Thank you for joining us today. Before I start my presentation, I would like to thank the organizers of the AANS for giving us the opportunity to present this work online. My work is entitled Predicting Tumor-Specific Survival in Patients with Spinal Metastatic Renal Cerebrocarcinoma. Also, we can analyze the most common cited scoring system in the literature and see which one of these is the most accurate at predicting survival of these patients. I have no disclosures related to this work. This research work aims to compare the performance of the most prevalent scoring systems and prediction models to determine both the overall survivor and tumor-specific survival for RCC patients undergoing surgical treatment for spinal metastatic disease. In fact, the test accuracy and their reliability in the context of RCC spinal metastasis of these scores have been questioned in the context of contemporary multidisciplinary treatments plans for RCC involving systematic therapy, surgery, and radiosurgery. This study aims to help spine surgeons risk stratify their patients and determine the risk-benefit ratio of invasive surgery by estimating prognosis and survival. Current surgical decision-making is guided by conceptual frameworks such as NOMPS that takes into account the creative spinal cord compression, spinal instability, radiosensitivity of the tumor, and the systemic condition of the patient. However, the benefits of a proposed treatment, which is often surgical resection, stabilization followed by adjuvant graduation, in conditions of limited survival and the overall risk-benefit ratio of an invasive procedure can be very difficult to assess. This figure here is brought from a previous work that we've done to bring together the prognostic scores for spine metastatic disease. For each model, the number of patients included, the method of data collection, the most common type of tumors included in the analysis, and the statistical tools used for building each model are listed. If you look more closely, you can see that mostly these prognostic scores were built from retrospective data of small sample sizes. The most common tumors that were included are lung, breast, and prostate cancer, less commonly renal cell carcinoma. As you can also see, mostly these scores were built using Cox or logistic regression models. However, the most recent ones are using more advanced logistic and statistical tools, which are support vector machine learning methods and neural network. In this table here, we summarize the predictors that are included in each scoring system for spinal metastatic disease. As you can see, the most common predictors that are found in these models are the performance or functional status of the patients, the primary tumor histology, the presence or absence of visceral metastases, the number of vertebral body involved, and also the presence of bone metastases. Now, going back to our study, this is a retrospective review that was done at the Massachusetts General Hospital and Regan and Holman Hospital from 2010 to 2019. We retrieved 86 patients who underwent spine surgery for metastatic renal cell carcinoma. We extracted clinical and surgical characteristics of all patients. We also calculated the preoperative prognostic scores. Also, we did a regression analysis of patient variables in association to one-year survival after surgery by using Cox proportional hazard models. Calibration and time-dependent discrimination analysis were tested to quantify the accuracy of each scoring system at three months, six months, and one year. We're going to touch on these later on during the course of this presentation. In this table here, we summarize our results. In the study, we showed that there is no difference in the proportion of prognostic factors commonly used in prognostic scores between the RCC survival groups. Our results demonstrate that there is no difference in gender, Frenkel grade, tumor pathology, Ferman grade, visceral metastasis, number of spine metastases, pathologic fracture, SIN score, degree of epidural spine compression, hemoglobin level, and thrombocytosis, which indicates that these factors have a limited prognostic value in predicting the survival amongst patients with metastatic RCC. We have to note here that the one-year survival rate was 72%. The prognostic variables that were associated with deaths before one year are poor performance status, neurologic deficit at presentation, and a low albumin level that we defined as an albumin level less than 3.5. Here, we present Kaplan-Meier curves and log-rank tests that were used to assess the significant difference between prognostic categories of each scoring system. The predicted one-year survival risk between prognostic classes of each score represented in this figure here. The red ones are those who showed no significant difference between their prognostic classes. However, a significant difference in survival was noted between the prognostic categories of the original Bayer score, the New England Spinal Metastatic score, the Sorg Classic, the Sorg Nomogram, revised Tokuhashi, and the original Tokuhashi score. Furthermore, to assess the performance of these prognostic scores, we calculated the AUC of the time-dependent ROC curves for the six prognostic scores with the p-value less than 0.05. We used an AUC sequentially at three months, six months, and one year. As you can see here, we showed that the New England Spinal Metastatic score had the best performance at the different time points, at three months, six months, and also at one year. In this study, we showed that currently standard prediction models for spine metastatic disease have a poor to fair performance in predicting the survival of contemporary metastatic RCC patients after spine surgery. We found that hypoalbuminemia, which reflects the overall health of cancer patients, is associated with survival and therefore should be assessed before surgery. Also, tumor-specific factors such as the primary tumor histology, in our case here, clear cell carcinoma versus other types, and the response to treatment should be further investigated to assess their importance in the decision framework for the treatment of spine metastasis. The New England Spinal Metastatic score demonstrated good performance at predicting the one-year survival after surgery because it incorporates factors that were highly associated with metastatic RCC survival, such as performance status and albumin level. Future models that incorporate genetic factors and response to treatments are expected to provide more accurate clinical decision tools to improve patient outcomes and quality of care. Before I wrap up my presentation, I would like to thank Dr. John Chin and Dr. Ganesh Shankar for their support and mentorship. Thank you everyone and stay safe!

tumor-specific survival

spinal metastatic renal cell carcinoma

scoring systems

prediction models

survival prognosis