The effect of adjuvant chemotherapy on survival in node negative colorectal cancer with or without perineural invasion: a systematic review and meta-analysis

Purpose It was aimed at assessing the benefits of adjuvant chemotherapy (ACT) for patients with node-negative colorectal cancer (CRC) either with or without perineural invasion (PNI). Methods We systematically searched PubMed, Cochrane Library, Embase, and Web of Science from database inception through October 1, 2023. Survival outcomes were analyzed using hazard ratios (HRs) and corresponding 95% confidence intervals (CIs). The methodological quality of included studies was assessed using the Newcastle-Ottawa Scale (NOS). Heterogeneity for the descriptive meta-analyses was quantified using the I2 statistic. Results Ten studies included in this review. ACT improved overall survival (OS) (HR 0.52, 95% CI 0.40–0.69) and disease-free survival (DFS) (HR 0.53, 95% CI 0.35–0.82) in PNI + patients but did not affect DFS (HR 1.13, 95% CI 0.72–1.77) in PNI- patients. A disease-specific survival (DSS) benefit with chemotherapy was observed in PNI + (HR 0.76, 95% CI 0.58–0.99) and PNI- patients (HR 0.76, 95% CI 0.57–1.00). And PNI decreased DFS (HR 1.94, 95% CI 1.52–2.47) and OS (HR 1.75, 95% CI 0.96–3.17) in node-negative CRC. Conclusions In conclusion, chemotherapy appears most beneficial for survival outcomes in node-negative patients with PNI, but may also confer some advantage in those without PNI. Systematic Review Registration Identifier INPLASY2021120103.

The effect of adjuvant chemotherapy on survival in node negative colorectal cancer with or without perineural invasion: a systematic review and meta-analysis Introduction Colorectal cancer is the third most common type of cancer in both men and women.Globally, almost 1.5 million new cases of CRC are diagnosed every year, of which more than a third are fatal (1).The most common cause of death is complications arising from metastasis (2).The primary treatment for stage I-II CRC is radical surgery (3).However, undetected micrometastases that persist after curative surgery may cause cancer recurrence (4).This micrometastasis is eradicated with ACT to enhance cure rates (5).
It is unfortunate that few reliable prognostic and predictive markers exist to identify patients at a high risk for disease progression during the early stages of CRC (6).Stage II CRC recurrence rates range from 7.9%-22%, whereas only 2%-5% of patients benefit from ACT (7)(8)(9)(10).Due to these reasons, the National Comprehensive Cancer Network (NCCN) does not recommend conventional ACT for stage II CRC unless certain risk factors exist.There were pT4 lesions, intestinal perforation, obstruction, 12-sample lymph nodes (LNs), lymph vascular invasion, PNI, poorly differentiated histology and margins that are positive, indeterminate, or close (11).Patients with these risk factors have a relatively poor prognosis (12).According to Lin et al., ACT was beneficial to patients with CRC and certain risk factors (13).In contrast, O'Connor et al. reported that ACT had no effect on any of these risk factors (14).Kumar et al. found that ACT was most effective for patients with pT4 in high-risk patients (15).Recent studies suggest, however, that ACT can benefit patients with PNI (16)(17)(18).PNI refers to tumor cells spreading through nerves.It was Bataskis who first described the prognostic value of PNI, which he defined as "tumor invasion around and through nerves (19)."PNI has been recognized as an unfavorable prognostic factor in CRC since it is associated with poor survival rates (20).
ACT, however, remains controversial because it is unclear whether these patients will benefit patients with PNI (21).This study was conducted to determine whether node-negative CRC patients with and without PNI receive different benefits from ACT.

Data extraction and quality assessment
The researchers (W.Yu and H. Ying) independently assessed the eligibility of all the studies and extracted the following information: The first author's name, the country in which the study was conducted, the sample size, the year of the study, the ages of the participants, the stage of their cancer, the chemotherapy regimen, and the period of follow-up.As well as OS, DFS, DSS, recurrence-free survival (RFS), and NOS.We consulted with a third reviewer (W.Hong) to resolve any discrepancies between the reviewers.In order to rate the quality of the articles, we used the NOS score.Articles that have an NOS score >6 (on a scale of 0-9) were considered to be of high quality (22).

Risk of bias analysis
Using non-parametric correlation tests, we examined the association between quality of reporting and HR.Begg and Egger tests were also conducted to determine whether publication bias was present (23).

Statistical analysis
Our analysis used HRs and 95% CIs to compare PNI with survival.When HRs and 95% CIs were not included, data were derived from survival curves according to Parmar et al. and Tierney et al. (24,25).Study heterogeneity was examined using I 2 statistics.Whenever there was obvious heterogeneity, as indicated by a p-value < 0.10 or I 2 exceeding 50%, a random effect model was used.In other cases, a fixed effect model was used.Our findings were further enhanced by performing meta-regressions and subgroup analyses in order to identify the sources of heterogeneity.We conducted sensitivity analysis to determine the stability of our combined results, and we assessed publication bias using the Begg and Egger test (26,27).Statistical significance was set at p < 0.05 using STATA 16.0 (Stata Corporation, College Station, TX, USA).

Search results and quality assessment
We conducted electronic searches of MEDLINE, Embase, Cochrane Library, and Web of Science, which yielded 743 studies.An additional 21 studies were identified from reference lists.After removing 442 duplicate records, 322 studies underwent title and abstract screening, of which 199 were excluded as Records excluded.The full texts of the remaining 123 studies were assessed; 17 studies could not be retrieved and 52 further studies were excluded based on predefined criteria.Ultimately, 10 studies met the inclusion criteria and were included in the systematic review and meta-analysis, comprising data on 118,529 patients in total.The study selection process is outlined in the PRISMA flow diagram (Figure 1).All patients underwent curative-intent resection of their CRC prior to ACT.Some studies also analyzed the highrisk factors after colon cancer surgery.Table 1 summarizes the 10 retrospective cohort studies included in the systematic review.These studies involved 118,529 patients with stage II CRC who underwent surgery.The studies compared ACT vs. no chemotherapy and reported on outcomes including OS, DFS, and recurrence.Follow-up times ranged from 5 to 10 years.We assessed the quality of ten articles by using the NOS score since they were retrospective cohort studies.A total of seven articles scored 7 points and six articles scored 8 points, with the main loss being the study controls for confounding factors (Table 2).

References
We compared OS, DFS and DSS between patients who received ACT and those who observation only, stratified by PNI status.For patients with node-negative CRC and PNI+, ACT was associated with significantly improved OS compared to observation (HR 0.52, 95% CI 0.40-0.69).There was moderate heterogeneity between the 3 included studies (I 2 = 41.3%, p = 0.130).In the PNI-subgroup, ACT also conferred an OS benefit over observation (HR 0.52, 95% CI 0.27-0.78).However, there was substantial heterogeneity between the 2 studies in this analysis (I 2 = 77.1%,p = 0.013).ACT appeared to improve OS regardless of PNI status.The OS benefit with ACT was similar between PNI + and PNI-patients (Figure 2).

Sensitivity analysis
Fixed effects and random effects models were compared to analyze prognosis (OS) in patients with PNI who were treated with ACT.
We analyzed the prognosis (OS) of patients with PNI who received ACT by comparing fixed effect and random effect models.OS did not differ significantly between the two models (fixed effect model: HR = 0.51, 95% CI = 0.43-0.61,random effect model: HR = 0.52, 95% CI = 0.40-0.69).In the sensitivity analysis, we arbitrarily deleted the OS and DFS literature, which did not affect the results of this study (Figure 6).

Publication bias
Our analysis included ten studies, but the subgroup studies were relatively few because they assessed different outcomes.There is an inherent risk of public bias in all reviews.According to Egger and Begg tests (Egger test: p = 0.189; Begg test: p = 0.308), DFS analysis did not detect a significant publication bias.In addition, the DFS analysis found no evidence of publication bias (Egger test: p = 0.925; Begg's test: p = 1.00).

Discussion
This systematic review and meta-analysis examined the efficacy of ACT for node-negative CRC stratified by PNI status.Our results suggest that chemotherapy improves overall and DFS in patients with PNI, but may not affect DFS in those without PNI.
OS was significantly improved with ACT vs. observation in the PNI + subgroup (HR 0.52, 95% CI 0.40-0.69),consistent with prior studies showing a survival benefit for high-risk stage II patients receiving chemotherapy (36,37).A recent cohort study of 500 colon cancer patients also found the addition of oxaliplatin to standard 5-FU chemotherapy prolonged OS and DFS selectively in the subgroup with PNI (33).The survival gain seen with chemotherapy in PNI + patients may be due to eradication of occult micrometastases not detectable on standard pathology (38).Interestingly, we also observed an OS benefit with chemotherapy in the PNI-subgroup (HR 0.52, 95% CI 0.27-0.78),although prior analyses have been conflicting (10, 39).The reason for improved OS with chemotherapy even for lower risk PNI-patients is unclear and warrants investigation.
DFS was significantly improved by chemotherapy in the PNI + subgroup (HR 0.53, 95% CI 0.35-0.82)but not in the PNIsubgroup (HR 1.13, 95% CI 0.72-1.77).These findings align with other studies demonstrating PNI is an independent prognostic factor for DFS (40).A potential explanation is that PNI + tumors are more aggressive and prone to early micrometastases or local recurrence after surgery that is eradicated by chemotherapy (41).The lack of DFS benefit with chemotherapy in PNI-patients highlights the need for risk-stratified treatment approaches to avoid over-treatment (42).Recent data suggest molecular profiling may help further stratify risk in node negative CRC (43).
This study has several limitations.The pooled sample size was relatively small for PNI subgroup analyses, particularly for secondary outcomes like DFS and DSS, warranting cautious interpretation.Publication bias remains a concern given the limited number of studies.There was heterogeneity between studies that may relate to differences in chemotherapy regimens, follow-up times, and underlying study populations.The retrospective observational nature of the included studies also has inherent biases compared to prospective trials.And this systematic review included studies published over a long timespan, ranging from 1993 to 2015.The inclusion of literature covering many decades could introduce bias, as changes in cancer treatments, staging modalities, and other factors over time may impact outcomes.Despite these limitations, this systematic review provides a comprehensive synthesis of current evidence regarding efficacy of ACT in early stage CRC with vs. without PNI.
ACT appears to improve survival outcomes primarily in nodenegative CRC patients with PNI.PNI may be an important factor to guide chemotherapy decisions in this population.Additional  Association between PNI and survival in node negative colorectal cancer patients.Diamond represents the pooled effect estimate of the overall analysis.Data are represented as HRs with 95% CIs.Inter-study heterogeneity quantified by I 2 with significance p < 0.10, HR, hazard ratio; ACT, adjuvant chemotherapy; PNI, perineural invasion; OS, overall survival; DFS, disease free survival overal; DSS, disease-specific survival; RFS, recurrence-free survival; PNI, perineural invasion.
Sensitivity analysis of overall high-risk factors receiving adjuvant hemotherapy on OS (A) and disease-free survival (B).

Conclusion
ACT improved OS and DSS in node-negative CRC patients regardless of PNI status.But DFS benefit with chemotherapy was observed only in patients with PNI.Overall, chemotherapy appears most beneficial for survival outcomes in node-negative patients with PNI, but may also confer some advantage in those without invasion.

FIGURE 4 ACT
FIGURE 4    ACT versus observation-only patients stratified by PNI, DFS.Diamond represents the pooled effect estimate of the overall analysis.Data are represented as HRs with 95% CIs.Inter-study heterogeneity quantified by I 2 with significance p < 0.10.HR, hazard ratio; DSS, disease-specific survival; ACT, adjuvant chemotherapy; PNI, perineural invasion.
″1″ means that the study is satisfied the item and "0″ means the opposite situation.AO, assessment of outcome; DO, demonstration that outcome of interest was not present at start of study; FU, follow-up long enough for outcomes to occur; REC, representativeness of the exposed cohort; SC, study controls for age; SNEC, selection of the nonexposed cohort.
AE, indicates ascertainment of exposure; AF, study controls for any additional factors; AFU, adequacy of follow-up of cohorts; PNI-patients (471 ACT) (15, 29).RFS was improved with ACT in PNI + patients (HR 0.79, 95% CI 0.42-1.46,I 2 = 0%), but RFS data were unavailable for PNI-patients (Table Ying et al. 10.3389/fsurg.2023.1308757Frontiers in Surgery 08 frontiersin.orgwell-designed prospective studies are needed to clarify the riskbenefit ratio of adjuvant treatment based on PNI status.Future research should also examine how emerging prognostic factors and individualized risk prediction models can optimize personalized adjuvant therapy for early stage CRC.