Medical science stands at the beginning of a digital revolution, that will fundamentally change its shape in the coming years and decades. Technologies that seemed merely science fiction are already a reality today. While terms like Medicine 2.0 and Big Data are still discussed among experts, some countries have already made connected, digital solutions, as well as tele- and internet medicine, part of their regular health care provision. Still, the advancement of artificial intelligence might soon overshadow all other future developments.
In the last decades, hardly any other discipline has been as strongly characterized by technical progress as radiology. There is an incredible jump between the development of the first x-ray films and the introduction of computer-tomography (CT) and Magnetic resonance imaging (MRI). The improvements in medical imaging have made diagnostics not possible, but also easier in many areas. At the same time, the amount of data recorded increased, together with the development of tomographers tremendously. Already today, data analysis is impossible without the proper software. At the same time the work of the radiologist has changed from pattern recognition to the proverbial search of ‘a needle in a haystack’.
在过去的几十年里，几乎没有任何其他学科像放射学一样具有技术进步的强烈特征。在第一批x射线胶片的发展和计算机断层扫描( CT )和磁共振成像( MRI )的引入之间有一个不可思议的飞跃。医学成像的改进使得诊断成为可能，但在许多领域也变得更加容易。与此同时，随着断层摄影技术的发展，记录的数据量增加了。如今，没有合适的软件，数据分析已经是不可能的了。与此同时，放射科医生的工作已经从模式识别转变为谚语中的“大海捞针”。
Computers are already able totake over parts of the radiological diagnostics and are not only faster but often also more accurate than humans. Moreover, they do not tire and as a result and the quality of the analysis remains consistent. While computers can perform an immense amount of complicated computing operations in a fraction of a second, they are still inferior to humans in other tasks, such as the differentiation of faces and objects.With the advancement of artificial intelligence, computers have stopped following rigidly programmed rules. The so-called "deep learning" is considered a breakthrough within the field of machine learning. This technology makes use of artificial neural networks and imitates the operating mode of the human brain. A special feature of this technology is that the quality of the analytical results increases with the amount of processed data. For all intents and purposes, deep learning is comparable to a person who learns and gains in experience.
Nowadays, deep learning is also used in medicine. The US-American Startup "Enlitic" uses this technology to evaluate x-ray images, for instance to recognize fractures as well as lung tumors independently.IBM's computer system "Watson" is based on this technology as well and attracted a lot of media coverage in 2011 when "Watson" won against humans in the American quiz show "Jeopardy!". However, the system is universally applicable and also capable of evaluating CT images and recognizing anomalies such as dissections and pulmonary embolisms.
In the coming years, artificial intelligence will play a growing role in many areas of professional and private life, and especially in medicine. Radiology is one of the first disciplines in which human and artificial intelligence will closely work together. While this technology might not replace radiologists in the foreseeable future, it certainly will fundamentally alter their work. Instead of diagnosing dozens of images every day, a radiologist might instead be able to access already processed data and place it within its clinical context more and more. Even if computers should be superior to humans in analytical speed and quality, ultimately humans will still decide whether a new physical development is of any pathological significance or relevance.