A multicentre phase II clinical experience with the novel aza-epothilone Ixabepilone (BMS247550) in patients with relapsed or refractory indolent non-Hodgkin lymphoma and mantle cell lymphoma


Owen A. O’Connor, Lymphoid Development & Malignancy Program, Lymphoma Service, Columbia University, Herbert Irving Comprehensive Cancer Center, 1130 St. Nicholas Avenue, Rm 216, New York, NY 10032, USA. E-mail: oo2130@columbia.edu


The epothilones represent a novel group of microtubule stabilization agents that appear to retain activity even in chemotherapy-resistant cell lines and animal models. Because of their ability to overcome chemotherapy resistance, we conducted a phase II study of Ixabepilone in patients with indolent non-Hodgkin lymphoma and mantle cell lymphoma (MCL). Ixabepilone was given at a dose of 25 mg/m2 weekly for three of four consecutive weeks. Patients were required to have received ≤4 prior chemotherapy regimens, with an interval of at least one month since the last treatment, 3 months from prior rituximab, and 7 d from prior steroids, an absolute neutrophil count >1 × 109/l and a platelet count >50 × 109/l. Dose reductions were allowed. The overall response rate in assessable patients was 27% in this otherwise heavily treated population. One patient with chemotherapy-refractory follicular lymphoma attained a complete remission that lasted approximately 8 months. Three responses were also seen in refractory MCL and one in small lymphocytic lymphoma. The duration of response ranged from 2 to 8 months. Major toxicities included fatigue, myelosuppression and neuropathy. These data suggest that Ixabepilone has activity in chemotherapy-refractory lymphoma.

The epothilones are naturally occurring cytotoxic macrolides synthesized by the myxobacterium Sorangium cellulosum (Gerth et al, 1996; Holfe, 1996). They represent a new class of non-taxane tubulin polymerization agents capable of causing mitotic arrest at the G2/M transition(Rose, 1992; Rowinsky et al, 1993). The total synthesis of epothilones A and B, as well as a host of different derivatives, has been accomplished(Bollag et al, 1995; Meng et al, 1997,Su, 1997). Based upon their early preclinical activity and microtubule binding affinity, at least five different analogs have emerged through early phase clinical trials, including: (1) Ixabepilone (formerly known as BMS247550); (2) 12,13-desoxyepothilone B (Epo D; KOS-862); (3) Patupilone (Epothilone B; EPO906); (4) BMS-310705, and (5) ZK-EPO(de Jonge & Verweij, 2005).

The epothilones exhibit a mechanism of action similar to paclitaxel, despite having a significantly dissimilar chemical structure. Both molecules bind to α,β-tubulin heterodimers in a 1:1 stoichiometric ratio(Bollag et al, 1995; Kowalski et al, 1997). Despite these mechanistic similarities, structural analyses of these two drugs have revealed no obvious homology, and a common pharmacophore has yet to be established. Emerging data is beginning to suggest that the epothilones and paclitaxel may bind to distinctly different, though close, allosteric binding sites on the tubulin protein. Interestingly however, when both substrate ([H3] paclitaxel) and inhibitor (epothilone A or B) concentrations were varied, a classical competitive inhibition pattern was obtained for both epothilone A and B (Kowalski et al, 1997). The epothilones stabilize cellular microtubule assemblies, thus inducing the formation of hyperstable tubulin polymers in both cultured cells and microtubule protein.

Early preclinical observations demonstrated that epothilones were considerably more potent than paclitaxel(Chou et al, 1998a,b, 2001). In addition, the epothilones are considered superior in their ability to retain marked activity against multi-drug-resistant (MDR) cell lines and tumours, scenarios in which many other natural products generally fail (Miller et al, 1991; Wilson et al, 1993, Chabner et al, 1994). These observations provide a strong rationale to explore these drugs in lymphoma where acquired drug resistance is a prominent feature of the disease.

The indolent lymphomas represent a very heterogeneous group of generally incurable diseases that include a diversity of non-Hodgkin lymphoma (NHL) sub-types. These diseases are characterized by a relentless penchant for relapse, with a trend towards gradually evolving acquired chemotherapy resistance over time. Similarly, mantle cell lymphoma (MCL) is characterized by short remissions to conventional cytotoxic therapy with no expectation for long-term cure. Because of the diversity of drugs that patients with these diseases often receive, relapsed indolent lymphoma represents an interesting opportunity to evaluate new therapeutic strategies designed to overcome drug resistance. The epothilones, given their potential to overcome many mechanisms of chemotherapy resistance, represent a novel class of drugs with the theoretical potential to treat drug-resistant NHL. We report here the first clinical experience with this novel class of drugs in patients with hematological malignancies.


Patient selection

Patients were required to have histologically confirmed lymphoma according to the World Health Organization (WHO) classification, including: chronic lymphocytic leukemia (CLL); B-cell small lymphocytic lymphoma (SLL); marginal zone lymphoma; follicular lymphoma, grades 1, 2 or 3; MCL and Waldenström macroglobulinemia. Prior history of transformed lymphoma was permitted as long as recent biopsies revealed no evidence of aggressive lymphoma. Eligibility required: (1) measurable disease (defined as ≥1 cm with spiral computed tomography (CT) scan). Leukemic patients required an absolute lymphocytosis >5 × 109/l with >30% bone marrow lymphocytes; (2) no more than four prior lines of cytotoxic therapy; (3) be off all chemotherapy for at least four weeks (6 weeks for BCNU (carmustine) or mitomycin C, and at least 7 d must have elapsed since steroids); (4) a period of at least 3 months since the last administration of any monoclonal antibody, unless there was objective evidence of progression following rituximab; (5) to be 18 years of age or older, have a life expectancy of 3 months or greater, and have a Karnofsky Performance Score >60%; (5) no signs of congestive heart failure. Patients were allowed febrile episodes up to 38·5°C as long as there was no evidence of active infection. Patients were eligible if they had at baseline a grade 1 or less sensory neuropathy. The protocol was approved by the Institutional Review Boards of the participating centres (Memorial Sloan Kettering Cancer Center, University of Vermont), and all patients signed an informed consent.

Patients were required to have: an absolute neutrophil count (ANC) >1 × 109/l (if known lymphomatous involvement of the bone marrow, then ANC > 0·5 × 109/l); platelet count ≥50 × 109/l, total bilirubin ≤1·5 times upper institutional limit of normal (ULN); aspartate transaminase (AST)/ alanine transaminase (ALT) ≤2·5 times ULN (4× ULN if liver involvement); and creatinine ≤2 time ULN, or creatinine clearance ≥50 ml/min. Patients were excluded if they were pregnant; had evidence of intracranial disease; had major surgery within four weeks of study drug; had uncontrolled illness including active infection, uncontrolled hypertension, unstable angina pectoris, cardiac arrhythmia, a myocardial infarction or cerebrovascular accident within 6 months; known human immunodeficiency virus (HIV) disease, or psychiatric illness/social situations that would limit compliance with study requirements.

Study design

This was a multicenter, single agent Phase II study of Ixabepilone in patients with relapsed, refractory indolent NHL and MCL. The major objectives of the study were to determine the frequency and duration of complete and partial response. The study employed a Simon two-stage design. Initially, 18 patients were enrolled in the first stage. If two or fewer patients responded, the trial would have been terminated. If at least three patients responded, then up to a total of 35 patients could be enrolled. These statistics were predicated on the assumption that, should 20% of patients respond, the drug would be declared as having promising activity. Assessable patients were required to complete at least two cycles of therapy.

Drug administration

Ixabepilone (BMS 247550; NSC #710428) was supplied by the Division of Cancer Treatment and Diagnosis (DCTD), National Cancer Institute. Ixabepilone (Bristol-Myers Squibb, CT) was administered at 25 mg/m2 IV once a week for three weeks out of every four, infused over 1 h. Ixabepilone was provided in 20 or 30 mg vials, which was reconstituted with 11 or 16·5 ml of provided diluent (ethanol plus polyoxyethylated castor oil [Cremophor EL] in a 1:1 mixture) respectively, to a concentration of 2 mg/ml. Once reconstituted, the required amount of drug was withdrawn and further diluted in Lactated Ringer’s Injection in a glass bottle or PVC-free IV bag. A “taxol-safe” tubing set was attached along with a 0·22 in-line filter. The final concentration range of drug was between 0·2 to 0·6 mg/ml.

All patients were monitored for at least 2 h following first drug administration, with hourly vital signs. Patients with isolated tachycardia (heart rate > 120 beats/min) had an electrocardiogram. There was no upper limit on planned therapy, and patients could continue to receive drug as long as there was evidence of clinical benefit without excess toxicity. All patients received primary prophylaxis (50 mg diphenhydramine PO or IV) and an H2 blocker (ranitidine 150 mg PO or 50 mg IV) 30 to 60 min prior to treatment.

Dose modification

Patients were treated at an initial dose of 25 mg/m2 weekly for three consecutive weeks on an every 4 week basis. Treatment was withheld for patients whose blood counts failed to meet the eligibility criteria for re-treatment. Use of antiemetics, erythropoietin and filgrastim were allowed if deemed necessary by the treating physician. Their use was dictated by standard institutional guidelines. While rare, a single oral dose of granisetron was administered in most cases where an antiemetic was required.

The criteria for re-treatment were the same as for initial treatment. Patients who failed to meet the criteria for re-treatment were delayed for one week. Initiation of subsequent treatments could be delayed for a maximum of three weeks to allow for recovery from any treatment-related toxicities. Patients who required two three-week schedule delays were removed from study. Patients who developed their first Grade 3 or 4 non-hematological toxicity during treatment could be retreated with the same dose of Ixabepilone. However, if the patient experienced continued Grade 3 or 4 toxicity of any nature, the dose was reduced to 20 mg/m2. Patients who experienced persistent toxicity or who did not tolerate the 20 mg/m2 dose level had their dose reduced to 17·5 mg/m2. Continued toxicity after this dose reduction resulted in removal from study.

Response criteria

Response criteria for patients enrolled on study followed the National Cancer Institute (NCI)-Sponsored Working Group guidelines (Cheson et al, 1999). Response criteria for patients with leukemic forms of NHL, including CLL, followed the NCI Sponsored Working Group guidelines (Cheson et al, 1996). Response was assessed every two cycles. Evaluable patients were required to have received at least two cycles of therapy. Imaging studies occurred after every two cycles, and then every 3 to 4 months off therapy.

Electrodiagnostic studies

Electrodiagnostic evaluations, including nerve conduction studies and needle electromyography (EMG) were completed on select patients. Motor and sensory nerve conduction and needle electromyographic studies were performed. Standard motor nerve conduction studies include evaluation of the median, ulnar, peroneal, and tibial nerves either unilaterally or bilaterally. Standard sensory nerve conduction studies include evaluation of the median, ulnar, radial, superficial peroneal and sural sensory nerves either unilaterally or bilaterally. For motor nerve conduction studies, onset distal latency, baseline to peak amplitude, and segmental conduction velocity of the compound muscle action potential (CMAP) was recorded. For sensory nerve conduction studies, onset distal latency, peak to peak amplitude, and conduction velocity of the sensory nerve action potential (SNAP) were recorded. F-response latencies were recorded on the median, ulnar, peroneal, and tibial motor nerves. Needle EMG was performed on selected muscles of the upper and or lower extremity and their corresponding paraspinal muscles.



This study reports our initial experience of 28 consecutive patients enrolled between May 2003 and January 2006. Table I presents the demographic features of all study patients. Twenty-eight patients were registered for treatment, of whom 22 were assessable for response (i.e. they completed at least two cycles of therapy). All patients were assessable for toxicity. Table II provides an accounting for all patients not considered assessable. Of the inevaluable patients, three came off study for rapid progression of disease (POD), one of whom experienced an unusual transformation of his MCL to a Burkitt histology. Two patients were removed from study after only two doses of drug for infectious complications, including cellulitis and pneumonia, and one was removed for grade 2 asthenia after three doses of drug.

Table I.   Patient demographics.
 n = 28
  1. *Cytotoxic therapies include all chemotherapy based treatment programs, radiation and radioimmunotherapies, and excludes all biologically based treatments including rituximab and interferon based treatments.

  2. †Pralatrexate is 10-propargyl-10-deazaaminopterin.

  3. (R)-CHOP, (rituximab) cyclophosphamide, doxorubicin, prednisone, vincristine; CVP, cyclophosphamide, vincristine, prednisolone; R-ICE, rituximab, ifosfamide, carboplatin and etoposide; EPOCH, cyclophosphamide, doxorubicin, etoposide, prednisone, vincristine; ProMACE, prednisone, methotrexate, doxorubicin, cyclophosphamide, etoposide; HyperCVAD, cyclophosphamide, vincristine, adriamycin, dexamethasone; CEPP, cyclophosphamide, etoposide, procarbazine and prednisone; HDC + ASCT, high dose chemotherapy + autologous stem cell transplantation.

Number of patients
 Male20 (71%)
 Female8 (29%)
 White non-Hispanic 27 (96%)
 African American/Hispanic1 (4%)
 Median age, years65 (42–81)
 Follicular lymphoma6 (21%)
  Grade 12 (7%)
  Grade 23 (11%)
  Grade 31 (4%)
 Mantle cell lymphoma15 (54%)
 Small lymphocytic lymphoma/chronic   lymphocytic leukaemia5 (18%)
 Marginal zone lymphoma1 (4%)
 Waldenstrom1 (4%)
Summary of prior treatment*
 Median number of prior therapies4 (1–6)
 Median number of prior   cytotoxic therapies3 (1–4)
 Median duration of time from   last treatment 3 months (1–11)
 Median time from diagnosis50 months (4–236)
Alkylator based therapy
 CHOP6 (21%)
 R-CHOP8 (29%)
 CVP +/− rituximab4 (14%)
 R-ICE2 (7%)
 EPOCH/ProMACE/HyperCVAD +    methotrexate/cytarabine1 (4%)/1 (4%)/2 (7%)
 IV/oral CEPP1/1 (4%/4%)
 Chlorambucil/cyclophosphamide2 (7%)/1 (4%)
 HDC + ASCT3 (11%)
Purine analog based therapy
 Fludarabine +/− rituximab8 (29%)
 Fludarabine/mitoxantrone/   dexamethasone2 (7%)
 Fludarabine/cyclophosphamide/   rituximab3 (11%)
 Pentostatin/cyclophosphamide   +/− rituximab4 (14%)
 Cladrabine (2-CDA)2 (7%)
Biological/experimental based treatment
 Rituximab alone19 (68%)
 I-131 Tositumomab (Bexxar)4 (14%)
 Pegylated interferon2 (7%)
 Thalidomide/rituximab1 (4%)
 Vaccine2 (7%)
 Bortezomib15 (54%)
 Pralatrexate†1 (4%)
Radiation therapy11 (39%)
Table II.   Accounting of patients treated with Ixabepilone not assessable for response.
Patient, age (years)DiseaseReason
Male, 75SLLPatient developed a rapid increase in his absolute lymphocyte count (ALC) from 111–150 × 109/l (5 d after drug) to 207 × 109/l (12 d after drug) after one dose of drug
Male, 75MCLPatient developed Grade 2 asthenia and voluntarily withdrew his consent after 3 doses of drug
Male, 71MCLPatient developed an unusual transformation of his MCL to a Burkitt histology with rapidly progressive disease after 4 doses of drug, prompting urgent combination chemotherapy
Male, 57MCLPatient developed a rapid progression of his leukemic MCL from ALC 72 to 141 × 109/l after one dose of drug, prompting the treating physician to remove him from study
Female, 78MCLPatient developed cellulitis after two doses of drug, requiring prolonged antibiotics, and could not be treated secondary to the infectious complication, though she attained a clear partial response to therapy
Female, 79CLL/SLLDeveloped neutropenic fever after two doses of drug, complicated by an acute inflammatory process in the lung treated with antibiotics and prednisone

Gender distribution was significantly biased towards males (71%), with an unintended bias towards a largely white non-Hispanic population (96%). The median age of 65 (range 42–81) years approximated the median age for patients with indolent lymphomas. Approximately 20% of all patients had follicular lymphoma, including grades 1–3. Fifteen patients (54%) had MCL, while five patients (18%) had SLL, and one patient each (4%) had marginal zone lymphoma and Waldenström Macroglobulinemia. The median number of all prior therapies was 4 (range 1 to 6), while the median number of prior cytotoxic chemotherapy regimens was 3 (range 1 to 4). All patients had been treated with at least one form of an alkylator-based treatment program, while 19 of 28 received a prior purine analog. Three patients received a prior autologous stem cell transplant, 15 of 28 had received prior bortezomib, and four had received prior radioimmunotherapy. All patients received prior rituximab, with 19 of 28 receiving multiple single agent courses.

Dose modifications

In general, the treatment was associated with specific toxicities in select patients. For the 28 registered patients, a total of 53 cycles of Ixabepilone was administered, which included 157 doses of the drug. The average number of cycles and doses received per patient was approximately 1·8 and 5·6, respectively. The median number of cycles administered was 2, with only 21% of patients receiving more than two cycles. Similarly, the median number of doses administered was 6, with only 21% of patients receiving more than six total doses of Ixabepilone. There was no significant difference in the amount of therapy administered to responders or non-responders (2·6 vs. 3·4 cycles/patient, and 7·3 vs. 5·3 doses per patient in responders and non-responders, respectively). Seven patients (25%) missed at least one dose of drug, while one missed two doses. Similarly, eight patients (29%) required a 1 week schedule delay in drug administration, while four (14%) required a 2 week delay. The most common reasons for missed doses or schedule delays were fatigue, neuropathy, and hematological toxicity (neutropenia and thrombocytopenia). Dose reductions to 20 mg/m2 and 17·5 mg/m2 were allowed; most patients tolerated the 20 mg/m2 dose well, with minimal complaints of fatigue and less neuropathy. Overall, only four patients required a dose reduction.


A summary of the toxicities observed during this study is presented in Table III.Overall, the most prominent advanced grade hematological toxicities were leucopenia and lymphopenia. Approximately 57% of patients experienced Grade 3 or 4 neutropenia, though none experienced febrile neutropenia. For many of these patients, once a neutropenic event occurred with a schedule delay, cytokine support was initiated. Thrombocytopenia was also fairly common, with 22% of patients experiencing advanced grade toxicity. Non-hematological toxicities were mostly Grade 1 or 2, with some of the more prominent toxicities including fatigue, (two patients experienced grade 3 or 4 fatigue), and neuropathy (most patients experienced reversible grade 1 neuropathy, but two patients experience a grade 3 toxicity). Because of the concern that many of these patients had prior bortezomib, those patients who experienced any early signs or symptoms consistent with a sensory or painful neuropathy, had their dose of Ixabepilone lowered to 20 mg/m2. Patients treated at the reduced dose of 20 mg/m2 appeared to exhibit less toxicity compared to the 25 mg/m2 dose level. The patients who received bortezomib did not appear to exhibit any more neurotoxicity than other patients on study. Gastrointestinal toxicity was also fairly common with some patients typically experiencing early grade diarrhea (39%) and a quarter of patients reporting grade 1 constipation.

Table III.   Major hematological toxicities (occurring in >10% of patients) and non-hematological toxicities for patients receiving Ixabepilone.
 Grade 1Grade 2Grade 3Grade 4
  1. Values in parentheses are expressed as percentages.

Hematological toxicity
 Leucocytes1(4)6(21) 7(25)7(25)
 Neutrophils/granulocytes3(11)5(18) 5(18)11(39)
 Prothrombin time4(14)0 2(7)0
 Partial thromboplastin time4(14)1(4) 2(7)0
 Thrombocytopenia19(68)2(7) 5(18)1(4)
Non-hematological toxicity
 Alopecia3(11)0 00
 Aspartate transaminase11(39)0 00
 Alanine transaminase6(21)0 1(4)0
 Alkaline Phosphatase8(29)0 1(4)0
 Bilirubin4(14)4(14) 00
 Creatinine7(25)2(7) 00
 Hyperglycaemia16(57)8(29) 1(4)0
 Hypernatraemia6(18)0 00
 Hypoalbuminaemia14(50)5(18) 00
 Hypocalcaemia1(4)10(36) 1(4)1(4)
 Hypokalaemia7(25)0 2(7)0
 Hypomagnesaemia4(14)1(4) 1(4)0
 Hypophosphataemia07(25) 4(14)0
 Hypoglycaemia4(14)0 00
 Hyponatraemia5(18)0 1(4)0
 Anorexia2(7)3(11) 00
 Constipation7(25)0 00
 Cough6(21)2(7) 00
 Diarrhoea7(25)4(14) 1(4)0
 Oedema4(14)1(4) 00
 Fatigue3(11)9(32) 1(4)1(4)
 Fever3(11)0 00
 Nausea7(25)1(4) 00
 Neuropathy–sensory6(21)0 2(7)0

The two patients that experienced grade 3 sensory neuropathy underwent electrodiagnostic evaluation for progressive neuropathic symptoms. One of these patients had been treated with bortezomib 2 years previously and was evaluated approximately one month following initiation of treatment with Ixabepilone. The second had been treated with CHOP (cyclophosphamide, doxorubicin, prednisone, vincristine) more than 5 years previously and bortezomib one month prior to Ixabepilone and underwent electrodiagnostic evaluation 2 months following its completion. Both patients demonstrated a moderate to severe sensory and a mild to moderate motor axonal polyneuropathy in a distal symmetric distribution, with acute and chronic changes noted on needle EMG. Both patients also demonstrated evidence for a superimposed radiculopathy based on spontaneous activity in paraspinal muscles.


Table IV presents the response data from the 28 registered patients, with 22 of those patients being evaluable for response. Overall, the response rate was 27%, including one patient who attained a complete remission. On an Intention to Treat (ITT) analysis, the ORR would be 22%. Responses were seen in each of the major NHL sub-types. In general, the duration of response was reasonable, ranging from 2 to 8 months, given the chemotherapy resistance of this population. Interestingly, the one patient who attained a CR had the longest duration of benefit of 8 months; the lymphoma in this patient had previously been refractory to bortezomib and standard R-CVP (rituximab, cyclophosphamide, vincristine, prednisolone). One patient with MCL achieved a partial remission with a 4-month duration of response despite having had disease refractory to intravenous CEPP chemotherapy (cyclophosphamide, etoposide, procarbazine and prednisone). All other patients had achieved roughly 5 to 6 months duration of benefit from their line of prior therapy that included rituximab, radioimmunotherapy or pegylated-interferon-Rituximab. While all patients came into study with progressive disease, 12 of the 28 achieved stabilization of their disease on study, though the duration of this was short once the patient stopped their treatment.

Table IV.   Response data for patients receiving Ixabepilone.
Disease (evaluable) CR/CRuPRSDPOD Response rate (%)Duration of response (months)
  1. *Small lymphocytic lymphomas include small lymphocytic lymphoma/chronic lymphocytic leukaemia, marginal zone and Waldenstrom macroglobulinemia.

  2. CR (u), complete response (unconfirmed); PR, Partial response; SD, stable disease; POD, progressive disease.

Follicular lymphoma (6/6)1131338 (patient in CR)
2 (patient in PR)
Mantle cell lymphoma (11/15)0353272
Small lymphocytic lymphomas* (5/7)0140205
Total (22/28)1512427 


Targeting the microtubule apparatus is a well established mechanism for treating many different kinds of neoplastic diseases (Rowinsky & Calvo, 2006). What remains curious is why certain anti-microtubule agents appear to exhibit very selective patterns of activity. For example, while the taxanes appear to be critical agents for the treatment of diseases like breast and ovarian cancer, they have not demonstrated much clinical activity in hematological malignancies. Similarly, while vincristine and vinblastine have activity in lymphoproliferative malignancies, they appear to exhibit little activity against select forms of epithelial or mesenchymal malignancies. The epothilones represent a novel class of antimicrotubule agents that appear to inhibit depolymerization, with the theoretical benefit that the epothilones appear to be effective in diseases that have acquired resistance to other natural product antineoplastic agents.

The results presented here suggest that Ixabepilone has activity in indolent lymphoproliferative malignancies, a set of diseases generally characterized by a low S-phase fraction. Intuitively, one would not necessarily expect high rates of response for a drug that is so cell cycle-dependent, unless the disease has become increasingly more aggressive with time. Our observations have been corroborated in another Phase II study that focused only on patients with aggressive lymphomas. Smith et al (2005) treated patients with Grade 3 follicular lymphoma, diffuse large B-cell lymphoma, MCL, primary mediastinal large B-cell lymphoma, and Burkitt lymphoma. Patients were treated with an initial dose of 20 mg/m2 on a weekly ×3 basis every 4 weeks. As of their last report, 24 patients were evaluable for response, of which eight patients (overall response rate of 30%) experienced a complete (n = 1) or partial remission (n = 7), with the complete remission being seen in a patient with Grade 3 follicular lymphoma, and other responses being seen in patients with DLBCL and MCL. The spectrum of toxicity noted in this study seemed to be less, with no Grade 3 or 4 thrombocytopenia, and only 16% of cases experiencing leucopenia. Interestingly, more cases of peripheral sensory neuropathy were seen in the study, reported by Smith et al (2005) including Grade 1 or 2 neuropathy in approximately one-half of all patients, and Grade 3 neuropathy in 33% of treated patients. This may be a function of the amount of therapy received, which was greater in the earlier study (Smith et al, 2005).

Collectively, these experiences would suggest that this epothilone derivative exhibits activity in diverse sub-types of NHL, though the optimal schedule remains unclear. Few (21%) patients in the present study could tolerate more than two cycles of treatment. One of the interesting features of these drugs has been the differences in dose-limiting toxicity (DLT) of the different analogs as a function of schedule(Larkin & Kaye, 2006). For example, Ixabepilone has been studied on a number of schedules, including: (1) a 1-h infusion every 3 weeks, where the maximum tolerated dose (MTD) was 40 mg/m2, and the DLT was primarily neutropenia, with lesser degrees of fatigue and peripheral neuropathy (at least Grade 1 or 2 in every patient) (Mani et al, 2004; Gadgeel et al, 2005; Ajani et al, 2006); (2) a 1-h infusion daily for five consecutive days every 21 d where the MTD was 6 mg/m2 and the DLT was neutropenia, with Grade 3 fatigue, stomatitis, and anorexia, though no peripheral neuropathy (Abraham et al, 2003; Low et al, 2005); (3) a 1-h infusion daily for three consecutive days every 21 d where the MTD was 8 mg/m2 and the DLT was neutropenia with other reports of Grade 3 fatigue, anorexia, stomatitis, ileus and vomiting (Zhuang et al, 2005); (4) 1-h infusion weekly for three of four consecutive weeks, as studied here, where the MTD was 25 mg/m2 and the major DLT was fatigue, with Grade 3 nausea, diarrhoea, myalgia and arthralgia, but with no grade 3 myelosuppression; and the (5) 3-h infusion at 40 mg/m2 every 3 weeks where the most frequent treatment related adverse events included fatigue (27%), sensory neuropathy (12%) and myalgia (12%) (Thomas et al, 2007). Though conducted in patients with solid tumors, these experiences suggest that the differences between studies with regard to toxicity may be a function of the pharmacokinetic parameters, with higher Cmax approaches being associated with more neuropathy, and greater area under the curve exposures being associated with more myelosuppression. While it is impossible to know a priori which schedule is going to be the most efficacious in lymphoma, there is certainly a need to try and understand the activity in lymphoma as a function of schedule. In patients with lymphoma, the threshold for tolerating myelosuppression may be much higher than the threshold for tolerating the neuropathy and asthenia observed on this schedule.

Neurotoxicity was a complication of treatment with Ixabepilone and affected most patients to some degree. Although only performed in a limited number of patients, electrodiagnostic evaluation demonstrated a motor and sensory axonal polyneuropathy in a distal symmetric distribution. This is consistent with the length dependent neuropathy anticipated from treatment with microtubule-stabilizing agents including Ixabepilone(Lee & Swain, 2006). Both of the patients who developed grade 3 toxicities had previously received bortezomib, which causes a similar pattern of neuropathy electrophysiologically, and both patients were noted to have evidence for radiculopathy electrophysiologically. Pre-existing neuropathy and radiculopathy, even if asymptomatic, may have predisposed them to the development of neuropathic symptoms from treatment with Ixabepilone.

From the available data, it is not possible to determine if Ixabepilone has any advantage compared to other tubulin-stabilizing agents like the taxanes, though its ORR is not dramatically different from a variety of new drugs being developed in lymphoma. For example, liposomal vincristine is associated with a 10% response rate in indolent lymphomas, and 47% for patients in aggressive lymphomas (Sarris et al, 2000). Similarly, drugs like bortezomib have produced response rates between 30–40% in patients with MCL, although they have demonstrated limited activity in other sub-types of lymphoma like Hodgkin lymphoma, CLL and diffuse large B-cell lymphoma (Belch et al, 2004,Goy et al, 2005; O’Connor et al, 2005a; Fisher et al, 2006; Strauss et al, 2006). Interestingly, the duration of treatment was found to be an important determinant in patients with follicular lymphoma (O’Connor et al (2005b)). Similarly, pralatrexate, a high affinity substrate for the reduced folate carrier has produced very high overall response rates in T-cell lymphoma, with little activity in B-cell lymphomas, at least when studied on a low dose weekly schedule (O’Connor, 2006). These experiences highlight a difficulty with drug development in the lymphomas that is distinctly different from other solid tumor malignancies. Namely, the diverse heterogeneity of diseases that comprise the lymphomas represents a vast spectrum of biology, within which there are likely to be significant differences in ORR. As such, empirical studies will continue to be necessary to identify potentially important new agents, and optimal schedules of administration based on both efficacy and toxicity. The major difficulty is determining when to ‘split’, and when to ‘lump’ in conducting single agent phase II studies in NHL. Obviously, early stage Phase II studies should adopt broad eligibility criteria to identify possible signals, which can be further interrogated in more focused Phase II trials. Nonetheless, these sorts of empirically-based studies in NHL present distinct challenges.

At present, the major issues facing the development of the epothilones in lymphoma revolve around schedule of administration. Future studies should explore the activity of the epothilones in discrete sub-types of lymphoma (for example aggressive versus indolent diseases) as a function of different schedules. For example, it may be prudent to focus on a comparison of schedules in a larger multi-arm study devoted to MCL or DLBCL for example, within which more definitive comparisons can be made. The activity of these compounds in preclinical models, and now consistent early phase II data suggesting activity in select sub-types of NHL, warrants more focused attention.


OAO is the recipient of the Leukemia and Lymphoma Society Scholar in Research Award. This work was supported under a NCI Phase II Grant (UO1 CA 69913) and the Vermont Cancer Center Support Grant P30CA22435.