Recent advances in understanding and managing T-cell lymphoma

Owing to the rarity of peripheral T-cell lymphoma (PTCL) and the heterogeneity of subtypes, there are no compelling data to guide the therapeutic approaches for such patients. Over the years, there have been remarkable advances in molecular subtyping and treatment of PTCL, although there are still many areas to be explored. In this review, we summarize recent updates on the evolution of understanding and treatment for PTCL.


Introduction
T/natural killer (T/NK) cell lymphoma represents a heterogeneous group of malignant lymphoproliferative diseases that arise from T-lymphocytes and NK cells. To distinguish it from immature T-cell neoplasms such as lymphoblastic leukemia/ lymphoma, mature T-cell lymphoma is often called peripheral T-cell lymphoma (PTCL). PTCL accounts for approximately 15% of all non-Hodgkin's lymphoma cases worldwide and shows regional differences in distribution.
Over the years, the molecular understanding of PTCL has advanced remarkably; as a result, the classification of PTCL was recently revised 1 . In addition, with new combinations of older drugs, novel agents such as checkpoint inhibitors, epigenetic modulators, and anti-folates have been introduced for the treatment of PTCL. Despite this progress, the prognosis for PTCL is worse than that for B-cell lymphoma, even than that before the era of rituximab 2,3 .
In this concise review, we provide an overview of the latest advances in the management of PTCL focusing on trials that have been carried out with novel agents. Of note, this review does not cover cutaneous T-cell lymphoma (CTCL) or extranodal NK-T cell lymphoma (ENKTL), because their biological features and treatment strategies differ from those of PTCL.

Brief review of the 2016 World Health Organization revision of nodal/extranodal peripheral T-cell lymphomas
Beginning with nodal PTCLs, up to 60-100% of angioimmunoblastic T-cell lymphoma (AITL) and up to 40% of PTCL not otherwise specified (PTCL-NOS) demonstrate surface markers of follicular helper T (TFH) cells 4-7 and have common genetic features, such as RHOA, TET2, DNMT3A, and IDH [8][9][10][11] . Two provisional entities have been introduced: follicular T-cell lymphoma and nodal peripheral T-cell lymphoma with the TFH phenotype. These entities can be diagnosed when their neoplastic cells express at least two or three TFH markers but do not display the clinicopathologic features of AITL, which would have been diagnosed as PTCL-NOS according to the 2008 World Health Organization classification. AITL and these two new entities are now categorized as nodal T-cell lymphoma with the TFH phenotype.
By definition, the PTCL-NOS designation encompasses all PTCLs lacking specific features that would allow categorization within any of the better-defined subtypes of PTCL, resulting in heterogeneity of this entity. A recent gene expression profiling study demonstrated that PTCL-NOS can be classified into two molecular groups according to the overexpression of either GATA3 or TBX21 12 , and the GATA3 subset shows inferior outcomes. However, as this study has not been applied in routine practice, the results have not led to the description of new entities.
ALK-negative anaplastic large-cell lymphoma (ALCL), which was a provisional entity in the 2008 classifications, has become a definite entity. This CD30-expressing neoplasm is not distinguishable from its ALK-positive counterpart on morphologic grounds, except for the absence of ALK expression. While the prognosis of ALK-negative ALCL is known to be less favorable 13 , the five-year survival rate is comparable to that of ALK-positive ALC when DUSP22 rearrangements are found 14 . After first being described in 1997 15 , breast implant-associated ALCL emerged as a distinct clinicopathologic entity, and it was proposed as a new entity in the 2016 classifications. All reported cases are ALK-negative, and the tumor is more frequently confined to the fibrous capsule. These cases show indolent clinical courses and respond well to implant removal and resection of the tumor. However, when the tumor presents with a mass discernible by radiologic or gross pathologic examination, it might be associated with a more aggressive clinical course 16,17 .
Major changes in extranodal PTCL cases have emerged from diseases that occur in the gastrointestinal (GI) tract. Enteropathyassociated T-cell lymphoma (EATL) is composed of two subtypes; type I EATL usually occurs following long-standing celiac disease (CD), showing large pleomorphic cells within an inflammatory background, and type II EATL occurs without antecedent CD, showing small, monotonous cells with epitheliotropism. In the 2016 classifications, diagnosis of EATL is to be used only for what was formerly type I EATL. Type II EATL has changed its name to monomorphic epitheliotropic intestinal T-cell lymphoma; it has been proposed as a new entity on the basis of its histologic, genetic, and molecular differences 18-20 . Whereas these two highly aggressive diseases show transmural growth, which often leads to GI bleeding or perforation, there is another one that grows superficially along the GI mucosa. Indolent T-cell lymphoproliferative disorder of the GI tract is a new provisional entity that usually presents with chronic diarrhea, weight loss, and malnutrition, mimicking the symptoms of inflammatory bowel disease 21 . The course of this disease is known to be indolent, although some cases of transformation have been reported 22 .

Current standard of care: conventional chemotherapy
Anthracycline-containing regimens such as a combination of cyclophosphamide, doxorubicin, vincristine, and prednisone (CHOP) or CHOP plus etoposide (CHOEP) are recommended as the front-line treatments for PTCLs 23,24 . Although clinical outcomes for these regimens vary according to histologic subtype, complete response (CR) rates range from 30% to 70%, and five-year overall survival (OS) rates range from 20% to 60%. Except in several small subsets, such as patients with ALKpositive ALCL or those with a low international prognostic index, the use of anthracycline-containing regimens failed to improve clinical outcomes 25-27 . Owing to the lack of efficacy of conventional CHOP, more intensive anthracycline-based regimens have been tried 28-32 ; however, given limitations of small numbers of patients, intractable toxicities, and unfavorable results, none is considered a standard option. CHOP-14 or CHOEP can still be considered in select patients 33,34 .
Various combinations of several non-anthracycline drugs have been evaluated in front-line settings. In the S0350 trial, a combination of cisplatin, etoposide, gemcitabine, and solu-medrol (PEGS) was tried in 26 patients with newly diagnosed PTCL 35 . The two-year progression-free survival (PFS) rate was 14%, and the two-year OS rate was 36%, which seemed to be no better than those of conventional combinations. Gemcitabine and etoposide were added to CHOP in 26 patients with newly diagnosed PTCL and this resulted in a CR rate of 62% and a median 215 days of event-free survival 36 . An Italian group reported outcomes of biweekly administration of six cycles of gemcitabine, ifosfamide, and oxaliplatin in 21 patients with high-risk PTCL. The CR rate was 67%, and the five-year event-free survival rate was 49%. In a recent report from the randomized phase II study of the UK group using the combination of gemcitabine, cisplatin, and methylprednisolone (GEM-P), objective response rates (ORRs) were 57.1% for the CHOP arm and 43.2% for the experimental arm 37 . Although grade 3-4 neutropenia or febrile neutropenia was more common in the CHOP arm, there was no difference in the two-year OS rate (53.1% versus 64.7%) or PFS rate (36.0% versus 39.0%). Outcomes of the recent trials for front-line treatment are summarized in Table 1.
Collectively, although there is limited evidence that CHOP or its variants improve the prognosis of patients with PTCL, there is also little evidence that a certain non-anthracycline combination can replace CHOP. A randomized phase II study is under way to compare a combination of ifosfamide, carboplatin, and etoposide (ICE) and dexamethasone with CHOP (NCT02445404).

Current standard of care: consolidation therapy
Except for ALK-positive ALCLs, which show favorable outcomes with chemotherapy alone, current guidelines recommend consolidative autologous stem cell transplantation (ASCT) in patients with chemosensitive PTCL 23,24 . However, this recommendation was not established based on prospective randomized trials.
In a report by Reimer et al., 83 patients with PTCL received four to six cycles of CHOP, and for patients in partial response (PR) or CR at the end of induction, ASCT was performed 38 . Fifty-five patients advanced to ASCT, and the three-year OS for patients who underwent ASCT was 71%. In the NLG-T-01 trial, 115 out of 160 patients with PTCL underwent ASCT after induction treatment of CHOEP-14 or CHOP-14 39 . The 5-year PFS and OS rates were 44% and 51%, respectively, and the recently announced follow-up data revealed that the 10-year PFS and OS rates were 38% and 41%, respectively 40 . Although it appears that prolonged survival can be achieved by ASCT, it should be noted that approximately 30% of patients are unable to receive it and this is mostly due to early progression. Several studies have compared the efficacy of front-line allogeneic stem cell transplantation (alloSCT) with ASCT 41,42 ; however, the outcomes of alloSCT were not significantly different from those of ASCT. Taken together, these data suggest that ASCT should be considered as the first treatment in patients with chemosensitive PTCL.

Current standard of care: relapsed or refractory peripheral T-cell lymphoma
Up to 20-30% of patients fail to achieve an initial response, and even after ASCT, over 50% will experience relapsed disease 38,39 . For those who have relapsed or refractory PTCL (rrPTCL), salvage chemotherapy with or without subsequent consolidation treatment can produce long-term remission. However, owing to refractoriness of the disease, overall outcomes remain poor 43 . Several novel agents are actively being investigated in patients with rrPTCL. The major clinical findings of the studies with novel agents are summarized in Table 2.

Monoclonal antibodies Alemtuzumab
Alemtuzumab is a humanized monoclonal antibody (mAb) against CD52. As a salvage treatment, it has been evaluated either as monotherapy 44 or as combination therapy with various backbones [45][46][47] . Although high CR rates were observed (36-54%),  58 . Several BV-based combinations have also been evaluated. In a phase I trial, 39 treatment-naïve patients with a diagnosis of CD30-positive PTCL were recruited to receive either sequential treatment (two cycles of BV, six cycles of CHOP, and eight cycles of BV) or combination treatment (six cycles of BV in combination with vincristine-omitted CHP and 10 cycles of BV) 59 . There were seven patients with non-ALCL PTCL-two PTCL-NOS, two AITL, two adult T-cell leukemia/lymphoma (ATLL), and one EATL-they were all allocated to a combination arm for which a CR rate of 100% was noted. A subsequent four-year follow-up analysis revealed that six out of seven of them were still alive 60 . Based on these favorable outcomes, a randomized phase III trial is ongoing to compare BV-CHP

Mogamulizumab
Mogamulizumab is a humanized mAb targeting the CC chemokine receptor 4 (CCR4), which is expressed in physiologic regulatory T cells. CCR4 is also expressed in nearly 90% of ATLL cases 61

Immune checkpoint inhibitors
It has long been recognized that the tumor microenvironment plays important roles in lymphomagenesis, proliferation, and immune evasion 66,67 . In T-cell lymphoma, the microenvironment defines the tumor itself (such as in AITL), and many subtypes manifest strong tissue tropism (such as in CTCL or primary GI lymphomas).
Given that PD-1 is expressed in a substantial portion of PTCL cases 68 and that PD-L1 is frequently expressed in certain virusassociated lymphomas 69 , this pathway is of major interest. Nivolumab is a fully human anti-PD1 mAb. In a phase I, doseescalating study, a total of 81 patients with lymphoid malignancy (B-cell lymphoma 31, T-cell lymphoma 23, and multiple myeloma 27) received nivolumab at doses of 1 or 3 mg/kg every two weeks 70 . Among five patients with PTCL, two PRs were observed, and the median PFS was 14 weeks. In the two responding patients with PTCL, a sustained duration of response (DoR) was observed. Several trials are ongoing to evaluate the role of this agent in patients with PTCL as a monotherapy (NCT03075553 and NCT02973113) or as combination therapy with BV (NCT0258163). Evaluation of pembrolizumab, another PD-1 mAb, in PTCL has been based on anecdote 71,72 , and two studies are under way for patients with rrPTCL (NCT03021057 and NCT02362997). In regard to mAbs targeting PD-L1 or PD-L2 or both, prospective studies with avelumab (NCT03046953) and durvalumab (NCT03161223 and NCT03011814) are taking place.
Ipilimumab is a fully humanized mAb targeting cytotoxic T lymphocyte-associated protein 4 (CTLA-4). In a phase I study, nivolumab combined with ipilimumab was given at 3 mg/kg or 1 mg/kg every three weeks for four doses, and this was followed by nivolumab monotherapy every two weeks 73 . Eleven patients with heavily pre-treated T-cell lymphoma were included, one PR and four standard deviations were noted, and one patient proceeded to alloSCT.

Anti-folates
Anti-folates demonstrate anti-tumor efficacy by inhibiting dihydrofolate reductase, which converts dihydrofolate to tetrahydrofolate. Depletion of tetrahydrofolate disrupts the synthesis of pyrimidines and amino acids such as serine, glycine, and methionine 74 .

Pralatrexate
In an early phase trial with pralatrexate, a higher CR was observed in patients with PTCL compared with those of B-cell lymphoma at the cost of significant mucositis 75 . This toxicity was later found to be alleviated by the administration of vitamin B 12 and folic acid.
In the PROPEL (Pralatrexate in Patients with Relapsed or Refractory Peripheral T-Cell Lymphoma) trial, a total of 111 patients with rrPTCL received pralatrexate at a dose of 30 mg/m 2 for six out of seven weeks along with vitamin B 12 and folic acid 76 . Among 109 evaluable patients, 12 CRs (11%) and 20 PRs (18%) were noted. The median duration of response and the median OS were 10.1 and 14.5 months, respectively. The most common grade 3-4 toxicities were thrombocytopenia (32%), mucositis (22%), neutropenia (22%), and anemia (18%), but pralatrexate was well tolerated, as the overall dose intensity was 80%. In the Japanese trial with the same dose and schedule, nine (45%) out of 20 evaluable patients achieved an objective response, and grade 3-4 thrombocytopenia occurred in 40% of cases 77 .
Based on the synergism pralatrexate has shown in preclinical analyses 78,79 , several combinations with it have been evaluated. In a phase II trial, a front-line combination of cyclophosphamide, etoposide, vincristine, and prednisone (CEOP) alternating with pralatrexate was evaluated in 33 patients with untreated PTCL 80 . Seventeen CRs (52%) and six PRs (18%) were achieved, and 15 patients received ASCT. The two-year PFS and OS rates were 39% and 60%, respectively, similar to outcomes with CHOP. Grade 3-4 anemia (27%), febrile neutropenia (18%), mucositis (18%), and thrombocytopenia (12%) occurred. In a phase I trial, pralatrexate followed by gemcitabine was administered in 34 patients with relapsed or refractory lymphoproliferative malignancies, among whom 11 patients with PTCL were included 81 . Of the 33 patients who were evaluable for response, seven (21%) showed a partial response, and two of the seven had T-cell lymphoma. In another case series, five elderly patients with rrPTCL received weekly administration of pralatrexate (15 mg/m 2 ) and bortezomib (1.3 mg/m 2 ) for three out of four weeks until progression 82 . One patient achieved CR after four cycles, which lasted over 12 months.

Histone deacetylase inhibitors
Histone deacetylase (HDAC) inhibitors demonstrate anti-tumor efficacy by upregulating the expression of genes for cell cycle regulators, cell type-specific differentiation, and pro-apoptotic proteins 83 Interestingly, an in vivo and in vitro study has suggested a role for romidepsin in the treatment of EBV-associated cancer, where it can induce the EBV lytic cycle 92 . In a Korean pilot study, however, three out of five patients with NK/T-cell lymphoma experienced EBV reactivation after romidepsin treatment 93 . This finding is consistent with the findings from the aforementioned two cases of EBV-associated lymphoproliferative disorder 88 .
Thus, EBV-reactivation should be taken into account when using romidepsin to treat EBV-associated lymphoma.
Several romidepsin-based regimens are being evaluated. In the front-line setting, romidepsin plus CHOP was administered in 37 patients with PTCL 65 . Although hematologic toxicities precluded the completion of the planned treatment in 18% of patients, the CR rate was 51%, the PR rate was 17%, and the OS rate at 30 months was 70.7%. Prospective trials with romidepsin plus CHOP (NCT01796002) or CHOEP-21 (NCT02223208) are ongoing.
In the salvage setting, romidepsin at 12 mg/m 2 (days 1, 8, and 15) with gemcitabine at 800 mg/m 2 (days 1 and 15) up to six cycles, followed by romidepsin maintenance (at 14 mg/m 2 ), was given to 20 patients with rrPTCL 94 . The clinical outcomes of the combination were a little better than those with monotherapy; the CR rate and the ORR were 15% and 30%, respectively. Another phase I study of romidepsin, gemcitabine, dexamethasone, and cisplatin was carried out by a Canadian group 95 , and of the 10 patients with PTCL, five responded. In addition, when romidepsin was added to ICE 96 , the CR rate was 64% (9/14) and the ORR was 78% (11/14). However, grade 3-4 thrombocytopenia and neutropenia occurred in 95% and 84% of the cycles, respectively.

Belinostat
Belinostat is a hydroxamic acid-derived, pan-HDAC inhibitor that demonstrates high affinity for the class I and II HDACs. In an early phase II trial, 24 patients with rrPTCL received belinostat, the ORR was 25% (6/24), and two patients experienced CR 97 .
In the phase II BELIEF (A Multicenter, Open Label Trial of Belinostat in Patients With Relapsed or Refractory Peripheral T-Cell Lymphoma) trial, 129 patients with rrPTCL were enrolled to receive belinostat at 1,000 mg/m 2 for five days every three weeks 98 . Among 120 evaluable patients, 13 achieved CR (11%) and 18 achieved PR (15%), and the ORR was 26%. In patients with CR, the median DoR exceeded 29 months. A subtype analysis revealed that, among 22 patients with AITL, the ORR was 45% (four CRs and six PRs) 99 . Most adverse events-nausea (42%), fatigue (37%), and pyrexia (35%)-were non-hematologic. Only 7% and 10% of patients experienced grade 3-4 thrombocytopenia and anemia, respectively.
These favorable hematologic toxicities provided momentum for initiating the next combination trial. In a phase I trial, belinostat at a dose of 1,000 mg/m 2 for five days every three weeks was combined with CHOP for six cycles in 23 patients with untreated PTCL 100 . The combination was well tolerated; 18 patients (78%) completed six cycles. Among 21 evaluable patients, the ORR was 86% (n = 18), and most patients comprising the ORR were CR (n = 14, 67%).

Chidamide
Chidamide is an orally administered benzamide class of HDAC that demonstrates selective inhibition of HDAC1, 2, 3, and 10 101 . In a phase II trial, 83 patients with rrPTCL took 30 mg of chidamide twice weekly 102 . Among 79 evaluable patients, 11 CRs (14%) and 12 PRs (15%) were noted, and the median time to response was 1.4 months and the median DoR was 9.9 months. Grade 3-4 thrombocytopenia occurred in 22% of patients, and neutropenia occurred in 11% of patients. On the basis of these results, the Chinese Food and Drug Administration approved the use of chidamide to treat rrPTCL in 2014. Subsequently, large-scale real-world data were recently released 103 . Chidamide monotherapy resulted in an ORR of 39% (100/256) and a CR rate of 9% 27/256). Chidamide-based combination therapies demonstrated the highest ORR (51%, 65/127) and CR rate (12%, 15/127). In the monotherapy group, grade 3-4 thrombocytopenia occurred in 10.2% of cases and neutropenia occurred in 6.2% of cases. The real-world data demonstrated clinical outcomes similar to those of the phase II trial. Several trials with chidamide-based combinations (NCT02809573, NCT02987244, NCT02856997, and NCT03023358) are ongoing.

Other histone deacetylase inhibitors
Vorinostat is an orally administered benzamide HDAC inhibitor of class I-II. In a front-line setting, a combination of vorinostat plus CHOP demonstrated 79% and 81% of two-year PFS and OS rate, respectively 104 . In a relapsed or refractory setting, a combination of vorinostat, lenalidomide, and dexamethasone (40 mg once daily) was explored 105 , but the outcomes were not remarkable.

Immunomodulatory drugs
Lenalidomide shows anti-lymphoma efficacy through immune modulation of the microenvironment and anti-proliferative and anti-angiogenic mechanisms 107 . Two trials have investigated single-agent lenalidomide for 21 out of 28 days in patients with rrPTCL 108,109 . The CR rate ranged from 8% to 30%, and the ORR ranged from 22% to 30%. A substantial portion of patients (26-35%) experienced toxicities and this led to treatment discontinuation. Outcomes in patients with AITL, compared with other subtypes, were unique 108 and this suggests a role for the microenvironment in this subtype. In a phase I/II trial, a combination of romidepsine, lenalidomide, and carfilzomib was tried in patients with rrPTCL 110 . In 16 evaluable patients, the CR rate was 31% (5/16) and the PR rate was 19% (3/16). Again, four out of five patients with AITL attained CR. A phase II study for untreated PTCL patients using lenalidomide plus romidepsin (NCT02232516) is ongoing.
Thalidomide, a prototype drug of this class, was evaluated for combination treatment. A combination of gemcitabine, cisplatin, prednisone, and thalidomide (GDPT) was compared with CHOP for patients with newly diagnosed PTCL 111 . In total, 103 patients were randomly allocated into two groups; 52 received GDPT and 51 received CHOP. The PFS, the primary end-point of the study, was significantly better in the GDPT arm; the two-year PFS rates were 57% versus 35% (P <0.05). Other outcomes, including the ORR (67% versus 49%, P = 0.046) and the two-year OS (71% versus 50%, P <0.05), were also favorable for the GDPT arm. Grade 3-4 myelosuppression occurred in 44% versus 41% of patients.

Proteasome inhibitors
In a phase II trial, the single-agent bortezomib induced an ORR of 67% in 15 patients with rrCTCL 112 . Because a chemosensitization effect was expected 113 , the efficacy of bortezomib was explored in combination with CHOP in patients with newly diagnosed PTCL/ENKTL/CTCL 114 . Of the 65 patients, 30 achieved CR (65%) and five other patients achieved PR. The ORR was 76% (35/46). When only three major subtypes of PTCLs (PTCL-NOS, AITL, and ALCL) were analyzed, the CR rate was 73% and the ORR was 87%. However, owing to frequent relapse after remission, the three-year OS and PFS rates were 47% and 35%, respectively. In another trial by the French group, the combination of bortezomib plus intensified CHOP-like regimen (ACVBP) was studied in 57 patients with newly diagnosed PTCL 115 . The outcomes did not appear to be higher than chemotherapy alone, as the CR rate was 45%.
A newer, irreversible proteasome inhibitor (carfilzomib) is currently under investigation in patients with PTCL (NCT01336920 and NCT03141203). Ixazomib, an oral proteasome inhibitor, is being assessed in a phase 2 study (NCT02158975).

Inhibitors of PI3K/mTOR pathways
PI3Ks transduce signals from various growth factors and cytokines into intracellular molecules by generating phospholipids, which then activate downstream effectors such as AKT or mTOR 116 .
Duvelisib, a dual inhibitor of PI3Kγ and δ, was studied in a phase 1 trial in patients with hematologic malignancies 117  Given that a substantial portion of patients with PTCL show phospho-AKT overexpression, which confers a poor prognosis 120 , AKT can be a reasonable target. In a phase II trial with MK2206, however, a frustrating result was observed; none of the three patients with PTCL responded 121 .
Activation of mTOR induces cell growth and survival in cancer, and especially in lymphoma, Myc activity is known to depend on the mTOR pathway 122 . In a phase II study of 30 patients with untreated PTCL who received everolimus plus CHOP 123 , CR was observed in 17 (57%) and PR was observed in 10 (33%). Despite these favorable tumor responses, frequent relapses were noted as the two-year PFS rate was 33%. Single-agent everolimus in 16 patients with relapsed or refractory T-cell lymphoma demonstrated an ORR of 44% and a median DoR of 8.5 months 124 .

Alternative agents Bendamustine
Bendamustine, which contains the structures of both alkylating agents and purine analogs, is one of the standard agents in indolent B-cell lymphoma. In the phase II BENTLY (Bendamustine in Patients With Refractory or Relapsed T-cell Lymphoma) trial, 60 patients with refractory or relapsed T-cell lymphoma received bendamustine at a dose of 120 mg/m 2 for two consecutive days every three weeks for up to six cycles 125 . With the majority of patients having AITL and PTCL-NOS (91%), the ORR was 50% and the median DoR was 3.5 months.

Aurora-A kinase inhibitors
Aurora-A is a mitotic kinase overexpressed in several subtypes of PTCL 127 . Alisertib, a small-molecule inhibitor of aurora-A kinase, has demonstrated favorable anti-tumor efficacy against rrPTCL in two phase II trials 128,129 . Given the promising results, a phase III LUMIERE (Alisertib or Investigator's Choice in Patients With Relapsed/Refractory Peripheral T-Cell Lymphoma) trial was performed in patients with rrPTCL to compare the efficacy of alisertib versus the investigator's choice, including pralatrexate, romidepsin, or gemcitabine 130 . With a planned accrual of 271 patients, interim analyses were carried out after 238 patients were recruited. ORRs, the primary end-points of the study, were 33% for alisertib and 43% for the investigator's choice 130 . In addition, no benefit was observed across the safety profiles. With these results, the trial was prematurely terminated.

Tipifarnib
Tipifarnib is an orally administered nonpeptidomimetic farnesyl transferase inhibitor. In a phase II trial, 93 patients, including 16 patients with T-cell lymphoma, received tipifarnib 300 mg twice daily for three out of four weeks 131 . A higher response rate was observed in the T-cell/Hodgkin's lymphoma cohort (31%) compared with the aggressive (17%) or indolent (7%) B-cell lymphoma cohorts. Among eight patients with PTCL-NOS, three achieved CR and one achieved PR. Based on these promising results, a phase II trial was carried out solely on patients with rrPTCL, and the preliminary results were recently reported 132 . A total of 18 patients received tipifarnib at a dose of 600 mg twice daily on days 1-7 and 15-21 in 28-day cycles, which demonstrated three PRs (17%) at the cost of grade 3-4 neutropenia (83%) and thrombocytopenia (61%).

CPI-613
CPI-613 is a novel lipoate derivative that inhibits mitochondrial metabolism in cancer cells 133 . In a phase I trial in patients with hematologic malignancies, the maximum tolerated dose was determined to be 2,940 mg/m 2 , a major toxicity was renal failure 134 . In another phase I trial in patients with relapsed or refractory T-cell lymphoma, bendamustine plus CPI-613 was administered 135 . Out of five evaluable patients, three CRs and one PR were observed.

Conclusions
Despite recent progress, there are hurdles to overcome in managing patients with PTCL, such as the poorly understood role of certain molecular features. Given the insufficient clinical outcomes out of the current standard of care, there are still unmet needs for the novel therapy.
Owing to the introduction of novel therapeutic agents however, recent outcomes are worthy of attention. Tailored clinical approaches regarding what drugs to initiate, when to consolidate patients, and how to best salvage patients require further investigation, including prospective trials.

Competing interests
The authors declare that they have no competing interests.

Grant information
The author(s) declared that no grants were involved in supporting this work.