The combination of pre-neoadjuvant chemoradiotherapy inflammation biomarkers could be a prognostic marker for rectal cancer patients

The neutrophil-to-lymphocyte ratio (NLR) and lymphocyte-to-monocyte ratio (LMR) have a strong association with prognosis in patients with Stage II/III rectal cancer (RC). We attempted to explore a new system combining these two ratios, named the NLM score, and examine its prognostic value in Stage II/III RC patients undergoing neoadjuvant chemoradiotherapy (NCRT). We retrospectively analyzed data of 237 stage II/III RC patients who underwent NCRT followed by standard TME in our hospital and defined the NLM score as follows: Score 2: pre-NCRT NLR > 2.565 and pre-NCRT LMR < 2.410. Score 1: (pre-NCRT NLR > 2.565 and pre-NCRT LMR > 2.410) OR (pre-NCRT NLR < 2.565 and pre-NCRT LMR < 2.410). Score 0: pre-NCRT NLR < 2.565 and pre-NCRT LMR > 2.410. Multivariate analyses implied that lower ypTNM stage (stage 0–I vs. II–III) (hazard ratio [HR] 0.420, 95% confidence interval [CI] 0.180–0.980 for OS; HR 0.375, 95% CI 0.163–0.862 for DFS) and an NLM score ≤ 1 (HR 0.288, 95% CI 0.134–0.619 for OS; HR 0.229, 95% CI 0.107–0.494 for DFS) could independently predict better overall survival (OS) and disease-free survival (DFS). The novel scoring system, which integrated pre-NCRT NLR and pre-NCRT LMR, was an independent prognostic factor in stage II/III RC patients undergoing NRCT and had better predictive values than these ratios alone.

Treatment, follow-up and endpoints. The multidisciplinary team meeting (MDT) conducted the treatment plan after considering recommendations of NCCN 12 , ESMO 13 , and JSCCR 14 colorectal cancer guidelines and the patient's physical condition. All included patients received 5-fluorouracil-based chemoradiotherapy, followed by the standard total mesorectum excision (TME) procedure and pathological assessment of specimen [12][13][14] .
Patients were followed up every three months in the first three years after surgery and every six months after that. Blood tests were accomplished at each follow-up. Chest and abdominal CT scans were conducted every six months, with total colonoscopy one year after the operation and every two years afterward [12][13][14] . Clinical, radiological, or histological findings were used to monitor tumor recurrence 13 .
Follow-up data was gained by telephone or directly from outpatient clinic records 15 . Disease-free survival (DFS) and overall survival (OS) were primary endpoints 15 . OS was the time interval from operation date to death date or last visit, DFS is defined as the time interval from operation date to recurrence, metastasis, or last visit date 15,16 . Patients were censored if alive at the last follow-up 16 . Statistical analysis. Statistical analysis was performed by SPSS 22.0. Continuous and categorical variables were expressed as median ± interquarteral range and patients' numbers (%), respectively. Association between NLR or LMR and CEA and CA19-9 were assessed by Kaplan-Meier analysis with log-rank test was used for OS and DFS evaluation 17 .
Cox proportional hazard regression modeling was conducted for univariate (UV) and multivariate (MV) analyses 18 . UV analysis was performed to evaluate potential risk factors that could be associated with prognosis based on previous research and clinical knowledge 19 . Variables with a p value < 0.10 were included in MV regression analyses 20 . To avoid NLR and LMR's influence on the NLM score in the MV analysis, two models excluding and including the NLM score were established.
All statistical tests were bilateral, and 5% was set as the level of statistical significance [18][19][20] .
Univariate and multivariate analyses for DFS and OS. On univariate analysis, ypTNM stage, pre-NCRT CEA, pre-NCRT CA19-9, pre-NCRT NLR, pre-NCRT LMR and NLM score were significantly associated with OS (Table 2) and DFS ( Relationship between NLR, LMR and CEA CA19-9 or other clinicopathological factors. Relationship between NLR, LMR and other clinicopathological factors were presented in Tables 4 and 5. Pre-NCRT NLR or LMR had no significant association between other factors.

Discussion
Our research focused on the relationship between pre-NCRT NLR, pre-NCRT LMR, NLM score and OS and DFS in RC patients receiving NCRT. The main finding was that the NLM score was an independent prognostic factor in RC patients and was superior to pre-NCRT NLR and pre-NCRT LMR alone for predicting prognosis. Previous studies explored the prognostic value of blood cell ratios for RC patients [21][22][23] . Shen et al. 24 investigated 199 patients with locally advanced RC treated with NCRT followed by surgery. They presented a high NLR (≥ 2.8) independently related to poor OS in MV analysis but not to DFS 24 . The relationship between NLR and prognosis and the cut-off value was similar to ours, but their median follow-up period was relatively short (31 months). Carruthers et al. 25 assessed 115 patients, showing that high NLR (≥ 5.0) was an independent prognostic factor for worse OS, decreased time to local recurrence, and shorter DFS 25 . Their results were consistent with ours in UV analysis. However, they only focused on the prognostic value of the single blood index, not comparing it to others nor combing it with others for better prognosis prediction.
Zhang et al. 26 analyzed 472 LARC patients undergoing NCRT and radical surgery. They indicated a high NLR (> 2.3) was an independent predictor for OS and DFS in MV analysis, while LMR was only in UV analysis 26 . We found that NLR was an independent predictor for only DFS, while LMR was neither. Although we had similar www.nature.com/scientificreports/ results, they did not combine NLR and LMR for a more powerful prognostic factor. Our study presented that the NLM score had a better prognostic value than both NLR and LMR for both OS and DFS. Although the mechanism of inflammation index affecting the prognosis of RC is not clear, some meaningful progress has been made. Tumor-associated neutrophils (TANs) origin from peripheral neutrophils 27,28 . They play an essential role in tumor progression since they could promote tumor growth, cause genetic instability and stimulate angiogenesis 27,28 . Tumor-associated macrophages (TAMs) are derived from circulating monocytic precursors and are essential in tumor progression's inflammatory microenvironment 29 . TAMs can produce angiogenesis and growth factors and protease enzymes, which promote extracellular matrix degradation, angiogenesis, tumor cell proliferation, and metastasis 30 . Different from monocytes and neutrophils, lymphocytes is essential in host cell-mediated immune regulation, which helps destroy residual malignant cells and related micrometastases 29 . Temporarily, tumor-infiltrating lymphocytes are related to improving the clinical prognosis of cancer [27][28][29][30] . All of the mechanisms we examined may explain why patients with a high NLR or a low LMR have poor prognosis.
Although similar studies focused on the prognostic value of immune indexes, they only focused on single factors and did not integrate these biomarkers nor compare their prognostic values 8,21,22,25,26,29 . This study established a novel scoring system combining pre-NCRT NLR and pre-NCRT LMR, named NLM score. The NLM score was superior to single ratios for predicting RC prognosis. Our research found that pre-NCRT NLR has a better prognosis prediction effect than pre-NCRT LMR, indicating neutrophils and monocyte could have different effects on tumor micrometastasis. More profound research in this area may be valuable.
This study had several limitations. Firstly, it was a retrospective study conducted at a single center. Secondly, in this analysis, the optimal cut-off points for pre-NCRT NLR and pre-NCRT LMR were 2.565 and 2.410, respectively. Nevertheless, the cut-off point in this study may only apply to our center's population. If doctors from other medical centers try to apply this prognostic scoring system, we recommend conducting their analysis to determine the optimal cut-off point for a particular patient group. Thirdly, we only enrolled the pre-NCRT NLR and pre-NCRT LMR into our prognostic scoring system. Other systemic inflammatory biomarkers still have high prognostic significance, such as the prognostic nutritional index.22,33 Regrettably, they were not available regularly in our department. Future research on the prognostic scoring system should include as many systemic inflammation indices as possible.

Conclusion
In conclusion, we established a novel prognostic system combing pre-NCRT NLR and pre-NCRT LMR for RC patients. We demonstrated an NLM score ≤ 1 could independently predict better survival. Further studies are necessary to verify its prognostic value.