LAUSR

Dear Editor, A 24-year-old woman presented in 2016 with superior vena cava obstruction due to an anterior mediastinal mass (Fig. 1), confirmed pathologically to be T cell lymphoblastic lymphoma. Treatment with an acute lymphoblastic leukaemia (ALL) protocol hyper-CVAD (cyclophosphamide, vincristine, doxorubicin, dexamethasone, methotrexate, cytarabine) failed. Salvage with high-dose methotrexate, ifosfamide, etoposide, dexamethasone and L-asparaginase resulted in a partial response. An HLA-matched sibling allogeneic haematopoietic stem cell transplantation (allo-HSCT) was performed in 2017, with cyclophosphamide (120 mg/kg) / total body irradiation (12 Gy) as the conditioning regimen. There was no graftversus-host disease (GVHD) post-HSCT. She relapsed 4 months afterwards, presenting as fever and recurrence of the mediastinal mass (Fig. 1). Bone marrow examination was normal, and analysis of short tandem repeats in the marrow aspirate showed complete donor DNA chimerism. All immunosuppressants were withdrawn. There was no radiological or symptomatic improvement after stopping immunosuppressants. GVHD did not occur. At that juncture, in view of the previous unsatisfactory response to conventional chemotherapy, the plan was to try immunotherapy for disease control, followed by a second allo-HSCT.With informed consent, an antibody against programmed death protein 1 (PD1) nivolumab was administered, at an intended schedule of 40 mg every 2 weeks. After the first dose, there was rapid-onset liver function derangement (peak alanine aminotransferase, 465 U/L, upper reference level, 36 U/L; peak alkaline phosphatase, 496 U/ L, upper reference level, 93 U/L), associated with non-pruritic macular rash, xerophthalmia and oral mucositis. Overall features were consistent with GVHD. She responded to treatment with prednisolone (1 mg/kg/day). Interestingly, chest X-ray performed about 2 weeks after the first dose of nivolumab showed resolution of the mediastinal mass (Fig. 2). In view of the radiological response and occurrence of GVHD, further nivolumab treatment was withheld. She continued to improve, and another chest X-ray 4 weeks later showed normal findings (Fig. 2). PET/CT scans performed 6 months and a year later (Figs. 1 and 2) showed continued metabolic complete response. At the latest follow-up 27 months after nivolumab treatment, she has remained in remission, with mild residual chronic GVHD, sicca symptoms and nail dystrophy (Fig. 3). She is on topical eye lubricants, sirolimus (0.5 mg per week) and thalidomide (50 mg per week). There are several important observations in this case. The withdrawal of immunosuppression was ineffective. Hence, anergy of effector T cells towards lymphoma cells was not due to the use of immunosuppressants. Instead, the high efficacy of PD1 blockade showed that effector T cell anergy was due to inhibition via the PD1 receptor. In immunotherapy based on checkpoint inhibition, a prevailing view is that mutations in the tumour genome generate neo-antigens, which serve as targets for effector T cells [1]. A high mutational load increases the array of tumour neoantigens targetable by effector T cells, thereby enhancing the efficacy of immune checkpoint inhibition. Among various malignancies, ALL has one of the lowest mutational loads [2]. Accordingly, immune checkpoint inhibition is not a recognized strategy in ALL [3]. Therefore, the excellent response of this case of lymphoblastic lymphoma (biologically similar to ALL) to nivolumab was extraordinary. It is all the more notable as the relapse occurred only 4 months after allo-HSCT. In a previous study, we showed that acute leukaemias relapsing within 11 months * Yok-Lam Kwong [email protected]