Good afternoon, everyone. My name is Andy Schilling and it is my pleasure to present a score that our team has been developing. Today, we'll be discussing a novel use of MRI to predict compression fractures in patients with spinal metastases. When a patient presents to clinic with a new diagnosis of spinal metastatic disease, it is essential to be able to accurately predict potential sequelae and develop a plan to prevent them. Previous studies have demonstrated that up to 30% of these patients will go on to experience a vertebral compression fracture or VCF. VCFs are painful, debilitating outcomes for these patients and they can lead to destabilization and neurologic compromise. Treatment for patients with spinal metastatic disease often requires an interdisciplinary approach that can involve radiation and surgery. If we believe that a patient is at particular risk of a VCF, this might inform our decision to pursue surgery to stabilize the spine. Therefore, we need predictive tools to help us make this determination. Currently, one of these tools is the Spinal Instability Neoplastic Score or SINS. It uses clinical and radiographic data to help us assess how a tumor is weakening the spine. SINS is an excellent tool for conveying how cancer has compromised the bone. However, it does not account for the patient's underlying bone quality, which might help us predict which patients are most at risk of a VCF. When we set out to predict VCFs in our metastatic spine patients, we knew that DEXs are used to measure bone density and diagnose osteoporosis. Unfortunately, we found that very few of our patients had previous DEXs, since most of our patients are not females over the age of 65. Furthermore, previous studies have shown DEXs lack sensitivity for predicting VCFs. On the other hand, nearly all patients with spinal METs have lumbar MRIs. So we sought to introduce an MRI-based scoring system to estimate trabecular vertebral bone quality, and then assess its ability to predict new VCFs in metastatic spine patients. We hypothesized that weaker trabecular bone will appear brighter on T1-weighted images due to fatty infiltrate. So brighter bone on T1 may be vulnerable and more likely to result in a VCF. In order to test this question, we conducted a retrospective review of 105 adults with spinal METs that presented at Johns Hopkins Hospital between 2012 and 2019. Roughly half of these patients went on to experience a VCF. The half that did not experience a VCF were used as controls. I'm excited to introduce the new score that we developed, the VBQ score, which stands for vertebral bone quality. This score is simple to calculate clinically, and I'd like to walk through an example now. First, we measure the median signal intensities of the L1 through L4 vertebral bodies on a T1 sagittal lumbar MRI. As a reminder, we hypothesized that higher signal intensity would correlate with fattier, weaker trabecular bone that is more prone to fracture. Next, we measure the CSF at L3 as an internal control to compare our score across different patients, MRI platforms, and clinical settings. We used L3 as a reference to ensure consistency across measurements. Finally, we divide the median of the L1 through L4 signal by the CSF signal to produce our VBQ score. I'd like to pause here to briefly review the SIN score, which we discussed earlier. As you can see, the SIN score is quite complex and integrates data about the tumor's clinical and radiologic features to predict the risk of instability. In contrast, the VBQ score is very simple, which makes it easy to calculate in a clinical setting. Our study design had two major components. First, we collected a variety of demographic, clinical, and radiologic data and performed multivariable analyses to determine which factors could help us predict a new VCF. Then, we calculated receiver operating characteristics to compare the predictive validity of our new VBQ score to the existing SIN score. Finally, we created a combined SINs and VBQ score in an effort to create the best tool we could for predicting VCFs. Here are the results of our univariate analysis. We found no significant difference in demographic characteristics of the patients that went on to experience a VCF and those who did not. There are two exceptions here. Current smokers and patients with a history of steroid use were more likely to experience a fracture, which is consistent with our understanding of the pathophysiology of VCFs. However, we did find that patients that went on to have a VCF had a higher SIN score and a higher VBQ score, which was promising for our study. As a quick summary, the factors that we associated with new VCFs in the univariate analysis were smoking history, steroid use, SIN score, and our new VBQ score. However, in a multivariate analysis, only the SIN score and the VBQ score remained statistically significant. Next, here are the results of our receiver operating characteristics analysis, which we performed to compare SINs and VBQ. You'll see that SINs had a predictive validity of 73%, and VBQ outperformed it with a predictive validity of 80%. And here's the combined score, VBQ and SINs, and together they reached a predictive validity of 89%. Hopefully, we've demonstrated that evaluating baseline bone quality is essential for predicting which patients with spinal metastases will experience a VCF. Our new VBQ score is a simple method that uses widely available MRIs to accomplish this. Currently, we are working on a correlation between VBQ and DEXA. We are also looking to validate the VBQ score in other patient cohorts, as well as at outside institutions. Finally, we'd like to conduct a perspective study with our new score. I would like to thank Dr. Shuba and the other members of our lab, as well as our colleagues in radiation oncology and radiology for their help with this project. And thank you to the AANS for the opportunity to present our work. Thank you very much for listening, and please send me any questions you may have.

VBQ score

vertebral bone quality

compression fractures

spinal metastases

MRI images