Abstract 4516: Prognostic value of PET/CT in newly diagnosed multiple myeloma: A Bayesian analysis with missing data imputation

Abstract Introduction: PET/CT assessment in newly diagnosed patients with multiple myeloma is a significant predictor of long-term clinical outcomes. Incorporating this imaging technique into a Bayesian analysis that addresses data gaps allows for more accurate prognostic estimates. This methodology enhances our understanding of the disease and guides informed clinical decisions, ultimately improving patient care. Objective: This study aims to evaluate the prognostic value of PET/CT in newly diagnosed patients with multiple myeloma, using Bayesian analysis to manage missing data imputation. We seek to determine how PET/CT findings predict long-term clinical outcomes, thereby enhancing treatment personalization. Hypotheses: A positive PET/CT at diagnosis may correlate with poor prognosis, indicating lower survival rates and worse disease control. A negative PET/CT at diagnosis may be associated with better prognosis, suggesting higher treatment response rates and longer survival. Study Design: An analytical study based on clinical and imaging data from PET/CT. Study Population: Patients diagnosed with multiple myeloma between 2017 and 2024 at an institution in Medellín, Colombia, approved by the ethics committee of Pablo Tobón Hospital. Variables of Interest: Clinical Variables: Age, sex, type of treatment. PET/CT Variables: Positive (+), negative (-). Other Variables: Cytogenetics, Beta 2 Microglobulin (B2M), Bone Marrow Biopsy, Albumin, Hemoglobin. Outcome Variables: Overall Survival, Progression-Free Survival, Treatment Response. Data Imputation Methodology: Using advanced R tools, including: MCMC: For parameter estimation despite incomplete data. SMC: To optimize sequential inferences and improve precision. Statistical Analysis: Kaplan-Meier survival curves for visualizing survival rates. Significance of covariates evaluated through 95% credibility intervals. Results: The Bayesian model highlights PET/CT and cytogenetics as key outcome variables. A negative PET/CT correlates with better survival and more accurate treatment response prediction, while a positive PET/CT is linked to higher disease progression rates. Limitations: Generalizability may be limited due to single-institution data. Missing data could introduce bias despite imputation methods. Short follow-up may restrict long-term outcome assessment. Variability in imaging protocols may affect result consistency. Conclusions: Prognostic Value of PET/CT: The study indicates that PET/CT evaluation is a significant predictor of long-term outcomes in multiple myeloma. A negative PET/CT is linked to better prognosis and treatment response. Clinical Decision Making: Integrating Bayesian analysis and data imputation into PET/CT assessment enhances prognostic accuracy, providing clinicians with valuable insights for personalized treatment and optimized patient care. Citation Format: Eliana Andrea González Artunduaga, Nelson Alirio Cruz Gutiérrez, Simón Sandoval Álvarez, Kenny Mauricio Gálvez Cárdenas. Prognostic value of PET/CT in newly diagnosed multiple myeloma: A Bayesian analysis with missing data imputation [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 4516.