Impact of Lymph Node Ratio on Oncologic Outcomes in ypStage III Rectal Cancer Patients Treated with Neoadjuvant Chemoradiotherapy followed by Total Mesorectal Excision, and Postoperative Adjuvant Chemotherapy

Purpose To evaluate the prognostic impact of the lymph node ratio (LNR) in ypStage III rectal cancer patients who were treated with neoadjuvant chemoradiotherapy (NCRT). Materials and Methods We retrospectively reviewed the data of 638 consecutive patients who underwent NCRT followed by total mesorectal excision, and postoperative adjuvant chemotherapy for rectal cancer from 2004 to 2011. Of these, 125 patients were positive for lymph node (LN) metastasis and were analyzed in this study. Results The median numbers of examined and metastatic LNs were 17 and 2, respectively, and the median LNR was 0.143 (range, 0.02–1). Median follow-up time was 55 months. In multivariate analyses, LNR was an independent prognostic factor for overall survival (OS) (hazard ratio [HR] 2.17, p = 0.041), disease-free survival (DFS) (HR 2.28, p = 0.005), and distant metastasis-free survival (DMFS) (HR 2.30, p = 0.010). When ypN1 patients were divided into low (low LNR ypN1 group) and high LNR (high LNR ypN1 group) according to a cut-off value of 0.152, the high LNR ypN1 group had poorer OS (p = 0.043) and DFS (p = 0.056) compared with the low LNR ypN1 group. And there were no differences between the high LNR ypN1 group and the ypN2 group in terms of the OS (p = 0.703) and DFS (p = 0.831). Conclusions For ypN-positive rectal cancer patients, the LNR was a more effective prognostic marker than the ypN stage, circumferential resection margin, or tumor regression grade after NCRT, and could be used to discern the high-risk group among ypN1 patients.


Introduction
The absolute number of metastatic lymph nodes (LNs) has been considered as an important prognostic factor for colorectal cancer [1][2][3]. In addition to the number of metastatic LNs, the number of examined LNs has been shown to be an independent prognostic factor for survival [4]. Meanwhile, neoadjuvant chemoradiotherapy (NCRT) followed by total mesorectal excision (TME) has become the treatment of choice for patients with LN-positive rectal cancer. This NCRT can result in a significant decrease in the number and size of examined LNs in the TME specimen [5]. Consequently, the number of examined LN is frequently below the recommended number of 12, regardless of the quality of TME and pathologic analysis. Hence, looking for a new method that can overcome the problem of overall LN retrieval was essential. The lymph node ratio (LNR), which is the ratio of metastatic to examined LNs, has recently been proposed as a prognostic factor in patients with stage III colorectal cancer [6,7]. Recently, several studies have also evaluated the prognostic value of LNR in ypN-positive rectal cancer patients who were treated with NCRT [8][9][10][11][12]. However, all of these previous studies did not evaluate the impact of LNR with the consideration of tumor regression grade (TRG) or circumferential resection margin (CRM). As TRG [13,14] and CRM [15,16] are being regarded as important prognostic factors nowadays and the association of the LNR and the distant metastasis has not been fully evaluated in these studies, the contemporary prognostic impact of LNR among rectal cancer patients who underwent NCRT has yet to be proven.
For that reason, we assess the impact of the LNR with the consideration of TRG and CRM in predicting survival and recurrence in ypStage III rectal cancer patients after NCRT.

Materials and Methods Patients
The institutional review boards (IRBs) of both Seoul National University Hospital (SNUH) and Seoul National University Bundang Hospital (SNUBH) approved this study. Because this study was carried out retrospectively, the IRBs waived the written informed consent from patients. And patient information was anonymized and de-identified prior to analysis. Between April 2004 and April 2011, 421 and 217 patients with rectal cancer received NCRT at the SNUH and the SNUBH, respectively. We selected patients who met the following inclusion criteria: pathologically confirmed primary rectal cancer, performance of TME, any ypT and ypN positivity, absence of distant metastasis at diagnosis, no history of other malignancies, and follow-up time of 12 months or more. A total of 125 patients remained and their medical records were retrospectively reviewed.

Pre-treatment evaluation
At the initial staging work-up, digital rectal examination (DRE), colonoscopy, chest radiography, computed tomography (CT) of the abdomen and pelvis, and magnetic resonance imaging (MRI) of the pelvis were performed for all patients. If patients were suspected to have distant metastasis, MRI of the liver or positron emission tomography was carried out. Cancer stages were scored according to the American Joint Committee on Cancer (AJCC) Staging System, seventh edition. A tissue biopsy of the primary lesion was performed for pathologic confirmation. Complete blood counts (CBC) and liver function tests were included in the initial workup. Carcinoembryonic antigen (CEA) levels were measured before NCRT and 4 weeks from the end of NCRT.

Treatment
Preoperative radiotherapy consisted of whole pelvic irradiation and primary tumor boost. The whole pelvis was irradiated with a dose of 45 Gy at 1.8 Gy/fraction, and the primary tumor received boost radiotherapy with doses of 5.4 Gy at 1.8 Gy/fraction. All patients received preoperative radiotherapy in the prone position. The treatment method of radiotherapy has been described previously [17,18]. During preoperative radiotherapy, all patients received chemotherapy concurrently. The chemotherapy regimens were 5-fluorouracil (5-FU; n = 67); 5-FU and leucovorin (FL) (n = 18); capecitabine (n = 33); capecitabine and oxaliplatin (n = 5); and cetuximab, irinotecan, and capecitabine (n = 2).

Pathologic examination
The entire TME specimens including mesorectal fat was serially sliced into 4-mm-thick sections and embedded in paraffin. TRG was assessed by Dworak system. The shortest distance from the outermost part of the tumor to the CRM was measured histologically. The CRM was considered positive, if tumor was located 1 mm or less from the surface of the specimen. All retrieved LNs were analyzed histologically. Metastatic LNs were defined as nodal tissue containing aggregates of tumor cells >0.2 mm in diameter.

Follow-up
Patients were followed up every 3 months for the first 2 years, every 6 months for the next 3 years, and yearly thereafter. Follow-up evaluations included a clinical examination, DRE, CBC, liver function test, and assessment of CEA level at each visit. Chest x-ray and abdominal and pelvic CT scan were conducted every 6 months, and a colonoscopy was performed at 1, 3, and 5 years after surgery. Locoregional recurrence was defined as recurrence detected in the pelvis. Recurrence outside the pelvis was considered as distant metastasis.

Statistical analysis
Overall survival (OS) was calculated as the time from the date of first treatment to the date of death. Disease-free survival (DFS) was defined as the time from the beginning of NCRT to the date of first disease recurrence, either locoregional failure or distant metastasis. Locoregional recurrence-free survival (LRRFS) and distant metastasis-free survival (DMFS) represented the interval from the first date of NCRT to the detection dates of locoregional recurrence and distant metastasis, respectively. Survival curves were generated using the Kaplan-Meier method. Log-rank tests were used to examine the univariate association between outcomes and the following clinicopathlogic factors: age, sex, type of surgery, pre-NCRT or post-NCRT CEA level, ypT stage, ypN stage, number of examined LNs, number of metastatic LNs, LNR, CRM, TRG, histologic grade, angiolymphatic invasion (ALI), venous invasion (VI), and perineural invasion (PNI). Multivariate Cox hazards analyses for OS, DFS, LRRFS, and DMFS were used to adjust comparisons for various factors including LNR, pN stage, CRM, and TRG.
LNR was defined as follows: the ratio of the number of metastatic LNs to the number of total examined LNs. The patients were categorized into two groups based on their LNR with a cut-off value of 0.152. This cut-off value was chosen with using Maxstat, the maximally selected rank method in R 2.13.0 (R Development Core Team, Vienna, Austria; http://www.R-project. org; [19]). All p values reported are two-sided, with p < 0.05 used to denote statistical significance.

Discussion
The main weakness of the number-based AJCC pN stage is that the prognostic accuracy can be profoundly influenced by the total number of LNs retrieved. The AJCC recommends a minimum of 12 LNs to ensure adequate LN retrieval and accurate staging. However, insufficient retrieval and examination of LNs are usual in clinical practice. In the circumstances of routine use of NCRT nowadays, patients often have < 12 LNs retrieved, despite the maintenance of all surgical standards [20]. This led to develop a new prognostic index, the LNR, that incorporates all the information about LNs in a single identifiable parameter. The LNR has been identified as a promising classification index in other malignancies including breast, pancreatic, and gastric cancer [21][22][23]. Furthermore, several studies have demonstrated that the LNR is superior to the pN stage in colorectal cancer [24][25][26]. In terms of rectal cancer, several studies have also analyzed the significance of LNR among patients who underwent adjuvant chemoradiotherapy [6,7,27] and NCRT [8][9][10][11][12] [9]. The results of NCRT studies and the present study are summarized in Table 4.
In the present study, the prognostic value of LNR was assessed in 125 ypN-positive rectal cancer patients treated with NCRT followed by total mesorectal excision and postoperative adjuvant chemotherapy. In a multivariate Cox model which also considered ypN stage, TRG, and CRM, the LNR was an independent prognostic factor and ypN stage was no longer significant. In addition, LNR has the potential to discriminate the high-risk group among patients with the same ypN stage. This finding is in line with that of Lee et al. [9] and Kang et al [12].
Adding the concept of LNR to the ypN stage will improve accuracy of predicting prognosis of rectal cancer. And the LNR can be used as a more useful indicator than the pN stage in terms of guiding the administration of intensified postoperative chemotherapy. Intensified neoadjuvant chemotherapy including oxaliplatin or cetuximab, has failed to improve complete response or survival [28]. A recently reported prospective trial named ADORE compared the effect of postoperative adjuvant FOLFOX with adjuvant FL in rectal cancer patients treated with NCRT [29]. After 38 months of median follow-up time, the FOL-FOX arm showed higher DFS than the FL arm (at 3 years, 71.6% vs. 62.9%; HR 0.657, 95% CI 0.434-0.994, p = 0.047). In particular, the benefits of FOLFOX were more significant for ypN1b patients (HR 0.356, 95% CI 0.132-0.960, p = 0.041) and ypN2 patients (HR 0.414, 95% CI 0.181-0.946, p = 0.037).
Meanwhile, the use of different cut-off values among reports is a limitation of LNR as a prognostic tool. Several methods have been used to determine cut-off values of LNR, including the receiver operating characteristic curve [10], mean or median value [11,12], and atypical selections among several cut-off points [6,8,9,27]. Our study used a maximal chi-square method in R software; this is meaningful in terms of minimizing subjectivity. However, our current study had several limitations, including its relatively small sample size and its retrospective design. A further large-scale prospective study is needed to determine the prognostic value of LNR and its optimal cut-off value in rectal cancer patients who underwent NCRT.

Conclusion
For ypN-positive rectal cancer patients after NCRT followed by TME, LNR had more prognostic value for OS, DFS, and DMFS than the ypN stage, TRG or CRM. The LNR may be used to discern a high-risk group who might benefit from more intensive adjuvant chemotherapy.
Supporting Information S1 File. Original clean data for analysis. (XLSX)