Background
Pancreatic ductal adenocarcinoma (PDAC), commonly known as pancreatic cancer, is the 10
th most common cancer type with an incidence of 10/100,000 but highly lethal (> 95%) and this is reflected by the fact that it is ranking as the 5
th most lethal cancer in absolute patient numbers after lung, colorectal, breast and prostate cancer [
1,
2]. Due to the declines in lethality in other major cancers, pancreatic cancer is predicted to become the fourth cause of cancer death in Europe [
2]. Dramatic progress was made during the past years to better understand the biology of this disease (reviewed in [
3]). Only 10-20% of the patients have resectable tumours at diagnosis and resection is a prerequisite for cure but even with adjuvant therapy median overall survival of resected patients is still as low as 20% after 5 years in randomised phase III studies (reviewed in [
4]). The large majority (> 80%) of patients with non-resectable disease at diagnosis can be subdivided into metastatic and locally advanced PDAC (LAPC) with both stages being about equally frequent. Compared with metastatic disease patients with LAPC have a better prognosis and - though often grouped together with metastatic disease not separated in randomised phase III trials - patients with LAPC should be separated from patients with metastatic disease.
Chemotherapy is an essential element in the treatment of LAPC to fight the high tendency of distant spread. But the combination of systemic with local treatment prolonged survival in a number of recent studies [
5,
6] compared with systemic therapy only. Of note, secondary resection after CRT was reported in a systematic review and meta-analysis in 1/3 of the patients leading to a median overall survival (mOS) rate of 20.5 months which is equally good as after primary resection [
7] and downstaging was also described [
8]. On the other hand, the inferiority of chemoradiotherapy (CRT) vs chemotherapy in a recent French trial [
9] can most likely be attributed to inadequate technique and quality of chemoradiotherapy highlighting the complexities of CRT for PDAC [
10]. Of note, 60 Gy were delivered in 2 Gy fractions to both the primary tumour and the elective lymphatics resulting in large planning target volumes (PTV) as 2 cm expansion margins were used from the clinical target volumes. Also, the FFCD-SFRO trial [
9] is the only randomised phase III CRT trial using 5-fluorouracil (5-FU)/Cisplatin as concurrent chemotherapeutic agents and this resulted in a very high rate of grade 3/4 toxicity for the adjuvant chemotherapy and prevented maintenance chemotherapy. Commonly, the combination of a fluoropyrimidine with radiotherapy is regarded to be the standard of care for CRT [
4] but a substantial number of gemcitabine based CRT trials was reported with encouraging results such as in the ECOG-4201 trial [
6]. The latter trial used IMRT together with 600 mg/m
2 gemcitabine weekly, a relatively high dose, resulting in a high rate of grade 3/4 toxicity.
The rationale for preferring gemcitabine over 5-FU in CRT regimens is its hypothesised superiority both, locally and systemically: in metastatic disease gemcitabine was able to prolong survival and to lead to higher clinical benefit compared to 5-FU [
11]. For the local effect when used with radiotherapy, gemcitabine is predicted to lead to higher tumour cytotoxicity than 5-FU because it is one of the most potent radiosensitising chemotherapeutic agents [
12]. Gemcitabine is an S-phase specific deoxycytidine analogue. It acts via competitive incorporation of dFdCTP and dCTP into DNA and results in DNA fragmentation and subsequent cell death. Furthermore, gemcitabine interferes with ribonucleotide reductase which is thought to have an impact on cell death by affecting DNA repair. Also, specific single-nucleotide polymorphisms in DNA the repair damage genes ATM, Chek1 and ATR were found to be significantly associated with OS after gemcitabine CRT especially when analysed for the combined effect of all three genes [
13]. In line with these observations, gemcitabine containing schedules were described to achieve a higher rate of pathologic response compared to 5-FU based protocols [
14]. The combination of gemcitabine with 5-FU or capecitabine which is commonly used as a chemotherapy combination was found to be too toxic for CRT in LAPC especially in terms of elevated gastrointestinal toxicity [
15]. Therefore we decided a different chemotherapeutic combination, gemcitabine and cisplatin, which had been investigated both preclinically and clinically: the synergism between the two drugs is attributed mainly to an increase in platinum-DNA adduct formation which is possibly related to changes in DNA due to dFdC incorporation into the DNA [
16‐
18]. The combination of the two drugs is clinically in use mainly in ovarian, non-small cell lung and pancreatic cancer and has been more effective than gemcitabine only in metastatic and locally advanced PDAC in the group of patients with good performance status [
19].
Despite of this rationale, gemcitabine initially was difficult to be combined with radiotherapy due to its acute toxicity profile depending profoundly on the absolute radiotherapy treatment volume [
18,
20]. This potential dangerous effect can now be more easily counter-balanced with highly conformal treatment planning and the use of IMRT/IGRT thereby increasing the tolerance of gemcitabine based CRT [
21]. In this analysis we compare the outcome and the toxicity of two CRT regimens in 93 patients with LAPC treated at our centre: One regimen was 5-FU/Mitomycin C (FM), the other gemcitabine/cisplatin (GC) given concurrently with radiotherapy. These two regimens have not been compared in the literature up to now but they both have been used in a number of trials in PDAC and other upper GI tumours [
22‐
25]. We report superior OS of the GC regimen with comparable high grade toxicity (grade 4 haematologic and grade 3/4 non-haematologic disease).
Discussion
Currently it is not clear which type of concurrent chemotherapy is best when combined with radiotherapy. While the standard of care for CRT is to combine a fluoropyrimidine with radiotherapy (5-fluorouracil or more recently also capecitabine) [
10], there is a tendency to use more and more gemcitabine based chemoradiotherapy. The combination of 5-FU and mitomycin C that we have used in this report was previously employed for CRT in a number of trials in PDAC [
22,
23,
32‐
34]. Mitomycin C was hypothesised to be useful in addition to 5-fluorouracil because of its predominant effectiveness in hypoxic conditions [
35] since severe hypoxia was shown to be present in pancreatic tumours [
36]. However, after the publication of a randomised phase III trial by Burris et al. [
11] showing the superiority of gemcitabine compared with fluorouracil for the treatment of patients with advanced pancreatic cancer, much effort has been made to combine gemcitabine-based regimens concurrently with radiotherapy. This is reflected by the fact that during the last decade a total number of 36 clinical trials using gemcitabine and chemoradiotherapy have been published in PubMed, the majority of them after 2005. The combination of gemcitabine and cisplatin concurrently with radiotherapy which we have used in this analysis was also tested in a number of CRT trials [
37‐
40] and chemotherapy trials [
41]. Preclinical studies have suggested a synergistic interaction between gemcitabine cisplatin being the result of gemcitabine incorporation into DNA and an increase of platinum -DNA adduct formation [
16,
17]. Therefore, this study reports the use of two chemotherapeutic regimens based on biological hypotheses.
One of the strengths of this retrospective comparison is the homogeneity of the treatment variables and of the selection process of the patients for definitive chemoradiotherapy within one single centre in a large number of patients. Both, median overall survival time (12.7 versus 9.7 months) and 12-month overall survival rate (53% versus 40%) were statistically significantly longer in the patient cohort treated with the GC regimen compared to the FM regimen. Comparing our survival results with other trials which have investigated the use of 5-fluorouracil versus gemcitabine chemoradiotherapy, Crane et al. showed a trend favouring gemcitabine (53 vs 61 patients) based CRT [
42] and Li et al. reported a statistically significant survival advantage for patients treated with gemcitabine concurrently to radiotherapy over those treated with 5-fluorouracil (16 vs 18 patients) [
43]. In contrast, no advantage of gemcitabine over 5-fluorouracil CRT was detected in two other trials. However, the trial reported by Brasuniene et al. was very small (10 vs 9 patients per arm) and therefore was substantially underpowered to be able to detect any difference [
44]. The second negative trial, reported by Wilkowski et al. [
37] compared 3 arms, 5-fluorouracil (30 patients), GC (31 patients) and GC chemoradiotherapy followed by GC chemotherapy (27 patients). As this trial is the only one using the GC combination as the here reported trial, it is worth to compare the two trials in more detail. The median overall survival rate for the GC arm and the arm with GC CRT followed by GC chemotherapy was 9.3 and 7.3 months, respectively in this study whereas we observed a median overall survival rate of 12.7 months. A hypothetical explanation for this difference is a higher total radiation dose to the primary tumour in our trial (55.8 Gy versus 50 Gy). The patient characteristics between the two studies were similar. However, it needs to be stressed that our analysis is retrospective in nature and therefore we cannot exclude factors such as selection bias or other inhomogeneities. We have tested known factors influencing survival as good as possible and these included TNM staging and performance status and did not observe any significant differences. Just very recently a meta-analysis on the use of gemcitabine based chemoradiotherapy compared to 5-FU including 229 patients from randomised controlled trials was published [
25]. This analysis described a survival advantage of gemcitabine based chemoradiotherapy compared to 5-Fu based for 12 month overall survival rates (RR 1.54, 95% CI 1.05 - 2.26, p = 0.03).
The toxicity analysis of the two regimens showed that the GC regimen led to a higher number of haematologic grade 3 toxicities, but interestingly not of grade 4 toxicities. Nausea and vomiting were the most frequent higher grade non-haematologic toxicities in both groups. Surprisingly, grade 3 nausea and vomiting were more frequent in the FM regimen despite of the emetogenic effect of cisplatin in the GC regimen. This might be attributable to the fact that antiemetic therapy has improved over time and FM being chronologically the first regimen used in this cohort. Comparing haematologic toxicity of the GC regimen in our trial with that reported by Wilkowski combined grade 3/4 leukocytopenia was comparable (48% versus 52%), grade 3/4 thrombocytopenia was less frequent in our trial (36% versus 52%) and grade 3/4 nausea was comparable (11% versus 13%). The addition of mitomycin C to 5-fluorouracil in our trial led to significant differences in haematologic grade 3/4 toxicity when comparing it with the Munich trial (thrombocytopenia: 20% versus 4%; leukocytopenia: 37% versus 4%) and nausea (20% vs 0%). The comparison of our study with the FFCD-SFRO [
9] and with the ECOG [
6] trials shows that our GC protocol resulted in lower GI toxicity compared to the ECOG regimen but more neutropenia which we attributed to the addition of cisplatin to gemcitabine. Compared with the FFCD trial our FM regimen led to a lower rate of non-haematological toxicity but a higher rate of thrombocytopenia which we attributed to the use of mitomycin C. The above mentioned recently published meta-analysis found significant differences of leukocytopenia, thrombocytopenia and gastrointestinal bleeding being more frequent in the gemcitabine group [
25]. This might be due to suboptimal radiation techniques used in the trials with the majority of the patients being treated about a decade ago. At that time the toxicity-volume relationship of gemcitabine chemoradiotherapy was not yet described as well as now and IMRT was not yet as commonly used as it is now.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
TBB: Performed the retrospective analysis. TBB, RS, RF: developed the chemotherapeutic protocols. All the listed authors have been involved in drafting or in revising the manuscript. All authors read and approved the final manuscript.