Linked Article: Di Raimondo et al. Br J Dermatol 2022; https://doi.org/10.1111/bjd.21063 In this issue of the BJD, Di Raimondo and colleagues provide new information about expression of checkpoint molecules in mycosis fungoides and Sézary syndrome. They established a combined checkpoint marker score covering expression of programmed death protein 1 (PD-1), programmed death ligand 1 (PD-L1) and inducible T-cell costimulator (ICOS) and found that a high score was associated with advanced-stage disease and lower overall survival.1 Furthermore, the results from nine samples of three patients with patch, plaque and tumour lesions are of particular interest. They were analysed with a seven-colour multispectral image to ascertain the expression of checkpoint markers in distinct cell populations.1 The authors found PD-L1 predominantly expressed on CD163+ macrophages and not on the T lymphocytes, which has also been reported in other types of cutaneous lymphomas.1, 2 Recent papers show that multistaining is critical to localize coexpression of immune checkpoint molecules and tumour cells or cell types of the tumour microenvironment (TME).1-3 Especially in lymphoma, it is difficult to differentiate neoplastic from reactive lymphocytes and from some other cells of the TME, based on morphological criteria alone.1-3 Even if it seems to be easier in solid tumours, differentiation is also often not possible, because in areas of close interaction of tumour cells and the TME (especially macrophages), the morphologies of the two cell populations can be quite similar.4, 5 Moreover, the TME is much more complex, and diverse cell types, such as plasma cells, neutrophils, eosinophils and dendritic cells, as well as fibroblasts and endothelial cells, are involved in tumour growth and immune response. Characterization of the TME as completely as possible contributes to a better understanding, and the use of seven-colour multispectral imaging is a step in the right direction. A recent study shows that immune cell topography predicts response to PD-1 blockade.3 Concerning cutaneous T-cell lymphomas, such as mycosis fungoides, we must consider an additional problem, because no immunohistochemical marker allows staining of tumour cells only or differentiation with certainty from reactive T cells. In addition, the tumour cell per se can exhibit different additional phenotypes, for example PD-1 and also other markers of follicular T helper cells, such as ICOS.6-8 In Sézary syndrome it has already been described that PD-1 is frequently expressed, and this marker is already used for diagnostics in daily routine.9 These different tumour cell phenotypes and the composition of the TME raise the need for personalized therapy tailored for each individual patient and might explain contrasting treatment response in immune checkpoint blockade. As mentioned by the authors, immune checkpoint therapy can be a promising option in T-cell lymphomas, but there are also reports about rapid and alarming progression after these therapies.10, 11 Determining the architecture of the TME and the tumour cells in cutaneous lymphomas is the first step in understanding the interaction between these different cell populations and will pave the way for further targeted and functional investigations. Knowing that checkpoint molecules are expressed in the TME, therapy directed against the TME should become the focus of future research.