Abstract
For resectable colorectal cancer (CRC), the standard treatment is perioperative adjuvant treatment and complete resection. For unresectable CRC, the standard treatment is systemic chemotherapy. The detection of promising biomarkers is necessary for optimizing the management of CRC and improving patient survival. If physicians can detect useful biomarkers, patients with CRC may benefit from more aggressive or less toxic treatment. Recent studies have shown that the inflammatory and nutritional status both influence the short and long-term oncological outcomes of patients with CRC during perioperative and/or chemotherapy. The utility of several tools for the evaluation of the inflammation and nutritional status has been reported. The introduction of such tools in the management of CRC could have a beneficial impact on postoperative surgical complications or adverse events of chemotherapy. An understanding of the characteristics of each of these evaluations is necessary for their introduction in daily clinical practice. The present report summarizes the background and current status of nutrition and inflammation evaluation tools and future perspectives on their application in the management of patients with CRC.
Colorectal cancer (CRC) is the third most common cancer and the second leading cause of cancer-related deaths worldwide (1, 2). Curative resection and perioperative adjuvant treatment are standard treatments for resectable CRC, while chemotherapy is the standard treatment for unresectable CRC (3, 4). The prognosis of CRC is gradually improving, and the survival of more than half of patients with CRC is limited and poor after diagnosis. Therefore, CRC patients with risk factors for recurrence require more aggressive treatment and management.
Recently, the nutritional and inflammatory status during treatment has been recognized as an important factor for both short- and long-term oncological outcomes in various malignancies (5-7). So far, various nutritional and inflammation assessment tools, such as the Glasgow Prognostic Score (GPS), Prognostic Nutritional Index (PNI), Controlling Nutritional Status (CONUT score), C-reactive protein Albumin ratio (CAR), and Albumin/Globulin Ratio (AGR), have been developed and evaluated in gastrointestinal cancers, including CRC. If physicians introduce these nutritional and inflammation assessment tools in daily clinical practice, CRC patients may be able to receive optimized treatments that are more aggressive in comparison to standard treatment. However, each nutritional and inflammatory evaluation tool has its own clinical characteristics. Therefore, physicians need to understand the characteristics of each nutritional and inflammation evaluation tool before their introduction into daily clinical practice.
This review summarizes the background, current status, and future perspectives of nutritional and inflammatory assessment tools for CRC treatment.
Clinical Impact of the Glasgow Prognostic Score (GPS) for CRC Treatment
The Glasgow Prognostic Score (GPS) was first reported by Forrest et al. in 2003. GPS was calculated using serum C-reactive protein level (cutoff value: 1.0 mg/dl) and serum albumin level (cutoff value: 3.5 g/dl). Subsequently, a modified GPS (mGPS) was developed. Forty-four studies have evaluated the clinical impact of the GPS (20 studies) and mGPS (24 studies) in patients with CRC (Table I) (8-51). Among them, 35 studies evaluated resectable CRC and 9 evaluated unresectable CRC. Ishizuka et al. first reported the clinical impact of the GPS in 315 patients with resectable CRC. The patients were divided into GPS 0 (n=183), GPS 1 (n=89), and GPS 2 (n=43) groups. There were significant differences in the mean survival time of the three groups. The mean survival time was 21.0 months in the GPS 0 group, 13.0 months in the GPS 1 group and 13.7 months in the GPS 2 group. In a multivariate analysis, the GPS was identified as an independent prognostic factor [hazard ratio (HR)=0.165; 95% confidence interval (CI)=0.037-0.732; p=0.0177]. Among 35 studies, the HR of the GPS for resectable CRC ranged from 1.28 to 12.06. Read et al. evaluated the clinical impact of the GPS in 51 patients with unresectable CRC. The patients were divided into GPS 0 (n=15), GPS 1 (n=26), and GPS 2 (n=7) groups. There were significant differences (p=0.036) in the mean survival times of the 3 groups. The mean survival time was 14.0 months in the GPS 0 group, 10.8 months in the GPS 1 group, and 7.9 months in the GPS 2 group. In a multivariate analysis, the GPS was identified as an independent prognostic factor (HR=2.27, 95%CI=1.09-4.73, p=0.028). Among the 9 studies, the HR of the GPS for unresectable CRC ranged from 1.62 to 7.603. The GPS is used as a predictive marker for anastomotic leakage in patients with CRC. Golshani et al. identified the GPS as a predictive marker for anastomotic leakage in 418 patients with CRC (52). The patients were divided into mGPS 0 (n=354), GPS 1 (n=45), and GPS 2 (n=19) groups. The incidence of anastomotic leakage was 15.8% in the mGPS 0 group, 22.2% in the mGPS 1 group, and 42.1% in the mGPS2 group. The odds ratios (reference to mGPS 0) were 1.09 in the mGPS 1 group (95%CI=0.53-2.25) and 4.11 in the mGPS2 group (95%CI=1.69-10.03). In addition, Iwasaki et al. used the GPS as a marker for conversion from laparoscopic surgery to open surgery (53).
Clinical impacts of Glasgow Prognostic Score (GPS) for colorectal cancer treatment.
Clinical Impact of the Prognostic Nutritional Index (PNI) on CRC Treatment
The PNI was first reported by Onodera et al. (1984). The PNI is calculated using the serum albumin and lymphocyte count. Twenty-eight studies have evaluated the clinical impact of the PNI in patients with CRC (Table II) (54-81). Among them, 23 studies evaluated resectable CRC and 5 evaluated unresectable CRC. In the resectable setting, Nozoe et al. first reported the clinical impact of the PNI in 219 patients with CRC. They divided CRC patients into PNI-low (n=23) and PNI-high (n=196) groups using a cutoff value of 40. When comparing the 1-, 3-, and 5-year overall survival (OS) rates between the two groups, there were significant differences. The 1-, 3-, and 5-year overall rates were 81.0%, 48.9%, and 32.6%, respectively, in the PNI-low group, and 96.2%, 87.5%, and 85.3% in the PNI-high group. In a multivariate analysis, the PNI was identified as an independent prognostic factor (HR=1.386; 95%CI=1.66-9.62; p=0.002). Among 23 studies, the HR of the PNI for resectable CRC ranged from 1.73 to 6.728. The reported cutoff values of the PNI ranged from 34.1 to 49.8. In the unresectable setting, Ikeya et al. evaluated the clinical impact of the PNI in 80 patients with unresectable CRC. They divided CRC patients into PNI-low (n=30) and PNI-high (n=50) groups using a cutoff value of 44.5. There were significant differences (p=0.036) in the mean survival between the two groups. The Medina CRC patient’s survival time was 37.0 months in the PNI-low group and 22.8 months in the PNI-high group (p=0.005). In a multivariate analysis, the PNI was identified as an independent prognostic factor (HR=2.373; 95%CI=1.355-4.148, p=0.002). Among the five studies, the HR of the PNI for unresectable CRC ranged from 2.373 to 2.46. The reported cutoff values of the PNI ranged from 40-51.35. The PNI has also been used as a predictive marker for postoperative surgical complications in patients with CRC. Tominaga et al. clarified that PNI is a predictive marker for postoperative surgical complications in 896 patients with CRC. They divided CRC patients into PNI-low (n=382) and PNI-high (n=514) groups using a cutoff value of 49.8. The incidence of postoperative surgical complications was 14.9% in the PNI-low group and 8.2% in the PNI-high group. The odds ratio was 1.913 (95%CI=1.264-2.897). Recently, differences in PNI during the perioperative period have been evaluated in patients with CRC. Lee et al. evaluated the prognostic impact of PNI changes during preoperative chemoradiation in 261 patients with rectal cancer (82). They calculated the PNI differences (dPNI) using the PNI status before and after preoperative chemoradiation. They found that almost half of the patients had moderate to severe dPNI. They demonstrated that patients with moderate-to-severe dPNI values had a poorer prognosis than those with low dPNI values. The hazard ratio (HR) for OS was 1.705-2.792.
Clinical impacts of Prognostic Nutritional Index (PNI) for colorectal cancer treatment.
Clinical Impact of the Geriatric Nutritional Risk Index (GNRI) on CRC Treatment
GNRI is calculated using the following formula: GNRI = [1.487 × serum albumin (g/l)]+[41.7 × actual/usual body weight) (kg)]. Fourteen studies evaluated the clinical impact of the GNRI in patients with CRC (Table III) (83-96). Among them, 12 studies evaluated the surgical setting, and two studies evaluated the chemotherapy setting. In the surgical setting, Iguchi et al. first reported the clinical impact of the GNRI in 80 CRC patients with liver metastasis. They divided patients with CRC into a GNRI-low group (n=30) and a GNRI-high group (n=50) using a cutoff value of 98. When comparing OS and RFS between the two groups, OS and RFS were significantly poorer in the GNRI-low group than in the GNRI-high group. In a multivariate analysis, the GNRI was identified as an independent prognostic factor for OS (HR=3.725, 95%CI=1.409-9.847, p=0.008) and RFS (HR=2.401, 95%CI=1.090-5.290, p=0.030). Among the 11 studies, the HR of the GNRI in the surgical setting ranged from 1.329 to 5.551. The cutoff value of the PNI ranged from 83.13 to 104.25. In the chemotherapy setting, Guc et al. evaluated the clinical impact of the GNRI in 185 patients with unresectable CRC. They divided CRC patients into a GNRI-low group (n=116) and a GNRI-high group (n=69) using a cutoff value of 107.28. When comparing the median survival between the two groups, there were significant differences (p<0.001) in the median survival between the two groups. The median survival time was 23 and 46 months in the low and high GNRI groups, respectively. In a multivariate analysis, the GNRI was identified as an independent prognostic factor (HR=2.22, 95%CI=1.55-3.17, p<0.001). In the two studies, the HR of the GNRI for unresectable CRC was reported to be 0.64 and 2.22, respectively. The cutoff values of the PNI were 97.3 and 107.28. The GNRI is also used as a prognostic marker for patients with early CRC who underwent endoscopic submucosal dissection. Kato et al. identified the GNRI as a prognostic marker in 729 patients with CRC. They divided CRC patients into a GNRI-low group (n=382) and a GNRI-high group (n=514) using a cutoff value of 96.3. In a multivariate analysis, the GNRI was identified as an independent prognostic factor for OS (HR=3.37, 95%CI=2.18-5.22, p<0.001). The GNRI is also a useful prognostic marker for patients with early CRC.
Clinical impacts of Geriatric Nutritional Risk Index (GNRI) for colorectal cancer treatment.
Clinical Impact of the Platelet-to-lymphocyte Ratio (PLR) on CRC Treatment
The PLR is calculated by dividing the platelet count by the lymphocyte count. Twenty studies evaluated the clinical impact of the PLR in patients with CRC (Table IV) (97-116). Among them, 15 studies evaluated the surgical setting, and five studies evaluated the chemotherapy setting. Neofytan et al. first reported the clinical impact of the PLR in 140 CRC patients with liver metastasis. They divided CRC patients into a PLR-low group (n=82) and a PLR-high group (n=58) using a cutoff value of 150. When comparing the OS and RFS between the two groups, OS and DFS were significantly poorer in the PLR-high group. The 3-, and 5-year OS rates were 62% and 40%, respectively, in patients with PLR >150, and 79% and 70%, respectively, in patients with PLR ≤150. In a multivariate analysis, PLR was identified as an independent prognostic factor for OS (HR=2.17, 95%CI=1.09-4.32, p=0.027) and DFS (HR=1.68, 95%CI=1.04-2.71, p=0.034). Among the 11 studies, the HR of the PLR for surgical settings ranged from 1.356 to 5.031. The cutoff value of the PNI ranged from 25.4 to 250. Ramos et al. evaluated the clinical impact of the PLR in 110 patients with unresectable CRC. They divided CRC patients into a PLR-low group (n=53) and PLR-high group (n=57) using a cutoff value of 107.28. When comparing the median survival between the two groups, there were significant differences (p<0.001) in the median survival. The median survival time was 25 and 19 months in the PLR-low and PLR-high groups, respectively. In a multivariate analysis, the PLR was identified as an independent prognostic factor for OS (HR=2.35, 95%CI=1.45-3.80, p<0.001) and RFS (HR=1.55, 95%CI=1.01-2.40, p=0.04). Among the three studies, the HR of the PLR for unresectable CRC ranged from 0.676 to 10.28. The cutoff value of the PNI ranged from 172.4 to 207.29. The PLR is also used as a predictive marker for the chemotherapy response of patients with CRC. Acikgoz et al. evaluated the PLR as a treatment response marker in 229 patients with CRC who received chemotherapy. They divided CRC patients into PLR-low (n=116) and PLR-high (n=113) groups using a cutoff value of 196.5. In a multivariate analysis for treatment response, the PLR was identified as an independent treatment response marker (HR=3.97, 95%CI=2.00-7.88, p<0.001).
Clinical impacts of platelet-to-lymphocyte ratio (PLR) for colorectal cancer treatment.
Clinical Impact of the Naples Prognostic Score (NPS) on CRC Treatment
The NPS was calculated based on the following four parameters: serum albumin (normal: ≥4 g/dl), total cholesterol (normal: >180 mg/dl), LMR (normal: ≤2.96), and NLR (normal: >4.44). Patients were divided into three groups. Patients with normal values for all four parameters were assigned a score of 0; those with one or two altered values were assigned a score of 1; and those with three or four altered values were assigned a score of 2. Four studies evaluated the clinical impact of the NPS in patients with CRC (Table V) (117-120). All studies evaluated patients in the surgical setting. In a prognostic factor analysis, Galizia et al. first reported the clinical impact of NPS in 562 patients with CRC who received curative treatment. They divided patients with CRC into NPS score 0 (n=117), NPS score 1 (n=274), and NPS score 2 (n=171) groups. When comparing the five-year OS rates among the three groups, there were significant differences. The five-year OS rate was 88% in the NPS score 0 group, 68% in the NPS score 1 group, and 34% in the NPS score 2 group. In a multivariate analysis, the NPS was identified as an independent prognostic factor (HR=2.90, 95%CI=1.57-5.35 for NPS score 1) (HR=8.01, 95%CI=4.38-14.65 for NPS score 2). The NPS is also used as a predictive marker for postoperative complications in patients with CRC. Sugimoto et al. identified the NPS as a predictive marker for postoperative complications in 235 patients with CRC. They divided CRC patients into an NPS 0-1 group (n=127) and an NPS 2 group. The incidence of postoperative complications was 17.3% in the NPS 0-1 group and 38.9% in the NPS 2 group. The odds ratio for postoperative complications (reference to NPS 0-1) was 3.04 in the NPS 2 group (95%CI=1.55-5.96).
Clinical impacts of Naples Prognostic Score (NPS) for colorectal cancer treatment.
Clinical Impact of the Lymphocyte to C-reactive Protein Ratio (LCR) on CRC Treatment
The lymphocyte-C-reactive protein ratio (LCR) was calculated as the absolute lymphocyte count divided by the absolute C-reactive protein level. Four studies evaluated the clinical impact of the LCR in patients with CRC (Table VI) (121-124). Among them, three studies evaluated patients in the surgical setting and one study evaluated patients in the chemotherapy setting. Okugawa et al. first reported the clinical impact of the LCR in 307 patients with CRC who underwent curative resection. They divided CRC patients into LCR-low and LCR-high groups using a cutoff value of 6676. When comparing OS and DFS between the two groups, OS and DFS was significantly poorer in the LCR-low group. In a multivariate analysis, the LCR was identified as an independent prognostic factor for OS (HR=5.21, 95%CI=2.42-11.2, p<0.001) and DFS (HR=1.88, 95%CI=1.07-3.31, p=0.02). Among the three studies in the surgical setting, the HR of the LCR ranged from 0.61 to 5.21. The cutoff value of the LCR ranged from 6,500 to 6,676. In the chemotherapy setting, Nakamura et al. evaluated the clinical impact of the LCR in 756 patients with unresectable CRC. They divided CRC patients into an LCR-high group (n=251), an LCR-intermediate group (n=253), and an LCR-low group (n=252). There were significant differences in the median survival of the three groups. The median survival time was 29.4 months in the LCR-high group, 19.3 months in the LCR-intermediate group and 13.1 months in the LCR-low group. In a multivariate analysis, the LCR was identified as an independent prognostic factor (HR=1.44, 95%CI=1.17-1.78, p<0.001 for LCR intermediate group; HR=1.96, 95%CI=1.58-2.42, p<0.001 for LCR low group).
Clinical impacts of lymphocyte to C-reactive protein ratio (LCR) for colorectal cancer treatment.
Clinical Impact of the Controlling Nutritional Status (CONUT) Score on CRC Treatment
The CONUT consists of three parameters: the serum albumin level, the cholesterol level, and the total lymphocyte count. Patients with CONUT scores of 0-1 have a normal nutritional status (CONUT 0), those with CONUT scores of 2-4 are at mild risk of malnutrition (CONUT 1), those with CONUT scores of 5-8 are at moderate risk (CONUT 2), and those with CONUT scores of 9-12 are at severe risk of malnutrition (CONUT score 3). Twelve studies evaluated the clinical impact of the CONUT score in patients with CRC (Table VII) (125-136). Among them, 10 studies evaluated patients in the surgical setting and two studies evaluated patients in the chemotherapy setting. In the surgical setting, Iseki et al. first reported the clinical impact of the CONUT score in 204 CRC patients who underwent curative surgery. They divided CRC patients into a CONUT score-low group (n=150) and a CONUT score-high group (n=54) using a cutoff value of 3. When comparing RFS and cancer-specific survival (CSS) between the two groups, RFS and CSS were significantly poorer in the CONUT score-high group than in the CONUT score-low group. The five-year RFS and CSS rates were 73.0% and 92.7%, respectively, in the CONUT score-high group and 53.6% and 81.0% in the CONUT score-low group. In a multivariate analysis, the CONUT score was identified as an independent prognostic factor for RFS (HR=2.04, 95%CI=0.862-3.989, p=0.0621) and CSS (HR=3.637, 95%CI=1.071-10.915, p=0.0393). Among the 10 studies conducted in the surgical setting, the HR of the CONUT score ranged from 1.792 to 10.2. The cutoff value of the CONUT score ranged from 2 to 7. In the chemotherapy setting, Daitoku et al. evaluated the clinical impact of the CONUT score in 211 patients with unresectable CRC. They divided CRC patients into a CONUT score low group (n=179) and a CONUT score high group (n=32) using a cutoff value of 5. When comparing the median survival between 2 groups, there were significant differences (p<0.001). Five-year survival rates were 21.4% to 22.4% in the CONUT score-low group and 9.1% in the CONUT score-high group. In a multivariate analysis, CONUT score was identified as an independent prognostic factor for OS (HR=2.01, 95%CI=1.26-3.12, p<0.05). In the two studies, the HRs of the CONUT score for unresectable CRC were 1.57 and 2.01, respectively. The cutoff values of the CONUT score were 4 and 5, respectively.
Clinical impacts of Controlling Nutritional Status (CONUT) score for colorectal cancer treatment.
Clinical Impact of the Albumin-Globulin Ratio (AGR) on CRC Treatment
The AGR was determined by assessing a preoperative blood sample and dividing the serum albumin level by the serum globulin level, which was calculated as the difference between the serum total protein level and the serum albumin level. Four studies evaluated the clinical impact of the AGR in patients with CRC (Table VIII) (137-140). Among them, three studies evaluated patients in the surgical setting and one study evaluated patients in the chemotherapy setting. Fujikawa et al. first reported the clinical impact of the AGR in 248 CRC patients who underwent curative resection. They divided CRC patients into AGR-low and AGR-high groups using a cutoff value of 1.32. When comparing OS and DFS between the two groups, DFS and OS were significantly shorter in patients in the AGR-low group. In a multivariate analysis, the AGR was identified as an independent prognostic factor for OS (HR=2.93; 95%CI=1.34-6.69, p=0.0072). Among the three studies, the HR of the AGR in the surgical setting ranged from 0.23 to 2.67. The cutoff value of the AGR ranged from 1.32 to 1.615. Shibutani et al. evaluated the clinical impact of the AGR in 66 patients with unresectable CRC. They divided patients with CRC into AGR-high (n=32) and AGR-low (n=34) groups sing a cutoff value of 1.25. When comparing the median survival between the two groups, DFS and OS were significantly shorter in patients in the AGR-low group. In a multivariate analysis, the AGR was identified as an independent prognostic factor for OS (HR=2.247, 95%CI=1.069-4.722, p=0.033).
Clinical impacts of Albumin-globulin Ratio (AGR) for colorectal cancer treatment.
Clinical Impact of C-Reactive Protein-to-Albumin Ratio (CAR) on CRC Treatment
The CAR was calculated as the serum CRP level (mg/dl) divided by the serum albumin level (g/dl). Twenty-three studies evaluated the clinical impact of the CAR in patients with CRC (Table IX) (134, 141-162). Among them, 20 studies evaluated patients in the surgical setting and three studies evaluated patients in the chemotherapy setting. Shibutani et al. first reported the clinical impact of the CAR in 705 CRC patients who underwent curative resection. They divided CRC patients into CAR-low (n=347) and CAR-high (n=358) groups using a cutoff value of 0.0271. When comparing RFS and CSS between the 2 groups, RFS and CSS were significantly poorer in the CAR-high group. In a multivariate analysis of RFS and CSS, the CAR was identified as an independent prognostic factor for RFS (HR=1.503, 95%CI=1.054-2.143, p=0.025) and CSS (HR=1.672, 95%CI=1.012-2.764, p=0.045). Among the 20 studies, the HR of the CAR in the surgical setting ranged from 0.199 to 5.09. The cutoff value of the CAR ranged from 0.024 to 4.33. Shibutani et al. evaluated the clinical impact of CAR in 99 patients with CRC who received palliative chemotherapy. They divided CRC patients into CAR-low (n=63) and CAR-high (n=36) groups using a cutoff value of 0.183. There were significant differences in OS between the two groups (p=0.0009). In a multivariate analysis, the CAR was identified as an independent prognostic factor for OS (HR=2.35, 95%CI=1.45-3.80, p<0.001) and RFS (HR=1.866, 95%CI=1.057-3.295, p=0.031). The CAR has been reported to be a predictor of adverse events associated with adjuvant chemotherapy. Tominaga et al. evaluated the clinical impact of the CAR in 136 patients with CRC who received adjuvant chemotherapy. They divided CRC patients into CAR-low (n=106) and CAR-high (n=30) groups using a cutoff value of 0.1. A multivariate analysis showed that CAR-high status was a significant determinant of severe side effects of adjuvant chemotherapy (odds ratio=7.06, 95%CI=2.51-19.88, p<0.01).
Clinical impacts of C-Reactive Protein-to-Albumin Ratio (CAR) for colorectal cancer treatment.
Future Perspectives on the Application of Nutrition and Inflammation Evaluation Tools in the Management of Patients With Colorectal Cancer
A number of reports have evaluated various nutrition and inflammation evaluation tools in the management of patients with colorectal cancer. Since various cutoff values have been reported, further studies are required to determine the optimal cutoff value of each tool before they can be applied in the clinical setting. The different cutoff values may be attributable to differences in patient background factors or heterogeneity in the methods of treatment and evaluation. Furthermore, the optimal timing for the application of each tool has not been reported. The time points at which each tool was applied in previous studies included the time of the diagnosis, first visit, before surgery, after surgery, and prior to the initiation of chemotherapy. Therefore, the ideal time to apply these tools must be established. The mechanisms underlying the effect of nutrition and inflammation on the prognosis of gastric cancer also remain to be determined. Recently, it was reported that the nutrition and inflammation status affect postoperative surgical complications, chemotherapy introduction, and chemotherapy adverse events. The survival of colorectal cancer patients has been reported to be affected by postoperative surgical complications and chemotherapy management. The underlying mechanisms by which the nutritional and inflammatory status (as assessed by these tools) influence the prognosis of patients with colorectal cancer remain to be determined.
Conclusion
Nutritional and inflammatory status may have some clinical influence on both short- and long-term oncological outcomes in colorectal cancer patients. However, the optimal cutoff values of each nutrition and inflammation assessment tool and the mechanism through which these parameters influence the prognosis are unclear. To optimize nutrition and inflammation assessment tools for patients with colorectal cancer, it is necessary to clarify these points in further studies.
Acknowledgements
This study was supported in part by the nonprofit organization Yokoyama Surgical Research Group (YSRG).
Footnotes
Authors’ Contributions
TA, NY, and AS substantially contributed to the concept and design. TA and NY made substantial contributions to the acquisition of data and the analysis and interpretation of the data. TA, NY, and AS were involved in drafting the article and revising it critically for important intellectual content. TA and NY approved the final version to be published.
Conflicts of Interest
The Authors declare no conflicts of interest in association with the present study.
- Received January 17, 2024.
- Revision received February 7, 2024.
- Accepted February 9, 2024.
- Copyright © 2024 The Author(s). Published by the International Institute of Anticancer Research.
This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC-ND) 4.0 international license (https://creativecommons.org/licenses/by-nc-nd/4.0).






