EFFECTIVENESS OF DIFFUSION WEIGHTED MRI DERIVED METRICS IN DIFFERENTIATION OF PEDIATRIC POSTERIOR FOSSA TUMORS

Background: The value of diffusion weighted imaging (DWI) in differentiation of pediatric posterior fossa tumors (PFTs)has been established previously, however most studies were limited only to the most common types of PFTs and their results were contradictory.. Aim of work: The aim of this study was to investigate the effectiveness of DWI in differentiation of PFTs taking care to include the less frequent types, and to clarify previously reported debatable findings. Patients and methods: We retrospectively analyzed 13 embryonal tumors (EMBTs), 4 ependymomas (EPNs),8 low grade gliomas (LGGs),2 mixed neuronal glial tumors (MNGTs), and 5 diffuse intrinsic pontine gliomas (DIPGs) . For all tumors, absolute
mean ADC of the non necrotic enhancing portions of the tumor and ADC ratios of the solid enhancing region of the tumor to the ADC of normal appearing cerebellum were generated. The ADC metrics of the various groups (designated according to histological subtype and WHO grade) were compared using the kruskal – wallis and student-ttests. Receiver operating characteristic (ROC) curve analysis was used to generate cutoff absolute ADC values and ADC ratios for differentiation of different types of PFTs Results: Mean ADC values and ADC ratios were significantly lower in high than in low grade tumors (P < 0.0001). Kruskal-wallis test showed statistically significant difference in the ADC metrics among the tumor groups (P=0.000).Mean ADC values were significantly higher in LGGs(1.71+/-0.13) than in ependymomas (1.13+/-0.13) and both were significantly higher than EMBTs (0.69+/-0.09). No overlap was seen in the range of ADC values and ratios of EMBTs, EPNs, and LGGs. DIPG mean ADC values and ratios overlapped widely with EPNs from which they could not be
differentiated using ADC metrics. DIPGs could be distinguished from EMBTs using a cutoff ADC ratio of>1.108× 10−3 (100% sensitivity and 92.3% specificity)but not with mean ADC values and from LGGs using a cutoff mean ADC value of ≤ 1.333 (100% sensitivity and specificity). No statistically significant difference (P >0.05) was seen between the mean ADC values of MNGTs (1.5+/-0.02) and LGGs (1.71+/-0.13) with wide overlap in their range of ADC values and ratios.
Conclusion: Apart from a few exceptions, absolute ADC values and ratios go a long way in enabling pre-operative differentiation of different types of PFTs.