Utility of diffusion-weighted imaging to differentiate benign and malignant solid orbital tumours.

ABSTRACT

OBJECTIVE: We assessed the utility of apparent diffusion coefficients (ADCs) derived from diffusion-weighted imaging to differentiate benign and malignant orbital tumours by oculoplastic surgeons in the clinical setting and sought to validate observed ADC cut-off values. DESIGN AND PARTICIPANTS: Retrospective review of patients with benign or malignant biopsy-confirmed orbital tumours. METHODS: Blinded graders including 2 oculoplastic surgeons, 1 neuroradiologist, and 1 medical student located and measured orbital tumour ADCs (10-6 mm2/s) using the Region of Interest tool. OUTCOME MEASURES: Nonradiologist measurements were compared with each other to assess reliability and with an expert neuroradiologist measurement and final pathology to assess accuracy. RESULTS: Twenty-nine orbital tumours met inclusion criteria, consisting of 6 benign tumours and 23 malignant tumours. Mean ADC values for benign orbital tumours were 1430.59 ± 254.81 and 798.68 ± 309.12 mm2/s for malignant tumours. Our calculated optimized ADC cut-off to differentiate benign from malignant orbital tumours was 1120.84 × 10-6 mm2/s (sensitivity 1, specificity 0.9). Inter-rater reliability was excellent (intraclass correlation coefficient = 0.92; 95% CI, 0.86-0.96). Our 3 graders had a combined accuracy of 84.5% (92.3%, 92.3%, and 65.4%). CONCLUSIONS: Our ADC cut-off of 1120.84 × 10-6 mm2/s for benign and malignant orbital tumours agrees with previously established values in literature. Without priming with instructions, training, or access to patient characteristics, most tumours were correctly classified using rapid ADC measurements. Surgeons without radiologic expertise can use the ADC tool to quickly risk stratify orbital tumours during clinic visits to guide patient expectations and further work-up.