1 INTRODUCTION

Adult T-cell leukemia-lymphoma (ATL) is a mature T-cell lymphoproliferative neoplasm caused by the human T-cell leukemia virus type-1 (HTLV-1) infection.1, 2 This peripheral lymphoma was first described in Japan,3, 4 where there is a high incidence in the Kyushu-Okinawa area.5 The geographic distribution of ATL corresponds with that of HTLV-1 carriers, with high incidence rates of ATL in HTLV-1 endemic regions.6, 7 For example, ATL accounts for ~25% of peripheral T-cell lymphomas in Asia compared with 2% in North America and 1% in Europe.8

Patients with ATL have poor survival outcomes despite intensive chemotherapy.9 ATL can be classified into four clinical subtypes: acute, lymphoma, chronic, and smoldering10; two large retrospective studies found a range of 4-year overall survival (OS) rates for each of these respective subtypes: 11–17%, 16–20%, 36–37%, and 52–60%.11, 12 Disease resistance to anti-cancer agents and the increased susceptibility of patients to various infections contribute to low OS rates with conventional chemotherapy (5–13 months).13, 14 Even for those who respond to treatment, the proportion of patients who relapse can exceed 40%,15 which often occurs within months of stopping treatment.16

Relapsed or refractory (r/r) ATL has extremely poor prognosis, with a median OS of less than 4 months after the first salvage therapy with conventional chemotherapy17 and an OS of 9 months in patients with intensive chemotherapy.18 However, treatment guidelines and therapeutic algorithms for r/r ATL are not well defined. According to NCCN guidelines, depending on the disease subtype, the recommended treatment options for r/r ATL include experimental therapy in a clinical trial, zidovudine and interferon, chemotherapy, as well as allogeneic hematopoietic stem cell transplantation (allo-HSCT).19

Mogamulizumab, a defucosylated humanized monoclonal antibody against C-C chemokine receptor 4 (CCR4) is another promising salvage targeted therapy for patients with r/r ATL, first approved in Japan in 2012 and not yet approved in other countries for r/r ATL.20 CCR4 is not common in other cancers, but is expressed in more than 90% of ATL patients21; its expression is a poor prognostic factor.22 Mogamulizumab monotherapy has been studied in r/r ATL patients with positive results.23 A recent study at a single center in Japan demonstrated a median OS and 1-year OS rate with mogamulizumab initiation to be 7.7 months and 42.0%, respectively.24

Another treatment option for r/r ATL, lenalidomide, is an oral immunomodulator with both antiproliferative and antineoplastic activity in B-cell lymphomas in preclinical studies.25 It was first approved in the United States in 2005 for myelodysplastic syndromes, and in 2017, its indication was expanded to r/r ATL in Japan. The median OS and median progression-free survival with lenalidomide treatment were 20.3 and 3.8 months, respectively.26 Several studies have assessed treatment outcomes for available therapies for r/r ATL, but there is a lack of consistency across studies in patient characteristics and survival outcome definitions for r/r ATL patients.

The primary objective of this study was to conduct a systematic literature review to synthesize the available evidence on survival outcomes for patients with r/r ATL treated with various systemic therapies.

2 METHODS 2.1 Literature search strategy

Articles published on r/r ATL treatment outcomes in English or Japanese between January 1, 2010, and January 31, 2020, were screened from PubMed and EMBASE literature databases. A detailed description of the search terms can be found in Tables S1, S2. This review was planned, conducted, and reported in adherence with the Preferred Reporting Items for Systematic Review and Meta-Analyses (PRISMA) guidelines for quality in reporting systematic reviews.27

2.2 Study selection criteria

The predefined eligibility criteria for inclusion were comprised of r/r ATL patients who received an approved systemic interventional therapy for r/r ATL. Prospective and retrospective observational studies, as well as randomized control trials (RCT), were assessed. Animal studies, in vitro/ex vivo studies, gene expression/protein expression studies, PCR/laboratory studies, editorials, non-systematic reviews, conference minutes, and case studies were excluded. Only studies that reported OS were included. Identified articles were assessed for eligibility by two researchers (WK and TM) independently using a two-step screening process to examine article titles and abstracts and then relevant full texts. Due to the small number of articles identified, especially the lack of available RCTs, no statistical integration, head-to-head comparisons, nor superiority testing was conducted to avoid biases and interpretation challenges. Instead, the eligible studies were synthesized as a systematic literature review with parameters extracted individually and described in the results.

2.3 Data extraction

Available data were extracted from the included articles into a predefined evidence summary template. Extracted data included: the reasons for exclusion (if applicable), characteristics of the included studies, target population, treatment characteristics, survival outcomes (including OS rate, median OS, and Kaplan-Meier [KM] plots), overall response rate (ORR), complete response (CR), partial response (PR), and adverse events. Data extraction was conducted by two independent researchers (TM and WK).

As part of an exploratory analysis, 30% OS was calculated to obtain estimates of comparative survival. Because eight KM curves from six studies showed a long tail below 30% survival probability, this statistic was selected for exploratory purposes. For 30% OS and studies without median OS reported, time-to-event outcomes were summarized using pseudo-individual patient data extracted from published KM survival curves. Graph Grabber (v2.0.2, Quintessa, Henley-on-Thames, UK) software application, a tool that enables users to extract data points from a graph image, was used to accurately reconstruct KM curves to obtain precise curve coordinates. The number of patients at risk for each arm during the follow-up was also extracted from identified studies, either presented beneath the published KM curves or extrapolated based on algorithms published in a prior study that assumes constant rate of censoring.28 Analyses were performed using R (version 4.0.1) with a published reconstruction algorithm.29

3 RESULTS 3.1 Search results

The systematic search identified 43 studies for title and abstract screening (Figure 1). An additional four studies were identified through a targeted literature search. After removal of duplicates and screening of the full-text articles, 14 studies were ultimately deemed eligible for extraction, which included a total of 21 treatment subgroups.

PRISMA flow diagram

3.2 Characteristics of studies and patients

As listed in Table 1, the final articles were 10 retrospective cohort studies, two single-arm phase II studies, one phase II RCT, and one prospective post-marketing surveillance study. Sample sizes ranged from 14 to 723 patients. Only the RCT evaluated r/r ATL patients from the United States/Europe/Latin America regions, whereas all other studies investigated patients in Japan. Eight studies included a mogamulizumab treatment arm: five were single-arm studies, and three were two-arm studies. One mogamulizumab single-arm study reported a subgroup analysis of mogamulizumab treatment prior to allo-HSCT.30 In each two-arm study, mogamulizumab treatment was compared with chemotherapy without mogamulizumab. Five studies examined allo-HSCT treatment, one single-arm study and four multiple-arm studies.15, 31-34 One multiple-arm study examined survival outcomes of allo-HSCT as well as with mogamulizumab prior to allo-HSCT.31 Another reported survival outcomes based on patient-level data for several treatment regimens,32 and two multiple-arm studies reported survival outcomes with and without donor lymphocyte infusion (DLI) among patients who received allo-HSCT.15 One study assessed lenalidomide as a single treatment arm,26 and one examined EPOCH regimen as a single treatment arm.35

TABLE 1. Summary of characteristics for studies of r/r ATL treatment and survival First author (year) [reference] Study design Treatment arm(s) Total number of patients for efficacy ATL subtype Proportion for male patients Age, median (years) Location Line of therapy Time to treatment, Median Response Ishitsuka (2019)30 Prospective post-marketing survey Moga 500 – 54% 67 Japan 59.8% (342/572) had 1 regimen of chemotherapy prior to moga –

ORR: 42%

(n = 523)

Tokunaga (2018)37 Retrospective cohort Moga 72

Acute: 50 (69%)

Lymphoma: 17 (24%)

58% 65 Japan Median 2 regimens [range:1–7] prior to moga 6.0 months [range: 0.1–92.9 months] from diagnosis to moga

ORR: 36%

CR: 24%

(n = 72)

Nakashima (2018)39 Retrospective cohort Moga 45

Acute: 32 (71%)

Lymphoma: 7 (16%)

60% 69 Japan Median 1 regimen [range:1–5] prior to moga 6.9 months [range: 1.0–127.0 months] from initial treatment to moga

ORR: 44%

CR: 18%

PR: 27%

(n = 45)

Ishida (2017)36 Phase II Moga 26

Acute: 14 (54%)

Lymphoma: 6 (23%)

42% 65 Japan 82% (22/27) received 1 prior regimen – – Kawano (2016)46 Retrospective cohort Moga 14

Acute: 10 (71%)

Lymphoma: 4 (29%)

43% 63 Japan Average 2 regimens [range: 1–4] of chemotherapy prior to moga –

ORR: 64%

CR: 43%

PR: 21%

(n = 14)

Fuji (2018)31 Retrospective cohort

1) Moga vs

Chemo w/o Moga

2) Allo-HSCT vs w/o Allo-HSCT

723

Acute: 500 (69%)

Lymphoma: 223 (31%)

56% 61 Japan 1st salvage therapy after r/r –

ORR:

Moga: 48%

Chemo: 23%

(n = 638)

Sekine (2017)40 Retrospective cohort Moga vs Chemo w/o Moga 164

Acute: 106 (65%)

Lymphoma: 56 (34%)

46% 68; 75 Japan Average 2 regimens overall prior to moga –

ORR:

Moga: 36%

CR:

Moga: 17%

(n = 87)

Phillips (2019)38 Phase II; randomized trial Moga vs Chemo (GemOx, pralatrexate, DHAP) 71

Acute: 33 (46%)

Lymphoma: 28 (39%)

42% 55; 51 USA, EU, Latin America

Median 2 regimens [range:1–6] prior to moga

Median 1.5 regimens [range:1–5] prior to chemo

9.1 months [range: 1.3–116.7 months] from diagnosis to moga

6.6 months [range: 1.3–150.6 months] from diagnosis to chemo

ORR:

Moga: 15%;

Chemo: 0%

CR:

Moga: 2%;

Chemo: 0%

PR:

Moga: 28%;

Chemo: 0%

(Moga, n = 47; Chemo (n = 24)

Toriyama (2018)35 Retrospective cohort EPOCH regimen 14

Acute: 12 (86%)

Lymphoma: 1 (7%)

50% 58 Japan 1st salvage therapy after r/r 2.8 months [range: 1.0–36.9 months] from initial therapy to EPOCH

ORR: 57%

CR: 7%

PR: 50%

(n = 14)

Ishida (2016)26 Phase II Lenalidomide 26

Acute: 15 (58%)

Lymphoma: 7 (27%)

54% 69 Japan Median 2 regimens [range: 1–4] prior to lenalidomide 2.1 years [range: 0.3–17.5 years] from ATL diagnosis to lenalidomide

ORR: 42%

CR: 19%

PR: 23%

(n = 26)

Kato (2019)34 Retrospective cohort Allo-HSCT + DLI vs allo-HSCT w/o DLI 252

Acute: 150 (60%)

Lymphoma: 65 (26%)

46% 54 Japan – –

CR: 37%

(n = 252)

Fujiwara (2017)33 Retrospective cohort Allo-HSCT 131

Acute: 62 (47%)

Lymphoma: 46 (35%)

– 54 Japan – 286 days [range: 53–3753 days] from diagnosis to transplant – Inoue (2018)32 Retrospective cohort Allo-HSCT 26

Acute: 55 (72%)

Lymphoma: 21 (28%)

46% 56 Japan – – – Itonaga (2013)15 Retrospective cohort Allo-HSCT + DLI vs allo-HSCT + cytoreductive 35

Acute: 29 (83%)

Lymphoma: 6 (17%)

51% 54; 51 Japan – – – Abbreviations: Allo-HSCT, allogeneic hematopoietic stem cell transplantation; ATL, adult T-cell leukemia-lymphoma; Chemo, chemotherapy; CR, completed response; DLI, donor lymphocyte infusion; MAC, myeloablative; Moga, mogamulizumab; ORR, overall response rate; PR, partial response; RIC, reduced intensity; and w/o, without.

The acute subtype was the most common classification of r/r ATL, ranging from 46% to 86% of patients in each study. The second most common subtype was lymphoma type, ranging from 7% to 39% of patients. The median age of r/r ATL patients ranged from 51 to 75 years old, and the proportion of male patients ranged from 42% to 60% (unweighted average: 50%).

Time to salvage therapy from diagnosis of r/r disease was not explicitly reported in any study. Two studies reported the study regimen as the first salvage therapy after r/r diagnosis,31, 35 and another two studies reported the proportion of patients with only one prior regimen (60–82%).30, 36 However, four studies reported the median time to study regimen initiation from initial diagnosis, which ranged from 6.0 to 25.2 months.26, 33, 37, 38 Two other studies reported the median time to study regimen initiation from patients’ first recorded therapy, which ranged from 2.8 to 6.9 months.35, 39

3.3 Overall response rate

ORR ranged from 34% to 64% in the seven mogamulizumab studies that reported ORR, whereas the ORR was 0% to 23% among the three studies that included chemotherapy without mogamulizumab arm. Only one study reported efficacy outcomes by site of disease for mogamulizumab treatment and found that the ORR was 85% (CR 75% and PR 10%) in the peripheral blood, 58% (CR 29% and PR 29%) in skin lesions, 45% (CR 20% and PR 10%) in lymph nodes, and 45% (CR 30% and PR 15%) in other extranodal lesions.39

3.4 Overall survival

OS was defined as the period from the date of the first dose of the study regimen to the date of death or the last follow-up for most (6/8, 75%) mogamulizumab, lenalidomide, and EPOCH studies, whereas OS was defined from the date of relapse for most (3/5, 60%) of the allo-HSCT studies. The median OS varied across studies that included mogamulizumab treatment arms (2.2–17.6 months), mogamulizumab treatment prior to allo-HSCT arms (4.5–7.4 months), allo-HSCT arms (3.8–6.2 months), and other chemotherapy arms (4.1–20.3 months) (Figure 2). One study reported in their subgroup analysis that the median OS for patients who received mogamulizumab treatment prior to allo-HSCT was longer than patients who received mogamulizumab treatment without transplant (7.4 months [95% CI: 5.2–9.9] vs. 5.5 months [95% CI: 4.8–6.4]).30 In another study, the OS for mogamulizumab treatment prior to transplantation and the OS for patients who did not receive mogamulizumab treatment prior to their transplantation were similar (4.5 months [95% CI: 1.3-NA] vs. 4.9 months [95% CI: 3.7–6.7]).31

Median overall survival of r/r ATL patients by treatment arm. *Studies that reported overall survival from diagnosis; †studies that reported overall survival from relapse; ‡studies did not report the timing of overall survival initiation; studies without those markings reported overall survival from the treatment initiation. §Studies reported without mogamulizumab. # Studies from which KM curves were reconstructed to calculate 95% CI; and ¶studies from which KM curves were reconstructed to calculate median OS and 95% CI

Out of 21 treatment arms, 13 treatment arms reached 30% OS during their study period. The exploratory 30% OS calculated for mogamulizumab treatment arms ranged from 8.7 to 27.1 months (Figure 3). Excluding estimates for the single mogamulizumab treatment prior to allo-HSCT (12.3 months), estimates were less varied for allo-HSCT arms (7.5–19.8 months) and other chemotherapy arms (7.1–17.0 months). Among the studies with available 30% OS, three mogamulizumab treatment studies, representing over half of patients (n = 169, 56.6%), had noticeably better survival relative to other studies (30% OS time between 24.2 to 27.1 months vs. 7.1 to 19.8 months).36, 38, 40

Exploratory 30% overall survival of r/r ATL patients by treatment arm (estimated from reconstructed KM curves). * Studies that reported overall survival from diagnosis; †studies that reported overall survival from relapse; ‡studies did not report the timing of overall survival initiation; studies without those markings reported overall survival from the treatment initiation; and §studies reported without mogamulizumab

Two studies evaluated OS by ATL subtype, one mogamulizumab treatment study and one allo-HSCT treatment study. In the mogamulizumab treatment study, patients with acute type were found to have shorter OS than patients with lymphoma type from mogamulizumab treatment initiation (9.7 months vs. 10.7 months).36 A similar relationship was found in the allo-HSCT treatment study, where the OS was 3.7 months in patients with acute type and 18.2 months from relapse in patients with lymphoma type.32

3.5 Adverse events

Cutaneous adverse reaction was the most common complication/adverse event reported from patients receiving mogamulizumab, while hematological events were prevalent among patients treated receiving the EPOCH regimen or lenalidomide (Table 2). The proportion of patients reporting cutaneous adverse reactions varied between 14% and 65%. In the two studies examining the EPOCH regimen and lenalidomide treatment, between 50% and 100% of patients reported hematological events (eg, neutropenia, thrombocytopenia, or anemia).26 Three mogamulizumab single-arm studies found that patients who did not develop skin problems had shorter OS than those who developed skin problems.36, 37, 39

TABLE 2. Major adverse events reported in the identified studies First author (year) [reference] Treatment Cutaneous adverse reaction CMV infection Lymphopenia Leukocytopenia Neutropenia Anemia Thrombocytopenia Nakashima (2018)39 Moga (single arm) 15 (33%) – – – – – – Ishitsuka (2019)30 Moga (single arm) 190 (33%) 47 (8.2%) – – – – – Kawano (2016)46 Moga (single arm) 2 (14%) 2 (14%) 6 (43%) 1 (7%) – – – Tokunaga (2018)37 Moga (single arm) 22 (31%) – – – – – – Ishida (2017)36 Moga (single arm) 17 (65%) – – – – – – Sekine (2017)40 Moga arm + Chemo arm 21 (24%) – – – – – – Toriyama (2018)35 EPOCH (single arm) – – – – 14 (100%) 14 (100%) 12 (86%) Ishida (2016)26 Lenalidomide (single arm) – – 18 (69%) 13 (50%) 19 (73%) 14 (54%) 20 (77%) Abbreviations: Allo-HSCT, allogeneic hematopoietic stem cell transplantation; and Moga, mogamulizumab; ATL, adult T-cell leukemia-lymphoma; Chemo, chemotherapy. 4 DISCUSSION

The largest number of studies in this review assessed mogamulizumab, followed by allo-HSCT and other chemotherapy. Although studies from all countries were eligible in this review in order to investigate the overall treatment landscape, all but one study included in final data extraction focused on treatment in Japan.38