Is concomitant radiotherapy necessary with gemcitabine-based chemotherapy in pancreatic cancer?

Abstract

AIM: To evaluate the efficacy and safety of gemcitabine (GEM) plus radiotherapy compared with GEM alone for pancreatic cancer (PC).

METHODS: A systematic search for eligible studies comparing gemcitabine plus radiotherapy with gemcitabine alone for PC was performed using MEDLINE, EMBASE, and the Cochrane Library. A quality assessment was performed in each study. Meta-analyses were performed to study the pooled effects of relative risk with 95% confidence interval (CI).

RESULTS: A total of 336 participants from four original studies were included. Gemcitabine plus radiotherapy resulted in comparable overall survival (HR = 0.84, 95%CI: 0.53-1.34, P = 0.48) and progress free survival (HR = 0.99, 95%CI: 0.97-1.01, P = 0.36) to gemcitabine alone. Moreover, concomitant radiotherapy was associated with a significantly higher incidence of severe (grade 3 or greater) toxicities, mainly anemia, leukocytopenia, thrombocytopenia, anorexia, nausea/vomiting, and asthenia/fatigue.

CONCLUSION: Radiotherapy is not beneficial with gemcitabine-based chemotherapy for PC. Further exploration for better radiotherapy approaches and therapeutic regimens for the treatment of PC is warranted.

Key Words: Pancreatic cancer; Radiotherapy; Chemotherapy; Gemcitabine; Meta-analysis

Core tip: We performed this meta-analysis to evaluate the efficacy and safety of gemcitabine plus radiotherapy compared with gemcitabine alone for pancreatic cancer (PC). A total of 336 participants from four original studies were included. Gemcitabine plus radiotherapy resulted in comparable survival results to gemcitabine alone. Moreover, concomitant radiotherapy was associated with a significantly higher incidence of severe (grade 3 or greater) toxicities, mainly anemia, leukocytopenia, thrombocytopenia, anorexia, nausea/vomiting, and asthenia/fatigue. Therefore, radiotherapy is not beneficial with gemcitabine-based chemotherapy for PC. Further exploration for better radiotherapy approaches and therapeutic regimens in the treatment of PC is still warranted.

INTRODUCTION

Pancreatic cancer (PC) is a highly malignant disease and the fourth cause of cancer-related death worldwide[1,2]. Radical surgery is regarded as the only possible cure. A high rate of locoregional as well as distant recurrence has been observed after surgical resection with curative intent in PC[3]. Therefore, adjuvant therapies, including chemotherapy (CT), radiotherapy (RT), and chemoradiotherapy (CRT), have been administered to improve the poor outcomes of PC[4].

A growing body of evidence establishes a central role for systemic CT in the treatment of PC. Appropriate CT might result in decreased recurrence and metastasis rates to improve the survival rate of patients[5]. Formerly, 5-fluorouracil (5-FU) was considered an effective CT agent for PC[6]. However, gemcitabine (GEM) has gained increasing acceptance since it was introduced to treat PC in 1997[7]. Numerous clinical trials and meta-analyses have demonstrated that GEM, compared with 5-FU, increases survival and clinical benefits in locally advanced or surgically resected PC[8,9]. Currently, GEM-based CT is the standard of care for PC.

In addition to CT, RT and CRT are regarded as acceptable treatment options. RT could palliate common symptoms through the intensification of local control and thus prolong survival for patients with locally advanced PC[10]. However, this conclusion was significantly challenged by studies showing that favorable outcomes could be achieved without radiation[11]. Considering the conflicting opinions, we conducted a meta-analysis to compare the efficacy and safety of RT in the management of PC, to provide objective information that might help guide clinicians who treat PC.

MATERIALS AND METHODS

Trial selection

Published and unpublished trials fulfilling the following selection criteria were included in the present meta-analysis: (1) study design - randomized controlled trials (RCTs) or other comparative studies; (2) population - patients with PC, at the resectable or locally advanced stage; and (3) intervention - comparing GEM-based chemotherapy with GEM-based chemotherapy plus radiotherapy. To minimize publication bias, we did not limit the publication language. The primary outcome measure of the study was overall survival (OS), which should include clear survival data.

Literature search

Databases, i.e., MEDLINE, EMBASE and the Cochrane Library, were searched to identify pertinent citations published since January 1997. The following searching strategy was employed: [radiotherapy (Text Word) OR chemoradiotherapy (Text Word)] AND [pancreatic cancer (Title) OR pancreatic neoplasm (Title) OR pancreatic carcinoma (Title) OR pancreatic adenocarcinoma (Title) and gemcitabine (Text Word)]. For unpublished data, trial registries, including clinicaltrial.gov, the national research register, and current controlled trials, were searched. Additionally, a manual search was performed for potential related articles. The manufacturers of relevant pharmaceutical agents were contacted for additional information.

Data extraction

The data extraction and analysis were conducted independently by two reviewers (Zhang X and Huang H) using a pre-designed form. If results from more than one follow-up analysis were published in one study, only those with the longest follow-up were included. For results not clearly described in the paper, the author was contacted for the necessary data. Disagreements were resolved by discussion.

Quality assessment

Two reviewers (Zhang X and Huang H) used standard criteria (e.g., allocation concealment, blinding, intention-to-treat analysis, and withdrawals) to assess the methodological quality of RCTs, in addition to the quantitative quality assessment using the scoring system developed by Jadad[12]. The quality scoring system was as follows: (1) allocation concealment, coded as adequate (1 score), or inadequate or unclear (0 score); (2) blinding, coded as double blind (2 scores), single blind (1 score), or open label (0 score); (3) intention-to-treat analysis, coded as used (1 score), or not used or unable to assess (0 score); and (4) withdrawals, coded as given (1 score) and not given (0 score).

Statistical analysis

The log hazard ratios (HR) and their variances for the time-to-event data were estimated using published methods, and when appropriate, summary statistics or Kaplan-Meier curves were reported[13]. An estimate of the relative risk (RR) was used for the dichotomous outcomes. The results were reported with 95% confidence interval (CI) on the test for overall effect, and heterogeneity was quantified using a χ² test with a P value < 0.1 considered statistically significant. For outcomes without heterogeneity, the pooled effects were calculated through the fixed effect model; the randomized effect model was employed if heterogeneity was detected among the studies. The funnel plot was applied to estimate the potential publication bias, with an asymmetric plot indicating a possible publication bias. Review Manager (Version 5.0. Copenhagen: The Nordic Cochrane Center, The Cochrane Collaboration) was used for the statistical analyses mentioned above, whereas Stata Statistical Software (Release 10, StataCorp LP, College Station, TX, United States) was used to perform the publication bias test as described previously[14].

RESULTS

Descriptions of studies

After an initial search in the databases, a total of 1062 potential studies were identified; 186 papers were retrieved for full-text review after excluding 887 articles on the basis of titles and abstracts. The data from one trial were published twice in different languages[15,16], and only the English version was included, which was labeled as Chauffert 2008[16]. In total, four studies including 336 participants were included in the meta-analysis[16-19], of which three were RCTs[16,18,19] and one was a retrospective follow-up study[17]. The specific flowchart identifying the qualified studies is shown in Figure 1. The demographics of the studies are shown in Table 1. The quality assessments of the studies are shown in Table 2.

Table 1 Patient demographics and characteristics of the included studies.

Study	Groups	Chemo regimens	RT regimens	No. of patients	Male	Age (yr)	MS (M)	Follow-up (M)	Clinical stage
Chauffert et al[16]	GEM-CRT	5-FU infusion, 300 mg /m2 per day, days 1-5 for 6 wk; cisplatin, 20 mg /m2 per day, days 1-5 during wk 1 and 5. Then, GEM 1000 mg/m2 weekly, 3/4 wk	Induction: 60 Gy, 2 Gy/fraction	59	31	60 (41-79)	8.6	31	Locally advanced unresectable PC
	GEM alone	GEM 1000 mg/m2 weekly for 7 wk. Then, GEM 1000 mg/m2 weekly, 3/4 wk	None	60	34	62 (38-80)	13	33	Locally advanced PC
Loehrer et al[18]	GEM+RT	600 mg/m2 per week for wk 1 to 5, then 4 wk later, 1000 mg/m2 for 3 of 4 wk	Starting on day 1, 1.8 Gy/Fx for a total of 50.4 Gy	34	19	66 (46.9-83.5)	11.1	NG
	GEM alone	1000 mg/m2 per week for wk 1 to 6, followed by 1 wk rest, then for 3 of 4 wk	None	37	18	69 (49.7-83.7)	9.2	NG
Van Laethem et al[19]	GEM-CRT	2 cycles of GEM (1000 mg/m2 on days 1, 8, 15, followed by 1 wk rest-a cycle), then GEM 300 mg/m2 once per wk for 5 to 6 wk	50.4 Gy in 28 fractions, 1.8 Gy per fraction for 5 to 6 wk	45	24	61 (44-75)	24.3	30.7	Curatively resected PC
	GEM alone	4 cycles of GEM (1000 mg/m2 on days 1, 8, 15, followed by 1 wk rest-a cycle)	None	45	27	58 (32-77)	24.4	33.3
Wang et al[17]	GEM-RT	2 cycles of GEM (1000 mg/m2 on radiation 1, 8, 15, 21-d a cycle)	50.4 Gy in 28 fractions on Monday through Friday over 5.5 wk	35	17	61 (44-71)	13	8-24	Locally advanced PC
	GEM alone	4-6 cycles of GEM (1000 mg/m2 on days 1, 8, and 15, 28 d a cycle)	None	21	13	61 (39-71)	8	8-24

GEM: Gemcitabine; RT: Radiotherapy; CRT: Chemoradiotherapy; PC: Pancreatic cancer; MS: Median survival; NG: Not given.

Figure 1 Flow of studies through the review process.

Table 2 Quality assessment of the studies included.

Study	AC	Blinding	ITT	Withdrawals	Score
Chauffert et al[16]	Unclear	Open-label	No	4.0%	1
Loehrer et al[18]	Unclear	Open-label	Yes	21.0%	2
Van Laethem et al[19]	Adequate	Open-label	Yes	0.8%	4
Wang et al[17]	Unclear	Open-label	No	NG	0

AC: Allocation concealment; ITT: Intention-to-treat; NG: Not given.

Meta-analysis

OS was reported in the four studies. The pooled analysis revealed that there was no statistically significant difference in OS between treatment with GEM plus RT and GEM alone (four studies, 336 participants, HR = 0.84, 95%CI: 0.53-1.34, P = 0.48, Figure 2). However, high heterogeneity was detected (I² = 78%).

Figure 2 Forest plot to show the pooled estimates of overall survival.

Information on progress free survival (PFS) or disease free survival was available in three publications. Because the definitions of the terms are similar[20], the results were summarized as PFS. Additionally, concurrent RT did not show obvious superiority to GEM alone in improving PFS (three studies, 280 participants, HR = 0.99, 95%CI: 0.97-1.01, P = 0.36, Figure 3) under homogeneity conditions.

Figure 3 Forest plot to show the pooled estimates of progression free survival.

The data regarding treatment-related toxicity were available in three RCTs. For the subsequent analysis, only severe (grade 3 or greater) toxicity (National Cancer Institute Common Toxicity Criteria; ctep.cancer.gov) was considered. The meta-analysis demonstrates that concomitant RT was associated with a significantly higher incidence of adverse events. Particularly, the incidence of anemia, leukocytopenia, thrombocytopenia, anorexia, nausea/vomiting, and asthenia/fatigue occurred more frequently in GEM plus radiation treatment than that in GEM treatment alone (Table 3). The heterogeneity for the evaluation of all toxicity variables was absent or acceptable except for neutropenia.

Table 3 Meta-analysis of adverse events.

Outcome	Studies	Participants	Statistical method	Effect size	95%CI	P	I2(%)
Anemia	3	240	RR, Fixed	5.32	(2.23-12.66)	0.0002	0
Leukocytopenia	3	240	RR, Random	2.31	(1.08-4.97)	0.03	49
Neutropenia	3	240	RR, Random	1.37	(0.60-3.16)	0.46	78
Thrombocytopenia	3	240	RR, Fixed	2.34	(1.04-5.29)	0.04	0
Anorexia	2	154	RR, Fixed	5.74	(1.05-31.39)	0.04	0
Diarrhea	3	240	RR, Fixed	2.58	(0.51-13.01)	0.25	0
Nausea/vomiting	3	240	RR, Fixed	3.58	(1.32-9.69)	0.01	0
SGPT abnormality	2	154	RR, Fixed	0.99	(0.34-2.88)	0.98	0
Asthenia/Fatigue	2	154	RR, Fixed	3.54	(1.23-10.18)	0.02	29
Weight Loss	2	154	RR, Fixed	4.02	(0.46-35.04)	0.21	0

RR: Relative risk.

Sensitivity analysis

To explain the heterogeneity among the included studies, a sensitivity analysis was performed first by excluding the non-RCTs. Notably, similar OS results were observed (three studies, 280 participants, HR = 1.06, 95%CI: 0.78-1.43, P = 0.72, Figure 4), and a marked reduction in heterogeneity was observed (I² = 33%), suggesting that the publication type contributes to the heterogeneity.

Figure 4 Similar overall survival results.

Another sensitivity analysis was conducted by including only the participants with locally advanced PC. Consistently, no significant difference in OS was observed between the GEM plus RT group and the GEM alone group (three studies, 246 participants, HR = 0.76, 95%CI: 0.41-1.40, P = 0.38, Figure 5), with high heterogeneity remaining (I² = 83%). These results indicate that although data derived from patients at different clinical stages were pooled, this type of data management might not be a definite cause of heterogeneity in this meta-analysis.

Figure 5 No significant difference in overall survival between the gemcitabine plus radiotherapy group and the gemcitabine alone group.

Publication bias

The funnel plot of the included studies for the OS analysis was largely symmetric (Figure 6; Begg’s test: P = 0.308, Egger’s test: P = 0.328), indicating the absence of publication bias. Because there were less than three studies included in the analysis of other variables, a test for publication bias was not performed.

Figure 6 Funnel plot of the included studies for the overall survival analysis is largely symmetric.

DISCUSSION

This systematic review provides a comprehensive overview of current evidence on the efficacy and safety of RT in combination with GEM-based CT for PC, showing that additional RT does not provide a significant increase in survival but rather results in an increased incidence of adverse events. Therefore, the use of RT is not recommended in PC patients in the setting of GEM-based CT.

Although surgery is central in a potentially curable case, a multidisciplinary team is typically required to guarantee the most coordinated treatment for PC[21]. Compared with other malignancies, the metastasis of PC occurs at a very early stage. Even in patients with apparently localized tumors, distant micrometastases most likely already exist[22,23]. Therefore, systemic CT has always been recognized as an essential therapeutic option to maximize survival[24]. In the last decade, there has been great progress in the treatment of PC following the advent of GEM, which is used as the first-line CT agent[4].

Advances in the field of radiation oncology, particularly those in radiation treatment delivery techniques, have led to improvements in local tumor control and a reduction in radiation toxicity[25,26]. Some clinicians hypothesize that although PC is dominated by distant metastatic spread, the importance of local progression should not be completely ignored because local recurrence in PC is very common, frequently with potential obstructions[27]. Controlling local progression in PC could decrease symptoms, whereas failure to control the primary tumor is associated with symptoms such as pain, gastric outlet and duodenal obstruction[28]. Therefore, RT to the postoperative bed and draining lymph nodes is frequently employed to ensure better locoregional control, and the data from clinical studies supporting the use of RT for PC were documented[29,30].

There have been several opposite opinions with regard to the contribution of RT to the improvement of survival. In a retrospective study, locally advanced PC patients receiving 5-FU-based CRT showed similar OS compared with those receiving GEM-based CT alone[31]. In a group of surgically resected PC patients, adjuvant 5-FU-based CT resulted in a significant survival benefit, whereas additional RT appeared to have a deleterious effect on survival[32]. In a previous meta-analysis by Ren et al[33], they concluded that RT appeared to provide benefits in terms of locoregional control; however, RT may not improve the OS for patients with resectable PC. In this meta-analysis, a benefit of RT for OS or PFS was not observed from the pooled results, which was consistent with the results from two sensitivity analyses, indicating that our primary findings were reliable and stable.

The related toxicities of therapeutic strategies in clinical practice are markedly important. The data from this meta-analysis indicate that concurrent RT resulted in a significant increase in toxic events compared to those associated with GEM-based CT alone. This finding could be attributed to the low sensitivity of PC cells to radiation and the narrow therapeutic window with GEM-based RT[34,35]. In a retrospective study in Japan, GEM-based CT showed comparable survival to CRT, with significantly less toxicity[36]. Recently, Takatori et al[37] published their results from an observational study in which it was shown that 49.4% of PC patients treated with GEM-concurrent proton RT developed radiation-induced ulcers in the stomach and duodenum that were detected by endoscopic examination. Moreover, given that endoscopic examination was only performed when indicated, the incidence of gastroduodenal ulcers caused by radiation might have been greatly underestimated in many clinical studies. Therefore, an unsatisfactory toxic profile might serve as the dominant barrier to the successful use of RT for PC.

Given the systemic nature of PC at diagnosis, the marked effect of radiation on quality of life (QOL) and the relatively low possibility of long-term survival, the decisions concerning therapeutic options should be entirely balanced. Simply highlighting small differences in survival without adjusting for QOL effects might lead to inappropriate therapeutic recommendations[38]. Among the four included original studies, QOL data are available in one study[18] in which the authors showed that no significant differences were observed between GEM alone and GEM plus RT treatment. However, considering that a small number of patients were included, future studies with larger samples of patients are needed to further confirm these results.

Another consideration of selecting a treatment option pertains to the medical cost, and the technology innovations in radiation oncology naturally lead to the question of cost-effectiveness[39]. Particularly when considering comparable efficacy, medical expense might be a deciding factor in the selection of a therapeutic regimen[40]. Murphy et al[41] evaluated the cost-effectiveness of modern RT techniques using a Markov decision-analytic model. The results demonstrated that RT had an incremental cost-effectiveness ratio of $126800 per quality-adjusted life years (QALY) compared to that of GEM alone, and the probability of cost-effectiveness at a willingness to pay of $50000 per QALY was much higher for GEM alone than that for any form of RT. Therefore, it is unlikely that RT would be cost-effective in diseases, such as PC, that have markedly limited survival, and this cost analysis might be an additional argument for not using RT.

Some limitations remain in this meta-analysis. First, the sample size of the included studies was fairly small, which impairs the interpretative power of the results. Second, this meta-analysis was performed independently of surgical considerations of resectability, and the conclusions might be further limited because surgery had a major effect on PC complications and survival. Third, the RT techniques used in the included trials might be outdated, and new radiation modalities, such as particle therapy, are likely to be safer and more effective than older radiation techniques. Thus, the results from this study should not be considered a total negation of RT for PC. Finally, the issue of therapy for PC is too complicated to be addressed merely by a meta-analysis; therefore, individualized treatment should not be ignored. These conclusions should be adopted with caution.

The findings from this meta-analysis indicate that the application of RT is not beneficial with GEM-based CT for PC. Further exploration for better RT approaches and therapeutic regimens for the treatment of PC is warranted.

COMMENTS

Background

Pancreatic cancer (PC) is a highly malignant disease and the fourth cause of cancer-related death worldwide. Radical surgery is regarded as the only possible cure.

Innovations and breakthroughs

The authors conducted a meta-analysis to compare the efficacy and safety of gemcitabine (GEM) plus radiotherapy compared with GEM alone in the management of PC, to provide objective information that might help guide clinicians who treat PC.

Peer review

This meta-analysis addresses the clinically important issue, whether treatment of advanced PC by gemcitabine chemotherapy alone or in combination with radiotherapy results in different outcomes, especially overall survival, progression free survival, treatment related toxicity and quality of life.