Primary cutaneous large B-cell lymphoma, leg type (PCDLBCL-LT) is a specific disease entity within the World Health Organization classification of lymphomas, accounting for 20% of all primary cutaneous B-cell lymphomas.¹ PCDLBCL-LT typically presents in the elderly, (median age of presentation, 76 years^2,3), has a female to male ratio of 7:2,⁴ predominantly involves the lower leg, and is often associated with poor clinical outcome.^3,5

Most PCDLBCL-LTs exhibit an activated B-cell diffuse large B-cell lymphoma (ABC DLBCL) phenotype, both immunohistochemically and on gene expression profiling. Malignant cells have a high proliferative fraction and typically express B-cell lymphoma 2 (BCL2), interferon regulatory factor 4 (IRF4), and B-cell lymphoma 6 (BCL6) on immunohistochemistry. Gene expression profiling is associated with constitutive nuclear factor (NF)-κB pathway activation and up-regulation of apoptosis-inhibiting genes, similar to findings in ABC DLBCL. Mutations of MYD88, resulting in constitutive NF-κB activation, are reported in PBDLBCL-LT in up to 69% of cases^6,7 and are associated with a worse clinical outcome.⁷ Apoptosis gene expression profiling in PCDLBCL-LT has shown not only high expression levels of antiapoptotic genes but also concomitant downstream inhibition of the intrinsic apoptosis pathway, as reported in chemotherapy-refractory nodal ABC DLBCL.^8,9

Treatment of relapsed/refractory PCDLBCL-LT remains a significant challenge, with many patients succumbing to uncontrolled disease.³ Current therapeutic options, including access to clinical trials, are limited because of both the rarity of disease and the often frail, comorbid patient population. Small case series of patients with MYD88-mutated PCDLBCL-LT have demonstrated responses to inhibition of B cell receptor (BCR) signaling via Bruton's tyrosine kinase (BTK) inhibition, although responses are of short duration (4 to 40 weeks) and associated with rapid emergence of resistance-conferring mutations.^10,11

The efficacy of the specific BCL2 inhibitor venetoclax in PCDLBCL-LT remains unexplored. In contrast to chronic lymphocytic leukemia, where virtually all cases exhibit extreme sensitivity to BCL2 inhibition, DLBCLs are much more heterogeneous in terms of BCL2 dependency; modulators of sensitivity to BCL2 inhibition are likely to include co-expression of other antiapoptotic BCL2 family members.⁸ Nevertheless, a subset of DLBCL may, like chronic lymphocytic leukemia, be exquisitely sensitive to BCL2 inhibition.^12,13 For example, preliminary results from the phase II CAVALLI study have shown that venetoclax in combination with rituximab, cyclophosphamide, doxorubicin, and prednisolone (R-CHOP) front line in DLBCL, when compared with matched controls in the Phase 3 study of Obinutuzumab in Combination With CHOP Chemotherapy Versus Rituximab With CHOP in Participants With CD20-Positive Diffuse Large B-Cell Lymphoma (GOYA), in BCL2-positive subgroups, particularly in the BCL2 fluorescence in situ hybridization–positive and double-hit subgroups, resulted in improved progression-free survival.¹⁴ The high-level BCL2 expression seen in PCDLBCL-LT similarly provides a rationale for investigation of BCL2 inhibition in this disease.¹⁵ We report a case of MYD88 wild-type, chemotherapy- and radiotherapy-refractory PCDLBCL-LT treated successfully with single-agent venetoclax, without major toxicities. Data are available in the European Genome-Phenome Archive (EGA; https://ega-archive.org/); EGA study accession number: EGAS00001003470; unique study name: ena-STUDY-LCRC-12-02-2019-19:06:32:037-306.

A previously well 77-year-old gentleman was diagnosed with PCDLBCL-LT of the left lower leg in 2006, with isolated cutaneous involvement on computed tomography (CT) scan. This study was approved by the University Hospitals of Leicester NHS Trust Research and Development (UK; 06/Q2501/122). Written informed consent was obtained for publication purposes. Immunohistochemistry showed co-expression of CD20, CD79A, paired box 5 (PAX5), BCL2, BCL6, IRF4, and CD10, with a Ki67 fraction of 90% (Figs 1A-1D). The patient was initially treated with four cycles of R-CHOP and involved-field radiotherapy (40 Gy, 15 fractions), with rapid and complete clinical response. In 2010, he relapsed again with left leg cutaneous involvement only, adjacent to the previously treated sites, and received additional treatment with four cycles of R-CHOP and radiotherapy (40 Gy in 15 fractions), again attaining clinical response. Two additional relapses in the left leg with progressively shorter disease-free intervals were treated with radiotherapy alone in 2011 and 2012 (4 Gy, single fraction). Subsequently, in 2016, an additional relapse, treated with rituximab, gemcitabine, cyclophosphamide, vincristine, and prednisolone, was refractory to chemotherapy.

FIG 1. (A-D) Immunohistochemistry of tumor biopsy specimen before commencing venetoclax (40× magnification). (A) Hematoxylin and eosin staining, (B) interferon regulatory factor 4 (IRF4), (C) B-cell lymphoma 2 (BCL2), and (D) Ki67 expression. Cells comprised large monomorphic blasts with open chromatin and multiple nucleoli. IRF4 and BCL2 expression was seen in all cells, whereas the Ki67 was greater than 90%. (E, F) ¹⁸F-labeled fluorodeoxyglucose–positron emission tomography/computed tomography images to show disease (E) before, and (F) after treatment with venetoclax. (G-I) Clinical photographs of disease (G) before venetoclax, (H) after 7 days of treatment, and (I) after 3 months of treatment with venetoclax.

Because of high expression of BCL2 and the development of chemorefractory disease, in January 2017 we obtained venetoclax through the UK AbbVie compassionate access scheme. Repeat biopsy confirmed persistent high-level BCL2 expression. Repeat staging by CT and ¹⁸F-labeled fluorodeoxyglucose–positron emission tomography/CT scans confirmed that disease remained localized to the left lower leg, as shown in Figure 1E. At previous relapse, bone marrow aspirate and trephine showed no evidence of a clonal B-cell population on flow cytometry or by immunohistochemistry on trephine. Venetoclax was commenced at 400 mg once daily and escalated at weekly intervals to 800 mg and 1,200 mg subsequently. The dose of venetoclax was escalated to 1,200 mg because of published data in DLBCL and follicular lymphoma confirming safety and tolerability of this dose.¹² However, 1,200 mg was not tolerated because of persistent grade 2 diarrhea and grade 2 fatigue, and thus 800 mg was adopted as a maintenance dose. After commencement of venetoclax, within 24 hours, a clinical response was noted. Clinical response is shown in Figures 1G-1I. Biopsy after 2 weeks of treatment showed a complete histologic and immunohistochemical response, although IGHV polymerase chain reaction showed persistence of clonal B cells (data not shown). A repeat ¹⁸F-labeled fluorodeoxyglucose–positron emission tomography/CT scan at 2 months showed a complete metabolic response (Fig 1F). Treatment with venetoclax was stopped 1 month after demonstration of metabolic remission. However, within 4 weeks of stopping venetoclax, disease recurred, once again confined to the left lower leg, but at different anatomic sites. An additional biopsy was performed, which confirmed relapsed disease and showed persistent high-level BCL2 expression (data not shown). Venetoclax was recommenced at a dose of 400/800 mg on alternate days, with only mild gastrointestinal (GI) toxicity (grade 1 Common Terminology Criteria for Adverse Events). Eighteen months later, with continued daily venetoclax, his disease remains in complete remission.

Whole-exome DNA sequencing was performed only on the last relapse sample because of a lack of suitable material from prior biopsies. Sequencing was performed on the Illumina HiSEquation 4000 platform to 100× target coverage. Variants were called using Genome Analysis Toolkit’s (GATK v4.0.3.0) Mutect2 variant caller. Variants were validated with a second caller, Variant Detection in Massively Parallel Sequencing Data (VARSCAN) v2.4.3. Call thresholds were a minimum of 10 variant reads, depth 25 reads, variant allele frequency 10%, scale-invariant feature transform score less than 0.35, Polyphen greater than 0.7. Somatic mutations were identified by paired analysis of tumor and germline DNA (Fig 2). Neither MYD88 mutations nor mutations in BCL2 or any other BCL2 family member were identified. Neither amplification nor translocation involving the BCL2 gene was detected. Potential genetic drivers of this malignancy included previously described ARID1A and NOTCH2 mutations (Table 1), as well as biallelic deletion of RB1 and CDKN2A and monoallelic loss of TP53 (Fig 2). The interplay between the NOTCH and NF-κB pathways is well described and similar to ABC DLBCL; constitutive activity of NF-κB may have an important role in PCDLBCL-LT. Alterations involving cyclin dependent kinase inhibitor 2A (CDKN2A), cyclin dependent kinase inhibitor 2B (CDKN2B), tumor protein P53 (TP53), and RB transcriptional corepressor 1 (RB1) resulting in deregulation of cell proliferation have been well described in DLBCL.¹⁶ Furthermore, genetic and epigenetic events that inactivate tumor necrosis factor alpha-induced protein 3 (TNFA1P3), resulting in transcriptional down-regulation, as shown in ABC DLBCL, may play a role in PCDLBCL-LT, but they were undetected on whole-exome DNA sequencing.

TABLE 1. Mutations Identified by Whole-Exome Sequencing of Primary Cutaneous Large B-Cell Lymphoma, Leg Type Tumor Biopsy Present in the COSMIC Database

View larger version (394K)

FIG 2. (A, B) Somatic copy number analysis of the tumor biopsy was carried out using Varscan and Sequenza programs against patient germline DNA. Frequency of copy number (log₂ ratio) obtained by pairwise comparison of read depths along the sequenced exome. Genes of interest are labeled. Biallelic loss of the tumor suppressor genes CDKN2A and CDKN2B as well as RB1 were observed along with monoallelic loss of TP53; however, no mutation of the remaining allele was detected.

In this report, we demonstrate that single-agent venetoclax induced a complete response without significant toxicities in a case of MYD88 wild-type PLDLBCL-LT. The factors contributing to successful outcome in this case remain unknown. Specifically, whether similar efficacy will be seen in PCDLBCL-LT with MYD88 mutation and whether high expression of BCL2 alone is a predictive biomarker remain unknown. However, it is interesting to note that a case of PLDLBCL-LT treated with single-agent ibrutinib developed chromosomal amplification of 18q21.3 containing the BCL2 gene at disease progression, again indicating a central role for BCL2 overexpression in the pathogenesis of this rare disease.¹⁰ Exploration of combination targeted treatment strategies are warranted in this patient population; however, the low disease incidence may present challenges in the design of appropriately powered clinical trials.