Introduction

Lymphomas are the 5th to 7th most frequently occurring malignant diseases in both sexes. It is not completely understood why there is still an increasing incidence of mature B‑cell lymphomas, yet a link to the broad use of immunomodulatory drugs, the increased prevalence of immunoderegulatory diseases and the exposure to environmental pesti-/herbicides is suspected [1, 2]. Fortunately, the progress in lymphoma therapy is noticeable: survival rates of patients continuously increase, at least when referring to the most common nodal lymphomas [1].

Here, we summarize the current standards in lymphoma diagnostics from the pathologist’s point of view, with a special focus on a multidisciplinary approach.

Gain of material

Histopathology is the keystone in lymphoma diagnostics, since this is the procedure of choice to distinguish between benign and malignant lymph node (LN) changes [3]. Importantly, a substantial number of benign LN disorders may both clinically and morphologically mimic lymphomas. Furthermore, genetic aberrations detectable in lymphomas as well as clonal outgrowths may be observed in reactive lymphoid processes, emphasizing the role of specifically trained histopathologists with regard to this type of diagnostics.

The trend towards less invasive procedures in the past years did not stop in front of hemato-oncology, leading to widespread use of fine needle biopsies (FNB) for diagnostic purposes. Though being informative in approximately 85% of cases (Fig. 1 and Table 1), there are several objective limitations and risks arising from the general use of FNB. For instance the distinction between angioimmunoblastic T‑cell lymphoma (AITL) and (e.g. drug-induced) paracortical hyperplasia [3] can only be achieved by the assessment of the total LN architecture, including the capsule and the surrounding soft tissue. These features cannot be evaluated on FNB as they merely contain a small detail of the LN. For obvious reasons, partial LN involvements by any processes are hardly possible to be diagnosed on FNB. In follicular lymphoma (FL), a statement about the grading is required, since grade 3B and—probably—grade 3A require a more aggressive clinical management, but this is barely obtainable on FNB. Additionally, some entities, such as Hodgkin lymphoma (HL) and AITL, go along with a very small amount of diagnostic tumor cells, which can easily be missed in a thin core cylinder (Fig. 1). Last but not least, lymphoma diagnostics require in almost all cases additional work-up, which sometimes cannot be applied to FNB due to lack of material or a too small amount of tumor cells. This also implies that there will be usually no material available for scientific questions after a diagnostic FNB work-up.

Fig. 1
figure 1

Lymphadenectomy specimen of a patient suffering from nodular lymphocyte predominant Hodgkin lymphoma with hyperplastic follicles and progressively transformed germinal centers. Note that the so called lymphocytic and histiocytic (L&H) tumor cells are only focally present in the respective zoomed-in zone (upper right part of the figure). A previous core needle biopsy of the patient (shown in the lower right part of the figure) was of no diagnostic yield; H&E stain

Full size image
Table 1 Diagnostic yield of lymphadenectomies compared to fine needle biopsies with respect to lymphomas
Full size table

For the reasons stated above, we strongly recommend—whenever possible—a total LN excision for lymphoma diagnostic purposes in order to avoid misinterpretations and unnecessary re-biopsies. If a positron emission tomography (PET) is preformed prior to the excisional procedure, the LN with the highest PET activity should be extirpated [4]. FNB should remain a tool for exceptional situations, e.g. in difficult surgically accessible regions such as the mediastinum and retroperitoneum. To avoid material waist in such circumstances, careful histotechnical handling is advisable, particularly production of sufficient amounts (at least 15) of unstained sections on adhesive glass slides for subsequent in situ histopathologic (immunohistochemistry, in situ hybridization) and in vitro molecular (gene expression profiling, sequencing on material scrapped-off those slides) analyses. In addition and minimizing the risk of expiry, parafilm wrapped unstained sections can be stored for a long time for scientific questions.

Processing of the specimen

Optimally, the gained material should be fixed immediately. If specific queries such as frozen section examination arise—which must be regarded as an exception in that consideration—the fresh specimen should be immediately brought to the histopathology lab and the remaining LN tissue timely fixed. Usually, buffered formalin (final formaldehyde concentration 4%) is used and should be applied in a specimen:formalin ratio of at least 1:10. The fixed material is best stored at a temperature between 4 and 8 °C. Due to crystallization artefacts, storage below 0 °C should be avoided. Depending on the size, sufficient fixation is obtained after 4 to 12 h. If frozen section is considered inevitable and immediate delivery for histopathologic examination is not feasible but manageable within 3 h, the material should be sent in a dry container for respective examination by means of a courier transport, in our opinion—taxi. Contact of the tissue with any fluids, such as saline, would induce considerable freezing procedure artifacts, and is therefore not recommended. If transportation within 3 h is not possible, a fast-track formalin/paraffin technical processing (fixation/dehydration) should be preferred to frozen section examination. If submission of unfixed material is considered for other requests, such as flow cytometry or cytogenetics, one part of the obtained tissue can be put into Roswell Park Memorial Institute (RPMI) medium and used for these examinations, while the other should be submitted in formalin and utilized for morphological diagnostics.

Clinical information

The delivery of sufficient clinical information is crucial for lymphoma diagnostics. Indeed, as indicated in Table 2, several lymphoma entities cannot be diagnosed outside specific clinical settings. In the following, we list the minimum clinical key-facts for rendering a correct diagnosis.

  1. 1.

    Age: Several lymphoma entities are defined by the patient’s age as they almost exclusively occur in young individuals and are thought to be associated with the immaturity of the immune system: e.g. pediatric type follicular lymphoma (PTFL), pediatric nodal marginal zone lymphoma and EBV-positive T‑cell and NK-cell lymphoproliferative diseases of childhood [5]. PTFL is an impressive example how the correct pathological classification influences the following treatment: in contrast to patients with conventional FL, who are often treated with multimodal chemotherapy, surgical excision alone seems to be sufficient for most PTFL patients [6, 7]. Information on sex and especially on ethnicity may be of upmost diagnostic importance in some virally driven lymphomas such as adult T‑cell lymphoma/leukemia [5].

  2. 2.

    Location: Some lymphomas manifest at characteristic locations and information about the site can be an important diagnostic component. For example, primary cutaneous follicle center lymphomas occur most often at the head or trunk, have an excellent prognosis and mostly require only localized therapy; without the knowledge of the clinical appearance, the differential diagnosis to cutaneous involvement by conventional FL might be difficult if not impossible [5]. The same applies to CD30+ cutaneous lymphoproliferations with respect to the differential diagnosis of cutaneous involvement by systemic anaplastic T‑cell lymphoma (ALCL) [5]. The rare entity of large B‑cell lymphoma with IRF4 rearrangement (Fig. 2) is almost exclusively found in the Waldeyer ring or head and neck region of younger individuals. It has to be distinguished from PTFL, on the one hand, as it probably needs a more aggressive treatment, and from DLBCL or FL grade 3B, on the other, as the latter two will need much more aggressive treatment [8].

    Special characteristics of the site of involvement can play an important diagnostic role. The presence of chronic inflammatory conditions such as chronic pyothorax, vascular and joint prosthesis, chronic skin ulcers or chronic osteomyelitis will render the correct diagnosis of a DLBCL associated with chronic inflammation [9, 10]. These cases should be recognized since they benefit from adjuvant surgical treatment of the underlying inflammatory condition and, despite being associated with Epstein–Barr virus (EBV) infection and displaying MYC amplifications and TP53 mutations, seem not to run a very aggressive course. Unquestionably, almost all patients with breast implant-associated ALCL, which is also a nice example of an integrative diagnostic entity, will be in need of excision alone [11].

  3. 3.

    Previous illnesses: Detailed anamnestic information about pre-existing illnesses are indispensable. For example, the diagnosis of post-transplant lymphoproliferative disorders cannot be set without appropriate background. The same applies for lymphoproliferative diseases associated with (primary) immunodeficiencies, such as e.g. Wiskott-Aldrich syndrome or ataxia-teleangiectasia [5]. Furthermore, the correct classification of DLBCL arising in/transforming from chronic lymphocytic leukemias (CLL), marginal zone lymphomas and FL can only be made with the clinical note of a previously existing small B‑cell lymphoma (SBCL), in case the low grade compound is not found in the diagnostic LN. This is of importance since DLBCL, high-grade B‑cell lymphoma (HGBCL) or HL transformed from SBCL show a more aggressive behavior than their de novo counterparts [12].

  4. 4.

    Medication: Considering the above-mentioned examples, it becomes also obvious that information about the patient’s previous medication is essential for the correct diagnosis of iatrogenic immunodeficiency-associated lymphoproliferative disorders, which not only virologically and genetically differ from their immunoproficient equivalents, but also often require a different management [5, 13, 14]. A number of applied drugs (e.g. carbamazepine, allopurinol, sulfonamides, methimazol, lamotrigine, gabapentin, nevirapine) and vaccinations against hepatitis B virus and smallpox can induce profound morphologic LN changes mimicking AITL, DLBCL or HL [3], the misdiagnosis of which will be most likely avoided in the setting of proper medication history knowledge.

Table 2 Lymphoma entities, establishing the diagnoses of which require integration of clinical, anamnestic, endoscopic, imaging or genetic data
Full size table
Fig. 2
figure 2

Large B‑cell lymphoma (Giemsa stain) with IRF4 rearrangement (insert). Without application of fluorescent in situ hybridization (FISH) this consultational case has elsewhere been diagnosed as diffuse large B‑cell lymphoma, illustrating that the sole possibility establishing this integrative genetico-pathologic diagnosis is specific testing for the respective recurrent aberration. The insert shows the typical appearance of a translocation/chromosomal break with a break-apart FISH probe with fused yellow-orange IRF4 locus signals corresponding to the non-rearranged allele at 6p25 and free red and green signals corresponding to the rearranged allele

Full size image

Morphology and phenotype

After paraffin embedding of the LN specimen, conventional histotechniques will mostly allow the correct differentiation between benign and malignant lesions or at least establishing a few differential diagnoses. The proper lymphoma classification concerning B‑ versus T‑cell origin, developmental stage of the lymphoma cells and specific entity assignment is mostly done via immunohistochemical stainings. The current WHO classification brought important changes in this regard, two of which will be exemplified here. Inevitably, now the “cell of origin” specification has to be given in all DLBCL, not otherwise specified: distinguishing these deriving from germinal center B‑cells from those displaying a program of activated B‑cells [15]. There are several techniques and algorithms to tackle the cell of origin in DLBCL, such as gene expression profiling and immunohistochemistry, each with different costs, prerequisites, feasibility and validity. No special technique has currently been recommended by the WHO [5], yet data suggest that stratification based on immunohistochemical algorithms for guiding therapy should be viewed very cautiously [16]. Nonetheless, it appears that stratification according to the so-called “Tally” algorithm, based on the evaluation of CD10, GCET1, FOXP1, MUM1p and LMO2, avoiding BCL6, may give the best results [17].

Also by immunophenotyping, a more precise classification of mature T‑cell lymphomas is now possible. Next to AITL, further entities of follicular T‑helper cell (TFH) origin have been identified (nodal peripheral T‑cell lymphoma with TFH phenotype (mostly T‑zone lymphoma), and follicular T‑cell lymphoma) and can be diagnosed using proper markers: PD1, BCL6, ICOS, CXCL13 and CD10 [18]. The detection of EBV by in situ hybridization has become extremely sensitive [19] and should be broadly applied to identify EBV-associated lymphomas and related lymphoproliferations [5, 19].

Genotype

Several lymphomas bear characteristic, to a part entity-specific, chromosomal aberrations (Table 2). This is also reflected in the recent WHO classification with the introduction of several genetically defined entities such as HGBCL with MYC and BCL2 and/or BCL2 rearrangements, large B‑cell lymphoma with IRF4 rearrangement, Burkitt-like lymphoma with 11q aberration, and ALK-positive lymphomas. Importantly, the (over)emphasis of genetics over other diagnostic parameters is particularly noticeable in HGBCL with MYC and BCL2 and/or BCL6 rearrangements, which are defined only by the presence of the respective rearrangements [20, 21]. Put into practice, this means that all mature blastic B‑cell lymphomas have to be tested for those translocations, e.g. by FISH, irrespective of their double-expressor score, since protein recognition by antibodies for MYC and, occasionally, of BCL2 may be abrogated due to somatic (hyper-)mutation of the respective genes [22]. Since interphase FISH with histopathological material is performed on sliced cells, cut-off scores to judge cases positive for rearrangements are needed. These scores depend on the probes utilized, the exact slide thickness and the size of the (tumor) cells studied, and are—in the case of commonly used break-apart FISH probes for MYC, BCL2 and BCL6 applied to 4 μm slides of DLBCL—4, 3 and 1.5% cells with break-apart signals, respectively [23].

Besides structural chromosomal anomalies, point mutations play an important role in the development of lymphomas and can be used for diagnostic or—yet to be broadly proven—theranostic purposes. For instance, diagnosing a lymphoplasmacytic lymphoma (LPL) used to be difficult and not always doubtlessly possible. Since the characteristic MYD88 L265P mutation (especially in combination with CXCR4 mutation) has been detected, the correct diagnosis can easily be set in many cases [24]. Importantly, in the setting of LPL and DLBCL (particularly of the ABC subtype) this MYD88 mutation seems to predict sensitivity towards ibrutinib [25]. The same applies for the BRAF V600E-mutation in hairy cell leukemia: although the mutation is not specific for this entity, its detection in the context of mature leukemic B‑cell lymphomas provides not only an important diagnostic tool, but also predicts sensitivity towards vemurafenib [26, 27]. Furthermore, quantification of both described mutations is valuable for monitoring disease progression or detection of minimal residual disease. Alterations in the TP53 gene are found in many malignant diseases, including different lymphomas. Identifying such mutations allows in small lymphocytic B‑cell lymphoma (B-CLL) very precise risk stratification, prediction of lack of response to fludarabine-containing therapy and sensitivity towards ibrutinib [28, 29]. In B‑CLL, the landscape of predictive mutations steadily grows (e.g. NOTCH1, ibrutinib-resistant BTK mutations), and we expect that similar data will soon emerge for other lymphomas [30, 31].

Finally, genetic testing is helpful in the diagnostics of recurrent disease: e.g. the differentiation between true late relapsing DLBCL or HL and metachronous de novo lymphoma can only be answered by proving a clonal relationship between both; same applies for the distinction of true SBCL transformation into e.g. DLBCL versus independent metachronous DLBCL [32,33,34].

Conclusion

The close collaboration between clinical oncologists and pathologists is indispensable for high standard lymphoma diagnostics. Equally, the pathologists depend on the information given by the clinical colleagues, as the clinician is dependent on a correct diagnosis provided by the pathologist. In the last few years, an enormous increase of knowledge considering tumorigenesis in general and lymphomagenesis in particular, has been recorded. This progress will make individualized, patient- and disease-tailored and, thus, more efficient and effective treatment possible.

Take-home message

For rendering a correct lymphoma diagnosis, a minimum of information provided by the clinician is needed: patient’s age (sex, ethnicity), tumor location, medical history and medication.

Besides accurate diagnostics, progress in molecular pathology makes statements on prognosis and theranostics in lymphomas possible.