Agents that interfere with microtubule structure and function continue to play a major role in cancer treatment. The vinca alkaloids represent the oldest class of drugs traditionally characterized as having a microtubule destabilizing function, whereas the taxanes—notably paclitaxel and docetaxel—are considered microtubule stabilizers. 1 The wide use of both classes of therapy is directly related to their broad activity and acceptable safety profile in most patients. The goals of improving efficacy and reducing toxicity of paclitaxel have resulted in the development of modified taxanes such as docetaxel, nanoparticle albumin-bound paclitaxel, and paclitaxel poliglumex. A series of analogs, such as larotaxel, milataxel, TPI-287, BMS-275183, and others, have also been investigated. 2 To date, there has been significant overlap between spectrums of activity and toxicities of these agents. None has clearly distinguished itself from paclitaxel, the first-in-class compound. However, nontaxane microtubule-targeting agents such as the epothilones are also on the rise. 3 Epothilones as a class have shown increased antiproliferative activity in vitro when compared with taxanes and shown activity in cell lines with natural taxane resistance indicated by multidrug resistance 4 or tubulin mutation. 5 Epothilone B (patupilone) and four of its synthetic derivatives (including ixabepilone) and epothilone D and one of its derivatives (KOS-1584) are under active investigation in many solid tumors. 3 Ixabepilone, the semisynthetic analog of epothilone B, was the first epothilone to get a US Food and Drug Administration indication in 2007 as monotherapy for patients with metastatic or locally advanced breast cancer whose tumors were resistant to anthracyclines and taxanes. 6 This approval was issued on the basis of two trials. The first was a single-arm phase II study 7 of 126 patients whose disease progressed within 6 months of prior chemotherapy, resulting in an independent radiology review response rate of 11.5% (95% CI, 6.3% to 18.9%). Treatment-related grade 3 to 4 toxicities included neuropathy (14%), asthenia (13%), and neutropenia (54%), with alopecia in 48% of patients. The second was a randomized phase III trial 8 comparing capecitabine with or without ixabepilone in 752 patients resistant to anthracyclines and taxanes. A statistically significant improvement in progression-free survival of 4.2 to 5.8 months (hazard ratio, 0.75; 95% CI, 0.64 to 0.88; P .0003) in favor of the combination was reported. Treatment-related grade 3 to 4 toxicities were similar and included neuropathy (21%), asthenia (16%), and neutropenia (68%), with alopecia in 31% of patients. The overall toxicity of ixabepilone, notably neuropathy, was not improved relative to paclitaxel, and responses in breast cancer were modest, albeit in patients clinically considered to be paclitaxel resistant. The relative efficacy of epothilone B when compared with paclitaxel or docetaxel in taxane-naive patients is not yet well characterized. As epothilones are evaluated in other tumor types, particularly for gynecologic cancers in which taxanes have major activity, paclitaxel remains the benchmark with which these agents must be compared. The questions are straightforward: first, is a particular epothilone more efficacious than paclitaxel in taxane-naive patients; second, is it significantly active in clinically defined taxane-resistant patients; third, does it have a favorable toxicity profile when compared with paclitaxel; and fourth, is it possible to predict who might respond? It is fair to say that preclinical attributes should drive the design of clinical trials, but the predictive value of some preclinical models has proven to be limited. 9 In this issue of Journal of Clinical Oncology, two trials evaluating epothilones in patients with gynecologic cancers are presented. These studies should be considered in the context of the aforementioned questions. First, the phase II study by Dizon et al 10 evaluated ixabepilone in a group of 52 patients with recurrent or persistent endometrial cancer with one prior chemotherapeutic regimen. The vast majority (94%) had received prior paclitaxel. The second-line response rate in these patients was 12%, falling short of the predefined benchmark in the study, and progression-free survival was only 2.9 months. Treatment-related grade 3 to 4 toxicities included neuropathy (18%), asthenia (13%), and neutropenia (52%), with alopecia in 46% of patients. This response rate was, as expected, less than that seen in the second-line setting with paclitaxel in taxane-naive patients (26%), but it was slightly improved when considered in the context of other agents in taxane-pretreated patients, including topotecan (9%), liposomal doxorubicin (9.5%), and docetaxel (7.7%). 11-14 The toxicity
