|
ИСТИНА |
Войти в систему Регистрация |
ИПМех РАН |
||
Brain structure segmentation and 3D printed individual craniometric rulers for cortex brain lesionsстатья
- Авторы: Okishev Dmitry, Abramyan Arevik, Konovalov Anton, Pilipenko Yury, Okisheva Elena, Eliava Shalva, Usachev Dmitry
- Журнал: Annals of 3D Printed Medicine
- Номер: 8
- Год издания: 2022
- Издательство: Elsevier Masson SAS
- Местоположение издательства: [Issy les Moulineaux], France
- Первая страница: 1
- Последняя страница: 6
- Номер статьи: 100079
- DOI: 10.1016/j.stlm.2022.100079
- Аннотация: Doi : 10.1016/j.stlm.2022.100079 IntroductionNeurosurgery remains the main treatment method for brain tumors, vascular lesions, and congenital anomalies worldwide. Particularly, neurosurgical interventions are required for superficial lesions that may be safely removed. The success of these operations requires a good understanding of the individual (for the patient) cortex anatomy. The high quality of such surgeries is also determined by using minimal but sufficient craniotomy. Frameless neuronavigation is commonly used to provide these two requirements1. Nevertheless, the navigation systems will not provide the necessary accuracy if there is a mistake in the patient's registration and data interpretation2. The preparation process for the navigation system takes some surgical time (≈30 minutes) thus increasing the duration of the intervention3. The high cost of an advanced navigation system also should be noted.Surgical guides and template printing for clinical usage has become simple and available nowadays4. Undoubtedly, 3D printing is most widely used in dentistry 5,6. 3D-printed guide templates are also actively used in orthopedics and traumatology7,8. A number of articles have been published on the use of surgical templates for pedicle screw placement9, 10, 11. Scalp templates are used for stereotaxic punctures of hematomas12,13. We performed a search in the published literature and did not find any articles describing the use of scalp templates for preoperative skin marking. This may be due to apparently complicated technique, but availability of 3D printer in the hospital and the skills of modeling and segmentation make this a quick and easy method.In our hospital, we have been investigating the clinical applicability of the individual anatomy segmentation and patient-specific surgical guides printing for more than two years in our hospital14. In this article (technical note), we propose a high-accuracy and low-cost navigation method for superficial brain lesions.Materials and methodsThe method consists of two elements: cerebral cortex segmentation in the target area (creation of a patient-specific anatomy model) and 3D printing of an individual craniometric ruler (ICR). Segmentation is performed according to a non-contrast T1-weighted MRI sequence using Inobitec Dicom viewer 2.6. An ICR is modeled in Blender 2.93 and printed using a rigid food-safe filament PET-G or PLA.The rulers are modeled on a previously segmented scalp. The model is a mesh that consists of several curved over the scalp cylinders with a diameter up to 20 mm. Usually ICR model consists of a ring where the inner part encircles the required craniotomy projection, and 3-4 rays oriented towards several reference points: e.g. the nasal bridge, the right and left tragus, outer corners of the eyes. After virtual plaicing, the model is cut off using the skin contours. A large contact surface with the patient's head provide unique positioning.We used two FDM printers: Geralkom Prusa i3 Steel Pro (0.6 mm nozzle) and Picaso Designer X PRO (0.5 mm nozzle). Ultimaker Cura 4.62 and Polygon X 2.3.0 were used to prepare g-code. Our usual printing settings: layer height – 0.4 mm, line with – 0.75 mm, wall line count – 2, infill – gyroid 15%, print speed – 60 mm/s, supports: normal zigzag or tree with 55 degrees overhang angle. Usually printing takes from 2 to 4 hours The printed ICR was considered as verified if it fitted the patient's scalp well. The process of ruler creation may be found in the attached video. The method is illustrated by the following example.The study was approved by the local ethics committee of the Federal State Autonomous Institution «N. N. Burdenko National Medical Research Center of Neurosurgery» (Approval No. 012 dated September 28, 2021). The study conforms to the ethical guidelines of the 1975 Declaration of Helsinki.A 33-year-old female was examined after two focal seizures in her face and arm. MRI revealed a small cavernoma in the right precentral gyrus laterally (Figure 1). Due to the obvious relationship between the patient's symptoms and cavernoma, a resection of the lesion was scheduled. The cavernoma was relatively small and located in an eloquent area; therefore, the navigation was used for its safe resection.Preoperative preparationsThe patient's cerebral cortex and the cavernoma were segmented based on MRI. According to well-known anatomical concepts, the frontal, precentral, and postcentral gyri in the model were marked with different colors (Figure 2). The chosen approach involved dissection and encephalotomy of the right central sulcus to minimize the damage of the adjacent structures. The meningeal arteries were also segmented based on a contrast-enhanced sequence. For the most accurate pre-craniotomy skin marking, an ICR was modeled and printed using PET-G filament. Four rays of the ICR were oriented towards four reference points: the nasal bridge, the right and left tragus, and the right outer corner of the right eye (Figure 3).Surgical managementA craniotomy was performed via a linear skin incision according to the markings. The meningeal arteries corresponding to the segmented ones were identified on the surface of the dura mater (Figure 4, A, B). The necessary matches between the exposed cerebral cortex and the segmented cortex model were easily found (Figure 4 C, D). The precentral gyrus and the planned dissection area in the lateral part of the central sulcus were identified. The motor evoked potentials revealed the face representation in this area; however, the previously planned approach was chosen as the safest one. Dissection of the sulcus and encephalotomy at the bottom of the sulcus revealed a relatively superficial cavernoma (Figure 4 E, F). The lesion was removed with no resection of the surrounding hemosiderin deposits. During the postoperative period, no increase in the focal neurological deficit was observed; however, the patient complained of a focal seizure in the left upper lip, therefore, anticonvulsants were prescribed.DiscussionIntraoperative navigation clearly represents substantial assistance for a neurosurgeon1. Frameless neuronavigation can be associated with some issues nonetheless2. First, it is necessary to register the patient's data, i.e. synchronization between uploaded images and the head position should be performed. It takes some time and prolongs operation duration up to 10-30 minutes depending on the user's experience3. Patient registration errors are relatively common. Commonly, many neurosurgeons allow a 0.5-1 cm discrepancy within the coordinates, but this shift would be significant in case of a small lesion search. The position of the anchor points may be displaced during some manipulations, e.g. due to head movements in the fixation device15. It is often necessary to change non-sterile to sterile tools, so there may be a loss of accuracy due to damaged or non-equal instruments16. The high price of high-quality advanced frameless neuronavigation systems and the need for regular specific maintenance are often overlooked. New users can make a mistake in skin marking for some lesion locations. This happens in the case of a non-standard (usually overextended) head position without Gantry Tilt correction17. The user finds and marks the displaced lesion projection on the head surface, thus shifting the whole approach (Figure 5). To avoid this situation, at least two projections should be used, and it is reasonable to use 3D data presentation. Also, it should be understood that the position of the structures could change significantly during the surgical intervention due to cerebrospinal fluid drainage, brain matter removal, and mechanical traction15.The best way of intraoperative navigation is to find permanent anatomical landmarks. For operations of the cerebral cortex, the clue reference point is the individual pattern of sulci and gyri18. Three-dimensional patient-specific reconstruction of the cerebral cortex can be performed quickly using watershed segmentation with a non-contrast 3D T1 sequence (e.g. T1 FSPGR BRAVO). The imaging for the cortex segmentation should be done before contrast enhancement as the meninges become isointense and blur with the brain matter. Post-contrast images may be used for the subsequent vessels segmentation.When planning an approach without frameless neuronavigation, common craniometric points and external landmarks (e.g. the auricle, eyebrows, skin texture irregularities) are used19. In this case, sufficient accuracy for a keyhole craniotomy is possible only when the lesion is located close enough to one of the above-mentioned landmarks. The accuracy decreases with increasing skin mobility in case of a soft tissue landmark. To expose a target area without neuronavigation, craniotomy in the frontal and parietal regions should usually be relatively large (5 cm or more), and the lesion often has to be removed "along the edge of the bone border".To ensure the most accurate marking, we have developed a method for creation of the individual 3D printed craniometric rulers. These models use skin features and the patient's head curvature for the ruler's unique positioning. Our algorithm allows the creation of such models within 3-7 minutes with subsequent printing within 2-4 hours. Both PLA and PET-G filaments are suitable for open chamber printers, have minimal shrinkage and high interlayer cohesion. We prefer PET-G filament as this plastic is less cooling-dependent, thus providing a faster printing. After a successful start of the printing process, no further effort is required as print is completely autonomous. Both the modeling and the printing time are not included in the "operation time", so the surgery duration is reduced. The ability to run print tasks remotely is also useful. After the print, it is necessary to separate the supports from the model without damaging the ruler. It requires experience, but the support may be left intact in some cases.The described method of preparing and printing individual craniometric rulers is cost-effective. When using the highest quality plastic, the maximum expenses are less than $2 for the most complex ruler. Equipment costs have also dropped recently. FDM printer with an acceptable quality of printing can be purchased up to $200; FDM printer with very high-quality components - up to $2000. Such equipment has a long service life and may work for years. FDM printers are easily repairable. We have already performed more than 20 operations using individual craniometric rulers, and the technique was effective in all cases so far.Figure 6Surgical guides are often used in dentistry and traumatology for surgical area marking, precise bone resections, and screw implantation. There are certified medical 3D printers and specific materials for this purpose. In neurosurgery, surgical guides are rarely used. We apply surgical guides for bone resections with the frontal sinus sparing (Figure 7) or for resections followed by the implantation of prefabricated custom implants (Figure 8)14.ConclusionThe proposed method of neuronavigation for the excision of superficial brain lesions is convenient and effective. We believe that this method of resection area determination is faster, cheaper than mapping based on a frameless neuronavigation system. It is for these reasons we consider using this navigation method even in clinics that routinely use frameless neuronavigation. The application of surgical guides is currently limited in the field of neurosurgery, but it has shown great potential.DisclosuresThe authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper.
- Добавил в систему: Коновалов Антон Николаевич
Прикрепленные файлы
| № | Имя | Описание | Имя файла | Размер | Добавлен |
|---|---|---|---|---|---|
| 1. | Полный текст | 1-s2.0-S2666964122000339-main.pdf | 3,5 МБ | 2 сентября 2022 [DOkishev_nsi] | |
| 2. | Preprint | main.pdf | 819,0 КБ | 23 августа 2022 [ANKonovalov] |