ALK+ Anaplastic Large Cell Lymphoma Diagnosed on Paediatric Pericardial Effusion Cytology

While rare, malignant paediatric pericardial effusions are crucial to recognise. The effusion will reaccumulate without proper treatment and, when it is the first presentation of cancer, misdiagnosis risks disease progression or acute decompensation. ALK-positive anaplastic large cell lymphoma (ALK+ ALCL), a rare cause of pericardial effusion, is a CD30-positive T-cell lymphoma driven by an ALK translocation, most commonly t(2;5)(p23;q35), with NPM1 as the partner gene [1]. The constitutively activated ALK kinase stimulates the RAS, JAK/STAT, AKT and other pathways, driving tumorigenesis [1]. ALK+ ALCL presents most frequently in patients < 1–30 years of age as Stage III–IV disease with bulky lymphadenopathy, mediastinal masses or skin lesions [1]. In children, ALK+ ALCL comprises 10%–30% of all lymphomas and responds well to cytotoxic chemotherapy (5-year event-free survival of 70%) [1, 2].

Here, we describe pericardial effusion as the primary presentation of paediatric ALK+ ALCL. This emphasises the diagnostic value of pericardiocentesis and highlights ALK+ ALCL in the differential diagnosis of malignant paediatric pericardial effusions.

An 11-month-old male with no significant past medical history was admitted from the emergency department to the intensive care unit with hypoxic respiratory failure following 2 weeks of cough and congestion unresponsive to steroids and albuterol. Oxygen saturation measured ~85%, and the physical exam demonstrated tachypnoea and subcostal retractions. Palpable lymphadenopathy was absent. Complete blood count showed 25.1 k/μL leucocytes (normal range 4.0–14.6 k/μL; 79.1% neutrophils; 14.7% lymphocytes; 4.4% monocytes; 0.5% basophils; 1.3% immature granulocytes), 545 k/μL platelets (normal range 150–400 k/μL) and 10.7 g/dL haemoglobin (normal range 10.5–13.5 g/dL). A complete metabolic panel was largely unremarkable. Laboratory tests for viral, fungal and bacterial infections were negative. Chest radiography demonstrated bilateral pleural effusions and cardiomegaly (Figure 1A). Echocardiogram showed a large circumferential pericardial effusion and mild right atrial collapse in diastole. EKG showed preventricular contractions and questionable right ventricular hypertrophy. Computed tomography scan, performed after ALK+ ALCL was diagnosed, demonstrated an indeterminate mass in the cervical paraspinal soft tissue and innumerable enlarged intrathoracic lymph nodes, some demonstrating encasement of surrounding vascular structures (Figure 1B). No cardiac mass was identified.

Due to concern for tamponade, emergent pericardiocentesis was performed using a 21-gauge needle and 102 mL serosanguinous fluid with a protein content of 4.4 g/dL was withdrawn. From this, cytospin slides were prepared. A cell block made from the centrifuged fluid was fixed in ethanol and post-fixed in 10% neutral-buffered formalin.

The cytospin and cell block preparations of the pericardial fluid were hypercellular and showed large abnormal cells with moderate eosinophilic cytoplasm, irregular and often multi-lobated nuclei, vesicular chromatin and multiple variably prominent nucleoli (Figure 1C,D). Mitotic figures were frequent, and some cells showed characteristic horseshoe-shaped nuclei with an eosinophilic paranuclear hof, consistent with hallmark cells. Background small lymphocytes and scattered neutrophils were present.

Immunohistochemical stains performed on the cell block (Figure 2) showed the large abnormal cells to be positive for CD45, CD30 (strong and diffuse), ALK1 (diffuse cytoplasmic, nuclear and nucleolar staining) and CD4 (weak) with a small subset positive for CD2, CD5, CD7 and granzyme (rare); and negative for CD3, CD8, TIA-1, CD20, CD68, CD138, CD1a, S100, MOC-31, BerEp4, calretinin, D2-40, WT-1 and HHV8. In situ hybridisation for EBV-encoded RNA (EBER) was negative. A GMS special stain for fungal organisms was negative.

Staging bone marrow biopsy demonstrated rare ALK-positive hallmark cells consistent with low-level involvement in the bone marrow. T-cell receptor gamma gene rearrangement analysis by polymerase chain reaction performed on the bone marrow sample was positive for a clonal population of T cells. Cerebrospinal fluid cytology was negative for lymphoma.

The patient was treated with an initial cytoreductive phase followed by brentuximab plus six alternating cycles of course A (ifosfamide, dexamethasone, methotrexate, cytarabine and etoposide) and course B chemotherapy (cyclophosphamide, methotrexate and doxorubicin). He has been disease free for 18 months since the completion of therapy.

Malignant pericardial effusion in infants is uncommon. When present, it is frequently due to hematologic malignancy. Of 799 inpatients aged 1–11 months with pericardial effusion identified from a database of roughly seven million hospitalisations, 43 (7.2%) and 11 (1.6%) had solid tumour and hematologic malignancies, respectively [3]. Drainage of the effusion, which allows for cytologic examination, was performed in 91 cases (11.4%), mostly in the settings of cardiac structural abnormalities and hematologic malignancy [3]. Furthermore, several studies have looked specifically at paediatric pericardial fluid cytology and collectively suggest that leukaemia/lymphoma is the most common aetiology when the fluid is positive for malignancy [4-6]. Importantly, most of the positive specimens in these studies were collected from patients with a known diagnosis. In a 40-year, single institution study of all paediatric pericardial serous effusion cytology specimens (N = 38), all three positive specimens (8%) were diagnosed as lymphoma, not otherwise specified [4]. In a separate 15-year, single institution study, 3/28 (10.7%) pericardial effusion cytology samples were positive for malignancy, two involved by rhabdomyosarcoma and one by a hematolymphoid malignancy [5]. Additionally, a 12-year study examining 104 pleural, 77 peritoneal and 2 pericardial specimens from two institutions found 16 (8.7%) specimens in total involved by hematolymphoid malignancy. These comprised 40% of those positive for malignancy in this study [6]. Therefore, when pericardial fluid cytology from an infant identifies atypical cells, leukaemia and lymphoma should be considered.

The differential diagnosis for atypical cells in paediatric pericardial fluid samples is wide, including reactive/inflammatory conditions (such as reactive lymphocytes responsive to infection or rheumatologic disease and therapy-related mesothelial/inflammatory changes) and malignancy (such as hematolymphoid malignancies, rhabdomyosarcoma, germ cell tumours, neuroblastoma, Ewing's sarcoma and thymoma) [7]. Given the notorious nature of pericardial mesothelial cells to demonstrate severe reactive atypia, particularly in the setting of physiological perturbations, care must be taken not to over-interpret these changes as malignant in nature, while simultaneously appreciating the possibility of a rare malignant process.

The vast majority of effusions are non-malignant, and primary pericardial presentation of paediatric neoplasms is rare [7]. Thus, clinical history is crucial for evaluation. Knowing the protein content of the fluid is helpful as well, as exudative effusions (protein content > 3.0 g/dL) are more associated with malignancy than transudative (< 3.0 g/dL) [7]. Our patient's recent history of upper respiratory symptoms suggested a viral aetiology for his effusion, although all infectious testing was negative. He had no history, signs, or symptoms to suggest rheumatologic disease. These pertinent negative findings and an effusion protein level of 4.4 g/dL raised malignancy on the list of differential diagnoses.

ALK+ ALCL has several histologic patterns such that, in some cases, diagnosis on cytomorphology alone may be challenging. Most commonly, ALK+ ALCL appears as variably sized and anaplastic lymphoid cells [8]. Alternative morphologies include small cell, lymphohistiocytic and Hodgkin-like, as well as other rare patterns, such as sarcomatoid [1]. Scattered so-called ‘hallmark cells’, large cells featuring horseshoe-shaped nuclei with a paranuclear eosinophilic hof and ample basophilic cytoplasm, are present in all patterns [1, 9]. On cytology, ALK+ ALCL features dispersed, pleomorphic lesional cells, about three times larger than a lymphocyte, with prominent single to multiple nucleoli [9]. Hallmark cells are mixed in, as well as wreath-like multinucleated giant cells [9]. There is typically a mixed inflammatory background [9]. By immunohistochemistry (IHC), CD30 is diffusely positive in a membranous and often paranuclear/Golgi pattern; additionally, Alk is positive in a nuclear and/or cytoplasmic or, rarely, a membranous pattern, depending on the alk translocation partner [8]. T-cell markers are often lost, and expression of cytotoxic molecules, such as TIA or granzyme B, is common [8].

If fresh material amenable to flow cytometry (FC) is received by the laboratory, analysis by this method is warranted when the cytomorphology is suggestive of a hematolymphoid process. On FC, ALK+ ALCL features moderate to high CD45 expression with moderate side scatter, an aberrant T-cell phenotype, and expression of myeloid antigens such as CD13 [8]. Fluorescent in situ hybridisation (FISH) studies using an alk break-apart probe will detect the translocation regardless of the partner gene [8]. However, ALK overexpression on IHC correlates with an alk translocation, and since atypical translocations (non NPM1-partnered) have not been shown to alter prognosis, FISH analysis is not typically necessary [8]. Our case featured diffuse nuclear, nucleolar and cytoplasmic ALK staining (Figure 2), corresponding to the typical NPM1 translocation [8]. This characteristic staining pattern results from heterodimerisation between the NPM1-ALK fusion protein and wild-type NPM1, a nucleolar protein that shuttles between the nucleus and cytoplasm [10].

Awareness of the various histologic patterns of ALK+ ALCL can help cytopathologists consider ALK+ ALCL within differential diagnoses [1]. For example, dispersed, anaplastic cytomorphology can mimic that of germ cell tumours, while small cell morphology can be mistaken for reactive lymphocytes [9]. Therefore, cytopathologists should have a low threshold to use IHC and FC and should not exclude ALK+ ALCL based on the absence of diffuse anaplastic morphology. Thus, it is important to obtain sufficient specimen volume to allow for the use of various ancillary studies, particularly when a malignant process is being considered in the differential diagnosis.

We could identify only two cases of paediatric ALK+ ALCL involving the pericardium without a cardiac mass in the English literature [11, 12]. Like our case, both previously reported patients presented with dyspnoea and pleural effusion and developed pericardial effusion. One case progressed to tamponade [11]. Both patients also had palpable lymphadenopathy. Diagnosis was made on pericardial biopsy in one case [12] and by skin rash biopsy plus pericardial cytology in the other [11, 12]. In contrast, the clinical presentation in the patient reported here was related predominantly to the pericardial effusion, with no other biopsy target, such as a rash, palpable mass, or lymphadenopathy, identified on physical exam. While the sensitivity of pericardial cytology to detect malignancy is variable [13], in this case it was the best diagnostic medium, since pericardiocentesis was relatively easily performed and excisional biopsy was not possible. Like others have suggested [5], effusion drainage offers a less invasive sampling technique that can also act as a therapeutic intervention.

In certain clinical scenarios, it can be important to distinguish systemic ALCL involving an effusion from primary effusion ALCL (PE-ALCL), as the latter typically has an indolent course [13]. The best-described PE-ALCL is breast-implant associated ALCL [14], which is not typically relevant to the paediatric population. Other rare PE-ALCLs have been reported, such as one characterised by IRF4/DUSP22 rearrangement [15]. Clinical findings and imaging studies can help distinguish effusion-based lymphomas from systemic disease.

In summary, this case highlights ALK+ ALCL as a rare cause of paediatric pericardial effusion. In effusion cases in which a malignant aetiology is suspected, pericardiocentesis can be considered for pathologic diagnosis, especially when other biopsy targets are unavailable. ALK+ ALCL can morphologically mimic other malignancies. When suspicious, the pathologist should readily use IHC and FC to avoid misdiagnosis, especially in cases with an indeterminate cytologic appearance. To that end, ample specimen should be collected for cytologic work-up.

Connor Hartzell conducted a literature review and primarily wrote the manuscript. Huiying Wang provided expertise and edited the manuscript. Emily Mason provided expertise, contributed data and edited the manuscript. Christopher J. O'Conor conceived the original idea, provided expertise, edited the manuscript and oversaw the project.

The authors have nothing to report.

The authors have nothing to report.

The authors declare no conflicts of interest.