Since 1978, the Japan Clinical Oncology Group (JCOG) has conducted six successive prospective clinical trials. All these trials were based on conventional chemotherapy, with various dose and administration modalities. The first trial JCOG 7801 used VEPA (a CHOP-like regimen that contained vincristine, cyclophosphamide, prednisolone and doxorubicin). The CR rate was only 17 % with a median survival time of 5 months. The second trial, JCOG 8101, was a randomized phase III study which included 54 patients and compared VEPA regimen with VEPA-M (VEPA plus methotrexate) [4]. Although the CR rate was improved in the VEPA-M group (36.7 %), no difference in median survival time (7.5 months) and overall survival (8 % at 4 years) was noted.

The third trial, JCOG 8701, was a phase II study with a more aggressive regimen (LSG 4), which combined three successive regimens: VEPA-B (VEPA plus bleomycin), M-FEPA (MTX, vindesine, cyclophosphamide, prednisolone, and doxorubicin) and VEPP-B (vincristine, etoposide, procarbazine, prednisolone and bleomycin). The CR rate was improved to 42 %. However, median survival rate and overall survival (OS) were poor with a median survival time (MST) of 8 months and overall survival rate of 12 % at 4 years. These trials enrolled also patients with other subtypes of NHL. MST was 44 months in NHL versus 8 months in the ATL group.

Following these initial trials, the JCOG designed specific regimens targeting ATL. The JCOG9109 trial (a phase II study conducted between 1991 and 1993) used pentostatin-containing regimen but did not show any improvement (MST 7, 4 months and 2 years overall survival rate: 15 %) [5].

The next protocol named JCOG 9303 was conducted between 1994 and 1996 and used more intensive multi-agent chemotherapy regimens [6]. Treatment was designed as follows: VCAP (vincristine, cyclophosphamide, doxorubicin and prednisolone), AMP (doxorubicin, ranimustine, prednisolone) and VCEP (vindesine, etoposide, carboplatin, prednisolone). It included intrathecal injection of methotrexate and aracytine. The use of granulocyte colony-stimulating factor (G-CSF) was systematic. Results were encouraging with a CR rate of 35 %, a MST of 13 months (versus 8 months in historical controls treated by CHOP-like regimens). The 2-year OS was 31 %. MCNU and carboplatin were used because their activities were not affected by the expression of P-glycoprotein, a product of MDR1, which is frequently expressed by ATL cells.

In order to confirm theses results, a phase III study (JCOG9801) was conducted between 1998 and 2003. This study compared two arms of treatment: VCAP-AMP-VECP versus biweekly CHOP. It included 118 patients (81 acute subtype and 26 lymphoma subtype) [7]. Response rate was higher in the experimental arm (40 % vs. 25 %). Progression-free survival at 1 year was 28 % versus 16 %, and overall survival was 24 % versus 13 % in the two arms respectively. There was a statistically significant difference only in a subgroup analysis (patients younger than 56 years, poor performance status).

As most of patients relapse after conventional chemotherapy, allogeneic stem cell transplantation (alloSCT) seems to be an attractive option as consolidation treatment. Most of the reports come from Japan. A number of retrospective studies have confirmed alloSCT using either myelo-ablative conditioning or reduced-intensity conditioning as a feasible treatment option for ATL patients. The largest retrospective study has been reported in 2010 [8]. This study included 386 patients who received alloSCT between 1995 and 2005, either with standard or reduced intensity conditionings. After a median follow-up of 41 months, 3-year overall survival was 33 %. Among patients who received family-related transplants, donor HTLV-I seropositivity adversely affected disease-associated mortality. Recently, the long-term results of a series of 30 patients who received reduced intensity conditionings were reported. Overall survival rate and progression-free survival rates were 36 % (95 % CI, 21–25 %) and 31 % (95 % CI, 17–45 %) respectively [9].

The response after donor lymphocyte infusion (DLI) is often considered being the best proof of a graft-versus-disease effect. Recently, the Nagasaki Transplant Group has reported objective responses for patients relapsing after alloSCT who have been treated with DLI. Moreover, the responses were durable, with three cases of long-term remission of more than 3 years [10]. A graft-versus-ATL effect is also suggested by the results of the national Japanese retrospective study, but results are less convincing [11].

However, the number of ATL patients eligible for alloSCT is very limited because of the low CR rate especially in the acute form, poor performance status, severe immunosuppression and low probability of finding a suitable donor in patients from ethnic minorities.

The combination of zidovudine (AZT) and alpha interferon (IFN) was first reported in two phase II studies [12–14]. High response rate was observed particularly in previously untreated acute ATL. The efficacy of this combination was confirmed in a French trial using AZT/IFN in 19 newly diagnosed ATL patients and in a UK clinical trial using AZT/IFN in 15 ATL patients [15, 16]. A meta-analysis was performed on ATL patient survival between 1995 and 2008 in three countries (the United States, the United Kingdom, Martinique and continental France) [17]. Two hundred and fifty-four patients have been enrolled in this study; they had various ATL subtypes: 116 acute ATL, 18 chronic ATL, 11 smouldering ATL and 100 ATL lymphoma. Different treatment strategies have been compared, namely, antiviral therapy alone (AZT-IFN), chemotherapy alone and chemotherapy followed by maintenance antiviral therapy. On the 207 patients for whom the first-line therapy was reported, 5-year OS rates were 46 % in 75 patients who received first-line antiviral therapy, 20 % in 77 patients who received first-line chemotherapy and 12 % in 55 patients who received first-line chemotherapy followed by antiviral therapy. Patients with leukaemic subtype significantly benefited from first-line antiviral therapy, whereas patients with ATL lymphoma had a better outcome with chemotherapy. In acute ATL, first-line antiviral therapy alone resulted in a significant survival advantage (5-year OS of 28 %) as compared with first-line chemotherapy with or without maintenance antiviral therapy (5-year OS of 10 %). Achievement of CR with antiviral therapy resulted in 82 % 5-year survival. In chronic and smouldering ATL, antiviral therapy resulted in 100 % 5-year survival. In ATL lymphoma, first-line antiviral therapy resulted in a significant survival disadvantage (median and 5-year OS of 7 months and 0 %, respectively) compared to first-line chemotherapy with or without maintenance antiviral therapy (median and 5-year OS of 16 months and 18 %, respectively). Finally, a multivariate analysis confirmed that first-line antiviral therapy significantly improved overall survival of ATL patients (HR 0.47; 95 % CI 0.27–0.83; p = 0.021). The presence of a functional p53 in tumour cells is associated with a better response [18].

A recent English study has reported the efficacy of the association of chemotherapy with AZT-IFN combination [19]. Seventy-three patients with aggressive subtypes (29 acute and 44 lymphoma subtypes) were retrospectively analysed. Sixty-seven patients received chemotherapy as first-line treatment, mostly CHOP-like regimen. Forty received the association of AZT-IFN including 27 as first-line therapy in association with chemotherapy (concurrent with chemotherapy or sequential).

Belonging to 65 patients for whom response was available, response rate was 81 % for patients who received combined regimen (chemotherapy plus AZT-IFN) versus 49 % for those who received chemotherapy alone. This seems to translate in a doubling PFS when compared patients treated with combined regimen to patients treated with chemotherapy alone (8 months versus 4 months respectively). This last result did not however reach statistical significance. Nevertheless, it was observed a better OS with prolonged median OS in acute (p = 0.0081) and in lymphoma subtypes (p = 0.001) in patients treated with combined regimen when compared to those treated with chemotherapy alone as first-line therapy. Moreover, exposure to AZT-IFN combination at any time was associated with a better prognosis for both ATL subtypes with reduced hazard ratio risk of death of 0.23 (95 % CI, (0.091–0.60), p = 0.002) in multivariate analysis. However, toxicity especially haematological toxicity was not recorded in this study. We know that AZT-IFN bears haematological toxicity by itself in addition to chemotherapy.

Arsenic trioxide is synergistic with IFN to induce cell cycle arrest and apoptosis in HTLV-I-infected and fresh ATL cells through rapid shut off of the NF-κB pathway and a delayed shut off of cell cycle-associated genes, secondary to Tax degradation by the proteasome [20–22]. Although it has been demonstrated that arsenic and IFN cooperate to cure murine ATL, derived from Tax transgenics through selective eradication of leukaemia-initiating cell (LIC) activity. This strongly suggests that LIC activity is dependent on continuous Tax oncogene expression. Hence, addition of arsenic to AZT/IFN, through elimination of LIC activity, may result in long-term disease eradication and potential cure [23]. A recent prospective phase II study evaluated the efficacy and safety of the combination of arsenic, IFN and AZT in 10 newly diagnosed chronic ATL patients. The response rate was 100 % including 7 CR, 2 CR but with more than 5 % circulating atypical lymphocytes, and one partial response. Side effects were moderate and mostly haematological [24]. We have also recently reported a series of 11 patients with ATL (3 lymphoma type, three chronic and five acute) treated with arsenic/IFN after induction chemotherapy [25]. At initiation of AsO3, four patients were in CR, two in PR and five in progression. Ten patients received AsO3 during 3–8 weeks. One progressed 3 days after starting AsO3 and six patients died. All were progressive at time of AsO3 initiation. Five patients survived: 3-lymphoma type in CR (25, 31, 46 months of follow-up), one acute subtype in CR (9 months follow-up) and one chronic subtype in PR (39 months follow-up). Tolerance was acceptable with peripheral neuropathy (n = 4), hand-foot syndrome (n = 3), skin eruption (n = 3, including two toxic epidermolysis). While preliminary, these observations nevertheless suggest that in ATL patients, arsenic/IFN efficiently targets ATL LIC activity and may be useful as a consolidation therapy for those patients achieving a satisfactory response after induction therapy.

Patients with smouldering or chronic ATLL subtypes have a better prognosis than patients with aggressive forms (acute and lymphoma) and have been considered as indolent forms. Many patients have been managed with a watch and wait policy until disease progression or treated with chemotherapy when poor prognostic factors were present. A recent published Japanese study reported on 90 patients with indolent subtypes (65 chronic and 25 smouldering) [3]. Forty-four (49 %) patients progressed to aggressive disease with a median time of transformation of 18.8 months (range 0, 3 months to 17.6 years) and 41 died. Median survival was 4, 1 year. No difference between the two subtypes (chronic and smouldering) was observed. The estimated 10-year survival rate was 25.4 % (95 % CI, 15, 3–36, 8 %). This study shows that even, in the indolent subtype, prognosis is poor. Moreover, patients who received chemotherapy had a worse prognosis and a shorter survival than patients who were managed by watchful waiting. These results underscore the need for further improvement in the treatment of patients with otherwise indolent forms of ATL.