Primary Mediastinal Large B-cell Lymphoma: A Diagnostic Conundrum Open Access

Primary mediastinal large B-cell lymphoma (PMBCL) accounts for 10% of all large B-cell lymphomas and up to 4% of all non-Hodgkin lymphomas. Median age at diagnosis is 35 years, with a female predilection (ratio 2:1). The presence of a mediastinal mass can cause dyspnea, or lead to compressive symptoms including superior vena cava syndrome. Extension into the thorax with consequent lung, pericardial, pleural or chest wall involvement can also be seen. Extra-thoracic disease, affecting the kidneys or adrenal glands, is usually found in less than 10% of patients [1]. Relapsed PMBCL can involve these organs in addition to the central nervous system. PMBCL was historically deemed as a subtype of diffuse large B-cell lymphoma (DLBCL) based on its clinicopathologic characteristics. However, molecular profiling unraveled a peculiar gene expression signature, substantially overlapping with that of classic Hodgkin lymphoma. Consequently, PMBCL was given the status of a distinct entity in the WHO classification due to its characteristic immunophenotypic and gene expression profile. Repeated genetic aberrations result in activation of NF kappa B and JAK-STAT pathways. Additionally, unique structural changes resulting in over-expression of PDL1 (CD274) and under-expression of major histocompatibility complex class 2 (MHC-2) create a distinct immune evasion phenotype [2].

Current guidelines vary considerably, and no consensus exists regarding the optimal frontline therapy. Evidence supports the utilization of either dose-adjusted etoposide, prednisone, vincristine, cyclophosphamide, doxorubicin, and rituximab (DA-EPOCHR) or rituximab, hyperfractionated cyclophosphamide, vincristine, Adriamycin, and dexamethasone (R-Hyper-CVAD) regimens owing to significantly better clinical outcomes than rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP) [3]. Presence of an immune-privileged phenotype in the disease biology of PMBCL served as the premise for utilization of PDL1 inhibitors in this setting. Majority of patients respond to rituximab chemoimmunotherapy, but up to 30% of patients relapse or have refractory disease. Studies are underway to evaluate whether inclusion of PD1 in the initial treatment regimen is superior to the current treatment approach. Novel radiomic biomarkers can also help risk-stratify patients and discern the subset that is more likely to respond to alternate therapies [4]. Here we report the case of a 39-year-old male with no past medical history who presented with cough and dyspnea secondary to a bulky mediastinal mass that had resulted in superior vena cava syndrome and cardiac tamponade. The tumor was mistakenly diagnosed as sarcoma based on available clinical and histopathological data. Eventually, next-generation sequencing (NGS) and immunohistochemistry (IHC) revealed a molecular signature that was consistent with PMBCL.

A 39-year-old male with no past medical history presented with a 2-month history of unremitting productive cough and exertional dyspnea. The week prior to presentation, there was a precipitous decline in his functional status wherein he had been feeling extremely light headed and short of breath upon walking just a few steps inside his house. He had trouble laying flat due to severe shortness of breath and cough, forcing him to sleep in a chair at a 90 degree angle. He had recently started carrying disposable sputum bags around due to his voluminous phlegm. He noted having generalized pruritus and drenching night sweats for a few weeks. He had unintentionally lost 10 pounds over the past 2 months. He had also experienced several episodes of epistaxis, severe headaches, and facial ruddiness during his bothersome coughing bouts. Family and psychosocial histories were noncontributory. Physical examination was notable for jugular venous distension and pulsus paradoxus. On admission, complete blood count and serum chemistries were within normal limits. EBV, HIV, and hepatitis panels were negative. LDH (282 U/L) and beta-2 microglobulin (4.1 mg/dL) were elevated. CT chest revealed a mediastinal mass (Fig. 1), measuring 10.6 × 9.5 cm, compressing the SVC with associated intraluminal thrombosis and a large pericardial effusion. The patient received decadron and underwent pericardiocentesis followed by a tissue biopsy. Anticoagulation was also started with subsequent symptomatic improvement.

Histopathologic examination revealed atypical spindle-shaped and pleomorphic cells in a background of dense fibrosis, interspersed with areas of necrosis and crush artifact (Fig. 2). Cytology was difficult to interpret due to the latter findings, but a tentative diagnosis of spindle-cell neoplasm was made while molecular diagnostics were pending. Comprehensive IHC staining (Fig. 3-7) revealed that the atypical cells were positive for CD20, PAX5, CD79a, CD30, CD23, MUM1 and weakly positive for MAL (myelin and lymphocyte) protein. They also tested positive for some myogenic markers (SMA and caldesmon). They were negative for CD15, BCL6, cytokeratin, S100, and CD34. EBER (Epstein-Barr encoding region) testing by in situ hybridization was also negative. NGS was performed using the Illumina NextSeq 550 system platform, which tested for a targeted transcriptome panel. Increased expression of TNFRSF8 and CD274 genes, which encode CD30 and PDL1 proteins, respectively, conformed to the immunoprofile of Hodgkin lymphoma. Low level somatic mutations were discovered in SAMD9L, BRD4, CBLB, MSH2, BCCRL1, SOCS1, and IRF4 genes with a VAF ranging from 3.8 to 16.87%. These findings collectively allowed us to arrive at a diagnosis of PMBCL. The patient was subsequently started on the R-Hyper-CVAD protocol (3A [hyperfractionated cyclophosphamide, vincristine, Adriamycin, and dexamethasone] alternating with 3B [cytarabine and methotrexate]). The patient has responded well to chemotherapy, and his cough and shortness of breath has improved markedly. He is now able to carry out activities of daily living and job duties without significant discomfort. His treatment course has been uneventful, and he has tolerated chemotherapy well thus far without developing any serious complications. Interim PET scan showed heterogeneous distribution of tracer within the anterior mediastinal mass that has marginally diminished in size [maximal SUV 4.5; Deauville score 3]. End of treatment-PET imaging is yet to be scheduled.

Our case had an interesting phenotype due to discordant histopathological and molecular profiles. On histopathological examination, PMBCL is typically characterized by diffuse proliferation of large cells, sectionalized by fine sclerotic bands. It can, however, present with varying morphologies of neoplastic cells, often exhibiting multilobulated cells resembling Hodgkin Reed-Sternberg cells. Diagnosis is challenging as biopsy specimens are often subpar due to tumor location, low yield, and/or crush artifact. Presence of compartmentalizing fibrosis can sometimes provide a false histopathologic impression of carcinoma or thymoma [1]. Histopathology was difficult to interpret in our case due to the presence of extensive fibrosis and crush artifact. Ample tissue sampling is therefore required to undertake comprehensive immunohistochemical analyses and prevent false-negative results, which frequently is not possible. PMBCL must also be differentiated from DLBCL with mediastinal involvement and nodular sclerosis subtype of cHL [Table 1]. Core biopsies are occasionally inadequate to distinguish PMBCL from mediastinal gray zone lymphoma, thereby requiring larger biopsies.

PMBCL is characterized by overexpression of genes in the JAK-STAT (IL13RA, JAK2, and STAT1) and NF-kB (TRAF1, TFNAIP3) pathways but underexpression of genes involved in B-cell receptor signaling. PMBCL tumors express B-cell lineage antigens, namely, CD45, CD19, CD20, and CD22. They lack surface/cytoplasmic immunoglobulin (Ig) but do possess the Ig coreceptor CD79a. MUM1/IRF4+ phenotype is commonly encountered with variability in BCL6 expression. CD10, CD15, and CD21 are negative, whereas CD30 is positive in up to 80% of cases. Positivity for CD200, CD23, TRAF1, nuclear cREL, and MAL, a lipid raft component, can distinguish PMBCL from DLBCL. MAL antibody (specificity 100%) is superior to CD200 (specificity 87%) for PMBCL diagnosis. PDL1 and PDL2 expression in PMBCL tumors is related to 9p24.1 alterations [2]. IHC staining showed that the tumor was positive for CD20, PAX5, CD79a, CD30, CD23, and MUM1. The atypical cells were, however, weakly positive for MAL and negative for CD15 and BCL6. NGS in our case revealed dense expression of TNFRSF8 and CD274 genes, which are responsible for production of CD30 and PDL1 proteins, respectively.

Treatment approaches vary considerably due to lack of unifying guidelines. Utilization of an intensive R-Hyper-CVAD regimen in our patient resulted in a favorable treatment response without any serious side effects. Studies have shown that R-CHOP is associated with subpar survival outcomes in patients with high-risk aggressive B-cell lymphomas, including PMBCL, when compared with R-Hyper-CVAD and rituximab, cyclophosphamide, vincristine, etoposide, and prednisone (R-CHOEP). Augmentation of R-CHOP with high-dose methotrexate and cytarabine results in greater overall survival (OS) and progression-free survival (PFS) rates when compared with R-CHOP alone. In the Swedish Lymphoma Register Study, the 2-year relative survival rates for PMBCL patients in DA-EPOCH-R, R-CHOEP-14, and R-Hyper-CVAD arms were 82%, 95%, and 100%, respectively. R-CHOP-14/21 arm, on the other hand, had the lowest survival rate of 74% [5]. Another Swedish study analyzing survival outcomes in 401 patients with high‐risk aggressive B‐cell lymphomas established the superiority of R‐Hyper‐CVAD over RCHOP/R‐CHOEP in both DLBCL (n = 285) and PMBCL (n = 40) subgroups. Five‐year OS and PFS in PMBCL patients receiving R‐Hyper‐CVAD were 95% and 89%, respectively. These survival estimates were significantly lower at 89% and 68%, respectively, in patients on R-CHO(E)P therapy [6]. Head-to-head comparisons among DA-EPOCHR and R-CHOP, with RT, have been carried out in three small retrospective studies. DA-EPOCHR was deemed superior in only one study [5]. No difference was found between these regimens in the other two studies [7, 8]. Evidence on the effectiveness of consolidative RT in PMBCL after R-CHOP is equivocal. DA-EPOCHR notably lessens the need for consolidative RT. Lowering RT-associated adverse sequelae, like secondary cancers and premature cardiovascular disease, assumes great importance, especially in younger patients. Long-term effects of PET-guided modern RT techniques remain unknown. PET scans have been recently employed to ascertain the need for consolidative RT after R-CHOP or R-CHOP-like therapies in PMBCL. Several studies have demonstrated high negative predictive values of negative end-of-treatment PET scans after R-CHOP without RT and DA-EPOCHR, ranging between 92–97% and 95–100%, respectively [9, 10]. It is unclear as to which regimen yields a higher negative predictive value, and more well-powered studies are needed to answer that question. The largest randomized non-inferiority trial (IELSG 37) evaluating the utility of consolidative RT with an end of treatment-PET negative scan after completion of rituximab chemotherapy found that omission of consolidative RT did not impair survival in PMBCL patients who had achieved complete metabolic response [11]. Majority of R-CHOP-like chemotherapy recipients can be effectively managed without RT. Nevertheless, use of RT in these patients occurs more frequently compared with DA-EPOCH recipients. DA-EPOCH is preferred at some centers in younger patients with bulky disease as it obviates the need for mediastinal RT. PET radiomics can help predict prognostic markers in PMBCL patients. The ILESG 26 study found that metabolic heterogeneity and high total lesion glycolysis values were associated with reduced PFS [4].

Most disease recurrences happen within the first 2 years, barring central nervous system relapses, which tend to affect up to 4.5% of patients and can occur after 2 years. Relapsed/refractory (R/R) PMBCL is treated similarly to DLBCL, wherein salvage therapy is followed by high-dose chemotherapy and autologous stem cell transplant (ASCT) in responders. Few studies have evaluated R/R PMBCL outcomes, especially after rituximab use was mainstreamed. A recent study from Memorial Sloan Kettering Research Center showed a 3-year PFS and OS rate of 57% and 61% in the intent-to-transplant population. These survival rates were marginally to 60% and 65% in patients who underwent transplantation [12]. ASCT has been shown to increase OS in other analyses as well [13].

CAR-T cell therapy has revolutionized the landscape of R/R PMBCL and is FDA approved for these patients after failure of two treatment lines, including ASCT. Both ZUMA-1 and TRANSCEND NHL001 studies demonstrated excellent response to CAR-T cell chemotherapy in PMBCL patients with objective response rates exceeding 70% [14, 15]. Pembrolizumab is also FDA approved for R/R PMBCL adult and pediatric patients who have failed two lines of therapy. Several PD1 inhibitors are currently being evaluated for upfront use in PMBCL.

PMBCL can present as a diagnostic conundrum due to its significant mechanistic and histopathologic overlap with other malignancies. Molecular classifiers serve as robust differentiation markers that help eliminate subjectivity in the diagnostic process. Molecular signature of PMBCL must be factored into determination of the optimal immunochemotherapy regimen.

This retrospective review of patient data did not require ethical approval in accordance with local/national guidelines. Written informed consent was obtained from the patient for publication of the details of their medical case and any accompanying images. The CARE Checklist has been completed by the authors for this case report, attached as online supplementary material (for all online suppl. material, see https://doi.org/10.1159/000545931).

The authors declare no conflicts of interest.

This study did not receive support from any sponsor.

Yazan Alrefai and Shruti Wadhwani: conceptualization, data curation, and original draft. Nikita Wadhwani: data curation, original draft, reviewing, and editing. Ayton Bangolo: conceptualization, supervision, and reviewing. Jason Mizrahi: reviewing and editing. Tatyana Feldman: conceptualization and supervision. All authors read and agreed to the final version of this manuscript.

Yazan Alrefai and Shruti Wadhwani: co-first authors.

All pertinent data obtained for the purpose of publishing this case report is included herein. Further inquiries can be directed to the corresponding author.