免疫球蛋白轻链 mRNA 原位杂交可细化 NLPHL–cHL 谱系分类Immunoglobulin Light Chain mRNA In Situ Hybridization Refines Classification Across the NLPHL-cHL Spectrum.

Abstract

Nodular lymphocyte-predominant Hodgkin/B-cell lymphoma (NLPHL/BL) and classic Hodgkin lymphoma (cHL) are biologically distinct entities, exhibiting preserved versus lost B-cell differentiation and immunoglobulin (IG) transcription. However, cases with overlapping morphologic and immunophenotypic features between NLPHL and lymphocyte-rich cHL (cHL-LR), here designated as NLPHL-OL, remain biologically ambiguous. We applied a novel, highly sensitive dual κ/λ mRNA in situ hybridization assay to assess IG light-chain expression across this spectrum. We analyzed 53 lymph node biopsies, including NLPHL-OL cases (n =15), cHL-LR (n = 11), along with typical NLPHL (n = 13) and cHL of other subtypes (n = 14) as controls. Distribution, intensity, and patterns of κ/λ mRNA expression in neoplastic cells was systematically evaluated and correlated with histology, architecture, immunophenotype, and Epstein-Barr virus (EBV) status. Diffuse, strong cytoplasmic light chain mRNA expression, representing the canonical pattern, was detected in the tumor cells of all NLPHL (13/13) and the majority of NLPHL-OL cases (12/15, 80%). In contrast, faint perinuclear or granular cytoplasmic signals (non-canonical patterns) were detected in 6/11 (55%) cHL-LR and 4/14 (28%) of other cHL subtypes. Canonical κ/λ expression was thus restricted to NLPHL and NLPHL-OL, confirming their classification as NLPHL with aberrant features, whereas cHL was negative or displayed exclusively non-canonical patterns. Non-canonical staining was also seen in 3/15 (20%) NLPHL-OL, including two EBV-positive cases. Rare aberrant light chain profiles in NLPHL and NLPHL-OL cases included single-cell κ/λ coexpression, and distinct LP-cell populations with mutually exclusive κ and λ expression. In summary, κ/λ mRNA ISH refines diagnostic classification across the NLPHL-cHL spectrum, is particularly useful in cases with overlapping features, and reveals canonical, non-canonical, and negative staining patterns that may reflect B-cell program integrity, placing these cases along a biological continuum.

Introduction

Classic Hodgkin lymphoma (cHL) and nodular lymphocyte-predominant Hodgkin/B-cell lymphoma (NLPHL/BL) are two entities recognized in the WHO 5 th Edition and ICC 2022 classifications. 1(p5),2 The historical use of the shared term “Hodgkin” reflects their morphological resemblance - large, atypical tumor cells embedded in a dense inflammatory milieu that both shapes tissue architecture and provides critical survival signaling cues to the neoplastic population. However, cHL and NLPHL show marked clinical, histologic, phenotypic and molecular differences and are regarded as biologically distinct. cHL exhibits a disrupted B-cell transcriptional program reflecting a “crippled” pre-apoptotic B cell, whereas NLPHL has the phenotype and gene expression profile of a germinal center B cell. The key distinction is considered the presence or absence of immunoglobulin transcription; the latter precludes physiologic post-germinal center light-chain maturation. 3–7 Evidence that a subset of NLPHL cases expresses B-cell receptors recognizing antigens of common bacterial pathogens such as Moraxella catarrhalis and thus are putatively antigen-driven supports the notion that lymphocyte-predominant (LP) cell immunoglobulin remains functional. 8

Despite their contrasting biology, a morphological and immunophenotypic gray zone exists between NLPHL and cHL, encompassing NLPHL with aberrant marker profile and the lymphocyte-rich subtype of cHL (cHL-LR). cHL-LR may closely resemble NLPHL morphologically and expresses B-cell transcription factors such as OCT2, BOB.1, and BCL6 more frequently than other cHL subtypes. 9 In addition, a follicular helper T-cell (TFH) microenvironment with rosetting by PD1+ or CD57+ T-cells can be seen in up to 50% of cHL-LR cases. 10 , 11 Conversely, NLPHL may express CD30 or CD15, exhibit variable downregulation of B-cell markers, or show Epstein-Barr virus (EBV) infection, features more typically associated with cHL. 12–15 Support for this gray zone also comes from cases with documented clonal relationships between synchronous cHL and NLPHL. 16

Although NLPHL and cHL are conceptually expected to differ in their immunoglobulin transcription, cases with overlapping features are rare and poorly studied. Flow cytometry, despite being an effective method for assessing immunoglobulin light-chain (IGL) expression in B-cell lymphomas, is poorly suited for cHL and NLPHL due to the scarcity and fragility of Hodgkin/Reed-Sternberg (HRS) and LP cells. 17 Likewise, in formalin-fixed paraffin-embedded (FFPE) tissue, assessment of light-chain restriction with immunohistochemistry is challenging because of background staining. 18

κ/λ mRNA in situ hybridization (κ/λ mRNA ISH) may provide a useful approach for assessing light-chain restriction in B-cell lymphomas, with diagnostic accuracy approaching that of flow cytometry. 19–24 Previous studies have shown that κ/λ mRNA ISH can help in distinguishing cHL from NLPHL, since LP cells in NLPHL retain functional B-cell differentiation and therefore show detectable κ/λ mRNA signals. 25–27 However, the expression of IGL mRNA has not been systematically explored in the context of cases with overlapping cHL-NLPHL features. 28 , 29 Making use of a recently available, highly sensitive dual κ/λ mRNA ISH, we specifically aimed to (i) validate previously reported differences in IGL expression between cHL and NLPHL; (ii) characterize light-chain restriction patterns; and (iii) explore the use of this novel technology for classifying challenging cases with overlapping cHL-NLPHL features.

Case Selection / Diagnostic Review

We retrospectively collected excisional lymph node and core needle biopsies, including cases submitted for second opinion, diagnosed either as showing overlapping features between NLPHL and cHL or as cHL-LR at the Institute of Pathology, University Hospital Tuebingen, Germany, between 2010 and 2025. Cases were included if representative FFPE tissue blocks or unstained slides were available for κ/λ mRNA ISH. Cases with overlapping features (NLPHL-OL) were defined as diagnostically challenging cases showing morphologic and/or immunophenotypic features atypical for conventional NLPHL. These included NLPHL with cHL-like features, such as CD30 and/or CD15 expression, partial loss of B-cell surface markers and/or transcription factors (CD20, CD79a, PAX5, OCT2, BOB.1), EBV positivity, preservation of germinal center architecture, purely interfollicular growth, or marked sclerosis. Additionally, to cover the full NLPHL–cHL spectrum, we included cases classified as cHL-LR, in which features suggestive of an intermediate phenotype - such as progressive transformation of germinal centers (PTGC)-like morphology, retention or unusually strong expression of B-cell markers, PD1+ T-cell rosettes, and a prominent TFH background - are well recognized. 10 , 11

Furthermore, cases of morphologically and phenotypically typical NLPHL and cHL of other subtypes were included as controls. All diagnoses were reassessed by three hematopathologists (JBS, FF, LQM) according to WHO 5 th Edition and ICC 2022 criteria. Cases were also subclassified according to current systems (cHL subtype, Fan et al. patterns, ICC grade in NLPHL). 1 , 2 , 30 , 31 Clinical data were retrieved from pathology request forms and institutional medical records. The study was conducted in accordance with the Declaration of Helsinki and approved by the local Ethics Review Committee.

Immunohistochemistry

Immunohistochemical studies were performed (if they had not already been part of the initial workup) on 2-3 μm FFPE sections using the Ventana Ultra automated staining system (Ventana Medical Systems, Tucson, AZ, USA) and Ventana reagents, following the manufacturer’s protocols. The complete antibody panel is provided in the Supplementary Methods. Staining intensity was semiquantitatively scored as negative (0), weak (1), moderate (2), or strong (3) in NLPHL-OL cases, and discrepant results were resolved by joint review.

κ/λ mRNA In Situ Hybridization

A detailed description of κ/λ mRNA ISH methodology and data interpretation is presented in the Supplementary Methods. Kappa and lambda light-chain mRNA expression was assessed by chromogenic dual in situ hybridization on FFPE tissue sections using the VENTANA Kappa and Lambda Dual ISH mRNA Probe Cocktail (Ventana Medical Systems, Roche Diagnostics), according to the manufacturer’s instructions, on a BenchMark ULTRA platform. Light-chain restriction was determined by evaluating the proportion of tumor cells expressing kappa or lambda mRNA. Mantle zone B-cells of a reactive on-slide tonsil, which show the lowest expression of IGL mRNA among mature reactive B-cells, served as external positive quality controls, while background plasma cells and B lymphocytes served as internal positive controls.

The percentage of positive LP or HRS cells, as well as intensity, light-chain mRNA type, and staining pattern of cytoplasmic κ/λ mRNA ISH signals were evaluated by multiple observers using both scanned digital slides and conventional glass-slide light microscopy. Only unequivocal cytoplasmic signals in clearly identifiable tumor cells were scored, and discordances were resolved through joint microscopic review. Signal intensity was scored as negative (0), weak (1), moderate (2), or strong (3). Light-chain mRNA expression was classified as kappa-restricted, lambda-restricted, or atypical. Staining patterns were categorized as (1) diffuse cytoplasmic in a high proportion of histologically defined tumor cells / canonical, (2) weak perinuclear rimming and/or faint granular punctate cytoplasmic in variable numbers of tumor cells / non-canonical, or (3) negative. Each case was classified into one or more of these patterns. κ/λ mRNA ISH was repeated in all cases with a non-canonical pattern for confirmation.

Results

Clinicopathologic Characteristics

Detailed clinicopathologic characteristics of the NLPHL-OL, cHL-LR, NLPHL and cHL groups are provided in Table 1 and Supplementary Tables S1-S4 .

∗ ) of available clinical data

A total of 53 lymph node biopsies were analyzed, predominantly excisional specimens, including NLPHL-OL (n = 15) and cHL-LR (n = 11) cases, as well as NLPHL (n = 13) and cHL (n = 14) controls. The median age was 41 years (range, 4-84), with an overall male predominance (male-to-female ratio approximately 2:1).

The NLPHL-OL cohort (n = 15) showed slight male predominance and a wide age range (median 47 years, range 19–84). The majority of cases with available clinical data presented in early stages (9/10). NLPHL-OL showed a broad range of Fan patterns, including pattern A and composite patterns combining low- and higher-risk features. In one case, classification according to Fan et al. was not feasible. Two cases (NLPHL-OL-4 and NLPHL-OL-7) had a history of prior cHL diagnosis or cHL-directed therapy. In NLPHL-OL-7, the prior diagnosis was revised to NLPHL; in NLPHL-OL-4, the previous biopsy was unavailable for review.

The cHL-LR group (n = 11) showed male predominance and a median age of 41 years (range, 15–80). The most frequent site of involvement was the cervical region (7/11). Mediastinal involvement was present in 2 of 6 cases with available clinical data. One patient was HIV-positive.

The NLPHL control cohort (n = 13) comprised non-consecutive cases selected to cover the morphological spectrum of NLPHL patterns. Patients showed marked male predominance and a median age of 37 years (range, 4–82). Most cases presented at early stages. Half of them predominantly displayed Fan pattern A (6/13, 46%), either alone or in combination with other low-risk patterns. Higher-risk patterns, particularly Fan D or E (6/13, 46%), were associated with splenic involvement and relapses. All cases showed strong CD20, OCT2, and MEF2B expression and PD1+ T-cell rosettes. One case (NLP-13) was positive for IgD.

The cHL control cohort (n = 14) showed male predominance and a median age of 35.5 years (range, 15–81). Nodular sclerosis (cHL-NS) was most frequent (8 cases), followed by 3 cases of mixed cellularity (cHL-MC). Both early- and advanced-stage disease was represented (5/10 each). Mediastinal involvement was common (8/11), particularly in cHL-NS (6/6), whereas splenic and bone marrow involvement were mainly seen in advanced disease. EBV positivity was present in all cHL-MC cases and occasionally in cHL-NS (3/8, 38%). Three cases were not subclassified. All 14 cases showed a typical cHL immunophenotype, with CD30-positive, MUM1-positive RS cells showing variable CD15 expression, weak PAX5 expression, and lack of CD20.

Histological and Immunophenotypic Features of NLPHL-OL and cHL-LR Cohort

The NLPHL-OL cohort showed a broad spectrum of architectural, cytological and phenotypical features overlapping NLPHL and cHL. The specific overlapping features of each case are detailed in Table 1 and Supplementary Tables S4 and S5 . PTGC-like changes were frequent (8/15, 56%), whereas preserved follicular architecture was less common and often focal (6/15, 40%). Sclerosis was present in a minority (3/15, 20%), two of the cases showing Fan pattern D. 32 T-cell rosettes and LP cells were seen in all cases (15/15, 100%). Hodgkin-like and Reed-Sternberg-like cells were present in 13/15 cases (87%) and 9/15 cases (60%) respectively, whereas mummified cells were present in a minority of cases (4/15, 27%).

Immunophenotypically, NLPHL-OL retained B-cell differentiation, with constant CD20 expression (15/15) and predominantly strong OCT2 (14/14) and MEF2B (12/12) positivity, whereas PAX5 was mostly weakly positive. CD79a and BOB.1 showed variable expression, while BCL6 was positive in all evaluable cases (12/12), with variable intensity. CD30 was expressed in the majority of cases (12/15, 80%), typically weak to moderate, and CD15 positivity was rare (2/15, 13%). EBV positivity was observed in 5/15 cases (33%); all cases expressed LMP1 (5/5, 100%) and were negative for EBNA2. Despite the unusual morphologic and immunophenotypic features, the strong CD20 and OCT2 expression together with MEF2B positivity favored the diagnosis of NLPHL in all cases.

The cHL-LR cohort showed PTGC-like changes in 1/11 (9%), a TFH-rich microenvironment in 2/11 (18%), and T-cell rosettes in 6/11 cases (55%). Mummified cells were present in 1 case. All cases uniformly showed strong CD30 expression (11/11, 100%), frequent CD15 positivity (9/11, 82%), and loss or marked downregulation of B-cell markers, with only weak PAX5 staining (9/9) and absence of CD20 expression in almost all cases (8/9, 89%). OCT2 was strongly expressed in 2 cases (18%). Strong MUM1 positivity was also common (9/9), whereas BCL6 was weakly positive in 5/8 evaluable cases. EBV positivity was observed in 4/11 cases (36%).

κ/λ mRNA ISH expression in NLPHL-OL, cHL-LR, NLPHL, and cHL

Detailed κ/λ ISH characteristics are summarized in Table 1 and Supplementary Tables S6-S9 and provide the framework for the analyses described below. The B-cell immunophenotypic profile of κ/λ mRNA ISH-positive cases is detailed in Table 2 .

Immunohistochemical staining intensity in large atypical cells was scored semiquantitatively as 0 (negative), 1 (weak), 2 (moderate), or 3 (strong). “–” indicates not assessed or not evaluable. * indicates rare single cells with weak positivity. Note that, despite non-canonical κ/λ mRNA ISH patterns, NLPHL-OL cases show stronger B-cell marker expression than cHL-LR and cHL cases.

κ/λ mRNA ISH was evaluable in all NLPHL-OL and cHL-LR, as well as in the control cohorts of NLPHL and cHL cases. Positive κ/λ ISH expression was observed in all NLPHL (13/13) and NLPHL-OL cases (15/15) ( Figure 1 ). κ/λ mRNA expression was identified in a high proportion of LP cells with a median of 100% in NLPHL and 90% in NLPHL-OL. In both groups the predominant staining pattern was diffuse cytoplasmic (canonical) - observed in 13/13 and 12/15 (80%), respectively - showing close alignment of κ/λ mRNA expression between these groups. NLPHL-OL cases with CD30 or CD15 positivity did not differ in their mRNA expression profile from conventional NLPHL ( Figure 2 ). Three cases of NLPHL-OL (20%) showed non-canonical staining, two of which were EBV-positive (NLPHL-OL-5 and -6). NLPHL-OL-5 was difficult to classify due to several aberrant features overlapping with cHL-LR ( Figure 3 ). NLPHL-OL-7 was also EBV-positive and showed κ/λ mRNA expression in only 50% of LP-cells; nevertheless, the tumor cells still had a diffuse/canonical pattern. Higher-risk or composite patterns retained canonical positivity but occasionally showed heterogeneity, including non-canonical staining in a subset of tumor cells and variable intensity.

In cHL-LR, 6/11 cases (55%) were positive for κ/λ mRNA; however, all of these showed a non-canonical staining pattern. By contrast, κ/λ mRNA ISH signals were completely negative in the majority of cHL (10/14, 71%), including cHL EBV-positive cases (5/6, 83%) ( Figure 4 ). cHL-NS and cHL-MC were less frequently positive than cHL-LR, with κ/λ mRNA expression observed only in a subset (3/8, 38% and 1/3, 33%, respectively). Overall, κ/λ mRNA ISH in cHL showed greater variability in the percentage of tumor cells and consistently displayed non-canonical staining patterns, including granular punctate and/or perinuclear rimming cytoplasmic signals; a canonical pattern was not observed. All cases with non-canonical patterns underwent repeat ISH and gave identical results.

Group comparisons using Dunn’s post hoc test showed no significant differences between NLPHL and NLPHL-OL (adjusted p = 0.312) or between cHL and cHL-LR (adjusted p = 0.440) ( Figure 5 ). In contrast, κ/λ mRNA ISH positivity differed significantly between NLPHL and cHL (adjusted p = 2.88 × 10 -6 ). The comparison between NLPHL-OL and cHL-LR was also significant (adjusted p = 0.0037). A summary of the findings in all cases is presented in Figure 6 .

Across all cohorts, κ light-chain restriction predominated (32/38, 84%), whereas λ light-chain expression was uncommon (4/38, 11%) and only present in NLPHL and NLPHL-OL. Atypical double κ/λ mRNA ISH profiles were also identified, including one EBV+ case with distinct LP-cell populations each expressing either κ or λ ( Supplementary Figure S1 ), and one case with κ/λ coexpression in LP cells. Four cases of κ-positive tumor cells revealed focal intranuclear granular λ signals, reflecting labeling of IGLL5 transcripts ( Figure 2 G-I ). Finally, assessment of the background B lymphocytes showed no secondary clonal population; however, rare follicles with PTGC-like changes showed skewing toward kappa light-chain mRNA expression, without definite evidence of light chain restriction.

Discussion

In this study, we investigated the expression of κ/λ mRNA as a surrogate marker for the functional transcriptional activity of the IG light chain locus in NLPHL and cHL, with particular focus on NLPHL with overlapping features and cHL-LR, aiming to refine the classification of these challenging cases. Our results demonstrate that the novel ultra-sensitive κ/λ mRNA ISH reliably distinguishes typical NLPHL from cHL and provides additional diagnostic support in overlap cases. In addition, we identified attenuated perinuclear and granular punctate staining patterns of currently unknown significance, referred to here as non-canonical patterns in a substantial subset of cases of cHL. These patterns may reflect varying degrees of preservation/alteration of the B-cell transcriptional program.

Previous studies using other assays have demonstrated that κ/λ mRNA ISH differentiates NLPHL from cHL, with uniform positivity in NLPHL and negativity in cHL. 25–27 In line with these findings, our results also indicate a clear distinction with 100% positivity in NLPHL, including cases with overlapping features, and complete absence of canonical staining in cHL. This distinction confirms the significant differences in the underlying B-cell transcriptional programs and offers further diagnostic support. In line with their preserved B-cell program, typical NLPHL revealed invariably canonical diffuse cytoplasmic staining. Architectural variation in NLPHL correlated with increased heterogeneity, although the canonical pattern was largely maintained.

Of note, in contrast to a recent study reporting negativity in all cHL cases, half of cHL-LR and a significant minority of cHL cases in our series showed limited κ/λ mRNA ISH positivity. 25 Although it is not completely clear whether these non-canonical staining patterns represent low-level immunoglobulin mRNA expression in the tumor cells, their recognition is of practical diagnostic relevance. To date, these patterns have not been specifically addressed in the literature. 25 , 26 The absence of this non-canonical pattern in previous studies may reflect differences in assay performance and sensitivity, including the possibility that the reagents used here may detect low-level light-chain signals not captured by other platforms.

Importantly, the constant expression of light chain mRNA in NLPHL-OL cases confirms their classification as NLPHL with aberrant features, with three cases displaying non-canonical patterns and one case showing reduced expression. Notably, three of these cases were EBV-positive. Despite morphological and immunophenotypic features that may complicate diagnosis, such as aberrant CD30/CD15 co-expression or partial loss of B-cell markers, κ/λ mRNA ISH yielded robust, reproducible results. In contrast, cHL-LR exhibited only non-canonical positivity (55%) or negativity (45%). κ/λ mRNA ISH positivity occurred more frequently in cHL-LR than in conventional cHL (55% vs. 29%), supporting a closer biological relationship to NLPHL and partial preservation of the B-cell program as indicated in previous studies. 9

However, the exclusively non-canonical nature of this positivity points more to a dysregulated immunoglobulin transcription than to a physiologic light chain expression as in NLPHL. 6 This may reflect reduced or residual immunoglobulin transcription, but could also result from enhanced degradation of aberrant transcripts through nonsense-mediated mRNA decay, a cellular surveillance mechanism that eliminates faulty mRNAs. 33 , 34

No significant differences were observed between NLPHL and NLPHL-OL, nor between cHL and cHL-LR. However, NLPHL-OL and cHL-LR cases differed significantly with respect to canonical versus non-canonical and the percentage of κ/λ mRNA ISH positivity. From a diagnostic standpoint, non-canonical κ/λ mRNA ISH patterns should not be interpreted as evidence of intact B-cell identity or be used as an argument for a diagnosis of NLPHL, but they might reflect a continuum of IG transcriptional activity. This distinction becomes particularly relevant in the NLPHL–cHL-LR gray zone, where conventional morphological and immunophenotypic criteria may be inconclusive, and where the final diagnosis has direct therapeutic implications, as treatment strategies differ substantially between entities. In such cases, only canonical κ/λ mRNA expression provides an additional layer of diagnostic information, but requires integration with architecture, B-cell marker and MEF2B expression, and EBV status. 35

EBV+ NLPHL-OLs more frequently showed reduced κ/λ mRNA ISH expression or non-canonical staining patterns when compared to EBV-negative NLPHL. EBV+ NLPHL is well-recognized across age groups, frequently expresses CD30, and typically shows latency type II, although rare cases of latency type I are also reported. 15 , 36 , 37 Whereas EBV likely represents an early driver event in cases with homogeneous EBV positivity, as in our cases, the relevance of a secondary/subclonal EBV infection of LP cells in these lesions remains uncertain. 38 , 39 In this context, our findings suggest that EBV may modulate κ/λ transcription in LP cells. EBV-associated perturbation of the B-cell program is well documented in cHL. 38 , 40 , 41

Mechanistically, EBV latent proteins such as LMP2A and LMP1 can interfere with B-cell differentiation programs by impacting key transcriptional regulators (e.g., PAX5) and inducing inhibitors of B-cell differentiation (e.g., ID2). 42–44 This has also been demonstrated in other EBV-driven B-cell lymphomas, 45 , 46 and may explain the non-canonical κ/λ profiles observed in EBV+ NLPHL-OL despite retained B-cell marker expression, including strong CD20 and OCT2.

Interestingly, κ/λ mRNA ISH revealed unconventional expression patterns that have, to our knowledge, not yet been described in LP cells. In 4 cases of NLPHL, intranuclear granular λ signals were detected in κ-positive LP cells, a pattern consistent with known cross-reactivity with IGLL5 , a λ-like pseudogene sharing 2 exons with the IGL constant region, whose transcripts are typically nuclear and detectable in lymphoid and non-lymphoid cells. 19–21 , 25 , 26 , 47 Recognition of this phenomenon is essential to avoid misinterpretation as λ light-chain mRNA expression.

In addition, two rare cases with aberrant patterns were observed: NLP-10 showed strong cytoplasmic coexpression of κ and λ within the same LP cells. True biallelic light-chain expression in the same cell has rarely been described in other B-cell malignancies, and functional mRNA transcripts of both light chains have been found in CLL with monotypic light chain expression, indicating that allelic exclusion can happen on the translational level. 26 , 47 , 48 NLPHL-OL-2, a case of relapsed EBV+ NLPHL-OL, exhibited two distinct LP-cell populations with κ and λ expression, respectively, without evidence of single-cell co-expression. Whether this represent true biclonality or the occurrence of a light chain switch within in a single clone would require higher-resolution approaches, including single-cell analyses.

This study has several limitations. Fixation and specimen age may affect RNA preservation and introduce preanalytical variability. Although positivity of both external and internal controls was strictly assessed, areas of suboptimal fixation or background staining heterogeneity may have reduced sensitivity in specific cases. Moreover, although non-canonical κ/λ mRNA ISH patterns were reproducible and showed clear non-random associations with diagnostic categories and EBV status, their biological basis cannot be fully resolved within the scope of this study, and non-specific background staining, although unlikely, cannot be completely excluded. Higher-resolution or single-cell approaches would be required to confirm that these signals reflect low-level IG transcription, and to investigate the impact of EBV infection on light-chain transcripts in LP cells. Finally, the relatively small number of control cases investigated may not reflect the complete spectrum of light chain mRNA expression in typical NLPHL and cHL, underscoring the need for validation in larger multicenter cohorts.

In summary, highly sensitive κ/λ mRNA ISH reliably distinguishes NLPHL from cHL and is a valuable adjunct to classify cases within the NLPHL–cHL-LR gray zone, as it provides a readout of IG transcription. Beyond a binary result, distinct non-canonical expression patterns in EBV-positive NLPHL and some cHL cases, especially cHL-LR, might reflect low-level IG mRNA transcription, although confirmation of this hypothesis will require more advanced approaches.