Surgical treatment of spinal deformities associated with neurological deficits using 3D modeling technologies

Full Text

BACKGROUND

Progressive spinal cord compression due to spinal deformities of various etiologies often leads to the development of neurological deficits. Patients with kyphoscoliosis associated with spinal cord compression are at high risk of disability and loss of working capacity, which underscores the social importance of this issue [1, 2].

Modern computer-based spatial modeling technologies and the creation of patient-specific implants using 3D printing offer new approaches to selecting surgical methods and techniques for managing patients with severe spinal deformities. The application of these technologies must be justified not only by preclinical safety and efficacy studies but also by clinical outcomes [3].

One of the most promising areas of additive technology development in vertebrology is personalized implants, which enable the management of extremely complex cases in individuals with spinal deformities, particularly in the context of congenital anomalies, spinal tumors, and post-traumatic defects. Full-size 3D spine models provide a more comprehensive assessment of the deformity and offer enhanced spatial understanding and orientation [4].

The use of patient-specific implants and full-size 3D models can improve outcomes and, in many cases, make surgical treatment possible for patients with deformities in spinal transitional zones [3]. This article presents clinical cases demonstrating the outcomes of treating patients with spinal deformities complicated with neurological deficits and explores the potential of 3D modeling in spinal surgery.

CASE DESCRIPTION

Clinical Case 1

Patient P., aged 15 years, was admitted to the Vertebrology Department at the N.N. Priorov National Medical Research Center of Traumatology and Orthopedics with complaints of a lack of independent ambulation and progressive weakness in the lower limbs, predominantly on the right. According to the medical history, the patient had been followed-up by an orthopedic specialist since birth for a congenital anomaly of the cervicothoracic spine. Neurological symptoms began 4 months prior to admission, with rapidly worsening lower limb weakness one month later, leading to complete loss of independent ambulation.

Neurologic examination on admission revealed that the patient was unable to stand, walk, or rise from bed unaided and could only sit with support. Sensation was reduced in the lower limbs. A kyphotic deformity of the cervicothoracic spine was observed from the posterior view at objective examination. There were no signs of cerebral or meningeal involvement. Tendon and periosteal reflexes were brisk and symmetrical in all extremities. Muscle strength of the upper and lower extremities was found to be as follows: finger extensors, 4 points; abductor pollicis longus, 4 points; iliopsoas, 2 points; quadriceps femoris, 4 points; right foot extensors, 0 points; right foot flexors, 3 points; right ankle flexors, 2 points; right gluteus medius and minimus, 3 points; right gluteus maximus, 2 points; mild paresis on the left side, 3 points. Abnormal reflexes were present, including Jacobsohn-Lask reflex, Rossolim sign (including its Venderovich modification), and Babinski sign with ankle clonus. The patient exhibited urinary and fecal incontinence. Neurological syndrome included cervicothoracic myelopathy, upper mixed distal paraparesis, and lower spastic paraparesis with muscle strength up to 3 points.

Diagnostic imaging included computed tomography (CT), magnetic resonance imaging (MRI), and X-ray of the cervicothoracic spine in anteroposterior and lateral views (Fig. 1).

 

Fig. 1. Computed tomography (a), magnetic resonance imaging (b), and radiography (c) of the cervicothoracic spine. CT and MRI revealed spinal stenosis at C4 to Th4 with MRI signs of myelopathy. Based on the spine CT and X-ray findings, the pattern of congenital spinal anomaly is as follows: a posterior-lateral wedge-shaped accessory hemivertebra at C7, localized angular kyphosis at the C6–C7 level, with a bony block involving the vertebral bodies and posterior elements of Th1–Th3, and a posterior-lateral wedge-shaped hemivertebra at Th4–Th5.

 

Based on clinical and imaging data (see Fig. 1), the diagnosis was congenital deformity of the cervicothoracic spine; severe congenital cervicothoracic kyphoscoliosis; spinal stenosis with anterior spinal cord compression at the C7–Th1 level; thoracic myelopathy; upper mixed distal paraparesis; lower spastic paraparesis; pelvic organ dysfunction (Frankel grade C, Ashworth 3, ASIA: motor 82, light touch 86, pinprick 84).

Due to the complexity of the deformity and worsening neurological status, staged surgical treatment was planned. Stage 1 involved placement of a halo fixator and gradual halo-pelvic traction. During the traction, partial neurological improvement was noted: reduced spasticity (Ashworth 3 to 2) and return of toe flexion. Stage 2 included spinal cord decompression and deformity fixation: laminectomy at C7 to Th2, posterior instrumentation at C2 to Th8, and posterior spinal fusion using autologous bone grafts (Fig. 2).

 

Fig. 2. Postural X-rays following the second surgical stage. Cervical and thoracic rods are connected with a dual-axis domino connector.

 

Postoperative recovery was uneventful, with wound healing by primary intention; sutures were removed on postoperative day 12. Within several days after the second stage of surgery, a positive trend in neurologic status was observed, manifested by a marked increase in lower limb muscle strength. By week 3, the patient was able to get out of bed independently with some caregiver assistance and ambulate within the room using a walker. By week 4, neurologic status improved to Frankel grade D.

To achieve anterior spinal cord decompression and complete 360° spinal fusion, the third stage of surgical treatment was performed one month later. Based on CT data, a patient-specific plastic 3D model of the spine and the sternocostal joints was created at a 1:1 scale. Using the model, the optimal surgical approach was determined, along with the extent of vertebral resection, fixation length, and the ideal working trajectory (“angle of attack”) for screw placement into the vertebral bodies, considering the sternal position. A customized anterior plate was designed (Fig. 3, 4).

 

Fig. 3. Design stage of a patient-specific model and anterior fixation plate.

 

Fig. 4. Patient-specific 3D-printed spine model (a) and custom plate for anterior spinal fixation (b).

 

Surgery involved anterior decompression and spinal cord revision. Th1–Th2 vertebral bodies were resected, and interbody fusion was achieved with a mesh cage and autografts. A patient-specific anterior plate was placed spanning C4–Th3 (Fig. 5).

 

Fig. 5. CT scan following anterior placement of the patient-specific plate. The green zone marks the projection of the modified Smith–Robinson approach, avoiding manubrial resection.

 

Postoperatively, the wound healed by primary intention, and sutures were removed on day 10. At 3-month follow-up, the patient demonstrated marked neurologic improvement, walking independently without support. Muscle strength had recovered to 4/5 points distally and 5/5 points proximally in the lower limbs, corresponding to Frankel grade E.

Throughout staged treatment, continuous improvement was observed (baseline: Frankel C; Ashworth 3; ASIA motor 82, light touch 86, pinprick 84), culminating in full neurologic recovery (Frankel E; Ashworth 1; ASIA motor 100, light touch 112, pinprick 112). The patient was followed for 5 years. Clinical examination revealed no signs of neurologic deterioration. Imaging confirms stable position of the metal implant and evidence of bone ingrowth.

Clinical Case 2

Patient S., aged 15, was admitted to the 14th Vertebrology Department of the N.N. Priorov National Medical Research Center of Traumatology and Orthopedics with complaints of a lack of independent ambulation and weight support by the lower limbs, as well as spinal deformity. The patient had been under orthopedic follow-up since the age of 2 for thoracic scoliosis (Fig. 6). At the age of 13, gait instability with episodes of falling first appeared, followed over the next year by foot deformity and progressive lower limb weakness. The patient was referred for genetic testing. Clinical exome sequencing of 6640 genes was performed, revealing a variant in the nucleotide sequence of exon 11 of the SH3TC2 gene in a heterozygous state. A novel heterozygous variant in exon 15 of the SH3TC2 gene was identified, and the SH3TC2 mutations found were deemed pathogenic. Such mutations are considered associated with Charcot–Marie–Tooth disease type 4C. The patient also had a history of obstructive sleep apnea with desaturation episodes down to SpO₂ = 88%. Electroneuromyography showed nerve conduction velocities below 22 m/s in the lower limbs.

 

Fig. 6. Postural spine X-ray at the age of 2 years.

 

Highlights of status localis: The patient is unable to get out of bed or walk. Muscle tone is normal in the upper extremities and increased in the lower extremities, with spastic characteristics and a tendency for leg “scissoring.” Equinus foot contractures and shortening of the Achilles tendons. Tendon and periosteal reflexes in the upper extremities are brisk and symmetrical; patellar and Achilles reflexes are absent. Lower mixed paraparesis: proximally up to 2 points, progressing to plegia in the feet. Hypesthesia is noted from the knee joint level, with impairment of superficial and deep sensation (including two-dimensional spatial perception). The patient has no control over pelvic organ functions, with incontinence.

Postural spinal X-rays (anteroposterior and lateral views) showed severe left-sided thoracic kyphoscoliosis with a Cobb angle of 90° and kyphotic angle of 115°, both with the apex at Th6–Th7 (Fig. 7).

 

Fig. 7. Postural spine X-ray at the time of hospital admission.

 

Helical CT, CT myelography, and MRI of the thoracolumbar spine were performed. MRI revealed that the spinal canal was filled with adipose tissue at the Th5–Th7 level, the spinal cord was displaced to the right, the cerebrospinal fluid space was not visualized at this level, and no signs of myelopathy were detected (Fig. 8).

 

Fig. 8. MRI of the thoracic spine at the apex of the deformity.

 

Based on the clinical presentation, genetic findings, neurological status, and imaging, the diagnosis was hereditary neuropathy, namely, Charcot–Marie–Tooth disease type 4C; severe left thoracic kyphoscoliosis of neuromuscular origin; spinal cord compression at Th6–Th9; severe mixed lower paraparesis; equinus cavovarus foot deformity of neuromuscular origin.

Given the complexity of the deformity, spinal cord compression, and progressive neurological deterioration, a patient-specific anatomical model of the spine and spinal cord was created based on CT myelography data (Fig. 9). The model was produced with separate printing of the bony structures and spinal cord, and designed to be disassembled, with components held together by magnets. This allowed for direct three-dimensional tactile assessment, enabling identification of the zone of maximal compressive myelopathy and radiculopathy at the Th6–Th9 level caused by the pedicles, costotransverse joints, and rib heads on the concave side of the deformity.

 

Fig. 9. An individualized anatomical model splittable into two parts along the sagittal plane using neodymium magnets, enabling visualization of the spinal cord (colored red) and the bony structures of the spine (colored yellow).

 

Considering the clinical presentation and diagnostic findings, surgical treatment with intraoperative neuromonitoring was planned. Posterior instrumentation at Th2 to L2 was carried out without further correction (in situ). Th6–Th9 laminectomy was performed, exposing dural sac without pulsation and spinal nerves at the same levels. The dural sac and spinal nerves appeared flattened, atrophic, and bluish in color. Visual inspection revealed compression of the dural sac at the Th6–Th9 level on the concave side of the deformity, caused by the vertebral pedicles. A high-speed burr was used to partially resect the pedicles of the Th6–Th9 vertebrae on the right side. Following the resection, the dural sac expanded. Tension of the dural sac was noted due to the Th8 nerve root on the left (convex side of the deformity). Rhizotomy of the Th8 nerve root on the convex side was performed to mobilize the spinal cord at this level. After resection of the Th6–Th9 pedicles and the Th8 rhizotomy, the spinal cord shifted toward the concave side, assuming a new position. In the newly formed space around the dural sac, its tight adherence to the costotransverse joints and rib heads at Th7–Th8 was observed, causing secondary compression. Therefore, the costotransverse joints and rib heads at Th7–Th8 were resected using a high-speed burr. The dura mater then expanded, and no signs of compression were observed. Satisfactory pulsation was observed 10 minutes after dural release. Posterior spinal fusion with autologous bone grafts was performed. The wound was irrigated, closed, and a drain was placed (Fig. 10).

 

Fig. 10. CT scan of the spine on postoperative day 3. Axial view at the deformity apex (Th8 level). The arrow indicates the decompression zone: resected pedicle of the Th8 vertebra, costotransverse joint, and rib head.

 

The drain was removed on postoperative day 2. On day 5, the patient developed a fever of 38.5 °C. Ultrasound showed fluid collection at the wound superior and inferior poles, which was treated with serial puncture, and 15–20 mL serous fluid was drained. By day 20, ultrasound revealed no residual hematoma, and temperature normalized.

In the neurological status during the early postoperative period, a reduction in spasticity was noted, with no lower limb scissoring. Proximal muscle strength was 2 points (the patient can actively flex the left leg at the hip and knee joints; on the right, only synergistic movements are possible). Active dorsiflexion and plantarflexion of the left foot emerged, limited by contracture. Babinski sign emerged (not seen preoperatively). Conduction hypoesthesia regressed from the L3 level, and right-sided hyperesthesia appeared from L2 downward. No focal signs from Th8 nerve root were noted.

Follow-up postural X-ray and CT confirmed proper positioning of the instrumentation, with no evidence of instability or malposition of the supporting elements. The patient was discharged on day 23 for outpatient follow-up.

At the 6-month follow-up, muscle strength in the proximal segments of the lower limbs was as follows: iliopsoas, 5 points on the left and 4 points on the right; quadriceps femoris, 5 points bilaterally. Ankle dorsiflexion and toe extension appeared, each rated at 3 points; ankle plantarflexion was 5 points on the left and 4 points on the right. The patient could walk in orthopedic footwear, though the tendency for right leg scissoring and foot pronation persisted.

The patient has been under observation for 3.5 years and continues to undergo periodic rehabilitation therapy in a specialized inpatient facility. As a result of staged surgical treatment, neurological status showed positive dynamics from the time of surgery (Frankel grade C; Ashworth 3; ASIA: motor 84, light touch 82, pinprick 84) to recovery of neurologic function (Frankel grade D; Ashworth 1; ASIA: motor 92, light touch 108, pinprick 110). The patient can independently stand, rise from bed, and walk with assistance while wearing orthopedic footwear. No neurological deterioration has been observed. Imaging confirms stable position of the metal implant and evidence of bone ingrowth.

DISCUSSION

Advances in surgical techniques and instrumentation in vertebrology have led to evolving strategies for treating spinal deformities complicated by neurologic impairment. Naturally progressing scoliotic and kyphotic spinal deformities disrupt global spinal balance and biomechanics [5]. One of the key challenges in managing complex deformities is progression of spinal cord dysfunction, manifested clinically by neurological deficits. Despite advances in surgical techniques for decompressing the spinal cord in such cases, the optimal surgical approach for deformities with a high risk of neurological compromise remains controversial [6].

Before the development and routine use of spinal fixation and correction systems, the literature described posterior (laminectomy) and posterolateral (costotransversectomy) decompression techniques, along with various modifications proposed by different authors [7, 8]. However, comparative studies showed that these methods failed to significantly improve neurological outcomes—more than half of patients experienced progression of neurologic deficits postoperatively [9, 10]. A technique known as transposition of the spinal cord has been proposed to expand the decompression zone in kyphoscoliotic deformities [11–14]. According to studies by international authors, this technique significantly improved surgical outcomes, resulting in recovery of neurological deficits following spinal cord decompression in patients with kyphoscoliotic deformities [15–17]. The evolution of segmental spinal fixation systems, including sublaminar hooks and pedicle screws, has enabled effective spinal cord decompression through a variety of bone resection techniques and their combinations [18].

In patients with severe congenital spinal deformities, anterior and middle column instability is often present, necessitating anterior stabilization [19]. When complete decompression of the spinal cord cannot be achieved via posterior or posterolateral access, anterior stabilization is combined with anterior decompression [20–23]. This approach was used in Clinical Case 1, where posterior deformity stabilization and anterior vertebral body resection at the deformity apex resulted in spinal cord decompression and stabilization. The procedure led to neurological improvement; however, due to the posterior vertebral column defect, there remained a risk of instrumentation instability and progressive deformity, potentially compromising neurological function.

In Clinical Case 1, it is noteworthy that halo-pelvic traction was used, despite being described in the literature as contraindicated in patients with spinal deformities complicated with neurological deficits. However, the use of halo-pelvic traction in the patient from Clinical Case 1 resulted in partial regression of the neurological deficit [8, 24]. During halo-pelvic traction, stabilization of the deformity occurs to a degree sufficient to reduce venous congestion in the spinal cord vasculature and improve spinal cord perfusion at the apex of the deformity. Restoration of perfusion and oxygenation in spinal cord tissue leads to the clinically observed regression of neurological deficits [10, 25].

Due to the complex geometry of the deformity and abnormal spinal anatomy, the use of standard anterior plates was not feasible in this case. This was attributed to the inability of standard implants to conform closely to the vertebral bodies, as well as to mismatches in screw hole locations and plate dimensions. An effective approach to spinal stabilization involved the manufacture of a personalized anterior plate with near-complete congruence to the contact surfaces. The design process allowed precise planning of screw placement and trajectories, ensuring avoidance of interference between surgical instruments and the sternum or clavicles, thereby eliminating the need for sternotomy. To plan all surgical steps and simulate implant placement, a full-size patient-specific 3D model of the spine was created. As a result of surgical treatment, complete regression of the neurological deficit was achieved along with spinal stabilization.

In the natural course of severe neurogenic spinal deformities, there is also a high risk of spinal cord compression, necessitating spinal stabilization and decompression of the spinal cord and nerve roots [26]. This scenario is exemplified by Clinical Case 2 described above. Charcot–Marie–Tooth disease type 4C often manifests within the first decade of life, frequently with scoliosis. The disease course varies—some patients remain ambulatory into their 50s, while others require a wheelchair by adolescence [27].

Despite comprehensive diagnostic evaluation, complete visualization of the compression zone was not achievable with either CT or MRI, making preoperative planning of the resection area inherently prone to error. Consequently, insufficient decompression would likely result in suboptimal clinical outcomes. A full-size 3D model of the spine and spinal cord revealed that the pedicles of the curve area contributed most to the spinal cord compression. Despite the similarity of the kyphoscoliotic deformities to those in Clinical Case 1, the optimal surgical approach in Clinical Case 2 was posterolateral decompression, which included not only laminectomy but also resection of the pedicles in the curve area and adjacent costotransverse joints. Resection of these bony structures created a new bed for the spinal cord. In this process, the spinal cord was not displaced but mobilized.

Using the modular 3D model with separately printed bone and neural components, the optimal length of decompression could be determined based on narrowing of the spinal cord model. The extent of pedicle screw fixation and the absence of anterior column defects limited the surgery to the dorsal approach alone. As in Case 1, thorough preoperative planning of the surgical technique and decompression zone led to a favorable clinical outcome—regression of neurological deficit and prevention of patient disability.

Over the past three decades, the published sources have established key surgical approaches for treating patients with kyphotic and scoliotic spinal deformities complicated by neurologic deficits. These can be categorized into the following surgery models:

• Stabilizing (in situ fixation and fixation with correction without direct decompression of the spinal cord and nerve roots);
• Decompressive-stabilizing (procedures involving instrumented fixation of the deformity with resection of spinal bony structures causing compression of the spinal meninges, performed via ventral or dorsal approaches);
• Decompressive-corrective (single- and multilevel vertebral column resection [VCR], pedicle subtraction osteotomy [PSO], and their various modifications).

Each approach has its strengths and limitations affecting spinal deformity correction and post-operative neurologic outcomes [1, 28–35].

The clinical cases presented demonstrate the efficacy of the decompressive-stabilizing model. An alternative surgical strategy is the decompressive-corrective approach, specifically VCR. Although VCR enables substantial spinal deformity correction with simultaneous spinal cord decompression, it presents significant limitations [36]. In Clinical Case 1, the deformity was caused by multiple congenital anomalies of the cervicothoracic spine (see Fig. 1), making it technically unfeasible to define the scope of and perform a three-column vertebrotomy. In Clinical Case 2, the deformity involved a long curve with a multilevel compression zone, which would have required VCR at three levels to achieve both correction and decompression (see Fig. 9).

The primary goal of surgical treatment for spinal deformities complicated by neurologic deficits is to preserve and/or restore spinal cord function. According to published data, VCR can achieve deformity correction of up to 54% in the coronal plane and 47% in the sagittal plane. However, neurologic deterioration occurs in 2–22% of cases, offsetting potential VCR benefits [36–39].

Therefore, avoiding three-column vertebrotomy—through the use of 360° spinal stabilization with minimal deformity correction and a limited decompression zone under halo-pelvic traction in Clinical Case 1, and dorsal-only spinal fixation combined with multilevel posterolateral spinal cord decompression in Clinical Case 2—allowed for regression of neurologic deficits and restoration of patient function with minimal risk of postoperative neurologic complications.

CONCLUSION

Patient-specific anatomical models play a crucial role in determining the surgical approach, resection area, and pedicle screw placement during planning. Their use enhances the precision of preoperative planning and improves accuracy in defining the extent of bone resection necessary for adequate decompression of the spinal cord and nerve roots. Three-dimensional spatial visualization of the spinal cord and nerve roots reduces the risk of injury during bone resection and implant placement, allowing for personalized decompression. Ultimately, this strategy supports the primary goal of surgery—neurologic deficit regression with a favorable functional outcome.

The advantages of 3D-printed models in surgical planning stem from their realistic anatomical representation of pathological zones, enabling better understanding of complex anatomy, screw trajectories, decompression sites, and the necessary decompression extent.

In severe cervicothoracic deformities, 360° spinal fusion is recommended. While standard implants are often optimal for posterior stabilization, their use for anterior stabilization can pose significant technical challenges or be entirely unfeasible. Custom anterior plates enable effective fixation in even the most complex deformities.

The main objective in treating spinal deformities complicated with neurologic deficits is preserving or restoring neurologic function. Less aggressive surgical models that minimize spinal cord injury risk, ensure safe decompression of the spinal cord and nerve roots, and incorporate additive technologies during preoperative planning and surgery allow restoring spinal cord function and preventing patient disability.

ADDITIONAL INFO

Author contribution. A.A. Kuleshov — data collection and analysis, writing the text of the article; A.G. Nazarenko — data collection and analysis, writing the text of the article; M.S. Vetrile — data collection and analysis, writing the text of the article; S.N. Makarov — data collection and analysis, writing the text of the article; I.M. Militsa — data collection and analysis, writing the text of the article; I.N. Lisyansky — data collection and analysis, writing the text of the article. All authors have approved the final version before publication and have also agreed to be responsible for all aspects of the work, ensuring that issues relating to the accuracy and integrity of any part of it are properly addressed and resolved.

Funding sources. No funding.

Disclosure of interests. The authors declare that they have no competing interests.

Statement of originality. In creating this work, fragments of my own text, published earlier, were used ([https://doi.org/10.17816/vto629012], distributed under the CC-BY 4.0 license).

Consent for publication. The authors received written informed voluntary consent from the patient to publish personal data in a scientific journal, including its electronic version (date of signing April 27, 2022, December 12, 2022). The scope of published data was agreed with the patient.

Data availability statement. Access to the data obtained in this study is closed due to confidentiality (the presence in the database of information on the basis of which the study participants can be identified and the lack of their consent to the dissemination of this information).

Provenance and peer-review. This paper was submitted to the journal on an initiative basis and reviewed according to the usual procedure. Two external reviewers, a member of the editorial board and the scientific editor of the publication participated in the review.

About the authors

Alexander A. Kuleshov

Priorov National Medical Research Center of Traumatology and Orthopedics

Email: cito-spine@mail.ru
ORCID iD: 0000-0002-9526-8274
SPIN-code: 7052-0220

MD, Dr. Sci. (Medicine)

Russian Federation, 10 Priorova str., 127299 Moscow

Anton G. Nazarenko

Priorov National Medical Research Center of Traumatology and Orthopedics

Email: nazarenkoag@cito-priorov.ru
ORCID iD: 0000-0003-1314-2887
SPIN-code: 1402-5186

MD, Dr. Sci. (Medicine), professor of RAS

Russian Federation, 10 Priorova str., 127299 Moscow

Marchel S. Vetrile

Priorov National Medical Research Center of Traumatology and Orthopedics

Email: vetrilams@cito-priorov.ru
ORCID iD: 0000-0001-6689-5220
SPIN-code: 9690-5117

MD, Cand. Sci. (Medicine)

Russian Federation, 10 Priorova str., 127299 Moscow

Sergey N. Makarov

Priorov National Medical Research Center of Traumatology and Orthopedics

Email: moscow.makarov@gmail.com
ORCID iD: 0000-0003-0406-1997
SPIN-code: 2767-2429

MD, Cand. Sci. (Medicine)

Russian Federation, 10 Priorova str., 127299 Moscow

Igor M. Militsa

Priorov National Medical Research Center of Traumatology and Orthopedics

Author for correspondence.
Email: igor.milica@mail.ru
ORCID iD: 0009-0005-9832-316X
SPIN-code: 4015-8113

MD

Russian Federation, 10 Priorova str., 127299 Moscow

Igor N. Lisyansky

Priorov National Medical Research Center of Traumatology and Orthopedics

Email: lisigornik@list.ru
ORCID iD: 0000-0002-2479-4381
SPIN-code: 9845-1251

MD, Cand. Sci. (Medicine)

Russian Federation, 10 Priorova str., 127299 Moscow

References

Supplementary files