. Conversely, a bad diagnosis usually means a bad outcome, no matter how skilled the physician — Andrew Weil (b.1942), an American celebrity doctor
There was a popular astrologer (in the pre-ultrasonography days, which is illegal in India for sex determination) to whom men used to throng with their pregnant wives to predict the sex of the foetus at birth. Knowing that about 50% chances of the impending child being either male or female, he would predict what the parents wanted to hear – male. In about 50% of the cases, his predictions were wrong. Some of the wronged parents went back to the astrologer to complain about the prediction becoming wrong and asking for a refund of the consulting fees. Then, the astrologer would go to a niche in the wall and bring a small chit of paper which had ‘girl’ written on it and say: “I had correctly predicted, but didn’t want to disappoint you.”
A similar charade seems to be happening in respect to diagnosis of patients by doctors. Against this background, it is interesting and relevant to note a news report filed by AFP from Miami, titled ‘AI beats doctors at detecting early signs of cervical cancer’ (11/1/19). Excerpts.
Artificial intelligence (AI) may be poised to wipe out cervical cancer after a study showed that computer algorithms can detect precancerous lesions far better than trained experts or conventional screening tests. By the look of things, AI seems poised to wipe out a whole lot more as well — jobs are on the line and quite a few careers appear to be on the verge of being wiped out of existence in the decades to come.
Despite major advances in screening and vaccination — which can prevent the spread of human papillomavirus, which causes most cases of cervical cancer — those gains have mainly benefitted women in rich nations. “Cervical cancer is now a disease of poverty, of low resources,” said senior author Mark Schiffman, a doctor who has been searching for a cure to cervical cancer for 35 years. “We are trying to find ways that are extremely cheap, extremely easy but very accurate, so that we can attack cervical cancer by vaccine and also a bit later, through a simple technique that is cell-phone-based,” he said.
The AI technique, called automated visual evaluation, found precancerous cells with 91 per cent accuracy, according to a report published in the Journal of the National Cancer Institute. In comparison, a human expert review found 69 per cent of precancerous cells, while conventional lab tests like Pap smears found 71 per cent.
Among women in the age bracket of 25 to 49, who face the highest risk of cervical cancer, the AI algorithm was even more accurate, finding 97.7 per cent of precancerous cells. It performed much better than humans looking at those same pictures. These are numbers that doctors cannot even imagine achieving unless they have some superhuman skill or possess great equipment.
On the same day as this news was released, there was another report of a doctor removing a good kidney thinking that it was a tumour.
So, will AI delete doctors from the lucrative business of medical diagnosis?