Value Of Magnetic Resonance Apparent Diffusion Coefficient In The Differential Diagnosis Of Posterior Fossa Brain Tumors In Children

Objective:To investigate the apparent diffusion coefficient（ADC）of magnetic resonance in the preoperative classification and differential diagnosis of pediatric posterior fossa tumors,so as to improve the understanding and diagnosis of pediatric posterior fossa tumors.Methods:1.The clinical data and head magnetic resonance imaging,including ADC,of 94 children with posterior fossa tumor in the First Hospital of Qinhuangdao City,Hebei Province were retrospectively analyzed.According to the 2018 WHO Classification Criteria for Central Nervous System Tumors,94 cases of posterior fossa tumors were classified into low-grade 32 cases（WHO Grade I 19 cases,all pilocytic astrocytoma;There were 13 WHO grade II cases,including diffuse astrocytoma in 8 cases,pilocytic myxoid astrocytoma in 2 cases,pleomorphic xanthoma astrocytoma in 2 cases,and ependymoma in 1 case）,and 62 high-grade cases（19 WHO Grade III cases were ependymoma with anaplasia;43 WHO Grade IV patients,including 36medulloblastoma and 7 atypical teratoid/rhabdomyoid tumors）.Mean ADC（ADCmean）,minimum ADC（ADCmin）and standardized ADC（tumor solid portion ADC/normal cerebellar white matter ADC,r ADC）were measured.receiver operation characteristic,ROC)analysis to determine the optimal threshold for distinguishing between high and low grade tumors.2.The ADC values（ADCmean,ADCmin and r ADC）of 36 cases of medulloblastoma,20 cases of ependymoma（1 case of ependymoma WHOⅡ,19 cases of intervariant ependymoma WHO III）,31 cases of low-grade astrocytoma,and 7 cases of atypical teratoid/rhabdomyoma were measured.ROC curve analysis was performed to generate thresholds for distinguishing different tumors.Results:1.The ADC values（ADCmean,ADCmin and r ADC）of high-grade tumors were all lower than those of low-grade tumors,and there was a statistical difference between the two groups（P<0.001）.ROC curve analysis showed that r ADC had higher diagnostic efficacy,and r ADC=1.52was used as the optimal threshold to distinguish high-grade and low-grade tumors.The sensitivity and specificity were 100%and 95.2%respectively.2.There was no statistical difference in ADCmean and r ADC between medulloblastoma and atypical teratoid/rhabdomyoma（P>0.05）,but there was statistical difference in ADCmin between them（P<0.05）.The optimal threshold was ADCmin=0.395×10^-3mm²/s to distinguish medulloblastoma from atypical teratoid/rhabdomyoma.The sensitivity and specificity were80.0%and 85.7%,respectively.The differences in ADCmean,ADCmin and r ADC of medulloblastoma,low-grade astrocytoma and ependymoma were statistically significant（P<0.001）.Compared with ADCmean and ADCmin,r ADC was more effective in differentiating medulloblastoma,low-grade astrocytoma and ependymoma.Through ROC curve analysis,r ADC=1.16was used as the optimal threshold to distinguish medulloblastoma from ependymoma,and the sensitivity and specificity were 95%and 97.2%,respectively;r ADC=1.74 was used as the optimal threshold to distinguish ependymoma from low-grade astrocytoma,and the sensitivity and specificity were 96.8%and 100%,respectively.Conclusions:ADCmean,ADCmin and r ADC have potential clinical application value in preoperative grading and differential diagnosis of common brain tumors in children’s posterior fossa,and r ADC is superior to ADCmean and ADCmin in diagnosis.