Poor Outcome of Adult T-Cell Leukemia/Lymphoma with Current Available Therapy: An Experience of Two Centers Free

Adult T-cell leukemia lymphoma (ATLL) is a mature T-cell malignancy of the CD4+ T cells. Since its discovery in 1976, it has been associated with the first oncogenic retrovirus identified – HTLV-1. The transmission of HTLV-1 primarily occurs vertically through breastfeeding, thus making it endemic only in certain population clusters, infecting 5–10 million individuals worldwide [1].

The disease occurs in only 5% of the HTLV-1-affected population. Southwestern Japan, the Caribbean islands, South America, and Tropical Africa are the areas bearing the brunt of the disease. In the USA as a whole, the incidence of ATLL is approximately 0.05 cases per 100,000 people and is more common in African Americans and Asian/Pacific Islanders with the average age at diagnosis in the sixth decade [2].

ATLL has been classified into four subtypes by Shimoyama – acute (60%), lymphomatous (20%), chronic (15%), and smoldering (5%). The different forms produce varying clinical presentations and prognoses. A provisional subtype of primary cutaneous tumoral has been described in the consensus meeting report [3]. This clinical spectrum led to trials of new treatment modalities with multi-agent chemotherapy regimens like cyclophosphamide, doxorubicin, vincristine, and prednisone (CHOP); vincristine, cyclophosphamide, doxorubicin, prednisone, ranimustine, vindesine, etoposide, carboplatin (VCAP-AMP-VECP); cyclophosphamide, vincristine, doxorubicin, prednisone, methotrexate, and ara-C (HyperCVAD); lenalidomide, antiviral therapies like zidovudine (AZT), and interferon alfa; and monoclonal antibodies such as defucosylated humanized anti-CCR4 (mogamulizumab). Unfortunately, the overall survival (OS) remained abysmal.

Florida has a diverse population and robust immigrant community from the Caribbean islands and South America with HTLV-1 endemic clusters providing a unique outlook on this rare disease. We reviewed the outcome of patients treated in two academic medical centers in Florida.

A chart review was carried out retrospectively in the two institutions in Central and South Florida (Moffitt Cancer Center and Memorial Healthcare System) where a large Hispanic population resides. The study was approved by the institutional board review committees in both institutions. Patients diagnosed from January 2000–January 2020 were included in the analysis.

All adults, age 18 or older, carrying the diagnosis of ATLL were included in the analysis. Demographics such as race and country of origin were based on self-reporting at the time of presentation. HTLV-1 infection testing was done with quantitative polymerase chain reaction (qPCR) and line immunoassay [4]. ATLL subtypes were defined [3] as follows. (1) Smoldering type; 5% or more abnormal lymphocytes of T‐cell nature in PB, normal lymphocyte level (<4 × 10⁹/L), no hypercalcemia (corrected calcium level <2.74 mmol/L), lactate dehydrogenase (LDH) value of up to 1.5 × the normal upper limit, no lymphadenopathy, no involvement of liver, spleen, central nervous system (CNS), bone and gastrointestinal tract, and neither ascites nor pleural effusion. Skin and pulmonary lesion(s) may be present. In case of less than 5% abnormal T‐lymphocytes in PB, at least one of histologically proven skin and pulmonary lesions should be present. (2) Chronic type, absolute lymphocytosis (4 × 10⁹/L or more) with T‐lymphocytosis more than 3.5 × 10⁹/L. LDH value up to twice the normal upper limit, no hypercalcemia, no involvement of CNS, bone, and gastrointestinal tract, and neither ascites nor pleural effusion. Lymphadenopathy and involvement of liver, spleen, skin, and lung may be present, and 5% or more abnormal T‐lymphocytes are seen in PB in most cases. (3) Lymphoma type, no lymphocytosis, 1% or less abnormal T‐lymphocytes, and histologically proven lymphadenopathy with or without extranodal lesions. (4) Acute type, remaining ATLL patients who have usually leukemic manifestation and tumor lesions but are not classified as any of the three other types. Lymphoma subtype was staged according to Lugano modification of Ann Arbor staging [5]. Response criteria were defined according to the proposed criteria from the international consensus meeting [6]. The choice of chemotherapy was done by the treating physician. The treatment was divided into three categories for uniformity of analysis. The first group included patients treated with CHOP. The second group included patients who received “intensive chemotherapy” including HyperCVAD [7] and etoposide, doxorubicin, vincristine, cyclophosphamide, and prednisone (EPOCH) [8], VCAP-AMP-VECP) [9], and cyclophosphamide, doxorubicin, vincristine, etoposide and prednisone (CHOEP) [10]. The third group of patients “others” received interferon and zidovudine (AZT/IF) or gemcitabine (1,000 mg day 1, 8), oxaliplatin (100 mg day 1) (GemOx), pralatrexate as a single agent, or gemcitabine (1,000 mg on day 1, 8), cisplatin (25 mg on day 1, 8), and dexamethasone (40 mg on day 1, 8) (GDP).

Descriptive statistics were analyzed for all treatment groups. χ² test, Fisher exact or Matt-Whitney-Wilcoxon test were used to examine significance of difference in baseline demographic, clinico-pathologic data for all groups and response to therapy. Patients we divided into “responders” as overall response rate (ORR) including those who had complete response (CR) or partial response (PR) to therapy versus “non-responders,” including patients with stable disease (SD) or progressive disease (PD) after therapy. Progression-free survival (PFS) was defined from the time of first dose of therapy till the first documented progression or death. OS was defined from the time of first dose of treatment till death of any cause. To examine the difference of PFS and OS between the treatment groups, the Kaplan-Meier method and log-rank test were utilized. Multivariate Cox regression was used to examine the impact of type of therapy on OS when adjusted for other clinically and statistically significant factors. Proportional hazard assumption was examined. All tests were two sided with a 0.05 significance level.

A total of 61 patients met the eligibility criteria and were included in the analysis. They were mostly females (66.7%). Most of the patients self-identified as African American (82.5%) while only 10.5% were Hispanic. Median age at diagnosis was 58 (range 32–75). The overall performance status on presentation was good as 71.9% had an Eastern Cooperative Oncology Group (ECOG) scale of 0–1 (Table 1).

Most of the patient identified their country of origin from the Caribbean. Jamaicans represented 34% of the patients. Those representing the USA were only 13% (Fig. 1).

Line immunoassay was done in 50 (82%) patients, the positive antigens in the line immunoassay were (P19-I/II = 71.1%, P24 = 78.7%, GP46 = 77%, GP21 = 42.6%, P19-I = 42.6%, GP46-I = 75.4%, GP46-II = 13.1%). PCR testing was done in 59 patients to confirm the presence of HTLV-1 infection.

The most common presentation seen was acute presentation (43.9%) followed by lymphoma presentation (36.8%) (Table 1). When the hematological parameters (white blood cells [WBC], absolute neutrophil count [ANC], hemoglobin [HGB], and platelets [PLT]) were compared between the different subtypes, only statistically significant difference was seen in the WBC between the acute and lymphoma types (22.2 × 10⁹/L) versus (5.6 × 10⁹/L) (p = 0.001). The median WBC count for the entire cohort was 8 × 10⁹/L, ANC (4 × 10⁹/L), HGB (10.7 × 10⁹/L), and PLT (195 × 10⁹/L) (Fig. 2). Only 2 patients had concomitant HIV infection. More patients had hepatitis B infection (13.1%, N = 8), while hepatitis C was documented in only 1 patient.

CD30 status was evaluated in 55 patients (90.2%). The patients were considered positive if they have an immunohistochemistry expression of CD30 >1%. A total of 47.3% had CD30 expression. The distribution of patients treated with different regimens was even.

OS for the patients with smoldering ATLL was 41.2 months (95% confidence interval [CI]: 8.8–73.5), chronic ATLL at 39.4 months (95% CI: 15.7–63.1), lymphomatous ATLL at 14.3 months (95% CI: 8.6–20), and acute ATLL at 8.3 months (95% CI: 5.6–11.1) (p = 0.15). CHOP was used to treat 19 patients (31.1%). The second most common approach included AZT/IF (N = 13, 21.3%). Intensified chemotherapy regimens were grouped together (EPOCH, CHOEP, VCAP-AMP-VECP, and HyperCVAD) and were used in 22 patients (38.6%) (Table 2).

Of the 25 patients with acute subtype, 10 (40%) were treated with frontline AZT/IF, 4 (16%) with CHOP, and another 11 (44%) with intensive chemotherapy. The lymphoma subtype (N = 21) had a different treatment approach as only 1 treated with AZT/IF (5%), CHOP (N = 9, 43%), and intensive chemotherapy (N = 10, 48%).

Patients were more likely to respond to CHOP (ORR: 72.2%) compared to intensified chemotherapy (ORR: 27.3%) or others (ORR: 25%) (p = 0.005). Initial response to frontline therapy translated into better OS with responding patients having OS (15.9 months) versus OS (7.2 months) in non-responders (p = 0.004) (Fig. 3).

There was a trend for improved PFS for the CHOP regimen (6.4 months) compared to (3.1 months) and (2.1 months) for the intensified regimens and others, respectively, but this did not reach statistical significance (p = 0.23). There was no statistical difference in the OS between the three regimens, 14.1 months versus 8.9 months versus 18.5 months for the CHOP versus intensified chemotherapy versus other options, respectively (p = 0.23) (Fig. 3).

Of the 61 patients, only 5 (8%) underwent allogenic hematopoietic stem cell transplant. OS for the transplanted patients was 15.9 months versus 10.2 months for no transplant (p = 0.127).

On multivariate analysis – including age, gender, ECOG, stage, subtypes, and first-line therapy as covariates – smoldering and chronic subtypes were the only predictive factors of OS with hazard ratio (HR) of 0.046 (95% CI: 0.003–0.774) and 0.031 (95% CI: 0.002–0.48), respectively.

Our study represents one of the largest cohorts of patients diagnosed with ATLL and treated in the USA. The disease is endemic in the Caribbean [11], leading to the increased prevalence in both of our referral centers. We have concluded that variable approaches in first-line management could have contributed to nonuniformity in outcome. Despite the diversified approach and lack of standard of care, CHOP regimen seemed to lead to best responses. Patients responding to first-line therapy have improved outcome likely related to better biology of the disease and sensitivity to chemotherapy.

Most of our patients self-identified as African American while the Hispanic patients represented around 10% which is consistent with previously reported cohorts from the East and South USA [2, 12]. Country of origin of our patients was comparable to similar cohorts from the state of Florida [13]. It is hard to conclude similarities to other experiences in the USA, however, New York, which is the second largest state with the incident of ATLL, showed prevalence of Jamaican origins but more patients reportedly from the Dominican and Trinidad [14, 15]. This is likely to represent different patterns of immigration to the USA between the two states from the Caribbean.

The increased incidence in females likely reflects the gender prevalence seen in Caribbean-origin ATLL [15, 16]. The subtypes of ATLL in our institutions were comparable to major datasets seen from the USA and Japan with the predominance of acute and lymphomatous types in contrast to indolent subtypes [14, 15, 17, 18]. Our institutional approach followed the overall recommendation on the frontline therapy with majority of acute ATLL treated with AZT/IF while only 1 patient with lymphoma ATLL was treated with combination antiviral/IF [19]. Frontline therapy with AZT/IF did not reveal promising ORR in our population compared to previous reports [20‒22]. Those initial reports included smaller number of patients treated in Europe mostly. This might be related to different biologies of the disease depending on the geographical origins [11, 16]. More recent reports from the USA have been conflicted on the outcome of frontline therapy with AZT/IF. The ORR had ranged 0–12.5% in smaller cohorts treated with AZT/IF [14, 15]. Those lower responses could be related to standard doses used and a higher dose might be needed to improve the outcome [13].

Our experience in CHOP is consistent with the published literature in this regard [9, 23, 24]. The lower response rate seen in the intensive regimen is likely related to the underlying patient population as it seems patients with acute ATLL were more likely to be treated with more intensive regimens and those patients likely to have worse outcomes than others [25]. Studies with intensive regimens such as EPOCH [8], VCAP-AMP-VECP [26], and HyperCVAD [27] showed an ORR of 50–70% in the frontline setting [23].

The PFS and OS were not different between the three regimens. This is likely related to the lack of an effective frontline approach and the difficulty in proceeding to a more definitive allogenic hematopoietic stem cell transplant [25].

More work is needed with large collaborative clinical trials given the rarity of this disease. ATLL therapy should focus on different approaches for different disease phenotypes given the lack of uniform presentation and outcome for those subtypes.

This study protocol was reviewed and approved by Institutional Review Board (IRB) at Moffitt Cancer Center, approval number 20275; date of decision, August 2, 2019. The protocol was granted exemption from requiring informed consent by the IRB.

Authors have not conflicts of interest to report.

No funding was provided for this study.

Hayder Saeed, Jose Sandoval-Sus, and Lubomir Sokol: design of the study. Amanda Brahim, Franco Castillo, Ning Dong, Stephanie Boisclair, Damaal Walker, Sanjay Bridgelall, and Ling Zhang: data collection. Hayder Saeed: data analysis. Hayder Saeed, Jerrin Joy, and Pullukkara: drafting the paper.

All data generated or analyzed during this study are included in this article. Further inquiries can be directed to the corresponding author.