- Email: [email protected]
Gemcitabine versus cisplatin, epirubicin, fluorouracil, and gemcitabine in advanced pancreatic cancer: a randomised controlled multicentre phase III trial
Articles
Gemcitabine versus cisplatin, epirubicin, fluorouracil, and gemcitabine in advanced pancreatic cancer: a randomised controlled multicentre phase III trial Michele Reni, Stefano Cordio, Carlo Milandri, Paolo Passoni, Elisa Bonetto, Cristina Oliani, Gabriele Luppi, Roberto Nicoletti, Laura Galli, Roberto Bordonaro, Alessandro Passardi, Alessandro Zerbi, Gianpaolo Balzano, Luca Aldrighetti, Carlo Staudacher, Eugenio Villa, Valerio Di Carlo
Summary Background Patients with advanced pancreatic adenocarcinoma have a poor response, progression-free survival, and overall survival with standard treatment. We aimed to assess whether a four-drug regimen could improve 4-month progression-free survival compared with gemcitabine alone. Methods In a randomised multicentre phase III trial, 52 patients were randomly assigned to 40 mg/m2 cisplatin and 40 mg/m2 epirubicin both given on day 1, 600 mg/m2 gemcitabine given intravenously over 1 h on days 1 and 8, and 200 mg/m2 fluorouracil a day given by continuous infusion on days 1–28 of a 4-week cycle (PEFG regimen), and 47 were assigned to 1000 mg/m2 gemcitabine given intravenously over 30 min once a week for 7 of 8 consecutive weeks in cycle 1 and for 3 of 4 weeks thereafter. The primary endpoint was 4-month progression-free survival. Secondary endpoints were overall survival, objective response, safety, and quality of life. Analyses were by intention to treat. Findings 51 patients assigned PEFG and 46 assigned gemcitabine alone had disease progression. 49 patients in the PEFG group and 46 in the gemcitabine group died from progressive disease. More patients allocated PEFG than gemcitabine alone were alive without progressive disease at 4 months (60% [95% CI 46–72] vs 28% [17–42]; hazard ratio [HR] 0·46 [0·26–0·79]). 1-year overall survival in the PEFG group was 38·5% (25·3–51·7) and in the gemcitabine group was 21·3% (9·6–33·0; HR 0·68 [0·42–1·09]). More patients assigned PEFG showed disease response than did those assigned gemcitabine (38·5% [25·3–51·7] vs 8·5% [0·5–16·5]; odds ratio 6·60 [2·11–20·60], p=0·0008). More patients in the PEFG group had grade 3–4 neutropenia and thrombocytopenia than in the gemcitabine group (p0·0001). Interpretation The PEFG regimen could be considered for treatment of advanced pancreatic adenocarcinoma.
Introduction Pancreatic adenocarcinoma is associated with a poor prognosis. More than 80% of patients present with advanced disease at diagnosis, and mortality is high because tumours infiltrate the abdominal blood vessels and spread to regional lymph nodes and liver at an early stage. Gemcitabine is regarded as the standard treatment for unresectable disease, and has led to an objective response in 4–26% of patients and a 1-year overall survival of 17–28% in phase III trials.1–11 Gemcitabine was licensed for treatment of advanced pancreatic cancer on the basis of improved clinical benefit and increased overall survival compared with fluorouracil.1 Several attempts to improve the efficacy of gemcitabine in advanced pancreatic adenocarcinoma by use of new agents or by addition of a second cytotoxic agent or other drugs to a standard dose and schedule of gemcitabine have not shown a survival advantage.2–11 However, a significant benefit in median progressionfree survival has been noted for gemcitabine combined with fluorouracil,2 cisplatin,7 or oxaliplatin,11 which suggests that platinum-based agents and antifolates have a role in treatment of pancreatic cancer. Moreover, http://oncology.thelancet.com Vol 6 June 2005
although combination treatment did not increase overall survival compared with gemcitabine alone, whether salvage therapy affects the relation between progressionfree survival and overall survival is not known. Strategies that have not included gemcitabine have been assessed, but have not shown better results.12,13 Findings from a phase II trial14 suggested that fixed-dose gemcitabine improves survival, but phase III data are not available. Gemcitabine combined with erlotinib improved 1-year overall survival compared with gemcitabine alone from 17% to 24%.15 Cisplatin, epirubicin, fluorouracil, and gemcitabine (PEFG)—an empirically derived treatment based on four agents that are active against pancreatic adenocarcinoma—showed promising activity with regard to response rate and survival in a previous phase II trial.16 We therefore did a multicentre randomised phase III trial to compare standard treatment of gemcitabine with that of PEFG by use of previously defined doses and schedules.16
Lancet Oncol 2005; 6: 369–76 Published online May 10, 2005 DOI: 10.1016/S1470-2045(05) 70175-3 See Reflection and Reaction page 352 Department of Radiochemotherapy (M Reni MD, P Passoni MD, E Bonetto MD, E Villa MD), Department of Radiology (R Nicoletti MD), Department of Statistics (L Galli MSc), and Department of Surgery (A Zerbi MD, G Balzano MD, L Aldrighetti MD, Prof C Staudacher MD, Prof V Di Carlo MD),S Raffaele H Scientific Institute, Milan, Italy Department of Oncology, S Luigi Currò H Catania, Italy (S Cordio MD, R Bordonaro MD); Department of Oncology, Pierantoni H Forlì, Italy (C Milandri MD, A Passardi MD); Department of Oncology, Azienda Ospedaliera and University, Verona, Italy (C Oliani MD); Department of Medical Oncology, Azienda Ospedaliera-Policlinico, Modena, Italy (G Luppi MD) Correspondence to: Dr Michele Reni, Department of Oncology, San Raffaele H Scientific Institute, via Olgettina 60, 20132 Milan, Italy [email protected]
Methods Patients Between April, 2000, and March, 2003, 104 patients from five institutions aged 18–70 years with histo369
Articles
104 patients randomised
54 allocated PEFG
50 allocated gemcitabine 3 ineligible 2 inadequate liver function 1 other cancer
2 ineligible (inadequate liver function)
52 analysed
47 analysed
Figure 1: Trial profile
logically or cytologically confirmed stage IVA (T4, N0–1, M0)17 or metastatic pancreatic adenocarcinoma were enrolled onto the study. Eligibility criteria included: Karnofsky performance status more than 40; at least one indicator lesion that was measurable in two dimensions and had not been irradiated previously; no previous chemotherapy; and adequate bone-marrow function (absolute neutrophil count 1500106 cells/L, platelet count 100109/L, and haemoglobin 100 g/L), kidney function (serum creatinine 132·6 mol/L), and liver function (total serum bilirubin 34·2 mol/L and serum transaminases 3 times upper normal limit). Patients who had ampullary tumours, other histological variants of pancreatic carcinoma, previous adjuvant chemotherapy, or previous malignant disease were ineligible for the study, with the exception of those who had had basalcell carcinoma of the skin, carcinoma in situ of the cervix, or other cancers for which the patient had been disease-free for at least 5 years. Eligibility criteria for assessment of clinical benefit were baseline use of analgesics of at least two nonsteroidal anti-inflammatory drugs, baseline pain
Age (years) Median (range) Sex Male Female Karnofsky performance status 70 70 Stage IVA Metastatic Previous treatment Pancreatic surgery Radiotherapy or chemotherapy Site of metastases Liver Lymph nodes Lung
Table 1: Baseline characteristics
370
PEFG (n=52)
Gemcitabine (n=47)
62 (37–69)
59 (25–69)
24 28
24 23
37 15
35 12
15 37
14 33
3 0
3 0
29 9 5
28 13 5
intensity score of 20 mm or more of a possible 100 mm on the Memorial Pain Assessment Card,18 or a Karnofsky performance status of 50–70. The study was done in accordance with the Declaration of Helsinki and approved by the local ethics committees of participating institutions; written informed consent was obtained from all patients.
Study objectives We chose 4-month progression-free survival as the primary endpoint to avoid the potential effect of salvage treatment after progression on overall survival and the potential effect of surgery or radiotherapy (allowed after 4 months in patients with stage IVA disease) on both overall survival and progression-free survival. Secondary endpoints were overall survival, objective response, clinical benefit, safety, and quality of life.
Randomisation Patients who fulfilled all inclusion criteria were registered by the attending physician at the coordinating institution. Baseline, treatment, and follow-up data were obtained at the coordinating institution. Randomisation was done by a secretary at a central location by a phone call. The random-allocation sequence had been generated previously by a statistician (LG) by use of a computer-generated random code. Patients were stratified by centre and stage; the trial was unblinded. Analysis was by intention to treat.
Treatment plan 54 patients were allocated to PEFG regimens that were repeated every 28 days16 and that consisted of 40 mg/m2 cisplatin (Cisplatino-TEVA®, Pharmachemie BV, Haarlem, Netherlands) and 40 mg/m2 epirubicin (Farmarubicina®, Pharmacia & Upjohn SpA, Milan, Italy) given intravenously on day one, 600 mg/m2 gemcitabine (Gemzar, Eli Lilly, Fegersheim, France) given in 1 h on days 1 and 8, and 200 mg/m2 fluorouracil (Fluorouracile-TEVA®, Pharmachemie BV, Haarlem, Netherlands) a day as protracted infusion for the duration of chemotherapy by use of an indwelling, implanted central venous catheter. Guidelines for dose reduction and treatment delay have been reported previously.16 50 patients were allocated to 1000 mg/m2 gemcitabine a week given intravenously over 30 min for 7 consecutive weeks followed by 1 week of rest, and then subsequently for 3 of every 4 weeks. Treatment was delayed if absolute neutrophil count was less than 1000106 cells/L, platelet count was less than 50109/L, or if the patient developed grade 3–4 non-haematological toxic effects. Patients were removed from the study if recovery was not evident within 2 weeks. For patients who developed grade 3–4 non-haematological toxic effects, treatment was delayed until recovery to grade 2 effects, at which http://oncology.thelancet.com Vol 6 June 2005
Articles
Outcome measures Toxic effects were graded by the National Cancer Institute Common Toxicity Criteria version 2.0.19 All scans were reviewed by a radiologist (RN) who was blinded to treatment assignment. Standard WHO response criteria were used to define the best antitumour effects recorded.20 Complete response was defined as the disappearance of all measurable and assessable disease in all disease sites; partial response as a decrease of 50% or more in the sum of the products of the perpendicular diameters of all measurable lesions, with no new lesions or progression of assessable disease; progressive disease as an increase of 25% or more of the product of the perpendicular diameters of all measurable lesions, the appearance of new lesions, or the reappearance of a lesion that had previously disappeared; and stable disease as that which did not
Number of cycles 0 1 2 3 4 5 6 Median (range) Total number of cycles Time between randomisation and treatment (days) Median (IQR) Duration of chemotherapy (weeks) Median (IQR)
PEFG (n=52)
Gemcitabine (n=47)
2 2 11 0 12 2 23 4 (0–6) 220
0 23 4 10 0 7 3 2 (1–6) 114
4 (1–8)
4 (1–7)
17 (9–25)
Table 2: Details of treatment
http://oncology.thelancet.com Vol 6 June 2005
10 (8–16)
Progression-free survival (months) 4-month (95% CI) Median (IQR) Overall survival 1-year (95% CI) 2-year (95% CI)
PEFG (n=52)
Gemcitabine (n=47)
Hazard ratio (95% CI)
p
60% (46–72) 5·4 (2·0–9·6)
28% (17–42) 3·3 (2·2–5·3)
0·46 (0·26–0·79) 0·51 (0·33–0·78)
0·001 0·0033
38·5% (25·3–51·7) 11·5% (2·8–20·2)
21·3% (9·6–33·0) 2·1% (0·0–6·2)
0·68 (0·42–1·09) 0·63 (0·42–0·96)
0·1119 0·033
Table 3: Progression-free survival and overall survival
meet the criteria for the above.20 The duration of complete response was defined as the time between the first documentation of complete disease resolution and the first documented observation of progressive disease, and the duration of partial response as the time between start of treatment and first observation of progressive disease. Progression-free survival and overall survival were measured from the initiation of treatment to the time of progressive disease, death, or last follow-up assessment. Clinical benefit was assessed on the basis of pain, Karnofsky performance status, and change in bodyweight as described previously.1 Briefly, patients were thought to have a clinical benefit if they had an increase of 20 points or more from baseline Karnofsky performance status score, a decrease of 50% or more in use of analgesics from baseline, or a reduction of 50% or more in baseline Memorial Pain Assessment Card score, without being negative for either of the other two criteria. This improvement had to be maintained for at least 4 weeks. Quality of life was assessed by use of the European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 3021 supplemented by the pancreatic cancer module (EORTC QLQ-PAN-26).22 Differences of more than 10 points on the transformed scales were classified as clinically important.23 The results of the quality-of-life analyses have been reported previously.24 100 Progression-free survival (%)
time the dose of gemcitabine was lowered by 25%. Both groups were allocated treatment for a maximum of 6 months—ie, six cycles of PEFG and five of gemcitabine. Treatment was continued up to this time unless unacceptable side-effects developed, disease progressed, the patient refused treatment, or a medical decision was made to stop treatment. In patients with stage IVA disease treatment was classified as complete after 4 months if no clinical benefit was seen, the reduction in tumour size was less than 25%, or if CA19.9 concentration decreased to less than 50% of baseline. At the end of treatment, patients with operable stage IVA disease were recommended for surgery and were allowed to receive radiotherapy. Assessment of disease, including spiral CT of the abdomen and chest, was done at baseline, every 8 weeks during chemotherapy, and then every 3 months. For assessment of clinical benefit, patients were asked to record pain intensity every day by completion of a Memorial Pain Assessment Card and a diary on the use of analgesics. Performance status and bodyweight were assessed once a week.
90 80
Log-rank p=0·0033
70 60 50 PEFG
40 30 20
Gemcitabine
10 0 0
2
4
6 8 Follow-up (months)
Numbers at risk PEFG
51*
36
31
20
Gemcitabine
47
27
13
6
10
12
15
10
6
6
2
1
Figure 2: Progression-free survival *One patient died 1 week after randomisation and before starting treatment.
371
Articles
Age (continuous variable) Group (PEFG [n=52] vs gemcitabine [n=47]) Stage (IVA [n=29] vs metastatic [n=70]) Karnofsky performance status (continuous variable)
Progression-free survival
Overall survival
Hazard ratio (95 % CI)
p
Hazard ratio (95% CI)
p
0·98 (0·96–1·00) 0·50 (0·33–0·77) 2·13 (1·33–3·41) 0·99 (0·98–1·01)
0·1080 0·0021 0·0020 0·4697
0·98 (0·95–1·00) 0·63 (0·41–0·96) 2·10 (1·29–3·42) 0·99 (0·97–1·00)
0·0717 0·0345 0·0035 0·1573
Table 4: Multivariate analyses
Statistical analysis On the basis of progression-free survival curves for gemcitabine1 and PEFG,16 and in view of the idea that a large difference in progression-free survival could be found only in the direction of the PEFG group and that the number of events would correspond to the number of patients because progressive disease or death for all patients was anticipated at the time of final analysis, a sample size of 100 patients was planned to attain 90% probability of detecting (with a one-sided 5% level) a 30% absolute difference in 4-month progression-free survival. Main analyses were by intention to treat. At univariate analyses, survival curves were estimated with the Kaplan-Meier method and compared by use of the log-rank test. Multivariate analysis by the Cox’s proportional hazard model was done to estimate independent risk factors that could affect progression-free survival and overall survival. Clinical characteristics, toxic effects, response, and clinical benefit were assessed with the 2 test or Fisher’s exact test for categorical variables. All probability values were from two-sided tests. Analyses were done with the Statistica 4.0 statistical package for Microsoft Windows). No interim analysis was planned or done, and no early-stopping rule was planned for this trial.
was discontinued before completion in 27 patients: 20 had progressive disease, four refused to continue chemotherapy (one with partial response and three with stable disease), and three discontinued treatment because of doctor’s instructions (two with partial response and one with stable disease). 114 cycles were delivered in the gemcitabine group. Three patients in the gemcitabine group received six cycles after the attending physician consulted the trial coordinator. Dose intensity was 734 mg/m2 a week (92% of intended dose). Treatment was discontinued before completion in 37 patients: 32 had progressive disease, two had stable disease and refused to continue chemotherapy, two patients with stable disease discontinued treatment because of pneumonia and febrile jaundice, and one patient with a partial response discontinued treatment because of doctor’s instructions. After chemotherapy, radiotherapy (median 54 Gy, range 54–56) was given to six of 15 patients in the PEFG group with stage IVA disease (one of whom was admitted to surgery), and to five of 14 patients in the gemcitabine group with stage IVA disease. One patient (allocated gemcitabine alone) had a reduction in tumour size of more than 25%, absence of clinical benefit, and a reduction in CA19.9 of more than 50% of baseline concentration after 4 months. This patient refused to continue chemotherapy and received radiotherapy.
Results Patients
372
100 90 80 Overall survival (%)
Final analyses were done on Feb 28, 2005, when all living patients had completed at least 24 months of follow-up (median follow-up 33·5 months [range 24·0–46·0]). Figure 1 shows the trial profile. No patient was lost to follow-up. 23 patients in the PEFG group and 20 patients in the gemcitabine group were assessable for assessment of clinical benefit, and 37 (19 and 18, respectively) did not meet the eligibility criteria, and 19 (10 and 9, respectively) were not compliant. Table 1 shows baseline characteristics of patients by treatment group. Table 2 shows treatment received by each group. 220 cycles of PEFG were delivered to patients in the PEFG group. Dose intensity was 9·5 mg/m2 a week (95% of intended dose) for epirubicin and cisplatin, 1140 mg/m2 a week (81%) for fluorouracil, and 255 mg/m2 a week (85%) for gemcitabine. Treatment
Log-rank p=0·047
70 60 50
PEFG
40 30 20 10
Gemcitabine
0 0
3
6
51* 47
46 41
35 30
9 12 15 Follow-up (months)
18
21
24
11 5
6 1
5 1
Numbers at risk PEFG Gemcitabine
25 20
20 10
12 6
Figure 3: Overall survival *One patient died 1 week after randomisation before starting treatment.
http://oncology.thelancet.com Vol 6 June 2005
Articles
Partial or complete response Stable disease Progressive disease
PEFG (n=52) % (95% CI)
Gemcitabine (n=47) % (95% CI)
38·5 (25·3–51·7) 28·8 (16·5–41·1) 32·7 (20·0–45·5)
8·5 (0·5–16·5) 25·5 (13·0–38·0) 66·0 (52·4–79·5)
Table 5: Best response during treatment
Table 3 shows survival by treatment group. 51 patients in the PEFG group and 46 patients in the gemcitabine group had progressive disease at followup. At the time of final analysis two patients were progression free, one patient in the PEFG group (at 23 months) and one patient in the gemcitabine group (31 months). 11 patients (six assigned PEFG and five assigned gemcitabine) had no radiological confirmation of progressive disease, and progressionfree survival was calculated as the interval between treatment initiation and death. More patients assigned PEFG were progression free at 4 months than were those assigned gemcitabine alone (table 3, figure 2). In the PEFG group, six patients had not progressed and were at risk at 12 months. These patients were free of progression at 12·9, 13·8, 15·2, 20·4, 23·4, and more than 23 months. In the gemcitabine group, one patient was free of progressive disease at 31 months. Multivariate analyses confirmed that treatment group was an independent predictor of progression-free survival (table 4). At follow-up, 95 patients had died of disease progression: three patients (one with metastatic disease) assigned PEFG and one patient with stage IVA disease assigned gemcitabine were alive at Feb 28, 2005 (n=2), Feb 3, 2005 (33 months’ survival for one patient assigned PEFG), and at Feb 9, 2005 (25 months’ survival for one patient assigned PEFG). The hazard ratio for death in the PEFG group compared with the gemcitabine group was 0·65 (95% CI 0·43–0·99, p=0·047, log-rank test). The proportion of patients with
1-year and 2-year overall survival was greater in the PEFG group than in the gemcitabine group (table 3, figure 3). In the PEFG group, five patients were alive and at risk at 24 months (these patients survived for 25, 29, 30, 33, and 46 months). In the gemcitabine group, one patient was alive and at risk at 24 months, and survived for 34 months. Multivariate analyses showed that treatment group was an independent predictor of overall survival (table 4). Radiological analyses showed that one patient had a complete response and 19 had partial responses after PEFG treatment compared with four patients who had a partial response after gemcitabine treatment (odds ratio 6·60 [2·11–20·60], p=0·0008; table 5). 15 patients in the PEFG group and 12 patients in the gemcitabine group had stable disease. A complete response that lasted for 17·5 months was recorded in one patient assigned PEFG who had more than ten liver metastases. Median duration of partial response was 9·5 months (IQR 6·3–12·9) after PEFG and 6·9 months (4·3–8·7) after gemcitabine; median duration of stable disease was 5·2 months (3·8–8·8) after PEFG and 5·6 months (4·6–9·2) after gemcitabine. 15 (65% [95% CI 45–81]) of 23 patients assigned PEFG had a clinical benefit compared with five (25% [11–47]) of 20 patients assigned gemcitabine (p=0·0139). Table 6 shows treatment-related side-effects. More patients had grade 3–4 neutropenia and thrombocytopenia in the PEFG group than in the gemcitabine group (p0·0001). One patient assigned PEFG was admitted to hospital for febrile neutropenia. Growth factors chosen by the attending physician were used for 3 days in five patients assigned PEFG (one patient at cycle one, one at cycle two, two at cycle three, and two at cycle five), and in two patients assigned gemcitabine (one patient at cycle one and one at cycle five). Two patients assigned PEFG and three patients assigned gemcitabine were admitted to hospital for nonneutropenic infections. Grade 3–4 anaemia and grade 3
PEFG (n=220)
Granulocytes Platelets Haemoglobin Stomatitis Nausea or vomiting Diarrhoea Anorexia Fatigue Hand-foot syndrome Liver Fever Febrile neutropenia Non-neutropenic infection
Gemcitabine (n=114)
Grade 0
Grade 1/2
Grade 3
Grade 4
Grade 0
Grade 1/2
Grade 3
Grade 4
55 (4) 71 (9) 91 (10) 176 (33) 66 (27) 190 (37) 213 (47) 203 (39) 212 (47) 207 (46) 211 (46) 219 (51) 218 (50)
71 (5) 86 (13) 121 (35) 35 (12) 141 (20) 29 (14) 7 (5) 13 (9) 7 (4) 9 (3) 7 (4) 0 0
68 (24) 62 (29) 7 (6) 9 (7) 13 (5) 1 (1) 0 4 (4) 1 (1) 4 (3) 2 (2) 1 (1) 2 (2)
26 (19) 1 (1) 1 (1) 0 0 0 0 0 0 0 0 0 0
60 (20) 91 (30) 58 (17) 106 (42) 60 (20) 97 (36) 112 (45) 100 (36) 114 (47) 112 (45) 87 (31) 114 (47) 111 (44)
32 (13) 21 (16) 51 (26) 5 (3) 45 (20) 17 (11) 0 10 (8) 0 1 (1) 27 (16) 0 0
16 (10) 2 (1) 5 (4) 3 (2) 8 (6) 0 1 (1) 3 (2) 0 0 0 0 2 (2)
6 (4) 0 0 0 1 (1) 0 1 (1) 1 (1) 0 1 (1) 0 0 1 (1)
Data are total number of cycles (number of cycles for which patients had worst toxic effects).
Table 6: Treatment-related toxic effects
http://oncology.thelancet.com Vol 6 June 2005
373
Articles
stomatitis were more common in patients allocated combination treatment than in those allocated gemcitabine alone (p=0·16). Red-cell transfusion was used in six patients assigned PEFG (one patient at cycle one, one at cycle two, two at cycle three, and two at cycle five), and in four patients assigned gemcitabine (four patients at cycle one and one at cycle six). Erythropoietin chosen by the attending physician was given to three patients in the PEFG group and to four patients in the gemcitabine group. There were no treatment-related deaths. Findings of the quality-of-life analyses have been reported previously.24 Our sample size was insufficient to obtain adequate statistical power to detect reliably differences between groups for multiple comparisons. Although conventional statistical testing was not done, we noted that patients allocated PEFG were 20–44% more likely to obtain a clinically relevant improvement in scores for emotional functioning, overall quality of life, cognitive measures, pain, fatigue, indigestion, dyspnoea, appetite loss, and flatulence than were those allocated gemcitabine. Patients in the gemcitabine group had better scores for sexual function and body image than did those in the PEFG group. 86 patients had radiological documentation of progressive disease. 22 of 45 patients assigned PEFG and 25 of 41 patients assigned gemcitabine received salvage therapy with more than ten different regimens. 17 patients in the gemcitabine group received salvage PEFG. The median interval between progression and death was 3·9 months (IQR 2·1–7·1) in the PEFG group and 3·8 months (2·7–8·2) in the gemcitabine group (p=0·3536). In the subset of patients who had salvage treatment, median interval between progression and death was 6·8 months (3·7–9·1) in the PEFG group and 7·0 months (4·5–9·2, p=0·4301).
Discussion We have shown that patients allocated PEFG had a more favourable outcome in terms of progression-free survival and overall survival than did those allocated standard treatment with gemcitabine. After a minimum follow-up of 24 months, the survival advantage of PEFG was maintained at 1 year and at 2 years, and was confirmed by multivariate analyses. The effect of radiotherapy and salvage therapy on disease progression did not differ between groups. The difference in treatment duration between groups was dependent mainly on disease progression, and was therefore associated with the difference in activity between regimens. Haematological toxic effects were significantly higher in patients assigned combination treatment than in those assigned gemcitabine, but were of short duration and manageable. Groups did not differ in frequency of febrile neutropenia, use of growth factors, red-cell transfusions, erythropoietin use, and nonhaematological toxic effects. 374
Gemcitabine was approved on the basis of data from a trial that used clinical benefit as a primary endpoint.1 However, this measure was not validated and its use has been criticised because it did not take into account quality of life.25 We have shown that PEFG is associated with a significantly improved clinical benefit compared with gemcitabine alone. Furthermore, although conventional statistical testing for quality of life was not done, when all the descriptive data were taken into account quality of life for patients assigned PEFG was at least equal to that for the gemcitabine group for most scales. This finding further supports the use of combination chemotherapy compared with gemcitabine in that improved clinical outcome with PEFG is not achieved at the cost of impaired quality of life. The main drawbacks of our study are with regard to sample size and the generalisability of results. The sample size, which might seem to favour the combination group, was calculated on the basis of previous data1,16 that suggested an unusually large difference in progression-free survival between PEFG and standard treatment. The choice to calculate sample size by a one-sided test was based on the idea that a large difference in outcome could be found only in the direction of the PEFG group, and that treatment-related or catheter-related complications in this group could not worsen the outcome.16,26–28 Our findings confirm the validity of both assumptions. However, if we had used a two-sided test, our sample size would have been about the same (ie, 49 patients per group) to achieve 80% probability of detecting, with a 5% level, a 30% difference from 35% to 65%. With regard to validity of the control group and generalisability of results, comparisons with different trials are difficult and remain speculative because of differences in selection criteria and imbalances in prognostic factors. Moreover, the potential effect of salvage therapy on overall survival is unknown in most reported series. The proportion of patients with a Karnofsky performance status of more than 70 or with metastatic disease in our study is consistent with those in previous phase III trials in pancreatic adenocarcinoma.1–11 By contrast, these trials included patients with a more favourable prognosis,2 such as those with stage I–II disease1–3,5,6,10 or patients older than 70 years,1–3,5,6,8–10 which could have positively biased the outcome, as could have a dropout of more than 20% of patients during the lead-in period.1 Nevertheless, the safety, dose-intensity, activity, and efficacy of our control group was consistent with other series,1–11 suggesting that such factors had a negligible effect on outcome. We chose 4-month progression-free survival as our primary endpoint. The European Agency for the Evaluation of Medical Products (www.emea.eu.int) and the US Food and Drug Administration29 classify progression-free survival as an appropriate endpoint for assessment of anticancer agents in human beings. http://oncology.thelancet.com Vol 6 June 2005
Articles
Unlike overall survival, progression-free survival is not affected by subsequent, or crossover, therapies or by supportive care, and needs smaller patient numbers to detect a difference between treatment groups than does assessment of overall survival. However, progression-free survival is associated with a risk of: assessment bias in unblinded trials; unequal time interval of progression assessment across treatment groups; asymmetrical censoring; and inappropriate crediting of an unknown amount of additional progression-free time in patients who die without documentation of progressive disease and who have the date of death recorded as the progression date. In our trial, the timing of tumour assessment was identical in the two groups, and a radiologist blinded to treatment group centrally reviewed CT scans to keep the effects of biases on progression-free survival to a minimum. Furthermore, about 10% of patients were not assessable for the primary endpoint because scans were not done in time to document progression, and the comparison between progressionfree-survival curves was repeated with these patients omitted, with similar findings (p=0·005, data not shown). Furthermore, the time between progression and death was much the same for both groups, suggesting that salvage therapy had a similar effect on outcome in both groups, and that progression-free survival might be an adequate prediction of overall survival. Previous trials have used 6-month progression-free survival as the primary endpoint, and on the basis of previously reported data for 6-month progression-free survival (52% for PEFG vs 22% for gemcitabine),16 if we had chosen this endpoint for our sample calculation, the number of patients would have been about the same (48 patients per group with =0·05 and =0·1, one-sided test). In conclusion, our findings suggest that a clinically relevant effect on the outcome for pancreatic cancer could be achieved with adequate chemotherapy. Because our trial has limitations in interpretation of secondary endpoints because of small patient numbers, a larger confirmatory trial may be needed before this combination regimen could be regarded as standard treatment. Nevertheless, our findings are important because PEFG had manageable toxic effects, did not negatively affect quality of life, and maintained a statistically and clinically relevant outcome advantage compared with standard treatment. Accordingly, PEFG might be a feasible and effective first-line treatment for patients with locally advanced or metastatic pancreatic adenocarcinoma. Contributors M Reni and L Galli were responsible for the study idea, study design, data analysis and interpretation, and drafting of the article. M Reni, S Cordio, C Milandri, P Passoni, E Bonetto, C Oliani, G Luppi, R Nicoletti, R Bordonaro, A Passardi, A Zerbi, G Balzano, L Aldrighetti, C Staudacher, E Villa, and V Di Carlo participated in data acquisition. All authors were involved in critical review of the article. M Reni had full access to all data in the study and had final responsibility for the decision to submit for publication.
http://oncology.thelancet.com Vol 6 June 2005
Conflict of interest We declare no conflicts of interest. Acknowledgments No funding was received for this study. References 1 Burris HA 3rd, Moore MJ, Andersen J, et al. Improvement in survival and clinical benefit with gemcitabine as first-line therapy for patients with advanced pancreas cancer: a randomized trial. J Clin Oncol 1997; 15: 2403–13. 2 Berlin JD, Catalano P, Thomas JP, et al. Phase III study of gemcitabine in combination with fluorouracil versus gemcitabine alone in patients with advanced pancreatic carcinoma: Eastern Cooperative Oncology Group Trial E2297. J Clin Oncol 2002; 20: 3270–75. 3 Moore MJ, Hamm J, Dancey J, et al. Comparison of gemcitabine versus the matrix metalloproteinase inhibitor BAY 12-9566 in patients with advanced or metastatic adenocarcinoma of the pancreas: a phase III trial of the National Cancer Institute of Canada Clinical Trials Group. J Clin Oncol 2003; 21: 3296–302. 4 Rocha Lima CMS, Green MR, Rotche R, et al. Irinotecan plus gemcitabine results in no survival advantage compared with gemcitabine monotherapy in patients with locally advanced or metastatic pancreatic cancer despite increased tumor response rate. J Clin Oncol 2004; 22: 3776–83. 5 Bramhall SR, Schultz J, Nemunaitis J, et al. A double-blind placebocontrolled, randomised study comparing gemcitabine and marimastat with gemcitabine and placebo as first line therapy in patients with advanced pancreatic cancer. Br J Cancer 2002; 87: 161–67. 6 Bramhall SR, Rosemurgy A, Brown PD, for the Marimastat Pancreatic Cancer Study Group. Marimastat as first-line therapy for patients with unresectable pancreatic cancer: a randomized trial. J Clin Oncol 2001; 19: 3477–355. 7 Heinemann V, Quietzsch D, Gieseler F, et al. A phase III trial comparing gemcitabine plus cisplatin vs gemcitabine alone in advanced pancreatic adenocarcinoma. Proc Am Soc Clin Oncol 2003; 22: abstr 1003. 8 Van Cutsem E, van de Velde H, Karasek P, et al. Phase III trial of gemcitabine plus tipifarnib compared with gemcitabine plus placebo in advanced pancreatic cancer. J Clin Oncol 2004; 22: 1430–38. 9 O’Reilly EM, Abou-Alfa GK, Letorneau R, et al. A randomized phase III trial of DX-8951f (exatecan mesylate) and gemcitabine vs gemcitabine alone in advanced pancreatic cancer. Proc Am Soc Clin Oncol 2004; 23: abstr 4006. 10 Richards DA, Kindler HL, Oettle H, et al. A randomized phase III study comparing gemcitabine + pemetrexed versus gemcitabine in patients with locally advanced and metastatic pancreas cancer. Proc Am Soc Clin Oncol 2004; 23: abstr 4007. 11 Louvet C, Labianca R, Hammel P, et al. GemOx (gemcitabine + oxaliplatin) versus Gem (gemcitabine) in non-resectable pancreatic adenocarcinoma: final results of the GERCOR/GISCAD Intergroup phase III. Proc Am Soc Clin Oncol 2004; 23: abstr 4008. 12 Maisey N, Chau I, Cunningham D, et al. Multicenter randomized phase III trial comparing protracted venous infusion (PVI) fluorouracil (5-FU) with PVI 5-FU plus mitomycin in inoperable pancreatic cancer. J Clin Oncol 2002; 20: 3130–36. 13 Ducreux M, Rougier P, Pignon JP, et al. A randomised trial comparing 5-FU with 5-FU plus cisplatin in advanced pancreatic carcinoma. Ann Oncol 2002; 13: 1185–91. 14 Tempero M, Plunkett W, Ruiz Van Haperen V, et al. Randomized phase II comparison of dose-intense gemcitabine: thirthy minute infusion and fixed dose rate infusion in patients with pancreatic adenocarcinoma. J Clin Oncol 2003; 21: 3402–08. 15 Moore MJ, Goldstein D, Hamm J, et al. Erlotinib improves survival when added to gemcitabine in patients with advanced pancreatic cancer. A phase III trial of the National Cancer Institute of Canada Clinical Trial Group (NCIC-CTG). Proceedings of the 2005 Gastrointestinal Cancers Symposium Jan 27–29, 2005; Hollywood, FL, USA, 3: abstr 77. 16 Reni M, Passoni P, Panucci MG, et al. Definitive results of phase II trial of cisplatin, epirubicin, continuous-infusion fluorouracil, and gemcitabine in stage IV pancreatic adenocarcinoma. J Clin Oncol 2001; 19: 2679–86.
375
Articles
17
18
19 20 21
22
23
376
Sobin LH, Wittekind C, eds. International Union Against Cancer. TNM classification of malignant tumours, 5th edn. New York: Wiley-Liss, 1997. Fishman B, Pasternak S, Wallenstein SL, et al. The Memorial Pain Assessment Card: a valid instrument for the evaluation of cancer pain. Cancer 1987; 60: 1151–58. Ajani JA, Welch SR, Raber MN, et al. Comprehensive criteria for assessing therapy-induced toxicity. Cancer Invest 1990; 8: 147–59. Miller AB, Hoogstraten B, Staquet M, Winkler A. Reporting results of cancer treatment. Cancer 1981; 47: 207–14. Fitzsimmons D, Johnson CD, George S, et al. Development of a disease specific quality of life questionnaire module to supplement the EORTC core cancer qol questionnaire, the QLQ-C30 in patients with pancreatic cancer. Eur J Cancer 1999; 35: 939–41. Fayers P, Aaronson N, Bjordal K, et al. EORTC QLQ scoring manual, 3rd edn. Brussels: European Organisation for Research and Treatment of Cancer, 2001. Osoba D, Rodrigues G, Myles J, et al. Interpreting the significance of changes in health-related quality-of-life scores. J Clin Oncol 1998; 16: 139–44.
24
25
26
27
28
29
Reni M, Bonetto E, Cordio S, et al. Quality of life outcome in advanced pancreatic adenocarcinoma: results from a phase III trial. Ann Oncol 2004; 15: abstr 911. Hoffman K, Glimelius B. Evaluation of clinical benefit of chemotherapy in patients with upper gastrointestinal cancer. Acta Oncol 1998; 37: 651–59. Reni M, Passoni P, Bonetto E, et al. Final results of a prospective trial of a PEFG (cisplatin, epirubicin, 5-fluorouracil, gemcitabine) regimen followed by radiotherapy after curative surgery for pancreatic adenocarcinoma. Oncology, in press. Passoni P, Reni M, Zanello A, et al. Preliminary results of a phase II trial of PEFG (cisplatin, epirubicin, 5-fluorouracil, and gemcitabine) in advanced gallbladder, biliary tract, and ampulla of Vater adenocarcinoma. Ann Oncol 2002; 13: abstr 724. Villa E, Reni M, Zanello A, et al. PEFG regimen (cisplatin, epirubicin, 5-fluorouracil continuous infusion, and gemcitabine) in advanced gastric adenocarcinoma. Ann Oncol 2003; 14: abstr D5. Johnson JR, Williams G, Pazdur R. End points and United States Food and Drug Administration approval of oncology drugs. J Clin Oncol 2003; 21: 1404–11.
http://oncology.thelancet.com Vol 6 June 2005
Gemcitabine versus cisplatin, epirubicin, fluorouracil, and gemcitabine in advanced pancreatic cancer: a randomised controlled multicentre phase III trial Michele Reni, Stefano Cordio, Carlo Milandri, Paolo Passoni, Elisa Bonetto, Cristina Oliani, Gabriele Luppi, Roberto Nicoletti, Laura Galli, Roberto Bordonaro, Alessandro Passardi, Alessandro Zerbi, Gianpaolo Balzano, Luca Aldrighetti, Carlo Staudacher, Eugenio Villa, Valerio Di Carlo
Summary Background Patients with advanced pancreatic adenocarcinoma have a poor response, progression-free survival, and overall survival with standard treatment. We aimed to assess whether a four-drug regimen could improve 4-month progression-free survival compared with gemcitabine alone. Methods In a randomised multicentre phase III trial, 52 patients were randomly assigned to 40 mg/m2 cisplatin and 40 mg/m2 epirubicin both given on day 1, 600 mg/m2 gemcitabine given intravenously over 1 h on days 1 and 8, and 200 mg/m2 fluorouracil a day given by continuous infusion on days 1–28 of a 4-week cycle (PEFG regimen), and 47 were assigned to 1000 mg/m2 gemcitabine given intravenously over 30 min once a week for 7 of 8 consecutive weeks in cycle 1 and for 3 of 4 weeks thereafter. The primary endpoint was 4-month progression-free survival. Secondary endpoints were overall survival, objective response, safety, and quality of life. Analyses were by intention to treat. Findings 51 patients assigned PEFG and 46 assigned gemcitabine alone had disease progression. 49 patients in the PEFG group and 46 in the gemcitabine group died from progressive disease. More patients allocated PEFG than gemcitabine alone were alive without progressive disease at 4 months (60% [95% CI 46–72] vs 28% [17–42]; hazard ratio [HR] 0·46 [0·26–0·79]). 1-year overall survival in the PEFG group was 38·5% (25·3–51·7) and in the gemcitabine group was 21·3% (9·6–33·0; HR 0·68 [0·42–1·09]). More patients assigned PEFG showed disease response than did those assigned gemcitabine (38·5% [25·3–51·7] vs 8·5% [0·5–16·5]; odds ratio 6·60 [2·11–20·60], p=0·0008). More patients in the PEFG group had grade 3–4 neutropenia and thrombocytopenia than in the gemcitabine group (p0·0001). Interpretation The PEFG regimen could be considered for treatment of advanced pancreatic adenocarcinoma.
Introduction Pancreatic adenocarcinoma is associated with a poor prognosis. More than 80% of patients present with advanced disease at diagnosis, and mortality is high because tumours infiltrate the abdominal blood vessels and spread to regional lymph nodes and liver at an early stage. Gemcitabine is regarded as the standard treatment for unresectable disease, and has led to an objective response in 4–26% of patients and a 1-year overall survival of 17–28% in phase III trials.1–11 Gemcitabine was licensed for treatment of advanced pancreatic cancer on the basis of improved clinical benefit and increased overall survival compared with fluorouracil.1 Several attempts to improve the efficacy of gemcitabine in advanced pancreatic adenocarcinoma by use of new agents or by addition of a second cytotoxic agent or other drugs to a standard dose and schedule of gemcitabine have not shown a survival advantage.2–11 However, a significant benefit in median progressionfree survival has been noted for gemcitabine combined with fluorouracil,2 cisplatin,7 or oxaliplatin,11 which suggests that platinum-based agents and antifolates have a role in treatment of pancreatic cancer. Moreover, http://oncology.thelancet.com Vol 6 June 2005
although combination treatment did not increase overall survival compared with gemcitabine alone, whether salvage therapy affects the relation between progressionfree survival and overall survival is not known. Strategies that have not included gemcitabine have been assessed, but have not shown better results.12,13 Findings from a phase II trial14 suggested that fixed-dose gemcitabine improves survival, but phase III data are not available. Gemcitabine combined with erlotinib improved 1-year overall survival compared with gemcitabine alone from 17% to 24%.15 Cisplatin, epirubicin, fluorouracil, and gemcitabine (PEFG)—an empirically derived treatment based on four agents that are active against pancreatic adenocarcinoma—showed promising activity with regard to response rate and survival in a previous phase II trial.16 We therefore did a multicentre randomised phase III trial to compare standard treatment of gemcitabine with that of PEFG by use of previously defined doses and schedules.16
Lancet Oncol 2005; 6: 369–76 Published online May 10, 2005 DOI: 10.1016/S1470-2045(05) 70175-3 See Reflection and Reaction page 352 Department of Radiochemotherapy (M Reni MD, P Passoni MD, E Bonetto MD, E Villa MD), Department of Radiology (R Nicoletti MD), Department of Statistics (L Galli MSc), and Department of Surgery (A Zerbi MD, G Balzano MD, L Aldrighetti MD, Prof C Staudacher MD, Prof V Di Carlo MD),S Raffaele H Scientific Institute, Milan, Italy Department of Oncology, S Luigi Currò H Catania, Italy (S Cordio MD, R Bordonaro MD); Department of Oncology, Pierantoni H Forlì, Italy (C Milandri MD, A Passardi MD); Department of Oncology, Azienda Ospedaliera and University, Verona, Italy (C Oliani MD); Department of Medical Oncology, Azienda Ospedaliera-Policlinico, Modena, Italy (G Luppi MD) Correspondence to: Dr Michele Reni, Department of Oncology, San Raffaele H Scientific Institute, via Olgettina 60, 20132 Milan, Italy [email protected]
Methods Patients Between April, 2000, and March, 2003, 104 patients from five institutions aged 18–70 years with histo369
Articles
104 patients randomised
54 allocated PEFG
50 allocated gemcitabine 3 ineligible 2 inadequate liver function 1 other cancer
2 ineligible (inadequate liver function)
52 analysed
47 analysed
Figure 1: Trial profile
logically or cytologically confirmed stage IVA (T4, N0–1, M0)17 or metastatic pancreatic adenocarcinoma were enrolled onto the study. Eligibility criteria included: Karnofsky performance status more than 40; at least one indicator lesion that was measurable in two dimensions and had not been irradiated previously; no previous chemotherapy; and adequate bone-marrow function (absolute neutrophil count 1500106 cells/L, platelet count 100109/L, and haemoglobin 100 g/L), kidney function (serum creatinine 132·6 mol/L), and liver function (total serum bilirubin 34·2 mol/L and serum transaminases 3 times upper normal limit). Patients who had ampullary tumours, other histological variants of pancreatic carcinoma, previous adjuvant chemotherapy, or previous malignant disease were ineligible for the study, with the exception of those who had had basalcell carcinoma of the skin, carcinoma in situ of the cervix, or other cancers for which the patient had been disease-free for at least 5 years. Eligibility criteria for assessment of clinical benefit were baseline use of analgesics of at least two nonsteroidal anti-inflammatory drugs, baseline pain
Age (years) Median (range) Sex Male Female Karnofsky performance status 70 70 Stage IVA Metastatic Previous treatment Pancreatic surgery Radiotherapy or chemotherapy Site of metastases Liver Lymph nodes Lung
Table 1: Baseline characteristics
370
PEFG (n=52)
Gemcitabine (n=47)
62 (37–69)
59 (25–69)
24 28
24 23
37 15
35 12
15 37
14 33
3 0
3 0
29 9 5
28 13 5
intensity score of 20 mm or more of a possible 100 mm on the Memorial Pain Assessment Card,18 or a Karnofsky performance status of 50–70. The study was done in accordance with the Declaration of Helsinki and approved by the local ethics committees of participating institutions; written informed consent was obtained from all patients.
Study objectives We chose 4-month progression-free survival as the primary endpoint to avoid the potential effect of salvage treatment after progression on overall survival and the potential effect of surgery or radiotherapy (allowed after 4 months in patients with stage IVA disease) on both overall survival and progression-free survival. Secondary endpoints were overall survival, objective response, clinical benefit, safety, and quality of life.
Randomisation Patients who fulfilled all inclusion criteria were registered by the attending physician at the coordinating institution. Baseline, treatment, and follow-up data were obtained at the coordinating institution. Randomisation was done by a secretary at a central location by a phone call. The random-allocation sequence had been generated previously by a statistician (LG) by use of a computer-generated random code. Patients were stratified by centre and stage; the trial was unblinded. Analysis was by intention to treat.
Treatment plan 54 patients were allocated to PEFG regimens that were repeated every 28 days16 and that consisted of 40 mg/m2 cisplatin (Cisplatino-TEVA®, Pharmachemie BV, Haarlem, Netherlands) and 40 mg/m2 epirubicin (Farmarubicina®, Pharmacia & Upjohn SpA, Milan, Italy) given intravenously on day one, 600 mg/m2 gemcitabine (Gemzar, Eli Lilly, Fegersheim, France) given in 1 h on days 1 and 8, and 200 mg/m2 fluorouracil (Fluorouracile-TEVA®, Pharmachemie BV, Haarlem, Netherlands) a day as protracted infusion for the duration of chemotherapy by use of an indwelling, implanted central venous catheter. Guidelines for dose reduction and treatment delay have been reported previously.16 50 patients were allocated to 1000 mg/m2 gemcitabine a week given intravenously over 30 min for 7 consecutive weeks followed by 1 week of rest, and then subsequently for 3 of every 4 weeks. Treatment was delayed if absolute neutrophil count was less than 1000106 cells/L, platelet count was less than 50109/L, or if the patient developed grade 3–4 non-haematological toxic effects. Patients were removed from the study if recovery was not evident within 2 weeks. For patients who developed grade 3–4 non-haematological toxic effects, treatment was delayed until recovery to grade 2 effects, at which http://oncology.thelancet.com Vol 6 June 2005
Articles
Outcome measures Toxic effects were graded by the National Cancer Institute Common Toxicity Criteria version 2.0.19 All scans were reviewed by a radiologist (RN) who was blinded to treatment assignment. Standard WHO response criteria were used to define the best antitumour effects recorded.20 Complete response was defined as the disappearance of all measurable and assessable disease in all disease sites; partial response as a decrease of 50% or more in the sum of the products of the perpendicular diameters of all measurable lesions, with no new lesions or progression of assessable disease; progressive disease as an increase of 25% or more of the product of the perpendicular diameters of all measurable lesions, the appearance of new lesions, or the reappearance of a lesion that had previously disappeared; and stable disease as that which did not
Number of cycles 0 1 2 3 4 5 6 Median (range) Total number of cycles Time between randomisation and treatment (days) Median (IQR) Duration of chemotherapy (weeks) Median (IQR)
PEFG (n=52)
Gemcitabine (n=47)
2 2 11 0 12 2 23 4 (0–6) 220
0 23 4 10 0 7 3 2 (1–6) 114
4 (1–8)
4 (1–7)
17 (9–25)
Table 2: Details of treatment
http://oncology.thelancet.com Vol 6 June 2005
10 (8–16)
Progression-free survival (months) 4-month (95% CI) Median (IQR) Overall survival 1-year (95% CI) 2-year (95% CI)
PEFG (n=52)
Gemcitabine (n=47)
Hazard ratio (95% CI)
p
60% (46–72) 5·4 (2·0–9·6)
28% (17–42) 3·3 (2·2–5·3)
0·46 (0·26–0·79) 0·51 (0·33–0·78)
0·001 0·0033
38·5% (25·3–51·7) 11·5% (2·8–20·2)
21·3% (9·6–33·0) 2·1% (0·0–6·2)
0·68 (0·42–1·09) 0·63 (0·42–0·96)
0·1119 0·033
Table 3: Progression-free survival and overall survival
meet the criteria for the above.20 The duration of complete response was defined as the time between the first documentation of complete disease resolution and the first documented observation of progressive disease, and the duration of partial response as the time between start of treatment and first observation of progressive disease. Progression-free survival and overall survival were measured from the initiation of treatment to the time of progressive disease, death, or last follow-up assessment. Clinical benefit was assessed on the basis of pain, Karnofsky performance status, and change in bodyweight as described previously.1 Briefly, patients were thought to have a clinical benefit if they had an increase of 20 points or more from baseline Karnofsky performance status score, a decrease of 50% or more in use of analgesics from baseline, or a reduction of 50% or more in baseline Memorial Pain Assessment Card score, without being negative for either of the other two criteria. This improvement had to be maintained for at least 4 weeks. Quality of life was assessed by use of the European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 3021 supplemented by the pancreatic cancer module (EORTC QLQ-PAN-26).22 Differences of more than 10 points on the transformed scales were classified as clinically important.23 The results of the quality-of-life analyses have been reported previously.24 100 Progression-free survival (%)
time the dose of gemcitabine was lowered by 25%. Both groups were allocated treatment for a maximum of 6 months—ie, six cycles of PEFG and five of gemcitabine. Treatment was continued up to this time unless unacceptable side-effects developed, disease progressed, the patient refused treatment, or a medical decision was made to stop treatment. In patients with stage IVA disease treatment was classified as complete after 4 months if no clinical benefit was seen, the reduction in tumour size was less than 25%, or if CA19.9 concentration decreased to less than 50% of baseline. At the end of treatment, patients with operable stage IVA disease were recommended for surgery and were allowed to receive radiotherapy. Assessment of disease, including spiral CT of the abdomen and chest, was done at baseline, every 8 weeks during chemotherapy, and then every 3 months. For assessment of clinical benefit, patients were asked to record pain intensity every day by completion of a Memorial Pain Assessment Card and a diary on the use of analgesics. Performance status and bodyweight were assessed once a week.
90 80
Log-rank p=0·0033
70 60 50 PEFG
40 30 20
Gemcitabine
10 0 0
2
4
6 8 Follow-up (months)
Numbers at risk PEFG
51*
36
31
20
Gemcitabine
47
27
13
6
10
12
15
10
6
6
2
1
Figure 2: Progression-free survival *One patient died 1 week after randomisation and before starting treatment.
371
Articles
Age (continuous variable) Group (PEFG [n=52] vs gemcitabine [n=47]) Stage (IVA [n=29] vs metastatic [n=70]) Karnofsky performance status (continuous variable)
Progression-free survival
Overall survival
Hazard ratio (95 % CI)
p
Hazard ratio (95% CI)
p
0·98 (0·96–1·00) 0·50 (0·33–0·77) 2·13 (1·33–3·41) 0·99 (0·98–1·01)
0·1080 0·0021 0·0020 0·4697
0·98 (0·95–1·00) 0·63 (0·41–0·96) 2·10 (1·29–3·42) 0·99 (0·97–1·00)
0·0717 0·0345 0·0035 0·1573
Table 4: Multivariate analyses
Statistical analysis On the basis of progression-free survival curves for gemcitabine1 and PEFG,16 and in view of the idea that a large difference in progression-free survival could be found only in the direction of the PEFG group and that the number of events would correspond to the number of patients because progressive disease or death for all patients was anticipated at the time of final analysis, a sample size of 100 patients was planned to attain 90% probability of detecting (with a one-sided 5% level) a 30% absolute difference in 4-month progression-free survival. Main analyses were by intention to treat. At univariate analyses, survival curves were estimated with the Kaplan-Meier method and compared by use of the log-rank test. Multivariate analysis by the Cox’s proportional hazard model was done to estimate independent risk factors that could affect progression-free survival and overall survival. Clinical characteristics, toxic effects, response, and clinical benefit were assessed with the 2 test or Fisher’s exact test for categorical variables. All probability values were from two-sided tests. Analyses were done with the Statistica 4.0 statistical package for Microsoft Windows). No interim analysis was planned or done, and no early-stopping rule was planned for this trial.
was discontinued before completion in 27 patients: 20 had progressive disease, four refused to continue chemotherapy (one with partial response and three with stable disease), and three discontinued treatment because of doctor’s instructions (two with partial response and one with stable disease). 114 cycles were delivered in the gemcitabine group. Three patients in the gemcitabine group received six cycles after the attending physician consulted the trial coordinator. Dose intensity was 734 mg/m2 a week (92% of intended dose). Treatment was discontinued before completion in 37 patients: 32 had progressive disease, two had stable disease and refused to continue chemotherapy, two patients with stable disease discontinued treatment because of pneumonia and febrile jaundice, and one patient with a partial response discontinued treatment because of doctor’s instructions. After chemotherapy, radiotherapy (median 54 Gy, range 54–56) was given to six of 15 patients in the PEFG group with stage IVA disease (one of whom was admitted to surgery), and to five of 14 patients in the gemcitabine group with stage IVA disease. One patient (allocated gemcitabine alone) had a reduction in tumour size of more than 25%, absence of clinical benefit, and a reduction in CA19.9 of more than 50% of baseline concentration after 4 months. This patient refused to continue chemotherapy and received radiotherapy.
Results Patients
372
100 90 80 Overall survival (%)
Final analyses were done on Feb 28, 2005, when all living patients had completed at least 24 months of follow-up (median follow-up 33·5 months [range 24·0–46·0]). Figure 1 shows the trial profile. No patient was lost to follow-up. 23 patients in the PEFG group and 20 patients in the gemcitabine group were assessable for assessment of clinical benefit, and 37 (19 and 18, respectively) did not meet the eligibility criteria, and 19 (10 and 9, respectively) were not compliant. Table 1 shows baseline characteristics of patients by treatment group. Table 2 shows treatment received by each group. 220 cycles of PEFG were delivered to patients in the PEFG group. Dose intensity was 9·5 mg/m2 a week (95% of intended dose) for epirubicin and cisplatin, 1140 mg/m2 a week (81%) for fluorouracil, and 255 mg/m2 a week (85%) for gemcitabine. Treatment
Log-rank p=0·047
70 60 50
PEFG
40 30 20 10
Gemcitabine
0 0
3
6
51* 47
46 41
35 30
9 12 15 Follow-up (months)
18
21
24
11 5
6 1
5 1
Numbers at risk PEFG Gemcitabine
25 20
20 10
12 6
Figure 3: Overall survival *One patient died 1 week after randomisation before starting treatment.
http://oncology.thelancet.com Vol 6 June 2005
Articles
Partial or complete response Stable disease Progressive disease
PEFG (n=52) % (95% CI)
Gemcitabine (n=47) % (95% CI)
38·5 (25·3–51·7) 28·8 (16·5–41·1) 32·7 (20·0–45·5)
8·5 (0·5–16·5) 25·5 (13·0–38·0) 66·0 (52·4–79·5)
Table 5: Best response during treatment
Table 3 shows survival by treatment group. 51 patients in the PEFG group and 46 patients in the gemcitabine group had progressive disease at followup. At the time of final analysis two patients were progression free, one patient in the PEFG group (at 23 months) and one patient in the gemcitabine group (31 months). 11 patients (six assigned PEFG and five assigned gemcitabine) had no radiological confirmation of progressive disease, and progressionfree survival was calculated as the interval between treatment initiation and death. More patients assigned PEFG were progression free at 4 months than were those assigned gemcitabine alone (table 3, figure 2). In the PEFG group, six patients had not progressed and were at risk at 12 months. These patients were free of progression at 12·9, 13·8, 15·2, 20·4, 23·4, and more than 23 months. In the gemcitabine group, one patient was free of progressive disease at 31 months. Multivariate analyses confirmed that treatment group was an independent predictor of progression-free survival (table 4). At follow-up, 95 patients had died of disease progression: three patients (one with metastatic disease) assigned PEFG and one patient with stage IVA disease assigned gemcitabine were alive at Feb 28, 2005 (n=2), Feb 3, 2005 (33 months’ survival for one patient assigned PEFG), and at Feb 9, 2005 (25 months’ survival for one patient assigned PEFG). The hazard ratio for death in the PEFG group compared with the gemcitabine group was 0·65 (95% CI 0·43–0·99, p=0·047, log-rank test). The proportion of patients with
1-year and 2-year overall survival was greater in the PEFG group than in the gemcitabine group (table 3, figure 3). In the PEFG group, five patients were alive and at risk at 24 months (these patients survived for 25, 29, 30, 33, and 46 months). In the gemcitabine group, one patient was alive and at risk at 24 months, and survived for 34 months. Multivariate analyses showed that treatment group was an independent predictor of overall survival (table 4). Radiological analyses showed that one patient had a complete response and 19 had partial responses after PEFG treatment compared with four patients who had a partial response after gemcitabine treatment (odds ratio 6·60 [2·11–20·60], p=0·0008; table 5). 15 patients in the PEFG group and 12 patients in the gemcitabine group had stable disease. A complete response that lasted for 17·5 months was recorded in one patient assigned PEFG who had more than ten liver metastases. Median duration of partial response was 9·5 months (IQR 6·3–12·9) after PEFG and 6·9 months (4·3–8·7) after gemcitabine; median duration of stable disease was 5·2 months (3·8–8·8) after PEFG and 5·6 months (4·6–9·2) after gemcitabine. 15 (65% [95% CI 45–81]) of 23 patients assigned PEFG had a clinical benefit compared with five (25% [11–47]) of 20 patients assigned gemcitabine (p=0·0139). Table 6 shows treatment-related side-effects. More patients had grade 3–4 neutropenia and thrombocytopenia in the PEFG group than in the gemcitabine group (p0·0001). One patient assigned PEFG was admitted to hospital for febrile neutropenia. Growth factors chosen by the attending physician were used for 3 days in five patients assigned PEFG (one patient at cycle one, one at cycle two, two at cycle three, and two at cycle five), and in two patients assigned gemcitabine (one patient at cycle one and one at cycle five). Two patients assigned PEFG and three patients assigned gemcitabine were admitted to hospital for nonneutropenic infections. Grade 3–4 anaemia and grade 3
PEFG (n=220)
Granulocytes Platelets Haemoglobin Stomatitis Nausea or vomiting Diarrhoea Anorexia Fatigue Hand-foot syndrome Liver Fever Febrile neutropenia Non-neutropenic infection
Gemcitabine (n=114)
Grade 0
Grade 1/2
Grade 3
Grade 4
Grade 0
Grade 1/2
Grade 3
Grade 4
55 (4) 71 (9) 91 (10) 176 (33) 66 (27) 190 (37) 213 (47) 203 (39) 212 (47) 207 (46) 211 (46) 219 (51) 218 (50)
71 (5) 86 (13) 121 (35) 35 (12) 141 (20) 29 (14) 7 (5) 13 (9) 7 (4) 9 (3) 7 (4) 0 0
68 (24) 62 (29) 7 (6) 9 (7) 13 (5) 1 (1) 0 4 (4) 1 (1) 4 (3) 2 (2) 1 (1) 2 (2)
26 (19) 1 (1) 1 (1) 0 0 0 0 0 0 0 0 0 0
60 (20) 91 (30) 58 (17) 106 (42) 60 (20) 97 (36) 112 (45) 100 (36) 114 (47) 112 (45) 87 (31) 114 (47) 111 (44)
32 (13) 21 (16) 51 (26) 5 (3) 45 (20) 17 (11) 0 10 (8) 0 1 (1) 27 (16) 0 0
16 (10) 2 (1) 5 (4) 3 (2) 8 (6) 0 1 (1) 3 (2) 0 0 0 0 2 (2)
6 (4) 0 0 0 1 (1) 0 1 (1) 1 (1) 0 1 (1) 0 0 1 (1)
Data are total number of cycles (number of cycles for which patients had worst toxic effects).
Table 6: Treatment-related toxic effects
http://oncology.thelancet.com Vol 6 June 2005
373
Articles
stomatitis were more common in patients allocated combination treatment than in those allocated gemcitabine alone (p=0·16). Red-cell transfusion was used in six patients assigned PEFG (one patient at cycle one, one at cycle two, two at cycle three, and two at cycle five), and in four patients assigned gemcitabine (four patients at cycle one and one at cycle six). Erythropoietin chosen by the attending physician was given to three patients in the PEFG group and to four patients in the gemcitabine group. There were no treatment-related deaths. Findings of the quality-of-life analyses have been reported previously.24 Our sample size was insufficient to obtain adequate statistical power to detect reliably differences between groups for multiple comparisons. Although conventional statistical testing was not done, we noted that patients allocated PEFG were 20–44% more likely to obtain a clinically relevant improvement in scores for emotional functioning, overall quality of life, cognitive measures, pain, fatigue, indigestion, dyspnoea, appetite loss, and flatulence than were those allocated gemcitabine. Patients in the gemcitabine group had better scores for sexual function and body image than did those in the PEFG group. 86 patients had radiological documentation of progressive disease. 22 of 45 patients assigned PEFG and 25 of 41 patients assigned gemcitabine received salvage therapy with more than ten different regimens. 17 patients in the gemcitabine group received salvage PEFG. The median interval between progression and death was 3·9 months (IQR 2·1–7·1) in the PEFG group and 3·8 months (2·7–8·2) in the gemcitabine group (p=0·3536). In the subset of patients who had salvage treatment, median interval between progression and death was 6·8 months (3·7–9·1) in the PEFG group and 7·0 months (4·5–9·2, p=0·4301).
Discussion We have shown that patients allocated PEFG had a more favourable outcome in terms of progression-free survival and overall survival than did those allocated standard treatment with gemcitabine. After a minimum follow-up of 24 months, the survival advantage of PEFG was maintained at 1 year and at 2 years, and was confirmed by multivariate analyses. The effect of radiotherapy and salvage therapy on disease progression did not differ between groups. The difference in treatment duration between groups was dependent mainly on disease progression, and was therefore associated with the difference in activity between regimens. Haematological toxic effects were significantly higher in patients assigned combination treatment than in those assigned gemcitabine, but were of short duration and manageable. Groups did not differ in frequency of febrile neutropenia, use of growth factors, red-cell transfusions, erythropoietin use, and nonhaematological toxic effects. 374
Gemcitabine was approved on the basis of data from a trial that used clinical benefit as a primary endpoint.1 However, this measure was not validated and its use has been criticised because it did not take into account quality of life.25 We have shown that PEFG is associated with a significantly improved clinical benefit compared with gemcitabine alone. Furthermore, although conventional statistical testing for quality of life was not done, when all the descriptive data were taken into account quality of life for patients assigned PEFG was at least equal to that for the gemcitabine group for most scales. This finding further supports the use of combination chemotherapy compared with gemcitabine in that improved clinical outcome with PEFG is not achieved at the cost of impaired quality of life. The main drawbacks of our study are with regard to sample size and the generalisability of results. The sample size, which might seem to favour the combination group, was calculated on the basis of previous data1,16 that suggested an unusually large difference in progression-free survival between PEFG and standard treatment. The choice to calculate sample size by a one-sided test was based on the idea that a large difference in outcome could be found only in the direction of the PEFG group, and that treatment-related or catheter-related complications in this group could not worsen the outcome.16,26–28 Our findings confirm the validity of both assumptions. However, if we had used a two-sided test, our sample size would have been about the same (ie, 49 patients per group) to achieve 80% probability of detecting, with a 5% level, a 30% difference from 35% to 65%. With regard to validity of the control group and generalisability of results, comparisons with different trials are difficult and remain speculative because of differences in selection criteria and imbalances in prognostic factors. Moreover, the potential effect of salvage therapy on overall survival is unknown in most reported series. The proportion of patients with a Karnofsky performance status of more than 70 or with metastatic disease in our study is consistent with those in previous phase III trials in pancreatic adenocarcinoma.1–11 By contrast, these trials included patients with a more favourable prognosis,2 such as those with stage I–II disease1–3,5,6,10 or patients older than 70 years,1–3,5,6,8–10 which could have positively biased the outcome, as could have a dropout of more than 20% of patients during the lead-in period.1 Nevertheless, the safety, dose-intensity, activity, and efficacy of our control group was consistent with other series,1–11 suggesting that such factors had a negligible effect on outcome. We chose 4-month progression-free survival as our primary endpoint. The European Agency for the Evaluation of Medical Products (www.emea.eu.int) and the US Food and Drug Administration29 classify progression-free survival as an appropriate endpoint for assessment of anticancer agents in human beings. http://oncology.thelancet.com Vol 6 June 2005
Articles
Unlike overall survival, progression-free survival is not affected by subsequent, or crossover, therapies or by supportive care, and needs smaller patient numbers to detect a difference between treatment groups than does assessment of overall survival. However, progression-free survival is associated with a risk of: assessment bias in unblinded trials; unequal time interval of progression assessment across treatment groups; asymmetrical censoring; and inappropriate crediting of an unknown amount of additional progression-free time in patients who die without documentation of progressive disease and who have the date of death recorded as the progression date. In our trial, the timing of tumour assessment was identical in the two groups, and a radiologist blinded to treatment group centrally reviewed CT scans to keep the effects of biases on progression-free survival to a minimum. Furthermore, about 10% of patients were not assessable for the primary endpoint because scans were not done in time to document progression, and the comparison between progressionfree-survival curves was repeated with these patients omitted, with similar findings (p=0·005, data not shown). Furthermore, the time between progression and death was much the same for both groups, suggesting that salvage therapy had a similar effect on outcome in both groups, and that progression-free survival might be an adequate prediction of overall survival. Previous trials have used 6-month progression-free survival as the primary endpoint, and on the basis of previously reported data for 6-month progression-free survival (52% for PEFG vs 22% for gemcitabine),16 if we had chosen this endpoint for our sample calculation, the number of patients would have been about the same (48 patients per group with =0·05 and =0·1, one-sided test). In conclusion, our findings suggest that a clinically relevant effect on the outcome for pancreatic cancer could be achieved with adequate chemotherapy. Because our trial has limitations in interpretation of secondary endpoints because of small patient numbers, a larger confirmatory trial may be needed before this combination regimen could be regarded as standard treatment. Nevertheless, our findings are important because PEFG had manageable toxic effects, did not negatively affect quality of life, and maintained a statistically and clinically relevant outcome advantage compared with standard treatment. Accordingly, PEFG might be a feasible and effective first-line treatment for patients with locally advanced or metastatic pancreatic adenocarcinoma. Contributors M Reni and L Galli were responsible for the study idea, study design, data analysis and interpretation, and drafting of the article. M Reni, S Cordio, C Milandri, P Passoni, E Bonetto, C Oliani, G Luppi, R Nicoletti, R Bordonaro, A Passardi, A Zerbi, G Balzano, L Aldrighetti, C Staudacher, E Villa, and V Di Carlo participated in data acquisition. All authors were involved in critical review of the article. M Reni had full access to all data in the study and had final responsibility for the decision to submit for publication.
http://oncology.thelancet.com Vol 6 June 2005
Conflict of interest We declare no conflicts of interest. Acknowledgments No funding was received for this study. References 1 Burris HA 3rd, Moore MJ, Andersen J, et al. Improvement in survival and clinical benefit with gemcitabine as first-line therapy for patients with advanced pancreas cancer: a randomized trial. J Clin Oncol 1997; 15: 2403–13. 2 Berlin JD, Catalano P, Thomas JP, et al. Phase III study of gemcitabine in combination with fluorouracil versus gemcitabine alone in patients with advanced pancreatic carcinoma: Eastern Cooperative Oncology Group Trial E2297. J Clin Oncol 2002; 20: 3270–75. 3 Moore MJ, Hamm J, Dancey J, et al. Comparison of gemcitabine versus the matrix metalloproteinase inhibitor BAY 12-9566 in patients with advanced or metastatic adenocarcinoma of the pancreas: a phase III trial of the National Cancer Institute of Canada Clinical Trials Group. J Clin Oncol 2003; 21: 3296–302. 4 Rocha Lima CMS, Green MR, Rotche R, et al. Irinotecan plus gemcitabine results in no survival advantage compared with gemcitabine monotherapy in patients with locally advanced or metastatic pancreatic cancer despite increased tumor response rate. J Clin Oncol 2004; 22: 3776–83. 5 Bramhall SR, Schultz J, Nemunaitis J, et al. A double-blind placebocontrolled, randomised study comparing gemcitabine and marimastat with gemcitabine and placebo as first line therapy in patients with advanced pancreatic cancer. Br J Cancer 2002; 87: 161–67. 6 Bramhall SR, Rosemurgy A, Brown PD, for the Marimastat Pancreatic Cancer Study Group. Marimastat as first-line therapy for patients with unresectable pancreatic cancer: a randomized trial. J Clin Oncol 2001; 19: 3477–355. 7 Heinemann V, Quietzsch D, Gieseler F, et al. A phase III trial comparing gemcitabine plus cisplatin vs gemcitabine alone in advanced pancreatic adenocarcinoma. Proc Am Soc Clin Oncol 2003; 22: abstr 1003. 8 Van Cutsem E, van de Velde H, Karasek P, et al. Phase III trial of gemcitabine plus tipifarnib compared with gemcitabine plus placebo in advanced pancreatic cancer. J Clin Oncol 2004; 22: 1430–38. 9 O’Reilly EM, Abou-Alfa GK, Letorneau R, et al. A randomized phase III trial of DX-8951f (exatecan mesylate) and gemcitabine vs gemcitabine alone in advanced pancreatic cancer. Proc Am Soc Clin Oncol 2004; 23: abstr 4006. 10 Richards DA, Kindler HL, Oettle H, et al. A randomized phase III study comparing gemcitabine + pemetrexed versus gemcitabine in patients with locally advanced and metastatic pancreas cancer. Proc Am Soc Clin Oncol 2004; 23: abstr 4007. 11 Louvet C, Labianca R, Hammel P, et al. GemOx (gemcitabine + oxaliplatin) versus Gem (gemcitabine) in non-resectable pancreatic adenocarcinoma: final results of the GERCOR/GISCAD Intergroup phase III. Proc Am Soc Clin Oncol 2004; 23: abstr 4008. 12 Maisey N, Chau I, Cunningham D, et al. Multicenter randomized phase III trial comparing protracted venous infusion (PVI) fluorouracil (5-FU) with PVI 5-FU plus mitomycin in inoperable pancreatic cancer. J Clin Oncol 2002; 20: 3130–36. 13 Ducreux M, Rougier P, Pignon JP, et al. A randomised trial comparing 5-FU with 5-FU plus cisplatin in advanced pancreatic carcinoma. Ann Oncol 2002; 13: 1185–91. 14 Tempero M, Plunkett W, Ruiz Van Haperen V, et al. Randomized phase II comparison of dose-intense gemcitabine: thirthy minute infusion and fixed dose rate infusion in patients with pancreatic adenocarcinoma. J Clin Oncol 2003; 21: 3402–08. 15 Moore MJ, Goldstein D, Hamm J, et al. Erlotinib improves survival when added to gemcitabine in patients with advanced pancreatic cancer. A phase III trial of the National Cancer Institute of Canada Clinical Trial Group (NCIC-CTG). Proceedings of the 2005 Gastrointestinal Cancers Symposium Jan 27–29, 2005; Hollywood, FL, USA, 3: abstr 77. 16 Reni M, Passoni P, Panucci MG, et al. Definitive results of phase II trial of cisplatin, epirubicin, continuous-infusion fluorouracil, and gemcitabine in stage IV pancreatic adenocarcinoma. J Clin Oncol 2001; 19: 2679–86.
375
Articles
17
18
19 20 21
22
23
376
Sobin LH, Wittekind C, eds. International Union Against Cancer. TNM classification of malignant tumours, 5th edn. New York: Wiley-Liss, 1997. Fishman B, Pasternak S, Wallenstein SL, et al. The Memorial Pain Assessment Card: a valid instrument for the evaluation of cancer pain. Cancer 1987; 60: 1151–58. Ajani JA, Welch SR, Raber MN, et al. Comprehensive criteria for assessing therapy-induced toxicity. Cancer Invest 1990; 8: 147–59. Miller AB, Hoogstraten B, Staquet M, Winkler A. Reporting results of cancer treatment. Cancer 1981; 47: 207–14. Fitzsimmons D, Johnson CD, George S, et al. Development of a disease specific quality of life questionnaire module to supplement the EORTC core cancer qol questionnaire, the QLQ-C30 in patients with pancreatic cancer. Eur J Cancer 1999; 35: 939–41. Fayers P, Aaronson N, Bjordal K, et al. EORTC QLQ scoring manual, 3rd edn. Brussels: European Organisation for Research and Treatment of Cancer, 2001. Osoba D, Rodrigues G, Myles J, et al. Interpreting the significance of changes in health-related quality-of-life scores. J Clin Oncol 1998; 16: 139–44.
24
25
26
27
28
29
Reni M, Bonetto E, Cordio S, et al. Quality of life outcome in advanced pancreatic adenocarcinoma: results from a phase III trial. Ann Oncol 2004; 15: abstr 911. Hoffman K, Glimelius B. Evaluation of clinical benefit of chemotherapy in patients with upper gastrointestinal cancer. Acta Oncol 1998; 37: 651–59. Reni M, Passoni P, Bonetto E, et al. Final results of a prospective trial of a PEFG (cisplatin, epirubicin, 5-fluorouracil, gemcitabine) regimen followed by radiotherapy after curative surgery for pancreatic adenocarcinoma. Oncology, in press. Passoni P, Reni M, Zanello A, et al. Preliminary results of a phase II trial of PEFG (cisplatin, epirubicin, 5-fluorouracil, and gemcitabine) in advanced gallbladder, biliary tract, and ampulla of Vater adenocarcinoma. Ann Oncol 2002; 13: abstr 724. Villa E, Reni M, Zanello A, et al. PEFG regimen (cisplatin, epirubicin, 5-fluorouracil continuous infusion, and gemcitabine) in advanced gastric adenocarcinoma. Ann Oncol 2003; 14: abstr D5. Johnson JR, Williams G, Pazdur R. End points and United States Food and Drug Administration approval of oncology drugs. J Clin Oncol 2003; 21: 1404–11.
http://oncology.thelancet.com Vol 6 June 2005