Successful treatment of central nervous system lymphoproliferative disorder in a kidney-pancreas and stem cell transplanted patient using intrathecal rituximab

Abstract

Central nervous system lymphoproliferative disorder (CNS-PTLD) after organ transplant is a unique clinicopathological entity and is associated with poor survival rates. When the CNS is involved, intravenous rituximab might not be the treatment of choice, due to its poor CNS penetration. However, intrathecal (IT) administration of rituximab has shown to be safe and efficient in small studies and in case series. We report here the case of a patient with late development of CNS-PTLD after kidney-pancreas transplantation who achieved complete remission after surgical resection and four cycles of IT rituximab and we provide a review of the literature for this treatment option.

Background

Post-transplant lymphoproliferative disorders (PTLDs) represent a spectrum of disorders that occur as a consequence of immunosuppression in recipients of solid organ or stem cell transplantation.

Several risk factors for development of PTLD have been identified, including age, type of allograft, degree of immunosuppression, concurrent infections and genetic factors.¹ The most important risk factor for Epstein-Barr virus (EBV)-driven PTLD is pretransplant EBV seronegativity.² Extranodal localisation is frequent and the central nervous system (CNS) is involved in 5%–30% of all PTLD cases.³ Isolated CNS-PTLD seems less frequent, with an incidence between 7% and 15% reported in the literature.^4–7 Isolated CNS-PTLD tends to appear late, with a median time to occurrence of 4–5 years in different case series.^{4 8–10} In a large case series of 84 patients with CNS-PTLD, the vast majority of PTLDs occurred later than 1 year from solid organ transplantation and one-third occurred later than 10 years after transplantation.⁴

When present, CNS involvement has been shown to be predictive of inferior survival. In a report of 230 renal transplant patients with PTLD from the French registry,² the authors showed that the 5‐year survival rate was 81% for patients with PTLD localised to the renal graft, whereas for recipients with cerebral PTLD, the 5‐year survival rate dropped to 53%.

The WHO 2016 divides PTLD into six distinctive subsets: plasmacytic hyperplasia PTLD, infectious mononucleosis PTLD, florid follicular hyperplasia PTLD, polymorphic PTLD, monomorphic PTLD (B-cell types and T-cell types/NK-cell types) and classical Hodgkin PTLD.¹¹ The majority of CNS-PTLD lymphomas are monomorphic, with a B-cell phenotype expressing CD20, and with diffuse large B-cell lymphoma (DLBCL) being the most frequent histological subtype.^{4 12}

Symptoms of CNS-PTLD are variable and can range from non-specific headaches and seizures to focal neurological symptoms depending on localisation of the brain lesions. Diagnosis can be supported by cerebrospinal fluid (CSF) analysis. EBV positivity in CSF is highly suggestive of CNS-PTLD¹³ and cytology of the CSF can show atypical lymphocytes. Flow cytometric findings suggesting clonality can also assist with diagnosis. Neuroimaging examinations are needed to detect brain lesions and associated complications such as brain herniation. MRI is superior to CT in terms of sensitivity and for detailed analysis of CNS-PTLD lesions.³ Positron emission tomography-CT (PET-CT) may be more useful for identifying PTLD lesions outside the CNS. However, biopsy remains the gold standard for definitive diagnosis of CNS-PTLD.

The goal of treatment is to cure the lymphoma, but also to preserve graft function. A minority of patients respond to a reduction in immunosuppressive drugs. However, additional treatment such as chemotherapy, immunotherapy, whole brain radiotherapy or surgery is required for most patients.

Case presentation

We report here the case of a Caucasian, 48-year-old male patient, diagnosed with type 1 diabetes and associated renal failure, treated with combined kidney-pancreatic transplantation from a cadaveric donor in 2001 for which he was receiving ciclosporin and mycophenolate mofetil.

In 2018, the patient was diagnosed with acute myeloid leukaemia (AML) with complex karyotype including monosomy 7. He attained a first complete remission (CR) after induction chemotherapy with a regimen based on cytarabine and daunorubicin, followed by allogeneic haematopoietic stem cell transplantation (HSCT) of peripheral blood cells from a female sibling Human Leucocyte Antigen (HLA)-identical donor. The patient received a reduced toxicity conditioning regimen adapted to his renal function, with fludarabine (15 mg/m²/day for 4 days) and busulfan (130 mg/m²/day for 3 days). Before transplant, the patient received immunosuppressive therapy with ciclosporin and during his HSCT received continuous mycophenolate mofetil. Higher doses of immunosuppression were introduced from day −1 for ciclosporin (3 mg/kg/day) and from day +1 for mycophenolate mofetil (2 g/day).

On day +1 after CSH transplantation, the patient presented with proximal left arm hemiparesis associated with slight dysmetria. Toxic as well as metabolic causes of hemiparesis were excluded.

Investigations

Cerebral MRI was rapidly performed, and revealed a right temporal mass surrounded by oedema. Contrast MRI was also performed, revealing a unique parietal lesion with perilesional oedema (figure 1). Differential diagnosis between a secondary metastasis and a primary glial tumour could not be made at this time. As infection could also not be excluded in this severely immunocompromised host, broad-spectrum triple antibiotic/antifungal treatment (posaconazole, amphotericin B liposomal, meropenem) was initiated. Abdominal and thoracic CT showed no other lesion.

Figure 1.

Axial MRI spin echo T1 sequence after administration of gadolinium illustrates a right parietal lesion, well delineated, round, with a ring enhancement, surrounded by oedema.

Lumbar puncture showed no signs of meningitis or cell atypia. A slight proteinorachia was noted (0.47 g/L), while CSF and cultures were negative. PCR was positive for EBV although at a low viral load (40 copies/mL). Of note, regular biweekly blood EBV PCR were performed and were always negative. Both donor and recipient were sero-positive for EBV before transplantation.

In order to make a definitive diagnosis, a complete mass excision was performed by guided craniotomy and after neutrophil and platelet recovery (on day +25). Histopathology revealed that 97% of the specimen analysis comprised necrotic cells. Nevertheless, a small portion of cells in the necrotising zone consisted of highly proliferating atypical B lymphocytes that were positive for EBV (Fluorescence in situ Hybridization (FISH)-positive for EBV-encoded RNA (EBER)). Morphology and immunochemistry confirmed a diffuse large B-cell type PTLD (PTLD-DLBCL type) (figure 2). Total body PET-CT showed no other lesion. Flow cytometry of the blood did not show any clonal B-cell population.

Figure 2.

Histology, immunohistochemistry and in situ hybridisation of the lesion. (A) H&E staining using standard techniques. (B) In situ hybridisation for the Epstein-Barr virus (EBV) -encoded RNA (EBER) transcript performed using the Ventana EBER1 800-2842 (750 ng/mL) (Ventana Medical Systems, Tucson, Arizona, USA) according to the manufacturer’s protocol. (C) CD20 staining using the antibody NCL-L-CD20-L26 from Novocastra (Leica Biosystems, Newcastle, UK). Staining was performed on a Ventana Benchmark ULTRA autostainer with antigen retrieval using the CC1 buffer for 120 min, incubation anti-CD20 (1/400) 60 min at 36°C followed by 3'3-diaminobenzidine HCl (DAB) revelation according to the manufacturer’s protocol. (D) The images show infiltration by large atypical cells positive for the B-cell marker CD20 and EBV EBER RNA. These cells were surrounded by necrosis that occupied >90% of the slide.

Treatment

In the absence of graft versus host disease (GvHD), immunosuppression was reduced, mycophenolate mofetil was stopped on day +24 and ciclosporin was presumed at lower therapeutic target levels (100–150 μ/L vs 200–250 μ/L) in order to prevent renal and pancreatic graft rejection. Steroids had been initiated after brain imaging for neurological symptoms to decrease cerebral oedema and were rapidly tapered after biopsy results. The patient was not eligible for systemic chemotherapy or immunotherapy with intravenous rituximab due to the recent stem cell transplantation and the high degree of immunosuppression.

Lacking other treatment options, intrathecal (IT) administration of rituximab was proposed. The patient received four cycles of 20 mg of IT rituximab on a weekly schedule. The first dose contained 20 mg of rituximab diluted in 4 mL of NaCl 0.9%, with consecutive doses containing 30 mg of rituximab diluted in 5 mL of NaCl 0.9%.

Tolerance was good, except for occasional headaches occurring in the first days after infusion. Seven days after the third administration of IT rituximab, the patient presented with epileptic seizures attributed to levofloxacin prescribed for a Pseudomonas aeruginosa pneumonia and possibly to brain lesion after mass excision. There was no relapse of epileptic seizures after discontinuation of levofloxacin and under prophylactic treatment of levetiracetam. The fourth infusion of rituximab was administrated after a 1-week delay. Brain MRI after the fourth cycle was negative for any suspicious lesion. Immunosuppression with ciclosporin was maintained at a slightly higher therapeutic range of 150–200 μ/L due to grade II GvHD that had occurred 2 months after the allogeneic HSCT. The patient remained in remission from CNS-PTLD with no signs of the disease on brain MRI performed 2 months after the final IT rituximab infusion.

Outcome and follow-up

Unfortunately, 7 months after HSCT, the patient presented with haematological AML relapse. Despite the transient efficacy of venetoclax associated with hypomethylating agent (5-azacytidine), remission was not sustained, and the patient died 1 year after HCST from progression of the leukaemia.

Discussion

Diagnostic challenge

Differential diagnosis of brain mass in an immunocompromised host includes primary CNS lymphoma, glioblastoma, metastatic disease, abscess or other infection, such as Aspergillus, Nocardia asteroides, Toxoplasma gondii, Mucorales, Myobacterium tuberculosis and less commonly, Cryptococcus.^{3 14}

Brain MRI findings of signal hyperintensity (reflecting haemorrhagic mass) hypo-intensity in T1 (reflecting hypercellularity) surrounding oedema, but also localisation of the mass, orientated the diagnosis to CNS-PTLD. CNS-PTLD lesions are often multiple, contrast-enhancing, associated with extensive oedema, and appear in deeper supratentorial areas and in the periventricular region and less frequently in the cerebellum and brainstem.¹ A particular element in our patient’s case was the solitary lesion presentation, while in most case series, the majority of CNSPTLDs manifest with multifocal disease.^{1 3 10}

EBV-positivity in CSF is highly suggestive of a PTLD. Nevertheless, the EBV viral load in the blood has a modest predictive value for identifying primary CNS-PTLD. Our patient’s serum was negative for EBV. Histology on a biopsy specimen of the mass, showing a DLBCL phenotype, and the absence of other lesions on PET CT confirmed the diagnosis of primary CNS-PTLD.

Although EBV-positive PTLD typically has an earlier onset (median time from transplantation to PTLD of 12 months versus 50–60 months in EBV-negative patients) and is less frequently monomorphic than EBV-negative PTLD,^{7 15} Evens and colleagues, in one of the largest series of 84 CNS-PTLD cases, reported that, similar to our patient, the great majority of patients with CNS-PTLD presented with late onset (>1 year) EBV-positivity and monomorphic subtype.⁴ This highlights the unique clinicopathological features of this disease compared with PTLD cases with secondary CNS involvement.

The patient presented a very late-onset PTLD of, 17 years after combined kidney-pancreas transplantation. Our hypothesis is that chemotherapy during induction and conditioning (busulfan) phases led to an extended mass necrosis, with surrounding oedema causing focal neurological symptoms. Indeed, before allogeneic HSCT, our patient did not have any brain imaging, either of which might have revealed an asymptomatic lesion. Steroids introduced after his brain MRI and before surgery likely further contributed to the extended necrosis of the mass revealed in the histopathological analysis.

Choice of IT rituximab treatment

The cornerstone of PTLD treatment is reduction of immunosuppression (RI). The following clinical factors are each associated with poor response to RI, such as late-onset PTLD, elevated lactate dehydrogenase, organ dysfunction and multiorgan involvement.⁷ Due to a lack of large prospective randomised trials, additional second-line treatment is often empirical and quite heterogeneous, often consisting of intravenous rituximab alone or in combination with chemotherapy. Tolerance to chemotherapy is generally poor in patients with PTLD, with high treatment-related mortality. Surgery has limited role in localised disease.¹⁶

A large multicentre analysis of 80 PTLD cases⁷ showed that rituximab-based therapy in conjunction with RI, as a first-line approach, improved survival compared with prior reports. Although intravenous rituximab has been shown to be an efficacious component of treatment for non-CNS-PTLD, the role of this treatment in CNS-PTLD is not well established. Patients with CNS-PTLD who are initially treated with RI alone have poor outcomes, demonstrating the need for combined-modality treatment options.⁴ High-dose methotrexate followed by high-dose cytarabine, intravenous rituximab alone or in combination with other agents, and whole brain irradiation are some of the available options. Overall response rates to these treatment options across different CNS-PTLD case series were modest, although 40% of patients reached long-term survival.⁴

Our patient presented with a solitary brain lesion that was completely removed by guided craniotomy. As this modality of treatment was not a classic approach for CNS-PTLD, further treatment options were considered in order to prevent progression or relapse. RI or chemotherapy were not considered given the short delay after allogeneic HSCT and the high risk of infections. Remaining options were IT immunotherapy with or without brain radiation.

Indeed, a challenge in CNS-PTLD treatment is that systemically administered rituximab has limited efficacy in treating CNS disease, likely due to poor CNS penetration.^17–19 When administered intrathecally, rituximab achieves more predictable and higher CSF concentrations compared with its intravenous administration, however the efficacy and safety of IT rituximab are unknown.²⁰ Few studies have evaluated safety and efficacy of administration of rituximab directly in the CNS (table 1).

Table 1.

Use of intrathecal or intraventricular ritixumab in patients with primary CNS-PTLD, peripheral PTLD with CNS involvement, primary CNS lymphoma, B-NHL with CNS involvement or B-ALL with CNS involvement

Studies	Number of patients and indication for IT R/I- vent R	Median age	Type of treatment (number of patients)	Dose of IT R/I-vent R (number of patients)	Response to treatment (number of patients)	Side effects (number of patients)
Ceppi et al17	25 patients B-NHL (15) B-ALL (3) PTLD (7)	12.8 years	IT R (17)/I-vent R (8) in monotherapy or combined with other treatment*	Median: 25 mg administrated two times per week (19) or weekly (6) Median number of doses: 6	CNS remission (18) PR (1)	Grade I–II peripheral neuropathy (5) Grade I–II headache (2) Grade I–II allergy (2) Grade III neuropathy (1) Seizures grade II (1) and grade IV (1)
Wu et al21	14 patients Peripheral PTLD with CNS involvement (10) Primitive CNS-PTLD (4) Received IT R (9)	25.7 years	IT R combined to intravenous R with or without systemic chemotherapy and addition of DLI if CR not achieved after 2 cycles of R based treatment	Sequential dose escalation (initial dose of 20 mg increased by 10 mg each week for a maximum dose of 50 mg)	CR (8/9)	Temporal headache and cauda equina syndrome immediately after administration of IT R (2) No long-term adverse events
Bonney et al25	2 patients PTLD	4 years	IT R with systemic chemotherapy and IT methotrexate and hydrocortisone (1) IT R with TIT (1)	20 mg two times per week (1) 20 mg weekly (1)	CR (2)	No adverse events
Czyzewski et al22	8 patients PTLD	16 years	IT R with intravenous R	Increasing doses from 10 mg up to 30 mg administrated weekly Median number of cycles: 7	CR (5) PR (2) Death from PTLD progression (1)	Short episode of seizures after IT R (1) No serious adverse events
Rubenstein et al20	10 patients (primary CNS lymphoma)	56.8 years	I-vent R monotherapy followed by IT R in 3/10 patients	3 dose cohorts including 3, 3 and 4 patients receiving 10, 25 and 50 mg two times per week of I-vent R, respectively 3 patients received 10 mg (1) and 25 mg (2) of IT R after I-vent R treatment	CSF clearance (6) Intraocular responses (2) Resolution of brain parenchymal lymphoma (1)	I-vent R: leucoencephalopathy grade I (1) Grade III hypertension (2) ITR: sacral paraesthesia with spontaneous resolution (1)
Schulz et al23	6 patients (B-NHL)	54.5 years	I-vent R monotherapy (4) IT R monotherapy (2)	I-vent R: 10–40 mg, 2–3 administrations per week IT R: 10–25 mg, 1–3 administrations weekly	CR and clearing of tumour cells in the CSF (4)	Grade III neuropathy (1) Grade I hypotension (1)
Jaime-Pérez et al24	7 patients (B-ALL)	10 years	IT R with systemic treatment †	10 mg weekly for 4 weeks	CR (5) Death from systemic relapse and GvHD (2)	No evidence of neurotoxicity

*Other treatment: systemic chemotherapy with intravenous R and triple intrathecal therapy (TIT) or IT methotrexate, systemic chemotherapy with TIT, intravenous R with TIT, intravenous R alone and in one case, intravenous R with I-vent methotrexate.

†Systemic treatment: in patients treated for B-ALL, during IT R treatment, only 6-mercaptopurine and weekly methotrexate were used as systemic treatment.

ALL, acute lymphoblastic leukaemia; B-NHL, B-non Hodgkin's lymphoma; CR, complete response; CSF, cerebrospinal fluid; DLI, donor lymphocyte infusion; GvHD, graft versus host disease; intravenous R, intravenous rituximab; PR, partial response; I-vent R, intraventricular rituximab; IT R, intrathecal rituximab; TIT, triple intrathecal therapy.

Patients with PTLD after allogeneic HSCT receiving IT or intraventricular (I-vent) rituximab seem to have high response rates (CNS CR at 70%–90%, reported in three case series^{17 21 22}). However, the specific role of IT/I-vent rituximab in achieving a remission is difficult to be established as patients in these studies received IT/I-vent rituximab combined to intravenous rituximab and in some cases, to systemic chemotherapy. On the other hand, Wu and colleagues²¹ showed a lower relapse rate in PTLD (only one of nine patients), suggesting a possible additive value of direct administration of rituximab in the CNS in the treatment of PTLD.

Patients receiving IT or I-vent rituximab for B-NHL and B-ALL with CNS involvement presented a CNS CR at 67%–93%^{17 23} and 33%–71%,^{17 24} respectively. The majority of patients included in these studies had relapsed or recurrent disease consisting a very difficult-to-treat population. In addition, in the study of Ceppi and colleagues,¹⁷ three of six patients (50%) receiving only IT/I-vent rituximab achieved a complete CNS remission, further proving efficacy of this treatment.

Rituximab IT/I-vent administration was well tolerated even in heavily treated patients, with no clinical evidence of long time neurotoxicity. The majority of adverse events were graded I–II, while grade III hypertension reported by Rubenstein et al²⁰ was associated with administration of I-vent R at 50 mg, consisting a dose limiting toxicity. Ceppi et al¹⁷ reported a grade IV neurotoxicity with prolonged seizures, probably related to the primary disease, in one patient, who was able to resume therapy with no further toxicities.

Our patient presented with mild headaches, with spontaneous resolution a few hours after rituximab infusion. Seizures occurred later than 48 hours after rituximab treatment and were more likely attributable to the brain lesion after surgery, precipitated by reduction of the epileptogenic threshold due to levofloxacin treatment.

Learning points.

• Intrathecal rituximab seems to be safe and efficient for isolated central nervous system lymphoproliferative disorder, but larger controlled studies are needed to confirm these findings.

Ethics statements

Patient consent for publication

Obtained.