Discrimination of paediatric brain tumours using apparent diffusion coefficient histograms.
Eur Radiol 2011;
22:447-57. [PMID:
21918916 DOI:
10.1007/s00330-011-2255-7]
[Citation(s) in RCA: 80] [Impact Index Per Article: 6.2] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 04/05/2011] [Revised: 07/08/2011] [Accepted: 07/28/2011] [Indexed: 11/27/2022]
Abstract
OBJECTIVE
To determine if histograms of apparent diffusion coefficients (ADC) can be used to differentiate paediatric brain tumours.
METHODS
Imaging of histologically confirmed tumours with pre-operative ADC maps were reviewed (54 cases, 32 male, mean age 6.1 years; range 0.1-15.8 years) comprising 6 groups. Whole tumour ADC histograms were calculated; normalised for volume. Stepwise logistic regression analysis was used to differentiate tumour types using histogram metrics, initially for all groups and then for specific subsets.
RESULTS
All 6 groups (5 dysembryoplastic neuroectodermal tumours, 22 primitive neuroectodermal tumours (PNET), 5 ependymomas, 7 choroid plexus papillomas, 4 atypical teratoid rhabdoid tumours (ATRT) and 9 juvenile pilocytic astrocytomas (JPA)) were compared. 74% (40/54) were correctly classified using logistic regression of ADC histogram parameters. In the analysis of posterior fossa tumours, 80% of ependymomas, 100% of astrocytomas and 94% of PNET-medulloblastoma were classified correctly. All PNETs were discriminated from ATRTs (22 PNET and 4 supratentorial ATRTs) (100%).
CONCLUSIONS
ADC histograms are useful in differentiating paediatric brain tumours, in particular, the common posterior fossa tumours of childhood. PNETs were differentiated from supratentorial ATRTs, in all cases, which has important implications in terms of clinical management. Key Points • MR based apparent diffusion coefficient histograms can help differentiate paediatric brain tumours • ADC histogram parameters correctly classified the great majority of posterior fossa tumours.
Collapse