World No Tobacco Day Is Approaching. How Can Technology Help Wang Yuan and Others Quit Smoking?
In 2018, the number of deaths from lung cancer reached 1.76 million. However, lung cancer has one of the highest survival rates among all cancers, and if detected early, the survival rate can be greatly improved. With the intervention of AI models, early prediction and screening of lung cancer can increase the survival rate of lung cancer by 40%, exceeding the expert level. In addition, AI is also working hard to help people quit smoking and stay away from tobacco.
On May 31 this year,The 32nd World No Tobacco Day.
The World Health Organization (WHO) originally chose May 31 as World No Tobacco Day because the next day is International Children's Day.We hope that the next generation will be free from the harm of tobacco.
The theme of this year's World No Tobacco Day is "Tobacco and Lung Health"
However, our world is still a long way from being smoke-free. Even in public places with "No Smoking" signs, there are still many people smoking.
Not long ago,Singer Wang YuanHe also became a hot topic for smoking in a restaurant. CCTV News reported on it, and even foreign media BBC reported on it.
Regarding this incident, some people directly called it the collapse of the "Three Good Youth" personality, and some people expressed that they could not accept public figures smoking in public. Of course, there were also many voices advising him that smoking was harmful to the body.I hope to quit smoking soon.
This incident also aroused people's attention to issues such as smoking in public places and the difficulty of smoking control.
Seeing the true face of tobacco
Why do people smoke? Smoking can eliminate fatigue, increase attention and regulate people's emotions, which are the main reasons for people to smoke. Once you have the habit of smoking, you will become dependent on tobacco and then addicted to it, which is difficult to quit. Under the pleasure brought by smoking, the harm of tobacco has long been forgotten by smokers.
But look at the following data and examples, and you will see the true face of tobacco:
Tobacco leaves contain the poisonous substance nicotine, also known as nicotine. One gram of nicotine can kill 300 rabbits or 500 mice.If 50 mg of nicotine is injected into a person, it will kill;
In most countries around the world 90% of lung cancer is caused by smoking;
Jiangsu Cancer Hospital found that among 1,000 lung cancer patients admitted 80% Is a long-term smoker.Female smokers are at a higher risk, 1.9 times that of male smokers;
Smoking not only causes lung cancer, but alsoOral cancer, laryngeal cancer, breast cancer, etc.;Mothers who smoke put the health of their next generation at risk, and their babies are more likely to have congenital heart disease than normal people 2 times...
Although the country has implemented a smoking ban since 2011 and "smoking is harmful to health" has become common knowledge, there are still 350 million smokers,Number of people who die each year from smoking-related diseasesAbout 1 million people,The global death tollOne fifth.
Lung cancer cells hidden too deep
Among all the harms caused by smoking,Lung cancer is undoubtedly one of the most serious hazards caused by smoking.
A global cancer statistics shows thatLung cancer is the malignant tumor with the highest incidence (11.6%) and mortality (18.4%) in the world.Especially in China, lung cancer has long been the number one killer of cancer.
(Blue: number of cases; Red: number of deaths)
Why are the incidence and mortality rates of lung cancer so high in China? One important reason is:Most lung cancer patients in my country are already in the middle to late stages of the disease when it is discovered.
If it can be detected and treated early, the survival rate of lung cancer patients can be greatly improved. The survival rate of advanced lung cancer is only 15.2%.But if detected early, the five-year survival rate can be as high as 80%.
But why is it that most lung cancer patients are only diagnosed in the middle or late stages?
On the one hand, most people do not take the initiative to do early screening for lung cancer, and wait until they have symptoms before doing so, which means it is too late when they find out.
On the other hand, becauseEarly stage small tumors are hard to see.The accuracy of early screening is not high enough, which is prone to misdiagnosis and missed diagnosis. More than 80% lung cancer cases cannot be detected early. A large number of lung cancers are not discovered until the late stage, which also brings high clinical and economic costs.
In order to improve the current situation, many technology companies are actively launching AI solutions cover all aspects of smoking cessation and lung cancer prediction and diagnosis.Google introduced its research results on using AI to predict lung cancer at this year's I/O conference.
AI outperforms experts in early diagnosis of lung cancer
Since late 2017, Google has made progress in predictive modeling for lung cancer using 3D modeling and NIH datasets from partners such as Northwestern University (located in Illinois, USA).
Recently, Google published a research report in Nature Medicine, saying that theyTrain AI to predict lung cancer, lower the screening threshold for lung cancer, and improve the success rate of early diagnosis.Moreover, the performance of the AI model exceeded that of six professional radiologists.
Typically, radiologists screen for lung cancer by looking at theHundreds of 2D images,It is difficult to detect small tumors. This method is pointed out to have a high false positive or false negative diagnosis rate.
Google builtEnd-to-end analysis approach,Using only the input CT images, they were able to predict lung cancer risk.Convolutional Neural Networks (CNNs)to train and thus obtain better results in image recognition.
With the help of AI models, Google can make predictions about lung cancer malignanciesOverall prediction(View in 3D) and can also identify subtle malignant tissue (lung nodules) that are usually difficult to find. In addition, when making judgments, AI will also combine past scan results, so it can better analyze the growth of suspicious tissue.
To test this approach, Google had the AI look at 45,856 copieschest CT scans (which included cancer cases), and then compared their diagnoses with those of six board-certified radiologists.
In the early stages of the study,AI has been able to detect 51 more cases of TP3T than doctors, while reducing the likelihood of false positive results by more than 11%.The method achieved an AUC of 94.4%(Model evaluation metrics in machine learning).
Of course, Google's model will still need a lot of clinical experiments and tests before it can be put into practical use. But Google has great confidence in the future prospects of this program.
A researcher from the team said at this year's Google I/O conference:Their artificial intelligence technology can detect lung cancer a year earlier than doctors, increasing the patient's chance of survival by 40%.And according to their survey, among eligible patients in the United States, only those with 2% to 4% have the opportunity to receive CT examinations. In the future, with the introduction of AI, this convenient tool may become a powerful tool for promoting early screening of lung cancer, and the threshold in this regard should be further lowered.
Quitting smoking with algorithms and neural networks
For lung cancer, we not only emphasize early screening, but also pay attention toImmune-based precancer prevention.
When it comes to lung cancer prevention, quitting smoking always comes first. Tobacco is the biggest preventable cause of cancer worldwide.Without tobacco, lung cancer would be a rare cancer.
However, it is very difficult for smokers to quit smoking, and there are countless examples of smokers who have tried to quit smoking but failed repeatedly.
Since you can’t rely on yourself to quit smoking, why not use the power of AI?
A Toronto-based kiwi.ai The company developed an app called Cue,It uses an algorithm to automatically detect when the user smokes.If the average time users spend smoking continues to decrease, there will be economic benefits.
Here’s how it works. The team created a model for tracking, built and licensedMotion recognition software that uses an embedded neural network to track movements with a high level of accuracy at low power consumption to detect when a user is smoking.
The team then makes the software available to OEM device manufacturers like IBM, robotics companies, and healthcare product providers, among others, so smokers can use the devices to help them quit.
During the test, the team found that after using the software, testers would wait longer than usual before smoking because of the time recommended by the software. Within 3 months, the test users' smoking time was reduced by 50%.
Since the app also encourages users to wait a few more minutes to earn points, the points will not only give them the opportunity to smoke next time, but can also be used to redeem reward coupons for Starbucks or Uber, thus achieving a win-win situation and allowing users to quit smoking happily.
Say no to tobacco and lung cancer
Today, AI is trying its best to fight lung cancer, but many smokers still turn a blind eye to the warnings that "smoking is harmful to health."
According to statistics from the National Cancer Center, the number of lung cancer cases and deaths in my country has ranked first among malignant tumors for 10 consecutive years.Every year, there are approximately 787,000 new cases of lung cancer and 631,000 deaths from lung cancer.
Faced with these shocking figures, AI has worked hard to make progress in the prevention and prediction of lung cancer, and I believe it will continue to improve and achieve better results in the future.
We ourselves should pay more attention to this number one cancer killer, understand the harm of tobacco, take preventive measures, and stop "Wang Yuans" from smoking, so that lung cancer has no chance to take advantage.
