Google KI diagnosed with lung cancer more successful than radiologists
How reliable is the Artificial intelligence for the diagnosis of cancer?
Google has now introduced a new System for artificial intelligence (AI), which is lung cancer can already be seen in early Tests, and its accuracy even for the diagnosis by radiologists exceeds.
In the current research work of the Google AI artificial intelligence was developed, which can be used for the diagnosis of lung cancer. The diagnosis of KI surpasses in its accuracy even human radiologist. The results of the investigation were published in the English scientific journal “Nature Medicine”.
AI was more reliable than human radiologists
The technology giant Google trained the algorithm on a 42,000-CT Scans of patients who had participated in a clinical study conducted by the National Institutes of Health. The diagnosis of cancer exceeded lied, the diagnostic accuracy of six participating radio. The AI was able to demonstrate diseases, five per cent more cancer, they also scored in the evaluation of individual Scans to eleven percent with fewer false positive diagnoses, which means that although there is a suspicion of cancer of the inventory, the knot, however is actually harmless was.
Improved precision is a major progress
The new technology of Google is able to improve the precision of the Screening enormously, according to the researchers. Even skeptics of lung cancer Screening said to the results that Google’s performance in the reduction of misdiagnosis, significant progress was. The new AI could prevent that in the future people will be negatively affected in their lives by a faulty diagnosis and wrong treatment.
The AI must be subjected to further testing
So-called Artificial intelligence (AI) is already considered for a long time as a way to improve the Screening of diseases, and to enable it to locate a number of malignant tumors with much greater accuracy. The Google System, however, must be previously subjected to even more stringent Tests before it can finally be used in the medical practice. Since the study was limited to already-treated patients, is not yet known whether the System would also result in new patients to similar good results.
Accuracy was 94.4 percent
As several other so-called Deep-Learning Algorithms, which have already been tested for use in the medical field, was trained by KI in this study, based on Scans of previous lung cancer Screening. Then the artificial intelligence has been commissioned to assess over 6,700 cancer screening. It should be noted that accuracy can be achieved in the diagnosis of cancer. The presence of cancer in these cases, the Doctors is already known. In the Test, the KI achieved an impressive accuracy of 94.4 percent, according to the authors of the study.
Collaboration between KI and Doctors to optimize the treatment
The statements of the algorithm were then compared with the diagnosis of six human radiologists, to determine how well AI and human cancer on the basis of the Scans were able to diagnose that you had never seen it before. If Screening records additional information in the Form of tomography Scans were available, it was the performance of the AI comparable with human experts. In cases where no additional tomography data were available, was able to diagnoses the AI of your human colleagues with an eleven percent positive failure, and five percent less false negative diagnoses clearly exceed. This may sound for human Doctors and physicians in the first Moment after a bad message, but it is not to replace specialists by the AI. In the diagnosis and treatment of cancer, Doctors will continue to play an important role. By combining the experience of a human doctor with the mind of a Deep-Learning algorithm, however, it is likely that fewer mistakes in the diagnosis will be made, which would lead to an overall better quality of life for all Concerned. (as)
