Many patients with recurrent, intermediate or high-grade non-Hodgkin lymphoma (NHL) may not respond to or are not candidates for aggressive salvage chemotherapy. Effective, less toxic regimens are needed. Although high-dose taxanes have not been reported to be very effective for the treatment of lymphoma, different delivery rates may allow for different mechanisms of action to be manifest and result in a different toxicity profile and response rate. The current study tested this hypothesis by using low-dose, weekly paclitaxel in patients with recurrent or refractory NHL.
Adults age > 18 years with refractory or recurrent, aggressive NHL who were not considered curable with standard high-dose therapy received paclitaxel at a dose of 80 mg/m2 weekly for 5 weeks for 2 cycles.
Thirty-four patients with refractory NHL and 4 patients with recurrent disease were treated. Approximately 45% of the patients had achieved a prior disease remission. The median number of prior regimens received was 3, 74% of the patients had an International Prognostic Index of ≥ 3 at the time of study entry, and 29% had failed high-dose therapy with autologous hematopoietic support. Only one patient encountered severe toxicity (sepsis). Myelosuppression was reported to occur in approximately 20% of patients. A total of 10 patients (26%) achieved a complete disease response and 4 patients (11%) achieved a partial response.
Patients with intermediate or high-grade lymphomas have a 55–70% chance of attaining a complete remission, but the majority develop disease recurrence. Despite aggressive salvage therapy, long-term disease remission is not obtained by many.1 Furthermore, because many patients may be too old or infirm for aggressive therapy and have only a 35–45% chance of 5-year survival,2 a regimen that is easily tolerated yet provides a reasonable response rate may provide an attractive alternative to more toxic regimens for salvage therapy in many of these ill patients.
Paclitaxel is a plant alkaloid.3 At standard doses (175–250mg/m2 every 3 weeks), paclitaxel is known to inhibit microtubule formation and lead to cell death,4, 5 and has been tested against ovarian, breast, and lung carcinomas.6 However, similar to many agents, paclitaxel has multiple mechanisms of action. It has been shown to inhibit angiogenesis in animal models,7, 8 which is known to play a role in non-Hodgkin lymphoma (NHL).9–11
Although high-dose paclitaxel has not been reported to be particularly efficacious in NHL, an alternate delivery dose and duration of exposure may allow a different mechanism of action to take precedence, such as the inhibition of angiogenesis, because to our knowledge it is not known whether this vascular inhibition ability has a dose-dependant effect. Although it has been shown to be safe and efficacious in some solid tumors,6 to our knowledge, low-dose weekly paclitaxel has not been reported in the treatment of NHL.
We conducted a Phase II trial of low-dose weekly paclitaxel in patients with advanced, aggressive NHL to assess the toxicity and efficacy of the treatment. Herein, we present preliminary data concerning some of the responding patients and note a change in stromal markers of disease (i.e., microvessel density and extracellular tenascin expression), suggesting one of the mechanisms of action manifested with this schedule of drug delivery may be antiangiogenesis.
MATERIALS AND METHODS
Patients age ≥ 18 years were eligible if they had biopsy-proven, measurable recurrent or refractory NHL that was International Working Formulation Grade D or higher, or mantle cell, anaplastic, or transformed NHL.2, 12 Patients who achieved a complete response (CR) or unconfirmed CR13 and who then developed progressive disease (PD) were included as having recurrent disease and had not been treated in that setting prior to entry on the current study. Those with refractory disease were patients who had never achieved a disease remission with at least one prior chemotherapy regimen or those who developed PD after attaining a disease remission but had not obtained at least a partial response (PR) with at least one salvage regimen prior to entering the current study. Patients were required to have a minimum leukocyte count of 2000/mm3, a minimum platelet count of 80,000/L, and 25% hematocrit. Patients were required to have an alanine transaminase, aspartate transaminase, alkaline phosphatase, and total bilirubin level that was less than three times the upper limit of normal, unless the elevation was related to the lymphoma. In addition, an estimated creatinine clearance of > 40 mg/mL was required, as well as a Cancer and Leukemia Group B performance status of 0–2. Patients were required to have a life expectancy of 12 weeks. Patients who were candidates for high-dose therapy with autologous growth factor support with curative intent were not eligible, although those who had failed prior high-dose therapy were considered to be eligible. This trial was approved by the local Institutional Review Board, and an informed consent document was signed by all patients.
Paclitaxel was administered at a dose of 80 mg/m2 over 3 hours, weekly for 5 weeks, with 1 week off for a complete 6-week cycle that could be repeated one time. Premedications included 50 mg of diphenhydramine, 300 mg of cimetidine, and 20 mg of dexamethasone. Patients who developed a ≥ Grade 3 liver, gastrointestinal, cardiac, blood pressure, or neurologic occurrence or any Grade 2 or higher renal or bladder toxicity had their therapy delayed until the side effect resolved. Grade 3 or higher blood or bone marrow toxicity resulted in a delay in therapy until recovery to ≤ Grade 2. If treatment was delayed, it was then to be readministered at a dose of 60 mg/m2. Growth factor support could be given as per the discretion of the treating physician. After these two cycles of therapy, patients were allowed to continue further consolidative therapy at the discretion of their treating physician.
Patient Monitoring and Toxicity Assessment
Within 6 weeks of registration, baseline assessments for height, weight, blood workup, and performance status were taken. Disease status was evaluated with a bone marrow examination, a gallium or positron emission tomography (PET) scan for functional imaging, and a computed tomography scan of the suspected sites of involvement.
During therapy, patients underwent at least weekly examinations, complete blood counts, and laboratory analyses. Toxicities were graded according to Common Toxicity Criteria Manual of the National Cancer Institute (version 2.0).14 Patient response was evaluated at the end of the first and second cycles of therapy using the standardized guidelines.15 A CR was indicated by the disappearance of all symptoms and signs of all measurable disease on physical examination, laboratory data, and radiographic studies. A PR was characterized by a reduction of > 50% in the sum of the products of the perpendicular dimensions of the 6 major dominant lesions. No new lesions could have appeared, and no existing lesions could have enlarged. Stable disease (SD) was characterized by a patient who had a < 50% increase or decrease in the sum of the perpendicular dimensions of the measurable lesions, and the appearance of no new lesions. PD was indicated by patients who had a > 50% increase in the sum of the perpendicular dimensions of the measurable lesions or the appearance of any new areas of malignant disease.
Factor VIII Interpretation for Microvessel Density
Slides from each sample were examined using a ×40 objective on a Zeiss microscope (Zeiss Inc., Thornwood, NY) to identify areas of the 10 highest vessel numbers (hot spots), which then were chosen for counting. Each area of highest vascularity selected in this way was reviewed at a magnification of ×200 with 10 fields counted per biopsy section. A grid ocular micrometer was used to count the microvessels per unit area; the grid was calibrated using a stage micrometer (Graticules Ltd., Tornbridge, Kent, U.K.). Microvessels for the current study were defined as those having a dimension < 10 microns. The minimum, maximum, mean, and median microvessel density (MVD) per mm2 were obtained for all biopsy specimens counted in this manner, reflecting similar methods reported by others.15–17
Thirty-eight patients with recurrent or refractory NHL were treated on this trial with 45% of the patients age > 60 years (Table 1). Seventeen patients (45%) had achieved a prior CR lasting a median duration of 9 months (range, 2–120 months). Eleven patients (29%) already had failed high-dose therapy with autologous hematopoietic support. Following the definitions presented earlier, 4 patients (11%) entered this study with recurrent disease, whereas 34 patients (89%) had refractory disease. Two patients (5%) had an International Prognostic Index (IPI) of 1, but had disease that was refractory to prior therapy as they had received three and four different regimens of therapy, respectively. Eight patients had anIPI factor of 2. Of these patients, all were found to have refractory disease with a median of 3 regimens (range, 1–5 regimens) and 8 cycles (range, 7–13 cycles) of therapy given, and 6 patients had Ann Arbor Stage IV disease. Fourteen patients (37%) had an IPI factor of 3, and 14 patients (37%) had an IPI factor of 4–5.
Table 1. Patient Characteristics at Study Entry
NHL: non-Hodgkin lymphoma; CALBG: Cancer and Leukemia Group B; LDH: lactate dehydrogenase; IPI: International Prognostic Index; CR: complete response.
No. of patients
Median age (yrs) (range)
Aged ≥ 60 yrs
Diffuse large cell
CALGB performance status
Elevated serum LDH
≥ 2 extranodal sites
IPI rating At study entry
Bone marrow involved at time of study entry
Median no. of prior chemotherapy regimens (range)
Median no. of prior chemotherapy cycles (range)
Median length of first CR (mo) (range)
Failed prior high-dose therapy with stem cell or bone marrow support
After the first cycle, 6 of 38 patients (16%) achieved a CR based on radiographic studies, including PET with 18F-fluorodeoxyglucose (FDG), which lasted a median duration of 11.5 months (range, 1–23 months) (Table 2); 7 patients (18%) achieved a PR. Fourteen patients (37%) were found to have SD after the first cycle of therapy. Eleven of the patients (29%) were found to have PD after the first cycle, 3 of whom died during this cycle.
After the second cycle, two of the five patients with a CR who received a second cycle developed PD whereas the other patients persisted in CR. Four of the six patients who achieved a PR from the first cycle and who received a second cycle achieved a CR with Cycle 2, whereas one of the patients who achieved a PR during the first cycle of therapy developed PD during the second cycle and the other patient maintained SD. Of the 14 patients with SD after the first cycle of therapy, only 1 patient achieved a PR after the second cycle.
Overall, the best response for all 38 patients was 10 CRs (26%) and 4 PRs (11%), for an overall response rate of 37% (95% confidence interval [95% CI], 0.22–0.52). Nine of the 10 patients who achieved a CR on this trial had achieved a prior CR that lasted a median of 12 months (range, 2–36 months); however, 9 of the 10 patients who attained a CR were found to have refractory disease at the time of trial entry. Of the 10 patients with a CR, there was a median of 2.5 prior regimens (range, 1–8 prior regimens) and 7 prior cycles (range, 2–24 prior cycles) before entry onto the current study. Furthermore, 6 of the 10 patients who achieved a CR (60%) had NHL involving the bone marrow, and 4 of the patients with a CR (40%) had failed prior high-dose therapy with hematopoietic stem cell support prior to entry onto the current study.
Consolidation therapy varied in the responders because only two of the patients with a CR were enrolled in a watchful waiting period after therapy and they remained in disease remission for 9 months and 12 months, respectively, before developing PD. Two patients were rechallenged with paclitaxel at the time of disease progression and one achieved a PR. Given the variety of posttherapy consolidative regimens, a meaningful duration of response and survival data reflective of this regimen alone are not available.
For the first cycle (Table 3), the only very severe toxicity encountered was sepsis in one patient. All other toxicities were ≤ Grade 3. Thirteen patients received hematopoietic growth factor support, with 5 patients receiving it as a primary prevention and 9 patients receiving it secondary to the occurrence of neutropenia, although 2 patients started therapy with an absolute neutrophil count (ANC) of 1000–1500 × 109/L, and 1 patient started with an ANC of 500–1000/ × 109/L. All patients who received growth factor support were given 300 μg per dose, with the exception of 1 patient who received 480 μg. Of those receiving growth factor support for secondary prevention, the median number of weeks per cycle that growth factor support was used was 2 (range, 1–4 weeks). The median number of doses per week was three (range, two to four doses per week). Those who received growth factor support for primary prevention received it weekly during therapy and for three to five doses per week depending on the physician's preference. None of these patients developed neutropenia. Three patients began the study with platelet counts of 50,000–<75,000 × 109/L, and 2 patients began therapy with platelet counts of 25,000–< 50,000 × 109/L whereas 6 patients were reported to have a platelet count of ≤ 75,000 × 109/L while receiving therapy. Only 2 patients had their therapy delayed approximately 1–2 weeks because of cytopenia and received subsequent doses of 60 mg/m2 of paclitaxel weekly without recurrence of toxicity.
Median duration of platelet toxicity (days) (range)
No patients died of toxicity, although one patient had a prolonged hospitalization for dyspnea.
Nineteen patients (50%) did not receive a second cycle of therapy, 11 because of early disease progression, 3 because of persistently low leukocyte counts, 1 because of infection, 1 because of patient withdrawal, and 3 for proceeding with other therapy. Of the 19 patients who received a second cycle, there were no toxicities > Grade 2 reported. Five patients received hematopoietic growth factor support for the second cycle, with three patients receiving primary prevention and two patients receiving secondary prevention.
Changes in MVD Noted in Responding Patients
A subset of responding patients had serial biopsies of the bone marrow performed with representative samples stained to evaluate changes in the MVD compared with changes in disease activity. Figure 1 denotes the change over time within the same patient from involved sites (pretherapy) to uninvolved sites (posttherapy). A consistent quantitated pattern was noted in these few patients, with an increase in MVD noted when the disease was present and a decrease noted after therapy. This pattern is depicted in Figure 2, which denotes a patient with lymphoma and easily identified microvessels as outlined by staining against factor VIII-related antigen. Involved sites typically have easily identified vessels prior to therapy (Fig. 2A) and there is a decrease in the density of these vessels that is noted on follow-up biopsy when the lymphoma responds to therapy (Fig. 2B) (P = 0.016 using a two-sided Wilcoxon signed rank test).
More toxic taxane regimens have been reported previously in patients with NHL. Press et al. used a 24-hour infusion of 175 mg/m2 of paclitaxel every 21 days.18 Casasnovas et al. used a dose of 250 mg/m2 every 3 weeks for a maximum of 6 courses.19 Toxicity was evident, and responses were limited (Table 4).
Table 4. Comparison of Higher Dose Regimens versus Low-Dose, Weekly Paclitaxel for NHL
Press et al.18 noted 64% of patients with ≥ Grade 2 toxicity, most commonly Grade 4 granulocytopenia. Casasnovas et al.19 reported that 100% of their patients were found to have toxicities of ≥ Grade 2, including pancytopenia. With regard to the current study, there were only 24% of patients in whom toxicities of ≥ Grade 2 were reported. It is clear that the low-dose weekly regimen evaluated in the current study resulted in a lower rate of toxicity in these patients.
The comparison of efficacy in these studies is interesting. The study by Press et al.18 reported that only 2 of their patients (5%) achieved a CR lasting a median of 3 months. The study by Casasnovas et al.19 reported 2 patients (5%) with a CR that lasted a median of 16.3 months. In contrast, low-dose weekly paclitaxel as employed in the current study was found to result in 26% of patients achieving a CR with a median duration of 11 months (range, 1–23 months) and 11% of patients achieving a PR. Responding patients nearly always demonstrated initial tumor shrinkage after the first cycle of therapy, with some improvement noted after the second cycle, whereas it was uncommon for those patients with SD after the first cycle of therapy to have an improvement in their antitumor effect by the end of the second cycle. This suggests that responding patients may benefit from continuing the regimen beyond two cycles to optimize the duration of response in future studies, whereas those who do not demonstrate signs of tumor shrinkage early in therapy should progress to alternative therapies.
The current study highlights the important issue that altering the dose schedule can favorably alter the toxicity as well as the response profile. The reason for the apparent improved response rate noted in lymphoma patients treated with the low-dose regimen compared with the higher dose delivered every 3 weeks is not clear. The dose density was similar to the higher doses provided every 3 weeks (range, 175–250 mg/m2). The results suggest that some mechanisms of action may be better exploited by different delivery schemas. Although to our knowledge altering drug delivery schedules has not been shown to be of benefit in many studies, some success has been reported. The Cancer and Leukemia Group B investigated the use of an infusional regimen of doxorubicin, vincristine, etoposide, prednisone, and cyclophosphamide (I-CHOPE) in patients who had already failed the CHOP regimen or bolus CHOPE. They reported a 48% response rate with a 17% CR rate and a 6-month progression-free survival rate of 24%.20
In the case of paclitaxel, mechanisms of action against tumor angiogenesis may be active with this low-dose, but more consistent-exposure regimen. Evidence that angiogenesis is important in the progression of lymphoma is mounting.8, 9
Preclinical evidence also supports the role of taxanes as antiangiogenic agents. In a murine model, paclitaxel given intraperitoneally at a dose of 6 mg/kg (n = 17) inhibited basic fibroblastic growth factor (bFGF) and vascular endothelial growth factor-induced neovascularization by 45% and 37%, respectively.7 Data from the current study (Figs. 1 and 2) are supportive of a change in the MVD correlating with disease response after exposure to paclitaxel, although we recognize that “correlation” does not necessarily indicate “cause.” This observation has led to a larger review of paired patient samples before and after therapy to assess stromal changes. The pattern noted in these few samples is consistent with a larger study we have reported in a preliminary fashion in which a decrease in MVD correlating with disease response and elevated MVD in nonresponding patients with lymphoma was noted.11
Future trials with these and other classes of agents should continue to focus on varying drug schedules to take advantage of all potential mechanisms of action.