Dr. Gregor Schmeiser // Schön Klinik Hamburg Eilbek
AI-based image analysis for cervical myelopathy
Challenge
Measuring imaging endpoints in the treatment of multilevel degenerative cervical myelopathy (mDCM) is a time-consuming and error-prone process.
Solution
In a study comparing two surgical treatments for mDCM, Dr. Gregor Schmeiser used RAYLYTIC as an independent imaging software laboratory to conduct highly precise and rapid analyses of complex radiographic outcomes.
Outcome
AI-assisted analysis of X-rays showed that despite differences in fusion rates, the two compared surgical approaches did not have significant differences in radiographic outcomes, particularly in sagittal balance and range of motion (RoM), with comparable clinical outcomes.
Excerpt
For Dr. Gregor Schmeiser, leading spinal surgeon at Schön Klinik in Hamburg Eilbek, fractions of degrees matter in his treatment of cervical myelopathy, so precise and reliable measurements are indispensable.
Surgical treatment of multilevel cervical myelopathy
Cervical myelopathy is a progressive, degenerative disease of the spine. In this condition, osteoarthritic changes in multiple segments of the cervical spine compress the spinal cord. This compression can impair fine motor skills and balance, cause severe neck pain, and even lead to paralysis. As symptoms increase, surgery is generally required to decompress the spinal cord and to stop this process.
Surgical treatment for mDCM can be classified into anterior and posterior approaches. Both categories of treatment are common, but there are few studies that compare their clinical and radiologic outcomes.¹
Imaging endpoint analysis in mDCM treatments
Given this gap in the research, Dr. Gregor Schmeiser and his colleagues aimed to compare two posterior surgical techniques: laminoplasty with unilateral screw fixation (uLP) and laminectomy with bilateral screw fixation (LC).
Because one problem with the dorsal approach is the high rate of postoperative infection and muscle atrophy, they wanted to determine whether there were differences in postoperative range of motion (RoM) and fusion rate, two common imaging endpoints, using a less invasive method, first as biomechanical work on cadavers1 and then as a clinical feasibility study.2
These results were then used to examine the extent to which radiographic outcomes influence clinical outcomes.
The radiologic outcomes focused on fusion rates as a function of range of motion (RoM) and parameters of cervical sagittal balance. According to Schmeiser, these variables are notoriously difficult to measure. “They are time-consuming as well as astonishingly error-prone and require the work of two to three radiologists,” he told us.
He added, “Especially if the images were not created in the same planes, manually evaluating these types of radiological parameters is extremely difficult.”
Several patient-reported outcome measures (PROMs) were used to measure clinical outcomes. To measure clinical outcomes, visual analogue scales (VAS) and the Neck Disability Index (NDI) were employed. Additionally, the severity of the myelopathy was assessed using the modified Japanese Orthopedic Association (mJOA) score.
Improving treatments for cervical myelopathy using AI
Schmeiser used RAYLYTIC as an independent imaging software laboratory to overcome these challenges. He acknowledged that the AI-supported, scientifically validated software was the most accurate way to describe the outcomes he needed for his study.
Dr. Schmeiser emphasized that the required analyses could be performed faster and more accurately because the AI does not tire out.
With the help of these analyses, the study found that although the two techniques compared resulted in different fusion rates, the sagittal balance parameters remained similar. These findings support the practicality of a unilateral approach as a less invasive treatment method for this serious condition.
¹ Schmeiser G, Schilling C, Grupp TM, Papavero L, Püschel K, Kothe R. Unilateral laminoplasty with lateral mass screw fixation for less invasive decompression of the cervical spine: a biomechanical investigation. Eur Spine J. 2015 Dec;24(12):2781-7. doi: 10.1007/s00586-015-4230-5. Epub 2015 Sep 8. PMID: 26350248.
2Schmeiser G, Bergmann JI, Papavero L, Kothe R. Surgical Treatment of Multilevel Degenerative Cervical Myelopathy: Open-Door Laminoplasty and Fixation via Unilateral Approach. A Feasibility Study. J Neurol Surg A Cent Eur Neurosurg. 2021 Dec 15. doi: 10.1055/s-0041-1739224. Epub ahead of print. PMID: 34911089.