Prognostic factors for primary cutaneous anaplastic large-cell lymphoma: a multicentre retrospective study from Japan

Background: The clinical implications of DUSP22 rearrangement and the association between DUSP22 rearrangement and lymphoid enhancer-binding factor 1 (LEF1) expression pattern in CD30+ cutaneous T-cell lymphomas (CTCLs) are unknown.

Objectives: This study assessed the incidence of DUSP22 rearrangement and its clinical and immunohistochemical implications in primary cutaneous anaplastic large-cell lymphoma (pcALCL), lymphomatoid papulosis (LyP) and CD30+ mycosis fungoides with large-cell transformation (MF-LCT), focusing especially on the association with the prognosis and LEF1 expression pattern. Prognostic factors of pcALCL were also examined.

Methods: We conducted a multicentre retrospective study including patients with pcALCL, LyP and MF-LCT diagnosed between 1 January 2000 and 31 December 2018 in Japan. Baseline data at diagnosis, treatment course, overall survival (OS) and disease-specific survival (DSS) were collected. Immunohistochemical analysis and fluorescence in situ hybridization to detect DUSP22 and TP63 rearrangement were performed using skin samples at diagnosis. We investigated the association between staining pattern and these gene rearrangements. We also assessed the prognostic implications of clinical status, immunohistochemical results and the presence of gene rearrangements.

Results: DUSP22 rearrangement was detected in 50% (11 of 22) of cases of pcALCL, but not in any cases with LyP (0 of 14) or MF-LCT (0 of 11). TP63 rearrangement was not detected in any case. Clinically, patients with pcALCL with DUSP22 rearrangement did not tend to develop ulcers (P = 0.081). There was no significant association between DUSP22 rearrangement status and immunohistochemical results, including LEF1 expression pattern. T3 stage and the presence of lower limb lesions were significantly associated with shorter OS (P = 0.012 and 0.021, respectively, by log-rank test). Similarly, they were significantly correlated with shorter DSS (P = 0.016 and 0.0001, respectively).

Conclusions: DUSP22 rearrangement is relatively specific to pcALCL among CD30+ CTCLs in Japan. Although the LEF1 expression pattern was not related to DUSP22 rearrangement in pcALCL, there was no rearrangement if LEF1 was not expressed. We confirmed that T3 stage and the lower limb involvement were significantly associated with decreased OS and DSS. The presence or absence of lower limb lesions should be included in T-stage subcategorization in the future.