Indeed, no correlation was observed between BM histology and outcome. Among the three patients with a second atypical BM, one died 1 year later due to septicaemia but without progression of his lymphoma, while the other two were still alive in partial remission after 48 and 72 months, respectively. On the other hand, five patients with progressive disease at the time of the second BM biopsy exhibited a normal BM histological pattern and three of them (patients 2, 11 and 12) died of lymphoma after 48, 2 and 6 months, respectively, with nodal (two cases) and meningeal (one case) involvement.No correlation was observed between histological pattern and molecular results in the second BM biopsy. The three patients with an atypical BM had either a polyclonal profile, an unrelated T-cell clone, or a medullary T-cell clone identical to skin and blood. Alternatively, a BM T-cell clone was found identical to skin and blood in four patients with normal BM histology. Interestingly, the molecular profiles of the first and second BM samples were identical in 10 of the 12 patients, including the five patients with an identical skin, blood and BM T-cell clone. Three of these five patients died of lymphoma and one had disease progression. As reported at diagnosis, 8 the monoclonal T-cell population detected in the second BM sample was already present in blood samples. The patients were already known to have progressive disease owing to clinical and biological status, except for one patient (patient 8) who is still in remission after 12 months. Such data would also suggest that monoclonal T-cells persist in consecutive blood and BM samples.Although the small number of patients with a second BM biopsy does not allow statistical conclusions to be drawn, the comparative analysis of a second BM specimen reinforces the conclusions of our first study. First, BM histological involvement is very rare in patients with epidermotropic lymphoma both at diagnosis and during follow-up. Second, BM histological and molecular staging does not provide any further information useful for patient management. Finally, we suggest that BM biopsy is of little value, not only at diagnosis but also during follow-up, unless specific haematological symptoms are present.