Objectives: To optimise radiological screening in von Hippel-Lindau disease (VHL) while minimising cost and morbidity.
Methods: A model of VHL was based on retrospective studies, and Bayes's theorem used to calculate the probability of the gene's presence and the likelihood of further lesions in affected families. A six year follow up was conducted to test the validity of the model.
Results: Follow up confirmed the accuracy and validity of the model. Posterior fossa haemangioblastomas occur in 79.2% of VHL cases, supratentorial, retinal and spinal haemangioblastomas in 6.9%, 42.8%, and 22.0%, phaeochromocytomas in 5.2%, and renal carcinomas in 14.5%. Population incidences are 1:15,700 live births (posterior fossa), 1:780,000 (supratentorial), and 1:116,000 (spinal). The birth rate of subjects with VHL is 1:43,000; new mutations occur in 1:178,000 live births. Penetrance is 90%; 40% present with multiple lesions and 6.4% die within two years after diagnosis.
Conclusions: For most patients presenting with a VHL-type lesion, with sufficient clinical and pedigree data, the presence or absence of the VHL gene, and the probability of further lesions occurring, can be assessed with a high degree of accuracy using the method described in this paper. Those cases in the non-VHL group do not require long term radiological follow up, nor do their relatives require radiological screening. Subjects in the VHL group should be screened for renal carcinoma indefinitely from the age of 20 years, and all clinically unaffected relatives should be screened genetically for the VHL gene. (Those found negative for the gene do not require further screening, but those found positive should be screened.)