Abstract
AIMS: Localised- and diffuse-type tenosynovial giant cell tumours (TGCT) are regarded as different clinical and radiological TGCT types. However, genetically and histopathologically they seem indistinguishable. We aimed to correlate CSF1 expression and CSF1 rearrangement with the biological behaviour of different TGCT-types with clinical outcome (recurrence). METHODS AND RESULTS: Along a continuum of extremes, therapy-naive knee TGCT patients with >3-year follow-up, mean age 43 (range = 6-71) years and 56% females were selected. Nine localised (two recurrences), 16 diffuse-type (nine recurrences) and four synovitis as control were included. Rearrangement of the CSF1 locus was evaluated with split-apart fluorescence in-situ hybridisation (FISH) probes. Regions were selected to score after identifying CSF1-expressing regions, using mRNA ISH with the help of digital correlative microscopy. CSF1 rearrangement was considered positive in samples containing >2 split signals/100 nuclei. Irrespective of TGCT-subtype, all cases showed CSF1 expression and in 76% CSF1 rearrangement was detected. Quantification of CSF1-expressing cells was not informative, due to the extensive intratumour heterogeneity. Of the four synovitis cases, two also showed CSF1 expression without CSF1 rearrangement. No correlation between CSF1 expression or rearrangement with clinical subtype and local recurrence was detected. Both localised and diffuse TGCT cases showed a scattered distribution in the tissue of CSF1-expressing cells. CONCLUSION: In diagnosing TGCT, CSF1 mRNA-ISH, in combination with CSF1 split-apart FISH using digital correlative microscopy, is an auxiliary diagnostic tool to identify rarely occurring neoplastic cells. This combined approach allowed us to detect CSF1 rearrangement in 76% of the TGCT cases. Neither CSF1 expression nor presence of CSF1 rearrangement could be associated with the difference in biological behaviour of TGCT.