Cutaneous T-cell lymphoma with CNS involvement: a case series and review of the literature

Cutaneous T-cell lymphomas (CTCL) are a rare group of hematologic malignancies characterized by malignant lymphocytes localized to the skin at presentation (cutaneous T-cell lymphomas represent a group of malignancies; however, for clarity, we refer to them in the remainder of the manuscript in the singular form ‘CTCL’). While incidence rates are estimated to be less than 10 per 100,000 per year globally, advances in technology, such as enhanced tools for genotyping and assessing clonality, have allowed more precise diagnosis and subclassification of this disease [1,2]. CTCL is a heterogenous disease, with varied presentations, histologic features and molecular distinctions [3]. Importantly, CTCL can result in a wide range of outcomes, ranging from a largely indolent and favorable course to an aggressive and morbid course [1,4,5]. The most common form of CTCL is mycosis fungoides (MF), which commonly presents as patches and plaques and follows a generally indolent disease course [4,6]. Approximately 25% of patients may progress to advanced stages including extracutaneous involvement [1,4,6]. Sezary syndrome (SS) is a syndrome of erythroderma and blood involvement accompanied by a poor prognosis with a median survival of <5 years. In addition to these two subtypes, numerous additional subtypes of CTCL have been defined, which differ in the biology and disease course [1,7].

Given the rarity yet heterogeneity among CTCL, there are limited data characterizing aggressive presentations such as central nervous system (CNS) involvement. Extracutaneous disease is associated with more extensive skin involvement [1,4,8–10]. Extracutaneous disease may manifest in lymph nodes, visceral organs (e.g., spleen, liver, lung) or blood/bone marrow [1,8,9]. A recent retrospective analysis of a large CTCL cohort confirmed the rarity and poor outcomes in patients with visceral CTCL disease despite treatment with novel agents [11]. Involvement of the CNS has been documented in the literature but is thought to be particularly rare [8,9,12–25]. CNS involvement is variable in presentation and progression [8]. Robust consensus guidelines for the evaluation and treatment of CNS involvement are lacking. Accordingly, given the critical need for additional understanding of extracutaneous manifestations of CTCL in the CNS, herein we characterize a series of cases that discusses the presentation, disease course, and lessons to be learned from historical cases of CTCL with CNS involvement at our institution.

Materials & methods

Case selection & chart review

Our retrospective database details a total of 545 CTCL cases seen at our institution with dates of diagnosis ranging from 1970–2021. We identified four total cases from our database with CNS involvement, representing less than 1% of all cases in our database. We then conducted a further review of their electronic medical records for data regarding presentation, treatment course, and outcomes. Study data were collected using REDCap (Research Electronic Data Capture) [26,27] tools hosted at Emory University. Here, we briefly present and discuss four cases of CTCL with CNS involvement and summarize the relevant details of their diagnoses, clinical courses, and treatment responses. One case predominantly displayed periorbital/orbital involvement (case 4), which varies among clinicians in its categorization as CNS metastasis. However, given the strong potential for periorbital/orbital metastases to locally invade the adjacent CNS, we found it appropriate and important to include this case, though our patient expired before confirmation of local invasion. Representative imaging was available for cases 1–3 but not for case 4.

Tables were generated using Microsoft Excel for Mac version 16.71 (Microsoft Corporation, WA, USA, https://www.microsoft.com/en-us/microsoft-365/excel) and graphical timelines were generated using GraphPad Prism version 9.5.1 for MacOS (GraphPad Software, CA, USA, www.graphpad.com).

Cases

Case 1

Case 1 was a female patient diagnosed with stage 1A MF at age 44. Prior treatments included topical steroids, psoralen + ultraviolet A (PUVA) radiation, and narrowband ultraviolet B (NB-UVB). Her disease was indolent for 7 years until she developed cutaneous progression with tumors. Biopsy demonstrated large cell transformation (LCT). She received localized radiation therapy with a satisfactory response. Subsequent treatment included topical and oral bexarotene, and several rounds of localized radiation for soft tissue tumors. She first established care at our institution at age 58, presenting with diffuse erythematous patches, plaques, and tumors, and with overall ∼50% body surface area involvement, including her face and scalp. Bexarotene was tapered, and she was started on extracorporeal photopheresis (ECP), nitrogen mustard, and focal radiation for ulcerating lesions. She experienced a continual recurrence of skin lesions and soft tissue tumors.

Two years later she developed bibasilar pneumonia but continued to decline despite adequate inpatient treatment with antibiotics. On hospital day 4, she had increasing lethargy, and computed tomography (CT) head revealed new hypodensities in the right caudate/corona radiata. Her neurologic exam was difficult to interpret due to the presence of symptoms of her previously diagnosed Parkinson's disease, and this prevented clear correlation of CT findings with clinical signs. Interestingly, these symptoms were later felt to be more consistent with multiple systems atrophy (MSA), which was supported by characteristic MSA imaging findings (Figure 1A).

Multiple imaging modalities (fluid attenuated inversion recovery (FLAIR), diffusion weighted imaging, MRI with contrast, PET-CT) revealed a right temporal lesion concerning for CNS involvement of her ongoing CTCL (Figure 1B–E). A diagnostic lumbar puncture was deferred, and a multidisciplinary team (including neurology and neuroradiology) determined her imaging findings to be most consistent with CNS-metastatic CTCL. She was started on systemic corticosteroids and referred for whole-brain radiation therapy but was not a candidate due to declining functional status. She died approximately three weeks later, living a total of 19 days following the identification of her CNS metastasis.

Case 2

Case 2 was a male patient diagnosed with stage 1B MF at age 41. Staging showed no lymphadenopathy or peripheral blood involvement by flow cytometry or T-cell gene rearrangement. He later developed small inguinal lymphadenopathy (<1.5 cm) on PET-CT, but bone marrow biopsy was unremarkable. Over the next 2 years, his disease continued to progress, and he was managed on various topical and systemic therapies (topical nitrogen mustard, gemcitabine, vorinostat, pralatrexate, brentuximab, and mogamulizumab).

Approximately 2 years after initial diagnosis, a skin biopsy revealed LCT with CD30 positivity, and PET-CT showed new involvement of the bilateral axillary lymph nodes. An excisional lymph node biopsy confirmed N3 nodal involvement. He was started on combination chemotherapy with brentuximab vedotin, cyclophosphamide, doxorubicin, and prednisone (BV-CHP). One week later, he was admitted to the hospital with unsteady gait and speech difficulties. MRI of the brain revealed a new left cerebellar mass with severe mass effect on the adjacent superior 4th ventricle (Figure 2A–C), suggestive of cerebellar parenchymal involvement with overlying leptomeningeal spread of his CTCL.

Over the following few months, the patient improved with chemotherapy. Interim PET-CT showed decrease in size, number, and metabolic activity of lesions throughout the body, and MRI showed resolution of previously seen leptomeningeal enhancement with persistent unchanged hyperintense signal in the left superior cerebellum (Figure 2D–F). To increase CNS penetration of therapy, he was switched to a combination of high-dose methotrexate (HD-MTX) with combination chemotherapy (cyclophosphamide, doxorubicin, vincristine and prednisolone (M-CHOP)).

Following initial improvement, his neurologic status declined, and a repeat MRI demonstrated progression of multifocal signal abnormality in the cerebellum, including an area of new restricted diffusion with peripheral enhancement in the left cerebellar hemisphere (Figure 2G & H). Lumbar puncture showed no evidence of CTCL involvement by cytology or flow cytometry. Repeat FDG PET CT confirmed disease progression (Figure 2I). Notably, this scan revealed intense abnormal increased metabolic activity affecting the cerebellum and periphery of the brain, which was highly suspicious for leptomeningeal disease. Given the limited treatment options and poor prognosis, the patient and his family opted to transition to palliative care. The patient expired 12 days after discharge, and in total survived ∼18 weeks after CNS symptoms and imaging first were identified.

Case 3

Case 3 was a female patient diagnosed at age 66 with stage 2B MF, presenting with erythematous patches in her upper and lower extremities and right abdomen for 1 year. Her disease was controlled with topical treatment. Three years later, she developed right monocular visual deficits and was referred to our institution. PET demonstrated isolated right intraocular involvement without other signs of systemic disease. Pathology taken from right vitrectomy demonstrated a T-cell population positive for CD8, CD30, and CD45 on immunohistochemistry. T-cell receptor on this specimen was clonal, matching prior skin biopsy. She was treated with a short course of ocular radiation with complete resolution. Approximately one year later, she presented with new left upper arm weakness and paresthesia and MRI brain revealed a new right parietal mass (Figure 3A–C). Subtotal resection of the right parietal mass demonstrated an atypical infiltrate consistent with T-cell lymphoma. She achieved a durable response following 6 months of HD-MTX and temozolomide and near-complete resolution of her CNS lesions (Figure 3D & E). However, she had persistent systemic disease, and approximately three years later, she relapsed with new plaques on the lower extremities and inguinal lymphadenopathy. At the time, her CNS imaging did not reveal evidence of persistent or relapsed disease. She was restarted on a course of HD-MTX but subsequently died from her disease 6 months later at the age of 74. In total, she lived approximately four years after her intraocular disease was identified.

Case 4

Case 4 was a male patient diagnosed with SS stage IVA2 at age 65. He was treated with interferon alfa-2b (IFNa), photopheresis, oral bexarotene, and topical triamcinolone. Approximately 1 year after treatment, he developed nodal progression despite cutaneous improvement. He was treated briefly with vorinostat and then transitioned to oral bexarotene due to poor tolerance. At age 68, due to new and worsening cutaneous lesions, he was started on romidepsin but continued to progress. He was started on total skin electron beam (TBSE) therapy, to which he had an excellent, near-complete response. He later had tumor stage progression and was started on pralatrexate with some response but was intermittently lost to follow-up. When he re-established care, his disease progression culminated in severe right-sided maxillary and bilateral periorbital involvement, causing significant surrounding edema and visual changes, for which he received palliative-intent radiation. His disease continued to progress and his performance status deteriorated. He was ultimately transitioned to hospice care dying shortly thereafter at age 72. For the majority of this patient's disease course, he experienced primarily cutaneous involvement, but ultimately developed CNS-adjacent (which likely progressed to CNS-invasive) disease, resulting in his death. He survived ∼27 weeks after his periorbital involvement first became apparent.

Discussion

We present four cases with CNS progression of CTCL. Our findings further demonstrate the poor prognosis associated with CNS metastasis (Figure 4), with a median survival of just 160 days following CNS-CTCL (range 19 days–4.4 years) in our four cases. Yet with prompt therapy with CNS-penetrating agents, durable remissions are possible. Case 3 enjoyed a survival of greater than 4 years following resection and successful CNS-directed therapy.

Less than 1% of cases in our database displayed evidence of CNS involvement. Though CTCL CNS metastases are rare, several cases have been reported to date [8,9,12–25]. We found that the median time in our cases from initial diagnosis to CNS metastasis was 5.4 years (range 3.4–15.5 years), which is consistent with recent reviews of published cases [9,24]. Given the rarity of CNS involvement, we would not advocate for routine CNS staging. However, patients with concerning neurologic symptoms should be screened for CNS metastases. Additionally, in our experience, diagnoses were heavily reliant on imaging findings, as biopsy is not always feasible and blood contamination may complicate diagnosis by CSF sampling. However, a combination of imaging and TCR gene rearrangement on skin and CSF biopsies has been successfully employed to establish this diagnosis [22].

Jensen et al. [12] described the effectiveness of radiation therapy as a palliative measure following CNS-CTCL. Unfortunately, two of our cases had rapid progression (cases 1 and 2) so radiation was not pursued, while a third (case 3) did not require radiation due to excellent response with systemic therapy alone. HD-MTX is the backbone of primary CNS lymphoma therapy [28,29] and is also effective in CTCL, as demonstrated by two of our cases. This is consistent with a preference for non-chemo-based, immunomodulatory approaches (including radiation therapy) in the management of CTCL when possible [30]. Our cases and others [8,9,12–14,16–21,23–25] mirror this prescribing pattern; though, further study is necessary to determine the superiority of immunomodulatory therapy in CTCL with CNS involvement. Though additional new, targeted therapies for CTCL are being tested (i.e., brentuximab (anti-CD30) [31], mogamulizumab (anti-CCR4) [32]) their use in CNS-involved disease remains in question as CNS-involved patients have often been excluded from major trials [32–36]. The degree of CNS penetration of these agents remains an ongoing area of investigation. Evidence to this end is equivocal and limited to a small number of case reports and series [37–40]. However, there is reason for optimism about this prospect, as antibody-drug conjugates have been proposed as a potential drug delivery method to treat CNS cancers [41]. Regardless, while it is tempting to speculate that such therapies could be utilized in the treatment of CNS-CTCL, formal clinical trials are warranted to further explore this option. Of note, our patient in case 2 did receive both therapies during his treatment course but continued to experience progressive disease.

More aggressive disease is generally associated with a higher risk of CNS progression. Extracutaneous disease and LCT are risk factors previously described [42,43]. Based on a small series of 10 patients with MF, Stein et al. [16] developed a risk calculator for defining the likelihood of CNS disease. Patients with high risk were defined as having ≥2 of the following risk factors: T stage ≥3, N3, M1, or B1. According to this model, high-risk patients had a 15.6% risk of developing CNS involvement versus 1.1% with no risk factors. In a large analysis of published cases, approximately 50% of patients had non-CNS visceral metastases prior to CNS involvement [9]. Moreover, Vu and Duvic [13] showed that LCT confers an increased risk for CNS metastasis, and a more recent study showed 24% of cases with LCT prior to CNS-CTCL [9]. Of the four patients presented here, three of them experienced LCT prior to CNS involvement (cases 1–3). Interestingly, none of the cases displayed other visceral metastases that preceded CNS metastasis.

Finally, it is important to note that the inclusion of case 4 as a CNS metastasis is presumed, but we believe the rapid clinical progression of this case may suggest that CNS involvement ultimately led to his demise. The progression of his case followed a path of visual disturbances accompanied by periorbital edema, which was treated with palliative radiation, and he subsequently showed symptomatic improvement of his visual deficits. Unfortunately, this patient was not seen by ophthalmology, nor did he receive imaging between the time of his orbital disease and his death; thus, it cannot be reported with complete certainty that his ocular disease included optic nerve involvement and resultant cerebral spread of his cancer. However, despite his visual improvement, he still experienced a rapid progression following this, which was accompanied by bradykinesia, psychomotor retardation, impaired vocal communication, and slowed cognition, which may have been suggestive of true CNS involvement. He also experienced interruptions in his care and was lost to follow-up for some periods following his diagnosis, which limited some of the records available to our investigation and could have contributed to his disease progression. Further, despite his relatively stable disease progression over the years prior to this presentation, he displayed rapid decline following the identification of periorbital involvement specifically. When examining the literature, it can be noted that there is a strong concordance between ocular/orbital disease and true CNS involvement in lymphomas and other tumor types [20,44–46]. Taken together, we believe it is reasonably likely that CNS-CTCL was a factor in his decline and ultimate death. Regardless, it is important to consider the practical implications of this case, which are that periorbital and ocular involvement should be closely monitored by an interprofessional team including ophthalmology, radiology, and neurology to detect the progression to true CNS disease.

Conclusion

Herein we present four cases of CTCL with CNS involvement from our institution. While varied in their presentation and course (Figure 4) (Table 1), we generally conclude from these cases the following points most relevant to clinical practice:

• CNS metastases of CTCL carry a particularly poor prognosis, and rapid progression toward death typically follows the identification of CNS metastases.

• However, some patients can achieve complete remission and recovery and enjoy substantial survival following CNS-directed therapy. While therapy is not always curative, it was employed here to achieve near-complete remission of CNS disease itself. Thus, radiation, surgical resection, HD-MTX, and adjunctive therapies play important roles as both palliative and curative strategies in patients who carry this diagnosis.

• When taken in the context of clinical findings and progression, radiologic imaging may be sufficient to make a diagnosis of CNS-CTCL (Figures 1–3), and CSF sampling may be excessive or even misleading due to blood contamination. However, clinicians should use their best judgment and discernment in combining appropriate tests to establish this diagnosis.

• Based on experiences in other tumor types [44,46] as well as in lymphomas [20,45], patients with CNS-adjacent disease (i.e., orbital/periorbital/intraocular) may be at high risk for progression to CNS metastases, and we suggest that early, aggressive treatment should be pursued if identified to prevent CNS involvement. Indeed, it may be prudent to approach the treatment of periorbital, orbital, or intraocular involvement with similar approaches to CNS disease, and clinicians should presume that the progression and behavior of this type of disease will be closely akin to CNS disease.

• While rare, CNS-CTCL should remain a suspicion for any physician following and managing patients with CTCL. Sufficient evidence exists to support the notion that CNS-CTCL can have a heterogenous presentation, as it did in these cases, and CNS disease can occur independent of LCT, advanced TNMB staging, cutaneous disease severity, SS versus MF, or any other prognostic indicator. Thus, basic neurologic surveillance should be a routine part of clinical visits given its rapidly progressing nature once identified, and providers should maintain a low threshold for involving interdisciplinary team members, such as colleagues in dermatology, medical oncology, diagnostic radiology, radiation oncology, ophthalmology, and surgery.

• Further reports of CTCL with CNS involvement and their respective presentations, treatment outcomes, and progressions will help the field move toward clear treatment guidelines for CNS metastases in patients with CTCL, and we encourage colleagues to continue sharing such cases. Additionally, the field would benefit from interinstitutional collaborative efforts that may allow more rigorous study design to inform evidence-based care for this at-risk patient population.