Prognostic value of imaging markers from 18FDG-PET/CT in paediatric patients with Hodgkin lymphoma

Abstract

Objective Identification of imaging prognostic parameters for early therapy personalisation to reduce treatment-related morbidity in paediatric Hodgkin lymphoma (HL). Our aim was to evaluate quantitative markers from baseline 2-[18F]fluoro-2-deoxy-d-glucose PET/CT as prognostic factors for treatment outcomes. Another goal was assessing the prognostic value of Deauville score at interim PET/CT. Methods Twenty-one patients were prospectively enrolled. Median age was 12 years (range 6–17); 13 were female. Patients underwent PET/CT for disease staging (bPET), at the end of two cycles of chemotherapy (iPET) and after chemotherapy. A total of 173 lesions were segmented from bPET. We calculated 51 texture features for each lesion. Total metabolic tumour volume and total lesion glycolysis from bPET were calculated for response prediction at iPET. Univariate and multivariate analyses were used for optimal cut-off values to separate responders at iPET according to the Deauville score. Results We identified four texture features as possible independent predictors of treatment outcomes at iPET. The areas under the ROC for univariate analysis were 0.89 (95% CI, 0.75–1), 0.82 (95% CI, 0.64–1), 0.79 (95% CI, 0.59–0.99) and 0.89 (95% CI, 0.75–1). The survival curves for patients assigned Deauville scores 1, 2, 3 and X were different from those assigned a score 4, with 4-year progression free-survival (PFS) rates of 85 versus 29%, respectively (P = 0.05). Conclusions We found four textural features as candidates for predicting early response to chemotherapy in paediatric patients with HL. The Deauville score at iPET was useful for differentiating PFS rates.

Clinical Perspectives for 18F-FDG PET Imaging in Pediatric Oncology: Μetabolic Tumor Volume and Radiomics

“…High “asphericity”, a textural parameter reflecting the lesion surface complexity was also correlated with the interim PET inadequate response. In a most recent prospective study in a small population of HL pediatric patients [ 152 ], four textural parameters, from the textural families of the grey-level co-occurrence matrix and neighborhood grey tone difference matrix were also predictors of interim PET treatment response ( Table 2 ).…”

PET/CT Imaging of Lymphoma Outside the Western World

“…The CUDIM participated in local research projects to reduce treatment-related morbidity in pediatric HL identifying four textural features in the bPET as candidates for predicting early response to chemotherapy at iPET (3). Analyses were used for optimal cut-off values to separate responders at iPET according to the Dauville score.…”

Baseline 18F-FDG PET/CT Radiomics in Classical Hodgkin’s Lymphoma: The Predictive Role of the Largest and the Hottest Lesions

“…Radiomics, an emerging field of research aiming to extract mineable, quantitative, high-dimensional data from clinical images, has been suggested as a solution [ 12 , 13 , 14 ]. Various studies have explored the possibilities offered by radiomics in HL [ 15 ] whether for computer-aided histological classification [ 16 ], staging [ 17 , 18 ], early metabolic response assessment [ 19 , 20 , 21 ], or refractory disease prediction [ 22 , 23 ].…”