Postoperative Adjuvant Treatment Strategy for Locally Advanced Rectal Cancer after Neoadjuvant Treatment

Background Neoadjuvant (chemo) radiotherapy is used as a standard treatment for locally advanced rectal cancer (LARC), but there is no general consensus on either the efficacy of postoperative adjuvant chemotherapy in patients with LARC after neoadjuvant treatment and surgery, or whether the addition of oxaliplatin to adjuvant chemotherapy provides survival benefits. Methods We performed a meta-analysis of data from the PubMed and Embase databases. We included patients with LARC who received neoadjuvant (chemo) radiotherapy and curative surgery. Overall survival (OS), disease-free survival (DFS), toxicity, and compliance were analyzed in the oxaliplatin/fluorouracil- (OX/FU-) based group compared with the FU-based group, and in the chemotherapy group compared with the observation group. Results Twenty studies were included in the analysis. Our results indicated that adjuvant chemotherapy prolonged OS (hazard ratio [HR] = 0.78, 95%CI = 0.67–0.91) in patients with LARC treated with neoadjuvant (chemo) radiotherapy and surgery compared with those in the observation group. Subgroup analysis showed the same results in both the ypStage II and ypStage III groups. Compared with those in the observation group, patients in the chemotherapy group also showed an increase in DFS (HR = 0.75, 95%CI = 0.60–0.93). No significant increase was observed in OS (HR = 1.04, 95%CI = 0.87–1.24) or DFS (HR = 0.98, 95%CI = 0.76–1.27) when oxaliplatin was added to FU-based adjuvant chemotherapy, as compared with the FU-based treatment, and subgroup analysis also indicated no survival benefits in the clinical stage II, clinical stage III, ypStage II, and ypStage III groups. Conclusions For patients with LARC who have already received neoadjuvant (chemo) radiotherapy and curative surgery, adjuvant chemotherapy improves OS over that in the observation group. Adding oxaliplatin to FU-based adjuvant chemotherapy does not confer survival benefits beyond those from FU-based adjuvant chemotherapy.


Introduction
Colorectal cancer is the second most frequent cancer in women and the third most frequent cancer in men [1]. To date, surgical resection is the main radical treatment. According to the NCCN guidelines, the standard treatment is postoperative adjuvant chemotherapy with or without oxaliplatin for patients with locally advanced rectal cancer (LARC) after neoadjuvant treatment [2]. However, according to the ESMO guidelines, the available evidence is insufficient for the use of postoperative chemotherapy after neoadjuvant chemoradiation [3]. Moreover, a consensus has not been reached on whether adjuvant chemotherapy should be used, and which specific chemotherapy regimens can be used after neoadjuvant treatment. Although some studies have demonstrated that adjuvant chemotherapy should be used for patients with LARC after neoadjuvant chemoradiation and surgery [4][5][6][7][8], the effects of adjuvant chemotherapy are unclear in patients with LARC receiving neoadjuvant chemoradiation.
An oxaliplatin-based adjuvant chemotherapy regimen for LARC patients who have received neoadjuvant treatment has been mainly extrapolated from results achieved with oxaliplatin-containing adjuvant chemotherapy in lymph node-positive colon cancer patients [9,10], a strategy that is also considered as an optional adjuvant chemotherapy for rectal cancer in the NCCN guidelines. However, whether oxaliplatin meaningfully affects the survival outcomes in patients with rectal cancer after neoadjuvant treatment remains unclear. In the German CAO/ARO/AIO-04 trial, the addition of oxaliplatin to adjuvant chemotherapy has been found to be beneficial in terms of disease-free survival (DFS) in patients with LARC after neoadjuvant chemoradiation with oxaliplatin as an additional radiosensitizer [11]. In the ADORE trial, oxaliplatin-based adjuvant chemotherapy was found to significantly increase DFS but not overall survival (OS) in patients with ypStage III rectal cancer [12]. However, the final results of the PETACC-6 trial, which added oxaliplatin to both neoadjuvant chemoradiation (as an radiosensitizer) and adjuvant chemotherapy, have indicated no survival benefits in patients with LARC receiving oxaliplatin/fluorouracil-based (OX/FU-based) adjuvant chemotherapy compared with FU-based adjuvant chemotherapy [13]. In addition, oxaliplatin inevitably results in toxicity, thus, leading to poor compliance in postoperative chemotherapy. In a random trial, more than 40% of the safety population who received neoadjuvant chemoradiation and subsequently at least one cycle of oxaliplatin-based adjuvant chemotherapy had grade 3/4 toxicity, and only 48.1% of patients completed all the planned cycles [14]. Therefore, the use of oxaliplatin should be comprehensively considered in terms of all aspects including efficacy, toxicity, and compliance. Regarding whether adjuvant chemotherapy should be used, although one study has reported that adjuvant chemotherapy improves OS among patients treated with neoadjuvant chemoradiation [15], some studies have found that adjuvant chemotherapy does not improve OS or DFS in patients with LARC [16]. Some trials have also indicated that adjuvant chemotherapy may be beneficial in ypStage III patients only but not in ypStage II LARC patients [12,17]. Furthermore, for patients with rectal cancer downstaged to ypStage 0 or I after neoadjuvant chemoradiation, the survival benefits of adjuvant chemotherapy also remain unclear, although adjuvant chemotherapy is typically considered. Some studies have shown that adjuvant chemotherapy confers survival benefits to patients with ypStage 0 and ypT2N0M0 (ypStage I) [15,18]; however, other studies have reported that adjuvant chemotherapy is not necessary for patients with rectal cancer achieving either ypStage 0 or ypStage I [19,20]. Therefore, it is of great significance in clinical practice to explore whether the survival benefits of adjuvant chemotherapy are related to pathologic tumor stage after neoadjuvant treatment in rectal cancer.
In this meta-analysis, we sought to summarize and further assess the survival benefits of postoperative chemotherapy following neoadjuvant treatment and curative surgery for patients with LARC and examine the role of oxaliplatin in adjuvant chemotherapy.

Materials and Methods
This meta-analysis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Guidelines (Supplementary File 1) [21], and the Newcastle-Ottawa Scale (NOS) criteria were used to assess the methodological quality of the included studies [22].

Eligibility Criteria.
The eligible studies met the following inclusion criteria: (i) patients were diagnosed with rectal cancer; (ii) all patients received preoperative (chemo) radiotherapy; (iii) all patients underwent surgical resection of rectal cancer; (iv) studies compared chemotherapy with observation or compared postoperative OX/FU-based chemotherapy with FU-based chemotherapy (OX/FU-based chemotherapy included 5-FU and leucovorin plus oxaliplatin [FOLFOX] or capecitabine plus oxaliplatin [CAPEOX] and FU-based chemotherapy included 5-FU alone or capecitabine alone, because capecitabine and 5-FU are homologous chemotherapeutic drugs and are recommended adjuvant chemotherapeutic drugs for LARC patients after neoadjuvant treatment); and (v) the outcomes of studies included OS, DFS, toxicity, or compliance. Studies were excluded according to the following criteria: (i) studies of patients diagnosed with diseases other than rectal cancer; (ii) studies of patients who did not receive preoperative (chemo) radiotherapy; (iii) studies that compared preoperative OX/FU-based therapy with preoperative FU-based therapy; (iv) reviews, meta-analyses, and case reports; and (v) studies without outcomes relevant to this analysis. Two reviewers assessed all studies independently, and the final selected studies were determined on the basis of agreement between the two reviewers.

Data Extraction.
Two researchers completed the extraction of suitable data after reviewing the full text of included studies, and all disagreements were settled by discussion. The detailed information extracted from studies included author, year of publication, country, study design, number of patients, age, follow-up duration, TNM stage, tumor location from anal verge, neoadjuvant treatment regimen, 2 BioMed Research International adjuvant treatment regimen, type of surgery, median interval from surgery to adjuvant chemotherapy, and study outcomes including OS, DFS, compliance (completion of planned number of cycles), and toxicity (e.g., vomiting, nausea, neuropathy, allergic reaction, diarrhea, and hand-foot syndrome).

Statistical
Analysis. The primary endpoints were DFS and OS. The secondary endpoints were compliance and toxicity. We evaluated the primary endpoint with the hazard ratio (HR) and 95% confidence interval (CI). Then, we calculated the pooled risk ratio (RR) with 95% CI to assess toxicity. If a study did not provide the HR or 95% CI directly, we used published data to obtain the statistics by using the methods reported by Tierney et al. [23]. We analyzed the overall OS and DFS for the OX/FUbased group versus the FU-based group, and the chemotherapy group versus the observation group. Additionally, we conducted subgroup analysis according to study design, pathologic tumor stage after neoadjuvant treatment (ypTNM), preoperative clinical tumor stage (cTNM), and neoadjuvant treatment regimen. Statistical analysis of compliance and toxicity was performed on the basis of data from the studies. According to the heterogeneity, we used a random-effects model when I 2 > 50% or p < 0:1; otherwise, we used a fixed-effects model [24]. Publication bias was assessed with funnel plots with Begg's and Egger's tests [25,26]. Findings were considered significant with a two-sided p value ≤ 0.05.

Study Research.
A total of 4741 studies were retrieved by the electronic search in total (977 studies were from PubMed, and 3764 studies were from Embase). A total of 4067 studies remained after the elimination of duplicates. Subsequently, 3914 articles were removed on the basis of the eligibility criteria according to the titles and abstracts, after which 153 articles were retained. After further reading and evaluation of the full text, 20 studies were ultimately included in this meta-analysis, including seven randomized controlled trials (RCTs) and 13 non-RCTs (nRCTs) [4-6, 8, 11-16, 27-36]. The detailed research steps are shown in Figure 1.

Characteristics of Included Studies.
Twenty studies on 30662 patients were enrolled in the meta-analysis. Among the included studies, eight were from Europe, four from the USA, three from China, three from Korea, one from Canada, and one from Israel. Regarding the interventions, four studies had two arms containing an OX/FU-based group and FUbased group, 13 studies contained a chemotherapy group and observation group, and the other three studies contained an OX/FU-based group, FU-based group, and observation group. In terms of the neoadjuvant treatment regimen, preoperative long-course radiation with chemotherapy was conducted in 18 studies, long-course radiation alone or with chemotherapy was conducted in one study, and shortcourse radiation alone or with chemotherapy was used in one study. The follow-up duration of most studies was more than 3 years, four studies had longer durations of more than 5 years, and only two studies had durations of less than 1                  Figure 3(c)). Furthermore, a longer DFS was observed in patients receiving chemotherapy than in patients in the observation group when patients were treated with preoperative radiotherapy and preoperative OX/FU-based chemotherapy as a radiosensitizer (HR = 0:41, 95%CI = 0:21 -0:78, p = 0:007, Figure 3(d)). Although preoperative neoadjuvant radiotherapy may affect the survival benefits of adjuvant treatment in patients with rectal cancer, there were insufficient data to conduct a corresponding subgroup analysis. In an RCT of 473 rectal cancer patients, 86% of the patients who received preoperative short-course radiotherapy and 14% of patients who received preoperative long-course chemoradiation were randomly assigned to a chemotherapy group and observation group, and the results showed a similar 5-year DFS in the chemotherapy group versus the observation group (62.7% versus 55.4%) [35]. However, there were no sufficient data to support other subgroup analyses. The detailed results are shown in Table 2.  Figure 5(a)), and adding oxaliplatin to FU-based adjuvant chemotherapy was not beneficial for OS, as compared with that in the FU-based group (HR = 1:04, 95%CI = 0:87 -1:24, p = 0:656, Figure 5(b)). Moreover, a tendency toward increased OS was found in the comparison between the chemotherapy group and observation group for patients receiving preoperative radiotherapy and preoperative OX/FUbased chemotherapy as a radiosensitizer (HR = 0:49, 95%CI = 0:20 -1:24, p = 0:133, Figure 5(c)). However, the effects of different neoadjuvant chemotherapy strategies on survival benefits could not be evaluated in the comparison between the OX/FU-based group and FU-based group, owing to insufficient data. For preoperative short-course radiotherapy, in the RCT in which most rectal cancer patients (86% patients) treated with preoperative short-course radiotherapy alone, the survival rates in the adjuvant chemotherapy group and observation group were similar (80.4% vs. 79.2%), and no significant increase in OS was observed with adjuvant chemotherapy compared with observation (HR = 0:93, 95%CI = 0:62 -1:39, p = 0:73) [35].

Discussion
In this analysis, we included 20 studies on 30662 patients. In the chemotherapy group versus the observation group, adjuvant treatment was found to be beneficial in terms of OS and DFS. Subgroup analysis based on neoadjuvant treatment strategy revealed similar results, in which the adjuvant chemotherapy group showed higher OS and DFS than those in the observation group in patients treated with preoperative long-course chemoradiation. In addition, the results showed an increase in OS in either ypStage II or ypStage III in the chemotherapy group versus the observation group. In the OX/FU-based group, compared with the FU-based group, the addition of oxaliplatin did not contribute to OS or DFS. The results were also demonstrated in the subgroup analysis of patients treated with neoadjuvant long-course chemoradiation, and there were no significant survival benefits in the clinical stage II, clinical stage III, ypStage II, and ypStage III groups. However, oxaliplatin clearly increased the incidence of neuropathy, allergic reaction, vomiting, and nausea. Regarding compliance in completing all adjuvant chemotherapy cycles, there was no significant difference in the OX/FU-based group versus the FU-based group. However, the OX/FU-based group had a much higher rate of dose reduction than the FU-based group.
It has been reported that patients with ypStage III can benefit from adjuvant chemotherapy [5,6], which is in agreement with the present results. Regarding patients with ypStage II who received neoadjuvant treatment, one study has shown no increase in OS or DFS [29], whereas other studies have indicated that patients with ypStage II benefit from adjuvant chemotherapy [5,6]. Present pooled analysis showed that adjuvant chemotherapy improved OS in ypStage II patients undergoing neoadjuvant treatment and surgery. Several potential explanations may account for the survival benefits from adjuvant chemotherapy in ypStage II patients. First, neoadjuvant chemoradiation may lead to downstaging, and a portion of ypStage II patients might be pN + before receiving neoadjuvant treatment. In addition, patients with ypStage II mainly had T3-T4 stage cancer, which might infiltrate deeper before neoadjuvant treatment was received. Therefore, the adjuvant chemotherapy regimens guided by ypTNM stage after neoadjuvant chemoradiation may be different from those administered according to pTNM stage without neoadjuvant chemoradiation. Therefore, these reasons may plausibly explain the improved survival in ypStage II patients, which was different from that in pStage II patients. Studies that explore the efficacy of adjuvant chemo-therapy following neoadjuvant treatment according to ypStage are needed for further validation.
Oxaliplatin-based adjuvant chemotherapy was provided for patients with LARC after neoadjuvant treatment and surgery. However, in our analysis, adding oxaliplatin to adjuvant chemotherapy did not confer a benefit in either OS or DFS. This result might have been due to the toxicity of oxaliplatin causing a dose reduction-a factor that would negatively affect survival. In some studies [4,11,15,29], oxaliplatin was added to neoadjuvant treatment in the OX/FU-based group. The toxicity of oxaliplatin had cumulative effects, and using oxaliplatin-based adjuvant chemotherapy after oxaliplatinbased neoadjuvant treatment increased the toxicity, thereby affecting the efficacy of adjuvant chemotherapy [11]. Our results showed that oxaliplatin could lead to toxicity, which was clearly reflected by neuropathy, acute reaction, vomiting, and nausea. These side effects may decrease patient compliance. In the ADORE trial, cycles with reduced doses were more frequent in the OX/FU-based group than in the FUbased group [12]. Our meta-analysis indicated no significant difference in compliance in the completion rate of the planned cycles between the two groups, but dose reduction decreased the efficacy of adjuvant chemotherapy. Therefore, the decision to add oxaliplatin in adjuvant chemotherapy should account for the balance between efficacy and toxicity. Finally, a study has found that adding oxaliplatin is not associated with OS or DFS in patients ≥73 years of age [37]. Therefore, in this study, some patients were above 73 years old, thus, possibly affecting the final results. However, we did not discuss this possibility further because of a lack of data.
In clinical practice, evaluating the effects of neoadjuvant treatment strategy on the survival benefits of adjuvant treatment in patients with LARC is meaningful. Given that most of the included studies used neoadjuvant long-course chemoradiation, we further evaluated the survival benefits of adjuvant chemotherapy in the setting of neoadjuvant longcourse chemoradiation. The results showed similar results to those of the overall analysis, indicating that adjuvant chemotherapy, compared with observation, contributes to survival benefit, and that adding oxaliplatin to FU-based adjuvant chemotherapy does not contribute to survival benefits beyond those conferred by FU-based adjuvant chemotherapy. In terms of neoadjuvant chemotherapy strategy, further subgroup analyses were conducted to evaluate the effects of neoadjuvant long-course chemoradiation with preoperative FU-based chemotherapy or neoadjuvant longcourse chemoradiation with preoperative OX/FU-based chemotherapy on survival outcomes. Similar results were observed, thus, demonstrating the reliability and accuracy of our results. However, subgroup analysis based on preoperative radiotherapy strategy (short-course versus long-course radiotherapy) could not be conducted, owing to an insufficient number of studies. Thus, future large-scale, prospective clinical studies are needed to explore the influence of neoadjuvant treatment strategy, including preoperative shortcourse radiotherapy, preoperative long-course radiotherapy, and long-course chemoradiation, on the relationship between adjuvant chemotherapy and survival benefits. Such

18
BioMed Research International an understanding may lead to more effective and suitable clinical treatment strategies for patients with LARC. There are several limitations to the current research. First, this was a retrospective analysis, and the potential for confounding on the basis of patient selection could not be eliminated. Second, we were unable to obtain the personal details of patients, although this information might have enabled better control of the confounding factors. Third, the heterogeneity in the study that could not be eliminated by subgroup analysis might have affected external authenticity to some extent. Moreover, some patients may have better pathological responses to neoadjuvant treatment, thus, achieving ypStage1 or even ypStage 0 after neoadjuvant treatment; however, the survival benefit of adjuvant chemotherapy still remains unclear for these patients in clinical practice. Unfortunately, we could not perform further analysis because the data for patients with ypStage 0 or I after neoadjuvant treatment were insufficient. In addition, we were unable to perform more subgroup analyses because of the insufficient number of studies. Finally, the recent guidelines for adjuvant chemotherapy regimens are mainly based on pStage. Differences exist between ypStage and pStage, and, thus, chemotherapy guided by ypStage might require further exploration. Future studies are needed to further investigate and resolve these problems.

Conclusions
Compared with observation, adjuvant chemotherapy was found to prolong OS in patients with LARC after neoadjuvant (chemo) radiotherapy and curative surgery. However, no survival benefit was observed with the addition of oxaliplatin to FU-based adjuvant chemotherapy compared with FU-based adjuvant chemotherapy.

Data Availability
The research data used to support the findings of this study were supplied by Zhenning Wang under license and so cannot be made freely available. Requests for access to these data should be made to Zhenning Wang, E-mail: josieon826@sina .cn.

Ethical Approval
The study did not violate the local regulations and was approved by the Research Ethics Committee of China Medical University, China.