Article Text
PDF
Abstract
Spinal muscular atrophy (SMA) is an autosomal recessive hereditary neurodegenerative disease causing progressive muscle atrophy, weakness and kyphoscoliosis. Nusinersen is a therapeutic agent for SMA that should be administered intrathecally. However, due to severe kyphoscoliosis, lumbar puncture can be challenging. Here, we present our experience of intrathecal administration of nusinersen in an SMA patient with severe kyphoscoliosis using a life-size three-dimensional printing (3D) skeletal model created with 3D printer. With this strategy, we were able to rapidly and safely perform the lumbar puncture.
- diagnostic techniques and procedures
- injections, spinal
- multimodal imaging
Statistics from Altmetric.com
Introduction
Spinal muscular atrophy (SMA) is an autosomal recessive hereditary neurodegenerative disease causing progressive muscle atrophy and weakness, and severe kyphoscoliosis.1 The survival of motor neuron (SMN) protein is thought to suppress the apoptosis of neuronal cells, and mutation or deletion of the SMN1 gene, which encodes the SMN protein, is thought to cause degeneration and loss of motor neurons in the anterior horn of the spinal cord. Estimated incidence of SMA is 1 in 10 000 live births.1 SMA is classified into types 1–4 based on its onset and severity, of which SMA type 1 has the earliest onset and presents the most severe symptoms.
Nusinersen, a therapeutic agent for SMA, was approved by the Federal Drug Administration in 2016 and subsequently by the Ministry of Health, Labor and Welfare of Japan in 2017.2 3 It is considered that nusinersen increases the expression of SMN protein and leads to suppression of apoptosis of neurons and consequently improves muscle strength.3 Nusinersen must be administered intrathecally to ensure direct delivery into the central nervous system. However, due to severe kyphoscoliosis in patients with SMA, the lumbar puncture procedure can be technically difficult.2 The challenges persist even if images, including three-dimensional (3D)-CT, are evaluated beforehand and the procedure is performed under fluoroscopic guidance using C-arm.4 Despite the procedural difficulty, repeated intrathecal administration of nusinersen is indicated every 4 or 6 months, which may be associated with patient discomfort due to prolonged duration of the procedure, radiation exposure and risk of complications.
Based on this clinical challenge, we present a case report summarizing our innovative approach to facilitate an intrathecal injection in a patient with SMA type 2 and severe kyphoscoliosis. With this technique, we aimed to improve the safety, reduce the procedure time and radiation exposure, and ultimately, increase the patients’ satisfaction.
Case report
On September 10 2018, a 38-year-old woman with SMA type 2 was referred for an intrathecal administration of nusinersen. The patient consented to reporting this case and signed the appropriate informed consent forms, which were reviewed by the ethics board of our institution. The patient’s height and body weight were 98 cm and 17 kg, respectively. The patient had severe muscle atrophy due to SMA type 2, and was bedridden; however, she did not need respiratory support. 3D-CT scans demonstrated severe kyphoscoliosis, with additionally imaging revealing the kidneys and intestine were located just below the atrophied muscles of the trunk (figure 1). Lumbar puncture appeared challenging and was accompanied by an anatomical concern for visceral injury.
(A) Three-dimensional (3D) CT scan showing morbidly severe kyphoscoliosis. Blue lines represent metal markers. (B) 3D-CT scan after removing the bones other than the vertebrae showing hypoplastic spinous processes and twisted vertebrae, thereby complicating the identification of the posterior midline of the trunk required for the landmark method of lumbar puncture. (C) CT scan at the fourth lumbar vertebra level showing atrophied muscles of the trunk and shift of visceral locations, which implies substantial risk of injury to the internal organs during lumbar puncture.
The first intrathecal administration of nusinersen was performed on September 13 2018 and was subsequently repeated four times until June 27 2019. All lumbar punctures were performed under fluoroscopic guidance using C-arm and the average time spent in the fluoroscopy room was 55 min. The presence of severe kyphoscoliosis and altered anatomy, as well as the vertebral deformities (eg, hypoplastic spinous processes, distorted pedicles and osteoporotic change), made it difficult to identify the midline and level of the vertebra.
Prior to the patient’s fifth injection on January 8 2020, we planned to create a life-size 3DP skeletal model of the patient and use it as a guide for lumbar puncture. We extracted Digital Imaging and Communications in Medicine (DICOM) data from a 3-month-old 3D-CT scan and used it for 3DP modeling. We used online 3DP service of Ricoh Japan to create a life-size skeletal model (40 ×30 ×18 cm) made of nylon resin (figure 2). Radiolucency of the 3DP model (made of nylon resin) was similar to that of the living body (figure 3). The cost for creation of the 3DP model was approximately US$2000, which was paid from our department operating expense.
(A) Due to hypoplastic spinous processes and twisted vertebrae, the identification of midline of the lumbar vertebrae is difficult in the X-ray. (B) Lumbar interlaminar space can be identified on the three-dimensional (3D) CT scan. Blue lines represent metal markers. (C) Simulation by tilting the 3D printing model by approximately 20° forward clearly revealing the interlaminar space between the fourth and fifth lumbar vertebrae.
Preprocedural simulation using the three-dimensional printing (3DP) model. (A) Determination of the body tilt and angle of the C-arm to obtain the appropriate fluoroscopic image of the interlaminar space between the fourth and fifth lumbar vertebrae. (B) Review of the previous fluoroscopic image of successful lumbar puncture. The hypoplastic spinous processes and distorted pedicles complicated the identification of the midline of the lumbar vertebrae despite fluoroscopic guidance in the previous attempt. (C) Appropriate image for lumbar puncture obtained using the 3DP model. Radiolucency of the 3DP model (made of nylon resin) is similar to that of the living body and practical for simulation. The lumbar puncture procedure was not challenging once the optimal fluoroscopic image was obtained.
Simulation of the procedure was conducted on January 7 2020 using the 3DP model. We placed the 3DP model on the fluoroscopic table and determined the body tilt and angle of the C-arm to obtain the appropriate fluoroscopic image of the interlaminar space between the fourth and fifth lumbar vertebrae by reviewing previous images (figure 3). Further, we confirmed that there were no internal organs in the path of the needle by reviewing the CT images (figure 1). On January 8 2020, we placed the 3DP model on the fluoroscopic table and re-evaluated the body tilt and angle of the C-arm for approximately 5 min. Following this, the patient was brought to the fluoroscopy room and positioned beside the 3DP model on the table. The C-arm was moved in parallel, and we could immediately obtain the appropriate fluoroscopic image for lumbar puncture. Over a 5 min procedure time, there was only one needle pass and nusinersen was successfully injected intrathecally. The time spent in the fluoroscopy room was 20 min, and the patient as well as the proceduralists expressed satisfaction regarding ease of injection.
The main difference between the current and previous procedures was the labor involved to ensure optimal trajectory of the fluoroscopic image using the C-arm. The lumbar puncture procedure had been technically straight forward once the appropriate fluoroscopic image was obtained. In the previous doses without the 3DP model, the time taken to ensure optimal trajectory of the fluoroscopic image was longer despite careful pre-procedural evaluation of images.
The technique of obtaining the optimal fluoroscopic image using the 3DP model was as follows: Identification of the level and midline of the lumbar vertebrae was difficult from the normal anteroposterior view (figure 4B1) due to severe kyphoscoliosis and unusual anatomy of the lumbar-sacral angle. Metal markers were attached on the third, fourth and fifth spinous processes of the lumbar vertebrae to obtain the optimal fluoroscopic image (figure 4A). The normal anteroposterior fluoroscopic image with metal markers revealed rotated vertebrae and misaligned midline (figure 4B2). Figure 4C1 appeared a normal anteroposterior view; however, the rotated vertebrae were visible in figure 4C2. With the markers on the third, fourth and fifth spinous processes of the lumbar vertebrae, the midline and interlaminar space between the fourth and fifth lumbar vertebrae could be easily identified (figure 4D2).
Simulation of fluoroscopic image using the three-dimensional printing model with and without metal markers on the spinous processes. (A) Metal markers attached on the third, fourth and fifth spinous processes of the lumbar vertebrae. (B1 and B2) Anteroposterior images. Difficulty in identification of the level and midline of the lumbar vertebrae without metal markers due to severe kyphoscoliosis and unusual anatomy of the lumbar-sacral angle. (C1 and C2) Oblique images (10° oblique angulation). The fluoroscopic image without metal markers appears a normal anteroposterior view. However, the fluoroscopic image with metal markers reveals rotated vertebrae. (D1 and D2) Oblique images (20° oblique angulation). Easy identification of the midline and interlaminar space between the fourth and fifth lumbar vertebrae with metal markers. In contrast, difficulty in identifying the interlaminar space without metal markers due to hypoplastic spinous processes and distorted pedicles.
In the previous procedures without the 3DP model, multiple needle passes or changes in the proceduralists were required due to misidentification of the midline and the level of lumbar vertebrae. However, the appropriate body tilt and angle of the C-arm could be determined in a shorter time by preprocedural simulation using the 3DP model compared with direct determination from the actual patient. Following the simulation, the only procedure required to obtain the optimal fluoroscopic image was to position the patient beside the 3DP model, and move the C-arm in parallel.
Discussion
We have demonstrated that a 3DP skeletal model can facilitate an intrathecal injection in a patient with severe anatomical pathology. A 3DP technology, which can freely create 3D models, has advanced in recent years. In the medical field, 3DP models are created using data obtained from imaging studies. The 3DP models enable physicians to understand anatomical features of individual conditions (eg, thoracic aortic dissection aneurysm, heart valve disease, facial fracture, tumors with surrounding vascular structures, and kyphoscoliosis) in a more thorough manner.5–9 Furthermore, the creation of such models from patient-specific imaging has a significant role in the pre-surgical planning and simulation of optimal surgical procedures. In particular, popularity of 3DP models has been increasing in spine surgeries.10–12 Vertebrae are surrounded by and located near major vessels, including the aorta, vertebral arteries, spinal cord and nerve roots, as well as the viscera, including the lungs and gastrointestinal tract. Studies have reported injury of spinal cord, thoracic aorta, esophagus or spleen from penetration of screws, as rare but critical intraoperative complications during spine surgeries.13 14 To prevent such complications, the use of patient-specific life-size 3DP models has been proposed, which would enable virtual surgical simulation to rehearse the drilling trajectory.15
The potential and unproven advantages of using the 3DP model for lumbar puncture in SMA patients with severe kyphoscoliosis are as follows:
Safety: Predetermination of the body tilt and angle of the C-arm might enable lumbar puncture along the needle route as planned, thereby reducing the risk of injury to the spinal cord and internal organs.
Reduction of procedure time and radiation exposure: Predetermination of the angle of the C-arm can reduce duration of the procedure, radiation exposure of both patients and medical staffs, patient discomfort and increase patients’ satisfaction.
Feasibility and repeatability: Simulation of the procedure in advance might enable to maintain the same puncture trajectory as in previous attempts, irrespective of the interval from the previous procedure or change in the proceduralist.
Saving costs: Nusinersen is extremely expensive. Simulation of the procedure using a 3DP model could reduce expenses by preventing failure of lumbar puncture.
Lumbar puncture under CT guidance is thought to be safer than under fluoroscopic guidance for patients with severe kyphoscoliosis.2 However, CT-guided lumbar puncture may not be practical in some clinical settings due to limited accessibility to the CT scanners. Further, radiation exposure of the patient in the CT-guided approach would be higher compared with that in fluoroscopic guidance.16 Considering that patients should receive administration every 6 months, reducing the number of CT-guided procedures might contribute to decreasing the radiation exposure of the patients.
However, application of the 3DP model to patients in the growth period warrants caution, considering the changes in the anatomy in relatively short periods. The 3DP model may be favorable, based on the safety and cost involved, in patients with non-progressive or fixed kyphoscoliosis.
Considering these potential advantages, we are excited about future studies and exploration of this technology. Whether or not the routine use of personalized simulation for interventional procedures is cost effective and practical is yet to be determined.
References
Footnotes
Contributors HA conceived the original idea, performed practical procedure and prepare manuscript. RI performed practical procedure. RT performed practical procedure. KA revised manuscript. KI supervised handling of 3D-CT data. MK handled 3D-CT data and created the skeletal model. JH handled 3D-CT data and created the skeletal model. MS acquired funding, revised manuscript and approved final manuscript.
Funding This study was funded by JSPS KAKENHI grant number 19H03749.
Competing interests None declared.
Patient consent for publication Obtained.
Provenance and peer review Not commissioned; externally peer reviewed.