Texture image analysis in differentiating malignant from benign adrenal cortical tumors in children and adults

To investigate the possible role of chromatin texture parameters, nuclear morphology, DNA ploidy and clinical functional status in discriminating benign from malignant adrenocortical tumors (ACT).

Forty-eight cases of clinically benign (n=40) and clinically malignant (n=8) ACT with a minimum of 5-years' follow-up were evaluated for chromatin texture parameters (run length, standard deviation, configurable run length, valley, slope, peak and other 21 Markovian features that describe the distribution of the chromatin in the nucleus), nuclear morphology (nuclear area, nuclear perimeter, nuclear maximum and minimum diameter, nuclear shape), and DNA ploidy. Nuclear parameters were evaluated in Feulgen-stained 5 mum paraffin-sections analyzed using a CAS 200 image analyzer.

Since ACTs present different biological features in children and adults, patients were divided into two groups: children (< or = 15 years) and adults (>15 years). In the group of children DNA ploidy presented a marginal significance (p=0.05) in discriminating ACTs. None of the parameters discriminated between malignant and benign ACT in the adult group.

ACTs are uncommon and definitive predictive criteria for malignancy remain uncertain, particularly in children. Our data point to DNA content evaluated by image analysis as a new candidate tool for this challenging task. Texture image analysis did not help to differentiate malignant from benign adrenal cortical tumors in children and adults.