Professional Documents
Culture Documents
Outline of Topics and Discussions 20230117
Outline of Topics and Discussions 20230117
By Tomoyuki Kakutani
Class started by lecture around 30 minutes, after that we discussed several questions regarding
Every year, about 367,000 new cases of cancer are diagnosed in the UK. Since the
1990s, the incidence has risen by 12–16% and is still rising. A patient's cancer
management depends more and more on medical imaging. The amount of imaging data
available to clinicians is rising as the incidence of cancer diagnosis rises. Because more
radiotherapy (RT) planning, the volume of data will keep growing. The quantitative
mining of imaging further expands the data's accessibility. Radiation oncology may
The development of autonomous systems that mimic human functions is the focus of
the branch of computer science known as artificial intelligence (AI). The process of
resolving real-world issues by creating statistical models from a dataset obtained in the
past is known as machine learning, a subfield of artificial intelligence that is the subject
of this article. These statistical models generate predictions by "learning" from collected
data. Deep learning is a subset of machine learning that uses neurons in neural networks
AI developments have increased efforts to automate jobs that are typically performed by
human observers. Automation has the potential to improve patient care by lowering
costs, boosting productivity, and minimizing errors. Planning for conformal brain
radiation (RT), for instance, can take two to four hours. Planning radiologists can
perform the contouring of Organs at Risk (OAR), whereas Advanced Practitioner and
Consultant radiologists have a specific set of skills for performing the contouring of
Gross Tumor Volume (GTV). While OAR contouring is not as time-consuming as GTV
To lessen this clinical strain and enable the redistribution of important healthcare
resources, administrative and routine clinical tasks like OAR contouring can be
used to improve healthcare outcomes and make more time available for tasks that are
most important to patients and people. As an illustration, AI has been used to partition
OAR and to prioritize individuals who are at high risk of getting colon cancer.
In radiation oncology, artificial intelligence (AI) has considerable promise for toxicity
prediction, automated RT planning and optimization, clinical trial patient selection, and
clinical burden reduction. To educate and train the radiation workforce in data
provenance, curation, and integration as well as the ethics of AI development and the
interpretation of AI technologies, resources are, nevertheless, needed. For instance, by
earlier in their careers, the appropriate degree of knowledge could be attained. The
can aid them in discovering novel applications for AI. By using these technologies,
clinicians may be able to identify patterns in local, regional, and global populations that
are not always obvious to the naked eye. However, patients should be included in the
development and deployment of these technologies, ensuring that the "do no harm"
clinical application of these technologies, even if AI still has a long way to go in terms
lack of prospective validation trials. Utilizing hype curves, Netherton et al.75 project
that 20% of clinics will be using AI within the next three years. To do this, however,
will call for more open and consistent reporting of AI model performance, vendor
integration of commercial technologies with existing NHS IT systems, and increased
AI could use the growing amount of data that clinical practitioners have access to. The
of sync.