Fixed dose-rate gemcitabine infusion as first-line treatment for advanced-stage carcinoma of the pancreas and biliary tree

Abstract

BACKGROUND

Gemcitabine infusion at the fixed dose rate of 10 mg/m² per minute (FDR-gemcitabine) has pharmacokinetic advantages and may result in improved therapeutic efficacy.

METHODS

Between April 2002 and September 2003, 40 patients with advanced-stage pancreatic adenocarcinoma (PDAC; n = 27) or biliary tree carcinoma (BTC; n = 13) were treated with weekly FDR-gemcitabine (1000 mg/m²). The primary end point was the response rate. The secondary end points were progression-free and overall survival (PFS and OS), tumor marker response, and clinical benefit response (CBR).

RESULTS

The overall response rate (ORR) on an intent-to-treat basis was 15% (95% confidence interval [95% CI], 4–26%). A positive CBR was obtained in 14 of 29 (48%) patients. Seventeen of 25 (68%) patients had a reduction in carbohydrate antigen 19-9 (CA 19-9) of > 25%. The median time to treatment failure and the median PFS were 17 weeks (95% CI, 13–22 weeks) and 19 weeks (95% CI, 15–23 weeks), respectively. The median OS was 40 weeks (95% CI, 36–45 weeks) and the 1-year actuarial survival rate was 25.8%. Multivariate analysis showed that a performance status score of 0–1 at study entry and locally advanced disease were the only independent predictors of longer PFS and OS, whereas a reduction in CA 19-9 serum levels > 75% was an independent predictor of longer PFS, but had no impact on OS. Toxicity was mild with Grade 3–4 neutropenia (according to the National Cancer Institute-Common Toxicity Criteria [version 2.0]) in 18 of 427 treatment weeks (4.2%), and Grade 3 anemia and thrombocytopenia in 6 of 427 treatment weeks (1.4%) and 9 of 427 treatment weeks (2.1%), respectively, and asymptomatic Grade 3–4 transaminase elevation in 21 of 427 treatment weeks (4.9%).

CONCLUSIONS

FDR-gemcitabine at the weekly dose of 1000 mg/m² demonstrated promising activity, despite negligible toxicity, in patients with advanced-stage PDAC and BTC. Cancer 2005. © 2005 American Cancer Society.

Pancreatic carcinoma, with an estimated 31,860 new cases and 31,270 deaths reported in 2004 in the U.S., is the 4th and 5th leading cause of cancer death for men and women, respectively.1 As highlighted by the finding that the mortality rate closely approaches incidence, pancreatic ductal adenocarcinoma (PDAC) has an extremely poor prognosis, with 1-year and 5-year relative survival rates of 24% and 4%, respectively. To our knowledge, there is no curative treatment besides surgery and even for those patients diagnosed with local stage disease, the 5-year relative survival rate is only 17%.1 Therefore, the majority of patients with PDAC are candidates for systemic treatment with essentially palliative intent. In this setting, weekly treatment with gemcitabine (Gemzar®; Eli Lilly and Company, Indianapolis, IN) at a weekly dose of 1000 mg/m² infused over 30 minutes provides a small but statistically significant advantage over bolus 5-fluorouracil (5-FU) in terms of both clinical benefit response (CBR) (23.8% vs. 4.8%; P = 0.0022) and survival (median survival duration of 5.65 vs. 4.41 months; P = 0.0025)² and is currently considered the treatment of choice for patients with inoperable disease. In addition to PDAC, gemcitabine has also shown antitumor activity in patients with advanced-stage biliary tree carcinoma (BTC).3–9 The relatively mild toxicity profile of gemcitabine has prompted the development of gemcitabine-based combinations in an attempt to improve these results. However, although randomized trials have now confirmed that such combinations are associated with a higher response rate, studies demonstrating that combination therapy results in a survival advantage are to our knowledge still lacking.10

Another approach aimed at improving gemcitabine activity has been pharmacokinetic modulation, achieved by prolonging the infusion time.11–13 Such an approach is justified by the observation that deoxycytidine kinase, the enzyme that catalyzes the conversion of gemcitabine to its active triphosphate metabolite,14 is rapidly saturated at the plasma concentrations achieved with the standard 30-minute infusion. Indeed, gemcitabine doses of > 300–350 mg/m² infused over 30 minutes have reportedly failed to result in an increased intracellular accumulation of gemcitabine triphosphate in peripheral blood mononuclear cells (PBMC).15–17 Conversely, infusion of the same gemcitabine doses over a prolonged period at the constant dose rate of 10 mg/m² per minute would avoid enzyme saturation and permit greater intracellular accumulation, possibly increasing gemcitabine antitumor activity. This strategy is currently being clinically tested with promising results in patients with leukemia, sarcoma, nonsmall cell lung carcinoma, and PDAC.18–21

In the current study, we investigated the toxicity and antitumor activity of weekly gemcitabine infused at the optimal dose rate of 10 mg/m² per minute in a population of patients with advanced-stage, inoperable PDAC or BTC. Unlike other trials testing the fixed dose-rate (FDR) infusion strategy to improve gemcitabine activity, we elected to use a dose of FDR-gemcitabine (1000 mg/m² infused over 100 minutes) that is 33% lower than the 1500 mg/m² dose infused over 150 minutes recommended in Phase I trials.22, 23 This decision was based on a current lack of evidence that increasing gemcitabine dose intensity has an impact on response rates, as well as on preliminary toxicity data from the trial of Tempero et al.21 available at the time of study design, which suggested a high incidence of hematologic toxicity with the 1500 mg/m² dose, and was justified further by the fact that 1000 mg/m² is the currently approved dose of 30-minute gemcitabine infusion in Italy.

MATERIALS AND METHODS.

Patient Population

Patients with previously untreated, histologically or cytologically confirmed, inoperable (locally advanced, recurrent, or metastatic) PDAC or BTC were eligible for the study. Eligibility criteria included age > 18 years; an Eastern Cooperative Oncology Group (ECOG) performance status (PS) ≤ 3; a life expectancy > 12 weeks; and adequate bone marrow (absolute neutrophil count [ANC] > 1.5 × 10⁹/L and a platelet count > 100 × 10⁹/L), cardiac, liver (bilirubin level < 5 mg/dL, aspartate aminotransferase/alanine aminotransferase level < 3 × the upper normal limit [UNL] in patients with no liver metastases and < 5 × UNL in patients with known liver metastases), and renal (creatinine level < 2 mg/dL) function. Exclusion criteria included uncontrolled central nervous system metastases, second primary malignancies, active infections, pregnancy or lactation, women of childbearing potential not using adequate contraception, or serious comorbidities. The study was reviewed and approved by the institutional review board of the Regina Elena National Cancer Institute (Rome, Italy), and written informed consent was obtained from all patients before entering the study.

Treatment and Study Design

The study was designed as a standard single-arm, open-label, Phase II trial. Study patients received gemcitabine at a dose of 1000 mg/m² by intravenous infusion at the FDR of 10 mg/m² per minute (100 minutes of total infusion time [FDR-gemcitabine]). Treatment was given weekly for 7 consecutive weeks and on Days 1, 8, and 15 of a 4-week cycle thereafter for a maximum of 6 cycles in the absence of progressive disease (PD) or unacceptable toxicity. Patients received full supportive care. Toxicity was evaluated by weekly assessment of the complete blood count, serum transaminase levels, history, and physical examination, and monthly assessment of complete serum chemistries. Body weight, PS (according to both ECOG and Karnofsky scales), pain intensity, and analgesic consumption were also recorded weekly. In patients with elevated levels at the time of study entry, serum carbohydrate antigen 19-9 (CA 19 -9) levels were assessed every 4 weeks. Toxicity was graded according to the National Cancer Institute-Common Toxicity Criteria, version 2.0. A 1-week treatment delay was applied in the event of an ANC < 0.5 × 10⁹/L, a platelet count < 50 × 10⁹/L, or nonhematologic toxicity ≥ Grade 3. A 25% gemcitabine dose reduction was applied in the event of an ANC < 1.5 × 10⁹/L, a platelet count < 75 × 10⁹/L, or any Grade 3 or 4 nonhematologic toxicity in the previous cycle.

Response was evaluated according to standard World Health Organization (WHO) criteria.24 The composite end point of the CBR was evaluated according to the criteria established by Burris et al.2 and included the assessment of pain (pain intensity and analgesic consumption) and functional impairment (assessed by Karnofsky PS) as primary measures and weight change (assessed by body weight) as a secondary measure. Each patient was classified as positive, stable, or negative for each of the primary CBR measures (pain intensity or PS). For all patients, positive indicated a sustained (≥ 4 weeks) improvement over baseline. If the patient was stable on both primary measures of clinical benefit, the patient was then classified as either positive or nonpositive on the basis of the secondary clinical benefit measure of weight. For patients to achieve an overall rating of a positive CBR, they had to be positive for at least one parameter without being negative for any of the others.

Tumor marker changes were considered significant when they were > 25% of the baseline levels and persisted for ≥ 2 subsequent determinations performed ≥ 4 weeks apart. Based on previous reports using cutoff values ranging from a > 20% to a > 75% decline from baseline levels for CA 19-9 response, we classified tumor marker response into 4 categories: a > 75% decrease, a 25–75% decrease, no change, and an increase.

Sample Size and Statistical Analysis

Response rate was the primary study end point. The secondary end points were progression-free and overall survival (PFS and OS, respectively), tumor marker response, and CBR. Sample size was computed according to the exact single-stage Phase II design described by A'Hern.25 The treatment was not considered to be of further interest if the true response rate was < 5% (p₀ = 5%). The alternate hypothesis assumed that a true response rate of ≥ 20% would be of considerable interest (p₁= 20%). With a 5% rejection error (α = 5%) and a power of 90%, ≥ 5 partial responses (PR) were to be observed of a total of 38 evaluable patients to declare the study positive. The Kaplan–Meier method was used to estimate PFS and OS. PFS was calculated from the first day of treatment to the date of disease progression and OS was calculated from the first day of treatment to the time of death from any cause. Cox proportional hazards models were used to compare survival among different patient/disease characteristics and treatment response groups and hazard ratios were appropriately derived from these models.

RESULTS

Between April 2002 and September 2003, 40 patients with advanced-stage PDAC (n = 27) or BTC (n = 13) were treated with FDR-gemcitabine. Patient characteristics are shown in Table 1. The median time elapsed from diagnosis to treatment start was 6 weeks. All patients completed ≥ 2 treatment weeks. The median and total number of treatment weeks administered was 8 weeks (range, 2–22 treatment weeks) and 427 weeks, respectively. Early (< 4 weeks) treatment discontinuation occurred in 3 patients for the following reasons: deterioration of clinical status (n = 2) and withdrawal for personal reasons (n = 1). Early death (< 8 weeks) was reported to occur in 2 patients (both with BTC), and was due to septic shock secondary to infectious cholangitis and was deemed to be unrelated to treatment.

Table 1. Patient Characteristics

Characteristics	No. of patients
PDAC: pancreatic ductal adenocarcinoma; BTC: biliary tree carcinoma; ECOG: Eastern Cooperative Oncology Group; PS: performance score.
Eligible patients	40
Male	21
Female	19
Median age in yrs (range)	63 (45–77)
PDAC	27
Locally advanced	6
Metastatic	21
BTC	13
Locally advanced	3
Metastatic	10
ECOG PS
0	16
1	17
2	7

Tumor Response

A total of 38 patients had measurable/evaluable disease according to standard WHO criteria.24 Although five patients could not undergo a formal disease restaging (early treatment discontinuation in three patients and early death in two patients), they were considered evaluable and to have PD at the time of treatment interruption or death for all subsequent analyses. Two additional patients (both with BTC) had histologic evidence of residual/recurrent disease after surgery and rising CA 19-9 serum levels, but they were not evaluable by conventional imaging techniques and therefore were considered only for tumor marker responses. Among the 38 evaluable patients, the best response to treatment was a PR in 6 of 38 patients (16%), a minor response (MR) in 7 of 38 patients (18%), stable disease (SD) in 13 of 38 patients (34%), and PD in 12 of 38 patients (32%) (Table 2). The overall response rate (ORR) in the intent-to-treat population was 15% (95% confidence interval [95% CI], 4–26%). In all patients experiencing a PR, response was confirmed at ≥ 4 weeks. Objective tumor shrinkage in response to treatment was observed almost exclusively in patients with PDAC, with 6 PRs and 6 MRs occurring in 27 evaluable patients. Conversely, only 1 MR was observed among 11 evaluable patients with BTC, with 3 additional patients experiencing SD (Table 2). PRs occurred in patients with locally advanced PDAC and those with metastatic PDAC (2 of 6 patients and 4 of 21 patients, respectively), as well as in patients with an ECOG PS of 0 (n = 3), 1 (n = 2), or 2 (n = 1) (data not shown). In the majority of patients, PD occurred at previously known sites of disease.

Table 2. Response to Treatment

Characteristics	No. of patients (%)
PR: partial response; MR: minor response; SD: stable disease; PD: progressive disease; PDAC: pancreatic ductal adenocarcinoma; BTC: biliary tree carcinoma; ORR: overall response rate: 95% CI: 95% confidence interval.
Total no. of patients	40 (100)
No. of evaluable patients	38
PR	6 (16)
MR	7 (18)
SD	13 (34)
PD	12 (32)
PDAC	27
No. of evaluable patients	27
PR	6 (22)
MR	6 (22)
SD	10 (37)
PD	5 (19)
BTC	13
No. of evaluable patients	11
PR	—
MR	1 (9)
SD	3 (27)
PD	7 (64)
ORR (95% CI)	15% (4–26%)

Clinical Benefit and Tumor Marker Response

Of the 16 patients with an ECOG PS of 0 at the time of study entry, PS remained stable in 15 patients (94%) and worsened in 1 patient (6%). Of the 24 patients with an ECOG PS of 1 or 2 at study entry, PS improved during treatment in 10 patients (41%), remained stable in 10 patients (41%), and worsened in 4 patients (17%) (Table 3). Twenty-three patients had pain requiring analgesic treatment at the time of study entry. Self-reported pain and/or analgesic consumption decreased during treatment in 12 patients (52%), remained stable in 10 patients (43%), and worsened in 1 patient (4%) (Table 3). Body weight improved in 5 of 40 patients (12%), remained stable in 30 of 40 patients (75%), and worsened in 5 of 40 patients (12%) (Table 3). Overall, 29 patients were evaluable for the composite endpoint of CBR according to the criteria of Burris et al.2 (18 patients with both ECOG PS ≠ 0 [Karnofsky ≤ 80%] and pain, 5 patients with ECOG PS = 0 [Karnofsky = 90%] and pain, and 6 patients with ECOG PS ≠ 0 [Karnofsky ≤ 70%] and no pain, respectively). Of these, 14 patients (48%) experienced a positive CBR, defined as an improvement in ≥ 1 parameter without worsening of any of the others (Table 3).

Table 3. Clinical Benefit Response

Characteristics	No. of evaluable patients	Improved (%)a	No change (%)	Worsened (%)
ECOG: Eastern Cooperative Oncology Group; PS: performance status; NA: not applicable. a The sum of percentages may not total 100, due to rounding.
ECOG
PS = 0	16	NA	15 (94)	1 (6)
ECOG PS ≠ 0	24	10 (41)	10 (41)	4 (17)
Pain intensity/analgesic consumption	23	12 (52)	10 (43)	1 (4)
Weight	40	5 (12)	30 (75)	5 (12)
Clinical benefit response	29	14 (48)	NA	NA

Twenty-five patients had elevated CA 19-9 serum levels at the time of study entry and were serially monitored every 4 weeks. As summarized in Table 4, 17 of 25 patients (68%) experienced a > 25% decrease in CA 19-9 serum levels during treatment (a > 75% decrease in 11 patients and a 25–75% decrease in 6 patients), 3 of 25 patients (12%) had stable levels, and 5 of 25 patients (20%) had a > 25% increase in CA 19-9 serum levels. The objective response to treatment in the 9 evaluable patients with a > 75% decrease in CA 19-9 levels was a PR in 3 patients, a MR in 2 patients, SD in 1 patient, and PD in 3 patients.

Table 4. Tumor Marker Response

Characteristics	No. of patients (%)
Evaluable	25
Decrease >75%	11 (44)
Decrease 25–75%	6 (24)
No change	3 (12)
Increase >25%	5 (20)

Toxicity

All 40 patients and 427 treatment weeks were evaluable for toxicity. All toxicities were reversible, no hospitalization due to severe adverse events was required, and no treatment-related deaths occurred. Hematologic and nonhematologic toxicities are summarized in Tables 5 and 6. The main toxicity was hematologic, with Grade 3–4 neutropenia occurring in 18 of 427 treatment weeks (4.2%) and in 5 of 40 patients (12.5%), respectively; Grade 3 thrombocytopenia occurring in 9 of 427 treatment weeks (2.1%) and in 4 of 40 patients (10%), respectively; and Grade 3 anemia occurring in 6 of 427 treatment weeks (1.4%) and in 2 of 40 patients (5%), respectively. No episodes of febrile neutropenia were observed and none of the patients required blood or platelet transfusions. The main nonhematologic toxicity was asymptomatic transaminase elevation, which reached Grade 3–4 in 21 of 427 treatment weeks (4.9%) and in 5 of 40 patients (12.5%), respectively, and promptly recovered after treatment interruption. Grade 3 diarrhea and Grade 1–2 nausea and emesis were reported in only 3 of 427 treatment weeks (0.7%) and in 9 of 427 treatment weeks (2.1%), respectively, and in 3 of 40 patients (7.5%) and in 7 of 40 patients (17.5%), respectively. Mild asthenia and gemcitabine-related nonneutropenic fever were reported in 5 of 40 patients (12.5%) and in 19 of 427 treatment weeks (4.4%), respectively. Overall, a 1-week treatment delay or a 25% gemcitabine dose reduction was required in 36 of 427 treatment weeks (8.4%) and in 41 of 427 treatment weeks (9.6%), respectively, mainly due to hematologic toxicity or elevated transaminase levels. Two patients discontinued treatment because of toxicity. Of these patients, 1 patient experienced a Grade 3 sensorimotor peripheral neuropathy, which developed after 6 treatment weeks and gradually resolved after treatment interruption, and 1 patient experienced a Grade 3 erythematous skin rash, with generalized desquamation and alopecia.

Table 5. Toxicity per Treatment Week (NCI–CTC, Version 2.0) (n = 427 Treatment Weeks)

Toxicity^a	Grade (% of treatment wks)
	1–2	3	4
NCI–CTC: National Cancer Institute Common Toxicity Criteria; AST: aspartate aminotransferase; ALT: alanine aminotransferase.
Neutropenia	51 (12.0)	15 (3.5)	3 (0.7)
Thrombocytopenia	26 (6.1)	9 (2.0)	—
Anemia	134 (31)	6 (1.4)	—
Nausea-emesis	9 (2.1)	—	—
AST	84 (19.7)	4 (0.9)	1 (0.2)
ALT	137 (32.0)	15 (3.5)	1 (0.2)
Diarrhea	2 (0.5)	3 (0.7)	—
Non-neutropenic fever	18 (4.2)	1 (0.2)	—

Table 6. Toxicity per Patient (NCI–CTC, Version 2.0) (n = 40 Patients)

Toxicity^a	Grade (% of patients)
	1–2	3	4
NCI–CTC: National Cancer Institute Common Toxicity Criteria; AST: aspartate aminotransferase; ALT: alanine aminotransferase.
Neutropenia	18 (45.0)	4 (10.0)	1 (2.5)
Thrombocytopenia	11 (27.5)	4 (10.0)	—
Anemia	29 (72.5)	2 (5.0)	—
Nausea-emesis	7 (17.5)	—	—
Asthenia	5 (12.5)	—	—
AST	25 (62.5)	1 (2.5)	1 (2.5)
ALT	29 (72.5)	2 (5.0)	1 (2.5)
Diarrhea	3 (7.5)	3 (7.5)	—
Non-neutropenic fever	12 (30.0)	1 (2.5)	—
Peripheral neuropathy	—	1 (2.5)	—
Skin rash	—	1 (2.5)	—

Survival Analysis

The median time to treatment failure and the median PFS period for the entire population were 17 weeks (95% CI, 13–22) and 19 weeks (95% CI, 15-23), respectively (Figs. 1A and B). The median OS was 40 weeks (95% CI, 36–45) (Fig. 1C) and the 1-year actuarial survival rate was 25.8%. Univariate and multivariate analyses, taking into account pretreatment characteristics (age, gender, ECOG PS at the time of study entry [0–1 vs. 2], diagnosis [PDAC vs. BTC], disease status [locally advanced vs. metastatic]), and response to treatment, demonstrated that ECOG PS at the time of study entry and disease status were the only significant predictors of PFS and of OS. With regard to PFS, the hazards ratio (HR) value was 0.21 (95% CI, 0.07–0.6; P = 0.004), and 0.24 (95% CI, 0.07–0.85; P = 0.027) for a PS of 0–1 and locally advanced disease, respectively; with regard to OS, the HR was 0.29 (95% CI, 0.1–0.81; P = 0.018) and 0.21 (95% CI, 0.04–1.04; P = 0.056) for a PS of 0–1 and locally advanced disease, respectively. A positive CBR did not appear to impact impact either PFS or OS significantly (data not shown). In the cohort of 25 patients who were evaluable for tumor marker response, a reduction in circulating tumor markers of > 75% was found to correlate with a significantly longer PFS (median PFS: 30 vs. 18 weeks; P = 0.01) (Fig. 2), but not OS (data not shown). In these patients, only ECOG PS at the time of study entry and tumor marker response entered the final multivariate model for PFS (HR = 0.09;95% CI, 0.09–0.54 [P = 0.008], and HR = 0.29; 95% CI, 0.08–1.03 [P = 0.056] for a PS of 0–1 and a tumor marker reduction of > 75%, respectively).

Figure 1. (A) Time to treatment failure, (B) progression-free survival, and (C) overall survival curves for patients with pancreatic ductal adenocarcinoma and biliary tree carcinoma treated with first-line fixed dose-rate gemcitabine (n = 40).

Figure 2. Progression-free survival curves for fixed dose-rate gemcitabine-treated patients with a > 75% decrease (n = 11, Curve A) or a < 75% decrease, no change, or increase (n = 14, Curve B) in carbohydrate antigen 19-9 (CA 19-9) serum levels during treatment (P = 0.01).

DISCUSSION

In the current study, we present our experience with a weekly FDR-gemcitabine regimen employed as first-line treatment for patients with advanced-stage PDAC or BTC. In this trial, FDR-gemcitabine yielded a 15% ORR; median PFS and OS periods of 19 weeks and 40 weeks, respectively; and a 48% CBR, despite mild and manageable toxicity that was comprised primarily of neutrocytopenia/thrombocytopenia and asymptomatic transaminase elevation.

In the successful registration trial, which randomly compared weekly gemcitabine (at a dose of 1000 mg/m² infused over 30 minutes) with bolus 5-FU, a 5.4% ORR; a median PFS and OS of 9 weeks and 5.65 months, respectively; and a 23.8% CBR were observed in the gemcitabine arm.2 Similarly, Storniolo et al.26 reported a 13% ORR; a median PFS and OS of 2.8 months and 5.1 months, respectively; and a 20% CBR in a population of > 1500 patients treated with a first-line regimen of weekly gemcitabine (at a dose of 1000 mg/m² infused over 30 minutes) in an investigational new drug treatment program. More recently, Tempero et al.21 reported the results of a randomized Phase II trial, in which patients with advanced-stage PDAC were treated with weekly gemcitabine administered using either a standard 30-minute or a FDR (at a rate of 10 mg/m² per minute) infusion at their respective maximum tolerated doses (2200 mg/m² and 1500 mg/m²), as determined in Phase I trials.22, 23, 27 The Tempero et al. study21 verified the hypothesis that the FDR infusion avoids deoxycytidine kinase saturation, thereby permitting increased intracellular accumulation of the active gemcitabine triphosphate metabolite. Indeed, the PBMC of patients who received the FDR infusions demonstrated accumulated gemcitabine triphosphate of approximately twice the concentration generated by the standard 30-minute infusion, using only 68% of the dose administered in the standard arm.21 Consistent with the assumption that the increased intracellular accumulation of gemcitabine triphosphate would result in better disease control, results of the study by Tempero et al.21 significantly favored the FDR arm in terms of median survival (5.0 months vs. 8.0 months for the standard and FDR arms, respectively; P = 0.013) and 1-year and 2-year survival rates (9% vs. 28.8%[P = 0.014] and 2.2% vs. 18.3% [P = 0.007], respectively).

With this background in mind, the results of our FDR-gemcitabine regimen (at the same dose of 1000 mg/m² employed in the studies of Burris et al.2 and Storniolo et al.26, but infused over 100 minutes to achieve the optimal dose rate of 10 mg/m² per minute) appear promising in terms of survival and CBR outcomes. From a toxicity standpoint, the use of a 33% lower FDR-gemcitabine dose, compared with that employed by Tempero et al.,21 resulted in a consistently lower hematologic toxicity, with Grade 3–4 neutropenia and thrombocytopenia observed in 12.5% and 10% of our patients, respectively (compared with 48.8% and 37.2% of patients, respectively, in the Tempero et al. trial21). Although a dose-response relation for gemcitabine antitumor activity has been proposed in preclinical models,28, 29 the considerable variation in dose intensity in studies that have employed the standard 30-minute infusion has not resulted in a consistent impact on reported response rates. The results reported in the current study suggest that this also may be the case for FDR-gemcitabine and indicate that, using this strategy, doses of gemcitabine > 1000 mg/m² may be consistently more toxic. This may be important in light of the essentially palliative nature of treatment for patients with advanced-stage PDAC, a setting in which consideration of toxicity and convenience for the patient, in addition to antitumor efficacy, plays an important role in the evaluation of different therapeutic options.

Based on the current evidence of gemcitabine activity in BTC, our trial included a small cohort (n = 13) of patients with advanced-stage BTC. In this small group of patients with BTC, the antitumor activity of the FDR-gemcitabine regimen was somewhat disappointing, with only 1 patient demonstrating objective tumor shrinkage, which did not meet the WHO criteria for a PR (Table 2). These results unfavorably compare with those reported in the literature with the standard 30-minute gemcitabine infusion, which achieves ORRs ranging from 12.5–36%.3–9 However, the median PFS and OS (range, 2.5–7 months and 6.5–11 months, respectively) reported in other studies are not dramatically different from those observed in our group of patients with BTC (PFS and OS of 21 weeks and 40 weeks, respectively), suggesting that the FDR-gemcitabine regimen we employed cannot be dismissed as inactive in patients with BTC based solely on the response rate obtained in such a limited cohort of patients. Therefore, we are continuing to accrue patients with BTC to better evaluate the therapeutic potential of FDR-gemcitabine in this disease.

Serial monitoring of serum CA 19-9 changes has been suggested to be useful as a biochemical response marker and an outcome prognostic parameter for patients with PDAC treated with gemcitabine-based chemotherapy30–33 or radiotherapy.34, 35 The interest in CA 19-9 kinetics as a surrogate endpoint stems from the difficulty of assessing tumor regression in a disease, such as PDAC, characterized by massive desmoplastic reaction at the primary tumor site, as well as from the failure of objective tumor response, CBR, or pretreatment characteristics to reliably predict PFS and/or OS.30, 32, 35 Survival analysis conducted in our cohort of patients with advanced-stage PDAC and BTC treated with FDR-gemcitabine confirms that a > 75% decrease in CA 19-9 serum levels is the only treatment-related factor that translates into a PFS advantage, whereas neither objective tumor response nor CBR show a significant predictive value. However, the CA 19-9 response occurred in the presence of radiographic tumor progression in three of nine patients, suggesting that CA 19-9 monitoring cannot routinely substitute for disease assessment with imaging techniques. Although this dissociation between biochemical and objective response is not readily explainable, it does not appear to be restricted to gemcitabine treatment in that it was also observed in one of three patients with PDAC treated with a second-line protocol employing a combination of celecoxib and infusional 5-FU.36

Overall, the current study results indicate that FDR-gemcitabine, at the dose of 1000 mg/m², is active and extremely well tolerated, providing effective symptom palliation despite negligible toxicity in patients with advanced-stage PDAC and BTC. Results of ongoing randomized trials are eagerly awaited to validate pharmacokinetic modulation as an approach to improve gemcitabine efficacy. Survival and CBR outcomes are encouraging and should prompt further evaluation of the FDR infusion strategy for gemcitabine, either alone or in combination with other agents.