Clinicopathologic consensus study of gray zone lymphoma with features intermediate between DLBCL and classical HL

Abstract

Gray zone lymphoma (GZL) is described as sharing features with classical Hodgkin lymphoma (cHL) and diffuse large B-cell lymphoma (DLBCL). However, there remains complexity in establishing diagnosis, delineating prognosis, and determining optimum therapy. Sixty-eight cases diagnosed as GZL across 15 North American academic centers were evaluated by central pathology review to achieve consensus. Of these, only 26 (38%) were confirmed as GZL. Morphology was critical to GZL consensus diagnosis (eg, tumor cell richness); immunohistochemistry showed universal B-cell derivation, frequent CD30 expression, and rare Epstein-Barr virus (EBV) positivity (CD20⁺, 83%; PAX5⁺, 100%; BCL6⁺, 20%; MUM1⁺, 100%; CD30⁺, 92%; EBV⁺, 4%). Forty-two cases were reclassified: nodular sclerosis (NS) cHL, n = 27 (including n = 10 NS grade 2); lymphocyte predominant HL, n = 4; DLBCL, n = 4; EBV⁺ DLBCL, n = 3; primary mediastinal large BCL n = 2; lymphocyte-rich cHL and BCL-not otherwise specified, n = 1 each. GZL consensus-confirmed vs reclassified cases, respectively, more often had mediastinal disease (69% vs 41%; P = .038) and less likely more than 1 extranodal site (0% vs 25%; P = .019). With a 44-month median follow-up, 3-year progression-free survival (PFS) and overall survival for patients with confirmed GZL were 39% and 95%, respectively, vs 58% and 85%, respectively, for reclassified cases (P = .19 and P = .15, respectively). Interestingly, NS grade 2 reclassified patients had similar PFS as GZL consensus-confirmed cases. For prognostication of GZL cases, hypoalbuminemia was a negative factor (3-year PFS, 12% vs 64%; P = .01), whereas frontline cyclophosphamide, doxorubicin, vincristine, and prednisone ± rituximab (CHOP±R) was associated with improved 3-year PFS (70% vs 20%; P = .03); both factors remained significant on multivariate analysis. Altogether, accurate diagnosis of GZL remains challenging, and improved therapeutic strategies are needed.

Graphical abstract

Figure 1.

Consensus pathology. (A) Case example of consensus GZL diagnosis (B-cell lymphoma with features intermediate between diffuse large B-cell lymphoma and classical Hodgkin lymphoma). Sheetlike proliferation of tumor cells resembling PMBL is seen in the top left panel. Focally, collections of eosinophils are noted (top right). The tumor cells are strongly positive for CD20, CD30, and CD15 (bottom). (B) Reclassified case: cHL NS2. Representative images from 2 cases of cHL NS2 are shown. (Top) Characteristic histology with a nodular growth pattern, fibrous bands, and a sheetlike growth of lacunar cells. (Bottom) Immunohistochemical stains from a second case show the nodular growth pattern with strong staining for CD20, but also positivity for CD30 and focal staining for CD15. (C) Reclassified case: primary mediastinal large B cell lymphoma. Diffuse proliferation of large neoplastic cells is seen (top left). The neoplastic cells have abundant clear cytoplasm and vesicular nuclei with basophilic nucleoli (top right). They are strongly, uniformly positive for CD20, with dim expression of CD30, and are negative for CD15 (bottom). (D) Synopsis of the immunohistochemical staining results from consensus confirmed GZL cases. The bar graph illustrates immunohistochemical staining for markers CD20, CD79a, PAX5, MUM1, CD30, CD15, CD3, and in situ hybridization for EBV (EBV-encoded small RNA [EBER]). For immunohistochemical studies, staining was considered to be positive if it had intensity of 2 or more (on +1 to +3 scale) and was distributed in more than 25% of neoplastic cells. Cases with weak/negative expression are not included in this graph. Hematoxylin and eosin staining, original magnification ×400 (A, B [top right panel], and C [top panels]) and ×200 (B [top left panel]). Immunohistochemical staining, original magnification ×400 (all other panels).

Figure 2.

Survival. The 3-year (A) PFS of 25 patients with GZL compared with 36 reclassified lymphoma cases was 39% and 58%, respectively (P = .19), and corresponding 3-year (B) OS was 95% and 85%, respectively (P = .15). Kaplan-Meier curves comparing PFS (C) and OS (D) for cHL NS2 vs cHL vs DLBCL; 3-year PFS rates were 43%, 65%, and 67% (P = .67), respectively, and 3-year OS rates were 83%, 88%, and 71%, respectively (P = .80). The outcome (E) for patients with GZL based on CD30 expression was 3-year PFS of 83% for Neg-1 vs 34% for 2-3 on immunohistochemistry; (F) the 3-year PFS for patients with hypoalbuminemia vs normal albumin were 64% vs 12% (P = .01). Kaplan-Meier curves (G) for patients who received CHOP±R therapy for frontline therapy vs not; 3-year PFS was 70% vs 20%, respectively (P = .03). The 2 latter findings persisted on multivariable Cox regression analysis.

Figure 3.

GZL diagnostic algorithm. The minimum diagnostic panel for workup of GZL should include B-cell markers (CD20 and PAX5), MUM1, CD30, and CD15. A broad panel of T-cell and cytotoxic markers is desirable to rule out anaplastic large cell lymphoma. EBV by in situ hybridization should also be a part of the diagnostic panel. (Top) Stepwise approach to the diagnostic evaluation of GZL. (Lower balloon) depicts potential pitfalls in the diagnostic evaluation of GZL to also consider. FNA, fine needle aspiration.