Search

BACKGROUND: Survival in mycosis fungoides (MF) is varied and may be poor. The PROCLIPI (PROspective Cutaneous Lymphoma International Prognostic Index) study is a web-based data collection system for early-stage MF with legal data-sharing agreements permitting international collaboration in a rare cancer with complex pathology. Clinicopathological data must be 100% complete and in-built intelligence in the database system ensures accurate staging. OBJECTIVES: To develop a prognostic index for MF. METHODS: Predefined datasets for clinical, haematological, radiological, immunohistochemical, genotypic, treatment and quality of life are collected at first diagnosis of MF and annually to test against survival. Biobanked tissue samples are recorded within a Federated Biobank for translational studies. RESULTS: In total, 430 patients were enrolled from 29 centres in 15 countries spanning five continents. Altogether, 348 were confirmed as having early-stage MF at central review. The majority had classical MF (81·6%) with a CD4 phenotype (88·2%). Folliculotropic MF was diagnosed in 17·8%. Most presented with stage I (IA: 49·4%; IB: 42·8%), but 7·8% presented with enlarged lymph nodes (stage IIA). A diagnostic delay between first symptom development and initial diagnosis was frequent [85·6%; median delay 36 months (interquartile range 12-90)]. This highlights the difficulties in accurate diagnosis, which includes lack of a singular diagnostic test for MF. CONCLUSIONS: This confirmed early-stage MF cohort is being followed-up to identify prognostic factors, which may allow better management and improve survival by identifying patients at risk of disease progression. This study design is a useful model for collaboration in other rare diseases, especially where pathological diagnosis can be complex.

Programmed death ligand 1 (PD-L1) is expressed by 20% to 57% of systemic diffuse large B cell lymphomas (DLBCLs). PD-L1 expression in primary cutaneous DLBCL (pcDLBCL) has not been studied so far. Sixteen paraffin-embedded tissue samples of pcDLBCL (13 leg type [LT], 3 others [OT]) were investigated for PD-1, PD-L1, and CD33 expression and the cellular composition of the tumor microenvironment, focusing on myeloid-derived suppressor cells (MDSCs) and tumor-associated macrophages. Membrane-bound PD-L1 expression by the tumor cells was observed in all samples, albeit to a variable extent (19.9%). As expected, most DLBCL-LT (10 cases) were classified as activated B cell like type, with a higher PD-L1 score (21.9%) compared with that of the germinal center B cell like type (7.7%). The surrounding infiltrate consisted predominately of CD163(+) M2 rather than CD68(+) macrophages (CD68:CD163=1:4 to 6). Moreover, a considerable proportion of CD33(+) MDSCs with PD-L1 coexpression was admixed. Tumor cells expressed CD33 to variable degrees (2% to 60%). The number of MDSCs or M2 macrophages did not correlate with pcDLBCL subtypes LT or OT. T cells were only a minor component of the tumor microenvironment. We propose that PD-L1(+) tumor cells and PD-L1(+) MDSCs shield the tumor against PD-1(+) tumor-infiltrating lymphocytes, consequently leading to inhibition and diminution of tumor-infiltrating lymphocytes. Moreover, we found a polarization to M2 macrophages, which may contribute to the poor prognosis of DLBCL patients. Thus, targeting of tumor cells and MDSCs using anti-PD-1/anti-PD-L1 or anti-CD33 antibodies might be a worthwhile new approach to treat this aggressive form of cutaneous B-cell lymphoma. © 2017 Wolters Kluwer Health, Inc. All rights reserved.

Background: Merkel cell carcinoma (MCC) is an aggressive neoplasm, which is often associated with Merkel cell polyomavirus (MCPyV). Programmed death-1 (PD-1) and its ligand PD-L1 are key players of the tumor microenvironment (TME). Methods: Fourteen paraffin-embedded tissue samples of MCC were stratified by their MCPyV detection. Apart from PD-L1 and PD-1, the TME was further characterized for the expression of CD33, FOXP3 and MxA. Results: We observed PD-1 in 2 of 12 tumors. PD-L1 expression by tumor cells was found in 7 of 8 MCPyV(+) samples and was detected particularly in the periphery. The tumor cells were surrounded by a shield of PD-L1/CD33 immune cells. Expression of PD-L1 by the tumor cells was higher in areas with a denser immune infiltrate. CD33(+) cells without direct tumor contact were PD-L1 negative. Only a low number of FOXP3(+) regulatory T-cells was admixed. Tumor cells of MCPyV(−) samples were mostly PD-L1 negative. Conclusions: Our data demonstrate that PD-L1 expression occurs in tumor and immune cells, in areas in which they are close in contact. Interferon seems to play a role in this interaction. We postulate that PD-L1(+)/CD33(+) cells shield the tumor against attacking PD-1(+) immune cells. Therefore, next to anti-PD-1/PD-L1 antibodies, blockade of CD33 seems to be a promising therapeutic approach. © 2017 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd