РАЗРАБОТКА МЕТОДА КОЛИЧЕСТВЕННОЙ ОЦЕНКИ АСФЕРИЧНОСТИ ОБЛАСТИ НАКОПЛЕНИЯ КАК НОВОГО ПРЕДИКТОРА НЕБЛАГОПРИЯТНОГО ИСХОДА У ПАЦИЕНТОВ С НЕЙРОБЛАСТОМОЙ

Abstract

The aim of this study was the development and implementation of algorithms for computer analysis of medical images to quantify the asphericity of the area of radiopharmaceutical accumulation in a primary tumor, as well as to study the prognostic significance of asphericity as a predictor of an adverse outcome in patients with neuroblastoma. This retrospective study included 142 primary patients (69 boys, 73 girls, median age 22.5 months, range 1-198 months) with newly diagnosed neuroblastoma. All patients underwent full-body scintigraphy with 123I. Cox regression analysis, ROC analysis, and Kaplan-Meier analysis with a log-rank test in assessing event-free survival were used to assess the predictive value of asphericity. The median follow-up period was 18 months (range 7 to 78 months; 17 patients had disease progression / relapse, 10 patients died). The asphericity of the accumulation area in univariate Cox regression analysis is a significant pre- dictor of poor outcome (p = 0.035; hazard ratio (HR) = 1.106 [1.007; 1.215] for an increase of one unit). The optimal cut-off point for the asphericity parameter in the ROC analysis was chosen equal to 28.4%, predictor sensitivity: 77.8% (95% CI: 65%; 92%), specificity: 65.1% (95% CI: 52%; 75%). According to the results of the log-rank test, statistically significant differences were obtained: in patients with high asphericity (more than 28.4%) and low asphericity, the mean event-free survival rate was 42.9 (95% CI: 38.9%; 46.8%) and 56.6 (95% CI: 48.3%; 64.8%) months, respectively (p = 0.044). For the first time in the Russian Federation, a method for assessing the asphericity of neuroblastoma has been developed and implemented in software. In this study, the asphericity of the radiopharmaceutical accumulation area in the primary tumor, assessed prior to treatment initiation, identified children with a high and low risk of progression / relapse and death, with a sensitivity of 77.8% and a specificity of 65.1%.