Clinical Features and CSF-Ctdna Characteristics in Extranodal Natural Killer/T‐Cell Lymphoma Patients with Central Nervous System Involvement

Background:

Extranodal natural killer (NK) T-cell lymphoma (ENKTL), is a rare subtype of NHL associated with an aggressive clinical course. Central nervous system (CNS) involvement [CNS(+)] in ENKTL is much rarer and there is a lack of clinical data from large samples, and the clinical implications of cerebrospinal fluid circulating tumor DNA (CSF-ctDNA) positivity in ENKTL remains largely unexplored. In this study, we analyzed the clinical features, characteristics of CSF-ctDNA, therapeutic strategies, and outcomes of CNS(+) in ENKTL.

Method:

We used database of Sun Yat-sen University Cancer Center for the diagnosis of ENKTL. Criteria for diagnosed CNS(+) were: (1) Brain parenchymal or leptomeningeal involvement confirmed by either brain MRI or CT scan or PET-CT; (2) Positive CSF cytology; (3) Brain biopsy confirmed histopathological diagnosis of ENKTL; (4) Presence of typical CNS symptoms with at least one mutant variant detected in CSF-ctDNA. Patients were identified as CSF-ctDNA positive (CSF(+)) based on the following criteria: (i) at least one mutant variant detected in CSF-ctDNA was shared with tissue; (ii) if tissue is not available, at least one mutant variant detected in CSF-ctDNA was canonical mutation as previously reported in ENKTL. Targeted next-generation sequencing (NGS) was performed at a Clinical Laboratory Improvement Amendments and College of American Pathologists-accredited testing laboratory (Nanjing Geneseeq Technology Inc, Nanjing, China). Sequencing libraries were constructed using the KAPA Hyper Prep Kit and sequenced on Illumina HiSeq 4000 using a panel of 475 leukemia- and lymphoma-related genes. The criteria of genotyping of tissue and plasma were according to the LymphGen algorithm. The primary endpoint was overall survival (OS) form CNS(+) to death.

Results:

Between 2006 and 2022, 1400 ENKTL patients were reviewed, a total of 47 (3.4%) patients with CNS(+). Median age at CNS(+) was 47 years (19-86). Female patients accounted for 38.3% (18/47). Thirty-six (76.6%) patients had ECOG scores of 2-3. Forty-four (93.6%) had secondary involvement. CSF protein was elevated in 57.4% (27/47) of patients. The positive rates of CNS imaging and CSF cytology were both 46.8% (22/47). The median number of prior lines of treatment was 2 (1-6).

Median time from diagnosis of ENKTL to CNS(+) was 8.0 months. Four patients were untreated due to rapid disease progression. Systemic therapy regimens after CNS(+) included: (1) high-dose methotrexate (HD-MTX)-containing regimens (n=14, e.g., single-agent, SMILE, HD-MTX+Temozolomide, etc). (2) PD-1 antibody (PD-1 mAb)-containing regimens (n=26, e.g., single-agent, PD-1 mAb plus chidamide, P-GemOx, etc.), and other regimens. Three (6.4%) patients received ASCT and 2 (4.3%) accepted whole brain radiotherapy. Seven patients received intrathecal nivolumab, and 36 patients received triple intrathecal therapy (MTX, Ara-C, dexamethasone). Median follow-up was 5.5 months. Median time from CNS(+) to death was 4.0 months (2.9-5.1). OS was better in patients received HD-MTX-containing regimens, with a median OS of 4.8 vs. 3.9 months (p=0.053, HR = 0.49). PD-1 mAb did not significantly improve OS, with a median OS of 4.6 vs. 4.2 months (p=0.80, HR = 0.92).

Baseline CSF cytology was tested in 40 patients and 25 baseline CSF-ctDNA sample were detected. The positive rates of imaging, CSF cytology and CSF-ctDNA were 46.8% (22/47), 55.0% (22/40) and 96.0% (24/25), respectively. Of the 18 neuroimaging negative patients, all patients (100%) were positive for CSF-ctDNA, and of the 13 patients with negative CSF cytology, 12 patients (92.3%) were positive for CSF-ctDNA. Three genes with the highest mutation frequency in CSF, plasma and tumor tissues were TP53, BCOR and FAT1. The abundance of the three genes in CSF, plasma and tumor tissues are as follows: TP53 (40.0%, 31.8%, 20.0%), BCOR (36.0%, 31.8%, 20.0%), FAT1 (20.0%, 27.3%, 20.0%). We compared the tumor tissue mutation results of 37 patients who did not present with CNS(+) and found that STAT3, KMT2D, TP53, ARID1A, DDX3X were common mutated genes.

Conclusion:

ENKTL with CNS involvement is rare and the prognosis is poor. Currently, there remains no effective treatment. CSF-ctDNA improves the probability of diagnosis of CNS involvement. TP53, BCOR, FAT1 are the most frequently mutated genes in ENKTL patients with CNS involvement. Further studies are needed to improve prognosis of this population.