Multi-level Fusion of the Imaging Data as the Basis for Individually Tailored Model of the Disease of the Brain/Spinal Cord.
Today every person can be tested by various types of imaging. Different imaging approaches allow for getting multi-level data about the brain/spinal cord from the novel computing imaging radiocomplexes, such as computed tomography (CT) of the brain and spinal cord, multi-slice spiral CT of brain and spinal cord, magnetic resonant imaging, positron emission tomography, magnetoencephalography. No matter where you pass your examination, you will receive the data about your brain/spinal cord on a contemporary data medium, such as discs or memory sticks. But how can these data be matched, how can we consider all the data and take everything into account? Can these data be integrated and used for the comprehensive analysis?
It is supposed that such data are analyzed by the clinician of the appropriate field (neurologist, psychiatrist, neurosurgeon) who should be knowledgeable in the method in the interpretation of the results. However, most of the clinicians do not have good expertise to accurately interpret the received multi-modal imaging data because they were obtained from imaging equipment of various manufacturers and are almost incompatible. Moreover, the imaging data of the different levels of the multi-functional state of the brain require extremely high expertise from the radiologists in every contemporary subdivision of radiology (CT, MRI, PET, MEG) which is hardly achievable. The clinician, such as a neurologist or a neurosurgeon, is not able to distinguish and to evaluate all information represented at the images. Hence, they make their evaluation by guts and have to rely on the conclusions of the radiologists. At a certain extent their opinion is the most valid and reasonable as they have the largest experience in the interpreting of the data of MRI, CT and PET. But they do not have the whole picture; they do not see the patient and cannot interpret the data of all tests. Moreover, they receive a large patient flow and they have no opportunity to correlate the data of one case. And they do not need it. These things are done by the researchers for their grants and protocols. But the researchers are away from the clinical practice and their approach is mechanistic. Neither they are able to integrate all the data received from various equipment in various medical institutions into one system.
This mission should belong to an attending doctor and its result conditions the diagnosis, the therapeutic strategy and prognosis. The attending doctor examines the results of all tests and imaging and makes up his own pint of view on the case. Is it objective? Most probably, it is rather superficial and short-term. Unfortunately, the visual analysis of the images of the brain and spinal cord is quite subjective. Human eye can hardly distinguish 256 hues of gray in which the images of CT, MRI, spiral CT and PET are represented. Even the eyes of the well trained experts distinguish only rough damages of the tissue and will miss delicate defects. The attempts to color the PET or MRI images by software is low effective and very far from reality, as it is still not clear what to look for and why. The received information about the pathology has no real meaning for a clinician. The size and localization of the site of injury in brain or spinal cord do not matter much for working out the therapeutic strategy. The data of neuroimaging are necessary for “correct” diagnosis when we determine the character of the process and possible etiopathogenesis of the site of injury (ischemic, hemorrhagic, demyelinated, inflammatory or other), as well as follow up.
However, to date we can use the fantastic personal data of neuroimaging to develop the novel strategies of bioengineering of the brain and spinal cord and to treat not only the patient but the disease itself at the level of tissues and cells. For that it is necessary to work out a 3D model of the specific disease of the brain. In other words, all available data should be correlated and fit into a format that is clear to a clinician. Hence, we have a task of integration of multi-level and multi-modal information about the brain of a specific person. And the solution is found in the available software of the imaging equipment. For example, the opportunity to correlate multi-level data of the inductivity of spins of the hydrogen atoms of water molecules in the neural tissue of the brain and spinal cord detected by MRI with the data of density of the cells of neural tissue as detected by the CT and with the data of glucose saturation of the cells of the PET. The human eye is hardly able to distinguish the variations in such a large scope of the data. Consequently, a large part of the available information is lost. But the use of the software and the technology of multi-level fusion of the data of MRI, CT and PET allows us seeing what was left without attention. As it was in the case with one of our patients who stayed in coma for 11 months and then developed the unresponsive wakefulness syndrome (apallic syndrome) after severe injury of the brain caused by the road traffic accident. We had to verify the main sites of injury that require restoration through bioengineering. Together with the CyberKnife experts we developed the technology of the multilevel fusion of the data of the imaging. The figures show the stages of the technology.
Figure 1. 3 Tesla MRI of the brain.
Figure 2. MRI-tractography of the conducting pathways of the brain
Figure 3. CT of the brain
Figure 4. PET of the brain.
Figure 5. Fusion of the data of the CT, MRI, MRI-tractography and PET.
Figure 6. Morpho-functional zones of the brain restoration.
Obviously, use of computed fusion of different methods of imaging permits finding morpho-functional zones for bioengineering of the brain and to use novel reconstructive techniques of radiobioengineering.