Can morphology predict 1p/19q loss in oligodendroglial tumours?

AIMS: To investigate the relationship between phenotype and genotype in oligodendroglial tumours and evaluate whether 1p/19q status can be reliably predicted from histological findings. METHODS AND RESULTS: Three neuropathologists reviewed the association between 10 histological variables, location and genetic losses at 1p, 19q and 17p13 in 63 oligodendroglial tumours (cohort 1). Based on these findings, a multiple logistic regression model for prediction of 1p/19q status was constructed. The ability of this model to predict 1p/19q status was tested on cohort 2, comprising 20 oligodendroglial tumours. Loss of heterozygosity at 1p, 19q and 17p13 was analysed using polymerase chain reaction. Combined 1p/19q loss and losses at 17p13 were mutually exclusive (P < 0.001). The variable H1a (more or <50% of cells with round, uniform nuclei and perinuclear halos) demonstrated the strongest association with 1p/19q status (odds ratio 11.9, 95% confidence interval 3.6, 39.6, P < 0.001). Calcifications, absence of gemistocytic cells and a non-temporal/non-insular location were also associated. The correct 1p/19q status was predicted in 80% of cases in cohort 2. CONCLUSIONS: There is a strong association between phenotype and genotype in oligodendroglial tumours. However, even when all significant variables are accounted for, perfect prediction (100%) of 1p/19q status cannot be obtained.