Blastic Plasmacytoid Dendritic Cell Neoplasm: Clinical, Immunohistochemical and Molecular Evaluation of 23 Cases with Primary Cutaneous Involvement,

Abstract 3565

Blastic plasmacytoid dendritic cell neoplasm (BPDCN) is a rare, haematologic malignancy. The disease may occur at any age, with strong male predominance. Skin presentation is very common. Histologically, tumour cells constitute a uniform population of round or more pleomorphic medium size elements. Clinically this tumour is very aggressive, with poor outcome in adults (median survival 13 months). Chemotherapy (CT), preferably with acute leukemia-type regimens, is the mainstay of treatment, when possible associated with stem cell transplantation (SCT). Few studies investigated copy number changes by array-based comparative genomic hybridization (array-CGH) and gene expression profiling, highlighting a peculiar set of chromosomal losses involving regulatory genes.

The present study focused on 23 cases of BPDCN, that were extensively investigated cytogenetically (array-CGH and SNP6 array) and immunohistochemically. Clinico-pathologic findings and molecular data were analyzed in order to identify prognostic markers and possible clinical subsets.

At onset 12 cases had diffuse, 7 multiple non contiguous and 4 localized skin involvement. Cases 16 and 18 were father and his daughter. Specific therapy was administered in 20 cases: 10 ALL-type CT; 4 NHL-type CT; 2 mono-CT with 1 radiotherapy; 2 radiotherapy alone; 2 SCT. Ten patients relapsed, in a median time of 10 months: 9 cases received further CT and 2 underwent allo-SCT. At last follow-up, 11 patients had died of disease and 2 of therapy-related causes. Median overall survival (OS) was 21 months. At last follow-up 10 patients were alive, 6 disease-free (8 to 28 months) and 4 with disease.

Immunophenotype was typical (CD4, CD56, BDCA2, Tcl-1, and CD123) in all but three patients: 1 negative both for CD4 and CD56, 1 for CD56 and 1 for CD4. All cases expressed CD45RA, 19 CD68/KP1, 10 CD2 and/or CD7 and, when tested, for BDCA2/CD303.

T-cell and B-cell receptor genes were germline with no evidence of EBV infection.

We observed chromosomal imbalances in all patients. Loss of CDKN2A locus (9p21.3) occurred in 15/23 patients, and in 5 cases a biallelic loss was found. FISH confirmed 9p21.3 loss in all 5 cases. Of note, patients with biallelic loss of locus 9p21.3, in respect to those with hemizygous loss, had reduced survival probability (hazard ratio 11.98; confidence interval [CI] 1.21–118.96; log-rank test, p=0.0349). Median OS was 20 and 35 months and median time to relapse 11 and 15 months, respectively. Furthermore, loss of RB1 (13q13.1-q14.3) and CDKN1B (12p13.2-p13.1) locus were observed in 13 cases. An additional, yet unreported, cytogenetic alteration was the loss of 7p12 in 5 of our patients. This is the locus of IKZF1 gene, which encodes the transcription factor Ikaros. Of note, biallelic and monoallelic deletions of IKZF1 has been shown to be associated with chemoresistance and very poor outcome in a subset of BCR-ABL1 positive acute lymphoblastic leukemias. Array-CGH and SNP6 in our series confirmed that BPDCN is characterized by a complex pattern of genomic losses, predominantly affecting chromosomes 9 (71%), 13 (61%), 12 (57%), 5 (19%), 7 (19%), 14 (14%) and 15 (14%).

In conclusion, this is the largest reported series of BPDCN patients, confirming high aggressiveness of the disease. We found negative correlation between age and time to relapse (p=0.0053) and a lower rate of bone marrow involvement in the group of patients with localized skin disease (2 of whom in CR after local skin radiotherapy). Furthermore, we observed absence of CD4 and/or CD56 expression in 3/23 cases. We are the first to report reduced OS probability among patients with homozygous loss of locus 9p21.3 in comparison to those with hemizygous loss, and the possible loss of 7p12-IKZF1 gene in BPDCN.