Objective To explore the diagnostic efficacy of ¹⁸F-fluorodeoxyglucose(FDG) PET/CT on intramedullary relapse of acute leukemia, and screening parameters that can be used to give the diagnosis.

Methods The ¹⁸F-FDG PET/CT imaging data of 81 patients (45 males and 36 females aged 3–55（25.14±15.07）years) who were clinically diagnosed with suspected intramedullary relapse of acute leukemia in Hebei Yanda Hospital from November 2016 to November 2019 were retrospectively analyzed. All the patients had received bone marrow transplantation. And the histopathological results of bone marrow biopsy were used as "golden standard" of intramedullary relapse. The patients were divided into relapse group (41 cases) and non-relapse group (40 cases). The maximum standardized uptake value (SUV_max) of all the suspected recurrent lesions and SUV_max of normal liver background (liver SUV_max) were measured and calculated. The ratio of appendicular skeleton SUV_max to that of axial skeleton and the ratio of axial skeleton SUV_max to that of normal liver background were calculated respectively. The different cut-off values and area under curve (AUC) based on receiver operating characteristic (ROC) curves were analyzed, and the diagnostic efficacy of ¹⁸F-FDG PET/CT on relapse of acute leukemia was explored preliminarily. Visual interpretation was applied as a reference criterion to explore the diagnostic efficacy of ¹⁸F-FDG PET/CT on the intramedullary relapse of acute leukemia based on different semiquantitative parameters. Differences in semiquantitative parameters between the two groups were compared by independent sample t-test for measurement data and Pearson chi-square test for qualitative data.

Results Axial skeleton SUV_max of the relapse group was higher than that of the non-relapse group (3.10±1.65 vs. 1.99±1.26, t=3.367, P=0.001). Meanwhile, compared with non-relapse group, the ratio of axial skeleton SUV_max to normal liver background SUV_max (1.20±0.56 vs. 0.89±0.74), and the ratio of appendicular skeleton SUV_max to axial skeleton SUV_max (1.58±1.38 vs. 0.79±0.37) were higher than that of relapse group. Both difference has statistics significance (t=2.186, 3.477; both P<0.05). The ROC curve analysis of the semiquantitative index of SUV_max revealed that the sensitivity, specificity, and accuracy of relapse diagnosis were 70.73% (29/41), 72.50% (29/40), and 71.60% (58/81), respectively (AUC=0.770), when axial skeleton SUV_max≥2.05 was used as the criterion of intramedullary relapse. When multifocal+diffuse increased ¹⁸F-FDG uptake was used as the criterion of intramedullary relapse, visual evaluation-based results indicated sensitivity, specificity, and accuracy of 85.37% (35/41), 70.00% (28/40), and 77.78%(63/81), respectively.

Conclusions Visual evaluation based on ¹⁸F-FDG PET/CT is a simple and reliable method for diagnosing acute leukemia relapse, and has high diagnostic efficacy. Based on SUV_max semi-quantitative parameter analysis is a useful supplement to visual evaluation.