Clinicopathological Features and Oncological Outcomes of Early and Late Recurrence in Stage III Colorectal Cancer Patients after Adjuvant Oxaliplatin-Based Therapy

Aims An adjuvant oxaliplatin-based regimen is the standard of care for patients with stage III colorectal cancer (CRC). Few reports have compared the clinicopathological features and oncological outcomes of such treatment between patients with early (≤1 year) and late recurrence (>1 year). Methods Between January 2012 and December 2019, CRC recurred in 128 (24.1%) of 531 patients with consecutive stage III CRC after they received curative resection and an adjuvant oxaliplatin-based regimen. The clinicopathological features and oncological outcomes of the 128 patients were analyzed retrospectively. Results The median follow-up period after the first chemotherapy cycle was 35.0 months (range, 7–100.9), and the median recurrence time was 16.1 months. Forty-seven patients (36.7%) had an early recurrence and eighty-one patients (63.3%) had a late recurrence. Compared with patients with late recurrence, those with early recurrence were mostly younger (median: 58 vs. 64 years, p=0.009), had less oxaliplatin-based therapy cycles (median: 8 vs. 12 cycles, p < 0.001), and had a shorter overall survival time (median: 23.3 vs. 39.7 months, p < 0.001). The area under the curve of patient age and chemotherapy cycles for predicting early recurrence was 0.629 and 0.705 (p=0.015 and p < 0.001), respectively. The receiver operating characteristic curve analysis demonstrated that the cutoff level for patient age was 57 years and the number of chemotherapy cycles was 8. A multivariate analysis revealed that patient age ≤57 years and oxaliplatin-based therapy ≤8 cycles were independent risk factors for early recurrence (odds ratio (OR) = 3.049, p=0.022; OR = 4.995, p=0.002). These factors were associated with an approximately 77.8% risk of recurrence within 1 year, compared with the 21.5% risk associated with patient age >57 years and oxaliplatin-based therapy >8 cycles (p = 0.003). Conclusion Patients with early recurrence had poorer survival than those with late recurrence. If >8 cycles of oxaliplatin-based therapy can be administered without disease progression, then patients with stage III CRC would have a lower risk of early recurrence.


Introduction
Colorectal cancer (CRC) is the third most commonly diagnosed cancer, with an estimated 1.9 million cases and 915,880 deaths reported in 2020 worldwide [1]. On the basis of fndings from an 8-yearfollow-up, the ACCENT group reported that 32.9% of patients with stage III CRC had cancer recurrence [2]. Moreover, 82% of patients with stage III CRC experienced recurrence within the frst 3 years of cancer diagnosis, and the incidence of recurrence peaked between 1 and 2 years after initial treatment [2,3].
According to the Ministry of Health and Welfare of Taiwan, CRC has become the most common cancer and the third leading cause of cancer-related death since 2006. In 2018, the CRC incidence was 41.8 per 100,000 individuals in Taiwan, with 14.9 deaths per 100,000 individuals [4]. Furthermore, a large proportion (25%) of patients with CRC in Taiwan have stage III CRC [4]. Te overall survival (OS) rate of patients with stage III CRC was 59.9%.
Adjuvant chemotherapy, in which oxaliplatin is combined with a fuoropyrimidine (FOLFOX or CAPOX) is the standard of care for patients with stage III CRC [5]. Even with curative surgery and adjuvant chemotherapy, the overall prognosis of stage III CRC remains unsatisfactory, with a 5-year survival rate of only 69% [4]. Moreover, despite adjuvant chemotherapy, patients with locally advanced CRC had an approximately 26.7% risk of relapse within 5 years [6]. Terefore, risk factors predicting the progress, relapse, and metastasis of CRC after adjuvant chemotherapy must be identifed.
Several risk factors for stage III CRC recurrence have been identifed, including rectal cancer, preoperative and postoperative serum carcinoembryonic antigen (CEA) level (>5.0 ng/mL), postoperative carbohydrate antigen 19-9 level, infltrative growth patterns, and >3 metastatic lymph nodes [6][7][8]. However, whether recurrence time would determine outcomes following an adjuvant oxaliplatin-based regimen in patients with stage III CRC who had undergone radical resection remains unclear. We here compared the clinicopathological features and oncological outcomes of such treatment between patients with early (≤1 year) and late recurrence (>1 year).

Patients.
Between January 2012 and December 2019, a total of 531 consecutive patients with histologically confrmed stage III CRC who had undergone surgical treatment at a single institution were analyzed. Of them, 162 patients were excluded because of the following reasons: 125 patients did not receive an adjuvant oxaliplatin-based regimen after surgery; 7 patients started receiving oxaliplatin-based therapy after recurrence; and 30 patients had other malignancies. Finally, 369 patients received the adjuvant oxaliplatin-based regimen after surgery. Of them, 128 patients had a recurrence. Figure 1 presents the fowchart of patient selection. Te present study was approved by the Institutional Review Board of Kaohsiung Medical University Hospital (KMUHIRB-E (I)-20210006). In accordance with the Declaration of Helsinki, this research study was performed in accordance with relevant guidelines, and informed consent was obtained from all the participants.
A 25% dose reduction was according to hematology toxicity such as absolute neutrophil count, platelet count, and nonhematological toxicity including skin symptoms, peripheral neuropathy, and acute laryngopharyngeal dysesthesia if patients were encountered with ≥Grade III adverse events (AEs) [11]. Administration was continued until any of the following criteria for discontinuation were fulflled [11,12]: (1) Disease progression was diagnosed clinically or by imaging (2) A Grade III AE occurred again even after dose reduction of chemotherapy regimen (3) A treatment course was delayed for more than 2 cycles of treatment owing to an AE (4) A Grade IV or V AE occurred (5) Te patient declined treatment (6) Te attending physician judged that continuation of the study was difcult for any other reason

Patient Follow-Up.
Te clinical stages and pathological features of primary tumors were defned according to the eighth edition of the UICC tumor-node-metastasis staging system [13]. Te clinicopathological features analyzed were sex, age, tumor size, tumor location, tumor invasion depth, lymphovascular invasion, perineural invasion, tumor grade, oxaliplatin cycles, and preoperative and postoperative serum CEA levels. Right-sided colon cancers were defned as those located in the cecum, ascending colon, hepatic fexure, and transverse colon, and left-sided cancers were defned as those located in the splenic fexure, descending colon, sigmoid, and rectum.

Journal of Oncology
Treatment responses were assessed using computed tomography, magnetic resonance imaging, or positron emission tomography, and the best responses were recorded. Te development of a new instance of local recurrence (tumor growth restricted to the anastomosis or the primary operation region) or distant metastatic lesions (distant metastases or difuse peritoneal carcinomatosis) during the postoperative surveillance period was defned as postoperative recurrence.
Recurrence within 1 year of the initial treatment with the adjuvant oxaliplatin-based regimen was defned as an early recurrence, and after 1 year was defned as a late recurrence [6]. Te progression-free survival (PFS) time was calculated from the date of the frst oxaliplatin-based therapy cycle to the date of recurrence. Te OS time was calculated from the date of the frst oxaliplatin-based therapy cycle. PFS and OS times were analyzed. All patients were followed up until their deaths, their last followup date, or May 2021.

Statistical Analysis.
Student's t test and a chi-square test were used to compare continuous and categorical descriptive variables, respectively, between the groups. Te univariate and multivariate logistic regression analyses were used to examine the relationships between clinicopathological features and recurrence. Cumulative PFS and OS rates were calculated using the Kaplan-Meier method, and the log-rank test was used to compare time-to-event distributions. Te predictive ability of patient survival was evaluated through receiver operating characteristic (ROC) curve analysis. Te area under the curve (AUC) was also calculated. Results were expressed as the mean ± standard deviation or odds ratio (OR) and 95% confdence interval (CI) were appropriate. A p value <0.05 indicated statistical signifcance. All data were analyzed using SPSS (Version 19.0; SPSS, Chicago, IL, USA).

Early Recurrence Compared with Late Recurrence.
Te median recurrence time after the frst oxaliplatin-based therapy cycle was 16.1 months. Forty-seven patients (36.7%) had relapse within 12 months (early recurrence) and eightyone patients (63.3%) had relapse after 12 months (late recurrence) ( Table 2). Between the two groups, sex, tumor size, tumor invasion depth, histology, and preoperative and postoperative CEA levels did not signifcantly difer (all p > 0.05). Te expression of recurrence did not difer signifcantly between the patients with local recurrence and those with distant metastasis (p � 0.449).
Te median PFS time was signifcantly shorter among the patients with early recurrence than among those with late recurrence (6.6 [95% CI: 5.    Figure 2(b)).

Association between Patient Age and Recurrence.
Compared with the patients with late recurrence, those with early recurrence were younger (median 58 years vs. 64 years, p � 0.009). Te predictive ability of patient age for early recurrence was evaluated through the ROC curve and AUC analyses. Te AUC of patient age for predicting early recurrence was 0.629 (p � 0.015). Te ROC curve analysis demonstrated that the cutof level was 57 years. Tus, patients aged >57 and ≤57 years were defned as "older adult patients" and "young patients," respectively. In the univariate logistic regression analysis, young age was revealed as a predictor for early recurrence (OR � 2.872, 95% CI: 1.330-6.200; p � 0.007, Table 3).

Association between Cycles of the Adjuvant Oxaliplatin-Based Regimen and Recurrence.
Compared with the patients with late recurrence, those with early recurrence had fewer cycles of adjuvant oxaliplatin-based therapy (median 8 cycles vs. 12 cycles, p < 0.001). Te predictive ability of these cycles for early recurrence was also evaluated through the ROC curve and AUC analyses. Te AUC of these chemotherapy cycles for predicting early recurrence was 0.705 (p < 0.001). Te ROC curve analysis indicated that the cutof level was eight cycles. Specifcally, when the patients received >8 cycles of adjuvant oxaliplatin-based therapy, the analysis result was defned as "positive," whereas when the patients received ≤8 cycles of adjuvant chemotherapy, the analysis result was defned as "negative." Forty patients (31.3%) received ≤8 cycles of adjuvant oxaliplatin-based therapy. Chemotherapy was terminated in 8 of the 40 patients because most of them were intolerant or oxaliplatin toxicity reluctant to further therapy. According to the univariate logistic regression analysis, the patients with a negative result had a 6-fold higher risk of early recurrence than those with a positive result (OR � 5.925, 95% CI: 2.626-13.371; p < 0.001, Table 3).

Risk Factors Infuencing Early Recurrence.
To identify the independent risk factors for early recurrence in patients with stage III CRC, we used a logistic regression model to perform the univariate and multivariable analyses (Table 3). In the multivariate analysis, early recurrence was signifcantly correlated with patient age ≤57 years (OR � 3.049, 95% CI: 1.171-7.941; p � 0.022) and oxaliplatin-based therapy ≤8 cycles (OR � 4.995, 95% CI: 1.806-13.815; p � 0.002). PFS was further analyzed on the basis of the relationship between patient age and adjuvant chemotherapy cycles (Figure 3). Each patient was classifed into one of the four following groups: (a) age >57 years and oxaliplatin-based therapy >8 cycles (N � 65); (b) age ≤57 years and oxaliplatin-based therapy >8 cycles (N � 23); (c) age >57 years and oxaliplatin-based therapy ≤8 cycles (N � 22); and (d) age ≤57 years and oxaliplatin-based therapy ≤8 cycles (N � 18). A signifcant diference in PFS was noted among the four groups (p � 0.003). Te 1-year PFS rate improved signifcantly to 78.5% in the patients aged >57 years with >8 cycles of oxaliplatin-based therapy. In the patients aged ≤57 years with >8 cycles of chemotherapy, the 1-year PFS rate was 56.5%. In the patients aged >57 years with ≤8 cycles of chemotherapy, the 1-year PFS rate was 36.4%. In the patients aged ≤57 years with ≤8 cycles of chemotherapy, the 1-year PFS rate was 22.2%.

Discussion
Oxaliplatin-based regimens have recently become the gold standard in postoperative adjuvant chemotherapy for patients with stage III CRC [14]. However, the choice of an appropriate treatment strategy and decision on therapy duration depend on the associated cumulative neurotoxicity. Studies have demonstrated that oxaliplatin-induced neurotoxicity usually presents at the 8th-10th cycle of FOLFOX [15]. Te treatment duration can be reduced from 6 (12 cycles) to 3 months (6 cycles) for patients with a low risk of recurrence; this reduction does not compromise efectiveness and may even signifcantly decrease the risk of cumulative sensitive neuropathy [5]. In the present study, 40 patients received ≤8 cycles of oxaliplatin-based therapy, and 8 (20%) of them developed intolerance to adverse efects and refused continuous therapy. Te period of therapy is a critical determinant of how well the therapy aids survival [16].
Recommendations are shown by the fndings of a recent pooled analysis of clinical trials that compared 6 versus 3 months of oxaliplatin-based chemotherapy [17,18]. According to the international, phase 3 trial conducted at 244 centers of Te Short Course Oncology Terapy study, Journal of Oncology 5 32.5% of patients with high-risk stage II and stage III CRC who received FOLFOX and were considered at a low recurrence risk may be treated efectively and experience less neurotoxicity with 3 months of oxaliplatin-based therapy compared with the standard 6-month regimen [19]. However, the International Duration Evaluation of Adjuvant Chemotherapy collaboration, France, in which 90% of patients received mFOLFOX6, could not establish noninferiority for 3 months of oxaliplatin-based chemotherapy [20]. Te results difered depending on the treatment regimen and patient risk group. However, our results revealed the signifcance number of cycles of treatments in reducing the risk of early recurrence in the adjuvant mFOLFOX6 regimen. For patients at a high recurrence risk (T4 and N2 subgroups), adjuvant chemotherapy should be ofered for 6 months [20]. For patients at a low recurrence risk (T1, T2, or T3 and N1 subgroups), 6 or 3 months of adjuvant chemotherapy may be ofered depending on whether a potential reduction in adverse events and no signifcant diference in PFS were observed with the 3-month regimen [5,17]. However, neither tumor invasion depth nor lymph node metastasis difered signifcantly between the early and late recurrence groups in the present study. Pathologic staging is not an efective predictor of early recurrence risk. Tsai et al. recommended at least seven cycles of FOLFOX for favorable PFS and eight cycles for favorable OS [16]. In the present study, 78 patients (60.9%) completed the full 12 cycles of adjuvant oxaliplatin-based therapy; however, 18 (23.1%) of them developed early recurrence. It is difcult to balance between overtreatment (which exposes the patient to unnecessarily high levels of chemotherapy toxins) and undergo treatment (which leaves a high risk of early recurrence unaddressed).
Among populations with recurrence, young patients had a 3-fold higher risk of early recurrence than the older adult patients (OR � 3.049, p � 0.022). Of the 41 patients aged ≤57 years in the present study, 18 (43.9%) patients received ≤8 cycles of adjuvant oxaliplatin-based chemotherapy. Although these young patients with ≤8 cycles of adjuvant chemotherapy had the worst 1-year PFS rate of 22.2%, they were more likely to receive second-line adjuvant chemotherapy or reoperation than the older patients were. Younger patients with stage III CRC might have a relatively lower risk of early recurrence (1-year PFS rate of 56.5%), if >8 cycles of adjuvant chemotherapy could be administered.
Compared with patients aged <70 years, patients aged >70 years have a higher beneft and safety profle of adjuvant fuorouracil-based chemotherapy than surgery alone [21,22]. Although only 29 patients were aged ≥70 years (22.7%) in the present study, the 1-year PFS rate improved signifcantly to 78.5% in the patients aged >57 years with >8 cycles of oxaliplatin-based therapy. In patients with stage III CRC who are older, who have multiple comorbidities, and who are less likely to receive chemotherapy, adjuvant chemotherapy might be associated with a lower death risk [23].
An abnormal preoperative and postoperative serum CEA level (≥5 ng/mL) was a signifcant independent negative predictive factor for postoperative recurrence [6,7]. In the present study, neither the preoperative nor the postoperative serum CEA level was a risk factor for early recurrence. However, patients with an abnormal postoperative serum CEA level might have a high risk of postoperative recurrence and mortality. For these patients, >8 cycles of adjuvant oxaliplatin-based therapy could be administered to lower the risk of early recurrence.
Although well-diferentiated adenocarcinoma is usually common in CRC, the histological subclassifcation has a great impact on the prognosis. Imai et al. reported poorly diferentiated adenocarcinoma of the CRC had signifcantly worse PFS and OS than well-to-moderately diferentiated adenocarcinoma [24], refecting that none of the recurrence were well-diferentiated in the present study.
Te 5-year OS rate of stage III CRC from recurrence was 13.5%, which was similar to the survival rate of 13.8% for the distant-stage disease of CRC according to data from the  Journal of Oncology Surveillance, Epidemiology, and End Results program [25,26]. In people with unresectable metastatic CRC, recent clinical trials have demonstrated that tailoring treatment to the molecular and pathological features of the tumor improves OS [1]. Terefore, genomic profling might be necessary to extend survival benefts and reduce drug toxicity.
In patients with stage III CRC after adjuvant chemotherapy, the expression of the epidermal growth factor receptor can be used to predict OS and PFS times and postoperative relapse [6,9]. Tis study has some limitations. First, this was a retrospective study conducted at a single center. Second, the toxicities of chemotherapy and subsequent lines of therapy were not discussed in the study. Tird, various biomarkers for stage III CRC with their possible association with the efcacy of adjuvant chemotherapy were not analyzed. A comparison of heterogeneous populations was not easy because the disease severity and therapeutic strategies between early and late recurrence groups were not completely comparable. Predicting the risk of early recurrence efectively for each patient individually is difcult. In addition, as to "cancer recurrence and metastasis," it indeed is a very complex progression. Emerging interesting studies show that it should be thought as a bidirectional process that has "self-seeding" in breast cancer, CRC, and nasopharyngeal carcinoma [27,28]. Moreover, when combined with 5fuorouracil and oxaliplatin, metformin potentially acts as an adjunctive agent to eliminate CRC in vitro and in vivo [10]. However, the present study used real-world data and

Conclusions
Patients with early recurrence had a poorer OS than those with late recurrence. Young age (≤57 years) and few cycles (≤8 cycles) of an adjuvant oxaliplatin-based regimen are independent risk factors for early recurrence. For patients at increased risk of recurrence from stage III CRC, if >8 cycles of adjuvant oxaliplatin-based therapy could be administered, people of all ages would have a low risk of early recurrence and might have a better OS.

Data Availability
Te results reported in the article can be found including, where applicable, hyperlinks to publicly archived datasets analyzed or generated during the study. By data, we mean the minimal dataset that would be necessary to interpret, replicate, and build upon the fndings reported in the article.