Gray zone lymfoom

WHO 2022 Mediastinal gray zone lymphoma (MGZL) is a B-cell lymphoma with overlapping features between primary mediastinal B-cell lymphoma (PMBL) and classic Hodgkin lymphoma (CHL), especially nodular sclerosis CHL (NSCHL).

Mediastinale gray zone lymfoom (MGZL)

Gedefinieerd als een B-cel lymfoom met kenmerken van een primair mediastinaal B-cel lymfoom (PMBCL) en klassiek Hogdkin lymfoom (HL).
MGZL komt, net als PMBCL en HL, typisch voor tussen de 30-40 jaar. Bij MGZL is 2/3 man.

Het MGZL heeft waarschijnlijk dezelfde precursor (thymic B‐cel) als PMBCL en HL, wat verklaart dat een relapse lymfoom zich ook als de andere entiteit kan uiten. Het exacte mechanisme voor de transformatie van de B-cel is niet opgehelderd. De gen-expressie van PMBCL en HL zijn gelijk, zoals gains in chromosoom 9p en 2p.

Diagnose

A spectrum of morphologies known with cHL and PMBCL can occur and divergent morphological areas can be seen within the same tumour, necessitating extensive sampling for the correct diagnosis. Similar to morphology, the immunophenotype of GZL has transitional features between cHL and PMBCL. Tumours resembling cHL will show prominent CD20, weaker/absent CD30 and absent CD15 whereas tumours resembling PMBCL will be strongly positive for CD30 and CD15 with negative CD20 and CD79a. As the B‐cell programme is preserved, B‐cell transcription factors, such as PAX5, OCT2 and BOB1, are positive in neoplastic cells. In comparison to a cohort of 51 patients with PMBCL, MGZL patients were more often male, expressed CD15, and had lower expression of CD20.

In a retrospective multicentre study of 112 GZL patients, the immunophenotype of MGZL appeared concordant with prior studies. In addition, there was no apparent significant difference in immunophenotype between MGZL versus NMGZL. The most prevalent immunophenotypes were: CD20+ (93%), CD30+ (89%), CD79a+ (78%), PAX5+ (98%), OCT2+ (96%) and MUM1+ (100%). CD15 (46%) and CD45 (67%) staining were more variable, whereas only 11% and 28% of patients were positive for CD10 and Epstein–Barr virus (EBV), respectively

Overall, the diagnosis of GZL remains challenging; workup and/or consultation at a centre with haematopathology expertise of this particular entity is highly recommended. We advocate detailed review of B‐cell markers (e.g. CD20, PAX5, CD79a, BOB1, OCT2) along with CD30, CD15 and MUM1 when GZL is suspected. Furthermore, for T‐cell markers, CD3 should be included as a minimum in the workup of every case presenting with sheets of large CD30 positive and CD20 negative cells to partly rule out anaplastic large cell lymphoma. EBER in situ hybridization should also be included. Collectively, the diagnosis of GZL should be considered if strong expression of CD20 is identified in cHL or strong CD15 expression is seen in an otherwise typical PMBCL.

Behandeling MGZL

Eerstelijns behandeling

There are no standard management guidelines for GZL patients in the up‐front or the relapsed/refractory setting. This is partly due to the small number of analyses examining therapeutic approaches for GZL as well as challenges in diagnosis, as discussed before. From the aforementioned prospective NIH study, 24 mediastinal GZL patients were treated with dose‐adjusted etoposide, prednisone, oncovin, cyclophosphamide, doxorubicin and rituximab (DA‐EPOCH‐R) and outcomes were compared with similarly treated PMBCL patients (Wilson et al, 2014). MGZL patients had significantly inferior outcomes compared with PMBCL [5‐year event‐free survival (EFS) of 62% vs. 93% and 5‐year overall survival (OS) 74% vs. 97%, respectively] (Dunleavy et al, 2011; Wilson et al, 2014).

Among the 112 GZL patients treated across 19 North American centres over a recent ten‐year period (Evens et al, 2015), the two most common therapeutic approaches were cyclophosphamide, doxorubicin, oncovin and prednisone (CHOP) +/− rituximab and doxorubicin, bleomycin, vinblastine and dacarbazine (ABVD) +/− rituximab, with only a handful of patients treated with DA‐EPOCH‐R. Approximately two‐thirds of patients received rituximab as part of their frontline therapy. The overall response rates (ORR) and complete remission (CR) rates for all patients were 71% and 59%, respectively, with 33% of patients having primary refractory disease (no difference identified between MGZL versus NMGZL). The ORR and CR rates for patients who received rituximab as part of frontline therapy were 82% and 73%, respectively, versus 59% and 43%, respectively, without rituximab (P = 0·02 and P = 0·008, respectively). At 31 months median follow‐up, the 2‐year PFS and OS for all patients were 40% and 88%, respectively. The 2‐year PFS rate for patients with stage I/II versus III/IV disease was 54% vs. 30%, respectively, and the corresponding OS rates were 94% vs. 76%. Interestingly, despite a comparative preponderance of early‐stage disease and lower IPI, the survival rates were not statistically different for patients with MGZL versus NMGZL.

In terms of frontline chemotherapy regimens in this retrospective series, PFS rate appeared significantly inferior for patients treated with ABVD +/− rituximab versus treatment with a DLBCL‐based regimen (i.e., CHOP+/‐R and DA‐EPOCH‐R) with corresponding 2‐year rates of 23% vs. 52%, respectively. It is important to highlight that this finding persisted on multivariate Cox regression, including when controlling for IPI and receipt of rituximab.


Rol voor radiotherapie?

There are minimal prospective, and no randomized, data regarding the utility of radiotherapy in GZL. Treatment paradigms for GZL in the series reported by Evens et al 2015 resembled typical PMLCL and cHL management with radiotherapy usually given to patients with bulky disease after R‐CHOP. Radiotherapy was not routinely given for patients treated with DA‐EPOCH‐R in the series reported by Wilson et al 2014, as is standard practice with this therapeutic regimen for patients with PMLCL (Dunleavy et al, 2013). In the series by Evens et al 2015, more patients with MGZL received radiotherapy versus NMGZL; this was probably dictated by the much higher frequency of early‐stage and bulky disease for MGZL patients in that analysis. When comparing response and survival rates, there did not appear to be a benefit to radiotherapy, however, caution should be applied to this finding given the retrospective nature and relatively small patient numbers.

Notably, relapses localized to the mediastinum may be curative. Wilson et al 2014 demonstrated that 44% (four of nine) patients achieved continuous remission with salvage radiation therapy alone. It is not known if outcomes would be improved for MGZL if adjuvant radiotherapy was more routinely used following DA‐EPOCH‐R therapy. However, given the currently defined outcomes for newly diagnosed GZL, in what appears to be a relatively chemotherapy‐resistant disease, we advocate use of consolidative radiotherapy after chemo‐immunotherapy for localized and/or bulky (>10 cm) disease.

Relapse/refractaire ziekte

Among available data, the ability to salvage GZL patients with relapsed/refractory disease appears fairly good.
In the series reported by Evens
et al 2015, the median time to relapse for GZL patients was 7 months (range, 1–64 months); most of these patients were treated with standard combination chemotherapy salvage regimens. Furthermore, the majority of patients were taken to haematopoietic stem cell transplantation (HSCT), usually autologous. The 2‐year OS rate for relapsed/refractory GZL patients who underwent HSCT was 88%, which compared with 67% for patients who did not have HSCT; this persisted on multivariate analyses controlling for IPI and response to pre‐HSCT salvage therapy. This improvement, however, probably reflected that fit patients with chemotherapy‐sensitive disease underwent SCT. Nevertheless, given these findings, there should be consideration for salvage combination chemotherapy followed by autologous HSCT for patients with relapsed/refractory GZL. Additionally, as described before for localized relapse in the mediastinum, there may be consideration for radiotherapy (+/− chemotherapy) without HSCT.

Clinical data regarding the use of biological agents for GZL is highly limited. Brentuximab vedotin is an attractive targeted therapy, given the CD30 positivity seen in many GZL cases. In a small subset of relapsed/refractory GZL patients, brentuximab vedotin demonstrated activity in a CD30+ relapsed/refractory DLBCL study (i.e., one CR and two partial remissions in six patients) (Jacobsen et al, 2015). Functionally active transcription pathways have not been clearly delineated in GZL, however data from PMBCL may be extrapolated. NF‐κB is constitutively active in PMBCL (Feuerhake et al, 2005) and thus inhibitors to this pathway, as well as treatment approaches targeting B‐cell receptor (BCR)‐mediated activation should be considered (Gebauer et al, 2014). Additionally, checkpoint inhibition may be a valid pathway to partially pursue, given the previously identified amplification of chromosome 9p24.1 in PMBCL, which was associated with activation of Janus kinase 2 (JAK2) and programmed cell death 1 ligand (Green et al, 2010). These data are also supported by findings from the NIH where alterations affecting the JAK2/PDL2 (also termed PDCD1LG2) locus in 9p24.1 were seen in the majority of MGZL cases (Eberle et al, 2011).

Prognose

Previous retrospective studies (including from 2 to 112 cases) showed that patients with an MGZL had a poorer outcome than patients with a PMBCL or diffuse large B-cell lymphoma (DLBCL). The poor prognosis was confirmed by the prospective study of the National Cancer Institute that included 24 MGZL cases treated with rituximab and an intensified chemotherapy regimen (namely DA-EPOCH-R: dose adapted etoposide, prednisolone, oncovin, cyclophosphamide, doxorubicin) with a shorter progression-free survival (PFS) and overall survival (OS) as compared with PMBCL. The largest previously reported series of GZL between diffuse large B-cell lymphoma (DLBCL) and CHL included 112 patients, 43% with a mediastinal GZL and 57% without mediastinal involvement. At a 31-month median follow up, 2-year PFS and OS rates were 40% and 88%, respectively. Furthermore, they observed that rituximab combined with chemotherapy could improve patient prognosis.

The prospective DA‐EPOCH‐R study identified serum absolute lymphocyte count, presence of tumour‐infiltrating dendritic cells, CD15 immuno‐expression on malignant cells and tumour morphology as prognostic makers for MGZL outcome (Wilson et al, 2014). Due to the smaller size of this study, these associations were identified on univariate analyses. Predominant PMBL‐like morphology and presence of CD68+ tumour‐associated macrophages were associated with EFS. Additionally, they showed that a gene signature encoding for dendritic cell–specific intercellular adhesion molecule‐3‐grabbing non‐integrin (DC‐SIGN, also termed CD209), which is a marker of dendritic cells and activated macrophages, was associated with OS. Furthermore, low absolute lymphocyte count and increased staining of CD15 in malignant cells were both associated with inferior EFS and OS (Wilson et al, 2014).

Analysis of clinical prognostic factors in the most recent series showed that poor performance status, increased lactate dehydrogenase (LDH), anaemia and advanced stage disease were associated with PFS on univariate analyses, while only stage was prognostic for OS (Evens et al, 2015). Both the IPI and IPS scoring systems were also prognostic for PFS and OS (as categorical and continuous variables). Multivariate analyses identified performance status as the most dominant prognostic factor for PFS, whereas increased LDH was borderline. For OS, advanced stage disease (i.e., III/IV) was the only significant prognostic factor (i.e., hazard ratio, 4·89; P < 0·05). Interestingly, CD20 positivity in malignant cells also independently predicted PFS on Cox regression controlling for receipt of rituximab (i.e., 88% of the CD20‐negative GZL patients experienced progressive disease).