**3. Gemcitabine: a new era for pancreatic cancer**

One decade later, gemcitabine, a nucleoside analogue that exhibits antitumor activity, started to show promising results in PC patients.

Conducted as a phase 2 trial by former concepts, Casper et al. reported their experience of treating 44 PC patients, all of them with confirmed adenocarcinoma, not amenable to curative surgical treatment, performance status 0–1, and measurable disease (32 patients with metastatic disease and 12 patients with local advanced PC). The initial dose of gemcitabine was 800 mg/m2 weekly for 3 weeks and 1 week off. Thirty-five patients received two or more cycles of Gemcitabine. Median TTP was 16 weeks. Five patients (11%) achieved a major response according to the radiological criteria used in the trial, the median duration of response for those patients was 13 months but only 5.6 months for all the treated population. Reported oneyear survival was 23%. Most of the patients presented mild toxicity to gemcitabine including hematological, cutaneous toxicity, alopecia, nausea, vomiting and diarrhea. Some of these patients were treated with increasing doses of Gemcitabine, 1000 mg/m2 then 1250 mg/m<sup>2</sup> and 2 patients up to 1500 mg/m2 ; however, those last patients needed a reduction of their doses due to flu-like syndrome [4].

In a phase 3 randomized clinical trial Gemcitabine was compared head to head with 5 FU. This study included 126 patients from Canada and the United States with locally-advanced-unresectable or metastatic PC. Despite the known limited benefit of 5-FU in PC the investigators that designed the trial decided to use it as control arm instead of a placebo arm. The primary end point for this trial was "Clinical Benefit" (considering analgesic consumption, pain intensity, performance status and weight); secondary end points included response rate, time to disease progression and survival. Gemcitabine was given as a 1000 mg/m<sup>2</sup> intravenous regimen weekly for 7 weeks and 1 week of rest, followed by 1000 mg/m2 weekly during 3 weeks with 1 week of rest during the rest of the treatment. 5-FU was given weekly in a fixed dose of 600 mg/m2 intravenously. Primary end point was met, and this resulted in a positive trial achieving a clinical benefit of 23.8% in the gemcitabine arm and only 4.8% in the 5 FU arm. Median OS and TTP were 5.6 months and 9 weeks for gemcitabine arm versus 4.4 months and 4 weeks for 5FU arm, reported 12-months survival was 18 and 2% for gemcitabine and 5FU respectively (p = 0.0025). Among gemcitabine treated patients only a 5.4% achieved a radiological response but none from the 5 FU arm did, 39 and 19% of SD was reported for both arms. When compared responders versus nonresponder median OS was 10.7 months versus 4.8 months regardless of the treatment arm. Grade 3–4 neutropenia was higher in the gemcitabine arm (25.9% versus 4.9% for 5-FU, p < 0.001), No severe infections were reported in either arms but grade 3 and 4 anemia was 9.7% in gemcitabine arm and 0% in 5 FU arm, grade 1–2 fever was higher in gemcitabine arm 30.1% versus 16% for 5-FU, rashes 23.8 versus 12.9%, grade 3–4 nausea vomiting 9.5 and 3.2% versus 4.8 and 0% respectively. For both arms there were no survivors after 19 months since starting treatment [10]. This trial led to FDA to approve Gemcitabine as a first line of treatment for unresectable or metastatic pancreatic cancer in 1996.

Due to the positive results of gemcitabine in patients with advanced PC many combinations of gemcitabine-based treatments were tested looking to improve OS and progression free survival (PFS), most of them did not show improvement in overall survival as shown in HU's meta-analysis [11]: gemcitabine—5-FU combination (conducted by J. D. Berlin); gemcitabine—irinotecan combination (trials conducted by G. Rocha Lima and by G.P. Stathopoulos); gemcitabine—oxaliplatin (trials conducted by C. Louvet and by E. Poplin); gemcitabine—pemetrexed

(conducted by H Oettle); gemcitabine—exatecan (conducted by G. Abou-Alfa); gemcitabine—cisplatin (trials conducted by V. Heinemann and by G. Colucci); gemcitabine—capecitabine (conducted by D. Cunningham); gemcitabine—bevacizumab (conducted by H. Kindler); gemcitabine—cetuximab (conducted by P. Philip); gemcitabine—axitinib (conducted by H. Kindler). Gemcitabine—sorafenib combination also resulted in negative trials in terms of overall survival [12].

Despite those several negative clinical trials, a meta-analysis conducted by Ciliberto aimed to evaluate the role of gemcitabine-based combination therapy when compared with gemcitabine alone. Including more than 10.600 patients from 34 randomized trials, the combination treatments showed marginal superiority in terms of survival, overall response, and disease control rate but with higher toxicity rates mainly diarrhea, nausea, neutropenia, thrombocytopenia. One of the interpretations of the authors was that combination regimens gemcitabine-based should be reserved only for well selected patient populations [13]. In an Asiatic population study, a 3-arms clinical trial was conducted to compare gemcitabine alone, gemcitabine plus S-1 combination or S-1 alone in local advanced and in metastatic pancreatic patients. The combination arm did not show superiority when compared with gemcitabine alone, however, S-1 showed noninferiority against gemcitabine with a good tolerability profile [14].

From PC biopsies Fjallskog et al. found and reported that 55% of tumor samples studied stained positive for Epidermal Growth Factor Receptor (EGFR) [15]. In a murine model of pancreatic adenocarcinoma adding erlotinib highly inhibited gemcitabine-induced MAP kinase signaling regardless of the activation of KRAS by maintaining high levels of ERBB2 protein [16]. A multicenter phase 3 double blind international trial assessed gemcitabine plus erlotinib combination versus gemcitabine plus placebo [5]. About 569 patients with unresectable or metastatic adenocarcinoma of the pancreas, ECOG 0–2 were randomized in a 1:1 ratio to receive either gemcitabine alone or gemcitabine plus erlotinib. Gemcitabine was given intravenously 1000 mg/m<sup>2</sup> weekly for 7 weeks and 1 week off, then 1000 mg/ m2 weekly for 3 weeks and 1 week off during the next cycles (28 days cycle). Erlotinib was orally given in a 100 mg dose and increased to 150 mg in a Canadian cohort. The primary end point was OS, secondary end points included PFS, ORR, duration of response, correlation of EGFR expression with outcomes, quality of life and toxicity. Reported median survival and one-year survival were 6.24 months and 23% for the gemcitabine—erlotinib arm versus 5.9 months and 17% for the gemcitabine-placebo arm. PFS was improved in the combination arm. Despite these positive results in statistical terms, the clinical value was marginal and the reported toxicity significantly higher in the combination arm, including 6 deaths protocolrelated all of them in patients in the gemcitabine-erlotinib arm including interstitial pneumonitis, sepsis, stroke and neutropenic sepsis. Immunohistochemical analysis and correlation with response did not show any improvement among EGFR positive patients. Interestingly, patients that developed skin rash grade 2 or higher lived longer when compared with whom did not (10.5 months for grade 2 versus 5.8 months and 5.3 months for grade 1 and 0 respectively), 1-year survival was 16% for rash grade, 0, 9% for rash grade 1 and 43% for rash grade 2 or higher (p < 0.001).

A German trial that compared erlotinib in combination with either capecitabine or gemcitabine as the front-line treatment for advanced pancreatic cancer patients allowing cross over after failure showed a low toxicity rate for both arms and not deaths treatment-related [17].

Despite that gemcitabine-erlotinib combination got FDA approval for metastatic or unresectable PC, considering its minimal benefit in terms of survival when compared with gemcitabine alone and also due to the higher toxicity profile of the combination, it is not considered as a "first option of treatment" in advanced pancreatic adenocarcinoma patient by different authors and international guidelines [18, 19].
