FOCUS
Personalized medicine: what it is and how to obtain it through diagnostic imaging
One of the objectives of medical science is the so-called personalized or ‘precise’ medicine, a very topical subject, especially in the field of oncology. It aims to enable differentiated therapeutic pathways for each patient based on the biological characteristics of their pathology, as well as aspects of their clinical history, characteristic elements of the person, and the environment in which they live in. As far as the treatment of tumors is concerned, there is already a form of personalized medicine today: together with the traditional therapeutic approach based on parameters such as cell type and presence or absence of lymph nodes and metastases, it is possible to set, for certain types of lesion, some specific and targeted therapies which are defined as ‘molecular targets’. Unfortunately, these therapies are based on the investigation of samples of tumor tissue, i.e. they require the use of invasive tests, and this contributes to the need to test new approaches to personalized medicine which, in addition to the efficacy of the therapy, cause less discomfort to the patients.
Personalized Medicine and Quantitative Approach:
In this context, you may ask what the role of imaging diagnostics is. Thanks to advanced methods of analysis such as radiomics and radiogenomics, common diagnostic tests (such as MRI or CT scan) can become tools for personalized medicine, not only to the benefit of patients, but also of structures and the health system. All this is already possible, but on one condition: that medical imaging is no longer regarded as a set of images to be interpreted visually (qualitative approach), but as a very rich wealth of data on which to perform quantitative analysis (data mining) in order to define the biological characteristics of the tumor which are completely invisible to the human eye and therefore can’t be detected by simple visual observation.
This method is called radiomics and relates, that is, associates, data acquired and processed by images (the so-called radiomic features) with clinical results, such as the response to a certain therapy, which the prognosis relies on. All this represents a real revolution: from a simple routine examination which would continue to be carried out, it is possible not only to define – as it regularly happens – the shape, size, location and margins of the lesion, but also the more suitable targeted therapy, according to the data obtained from the imaging itself. Despite being absolutely central to the process, the data from biomedical imaging can then be enriched with other characteristics (that are quantitative, i.e. measurable) of the patient and his pathology: data from his clinical history, but also information related to his environment, lifestyle, and much more. The enrichment of data, the use of Machine Learning techniques and – in a broad sense – of Artificial Intelligence allows the creation of predictive models that act as decision support for setting the best possible therapeutic procedure, that is – the one with the highest probability of a positive response.
Radiogenomics: one step ahead
Closely related to radiomics is radiogenomics, which is a sort of evolution, the immediate next step. The concept is similar, namely the ability to identify, directly from routine exams, data that are completely invisible to the eye and when processed with data mining techniques, can provide essential information in order to set the best possible therapy. Radiogenomics, in particular, associates the quantitative data obtained through the above techniques with the genetic analysis of tumor lesions, which is extremely useful since it is known that some tumors present genomic alterations and, depending on whether they are present (or not), a personalized therapy may be defined. All this becomes possible without imposing those invasive examinations that are indispensable today and subject the patient to obvious suffering; furthermore, obtaining this information from the quantitative analysis of biomedical imaging allows the survey to be repeated several times during the treatment in order to also verify the changes induced by it. It has immense benefits for the patient.