Receiver Operating Characteristic Prediction for Classification: Performances in Cross-Validation by Example

Segmented neutrophil-to-monocyte ratio and systemic immune-inflammation index associated with the severity and functional prognosis of acute ischemic stroke

PURPOSE/AIM OF THE STUDY

To identify the inflammation indexes associated with the severity and functional prognosis in ischemic stroke.

MATERIAL AND METHODS

A prospective study was conducted with ischemic stroke cases included in the i-ReNe clinical registry. Patients were divided into groups according to the severity on admission measured by the National Institutes of Health Stroke Scale (NIHSS) and the functional prognosis at 30 and 90 days of discharge measured by the modified Rankin Scale (mRS).

RESULTS

We included 145 patients with a mean age of 61.5 ± 12.75, 97 (66.9%) were men. The leukocyte and neutrophil counts, Neutrophil-to-Lymphocyte ratio (NLR), Derived Neutrophil-to-Lymphocyte ratio (dNLR), Platelet-to-Lymphocyte ratio (PLR), Segmented Neutrophil-to-Monocyte ratio (SeMo ratio), and Systemic Immune-inflammation index (SII) were higher in moderate-to-severe stroke (NIHSS ≥6). NLR, PLR, SeMo ratio, and SII were higher in the group with severe disability and death at 30 days (mRS ≥4). In the multiple logistic regression analyses, SeMo ratio >14.966 and SII >623.723 were associated with moderate-to-severe stroke (NIHSS ≥6). In addition, SeMo ratio >7.845 was associated with severe disability and death at 30 days (mRS ≥4).

CONCLUSIONS

Systemic inflammation indexes could be rapid and low-cost markers used in the initial evaluation of ischemic stroke, whose values could help to stratify patients according to their severity and functional prognosis. This is the first study to establish a relationship between ischemic stroke and the SeMo ratio.

Abilities of Pre-Treatment Inflammation Ratios as Classification or Prediction Models for Patients with Colorectal Cancer

Background: Systemic inflammatory status is known as an important factor of colorectal cancer prognosis. Our study aimed to evaluate the performances of inflammation biomarker ratios as classification models of seven outcomes in patients with colorectal cancer. Methods: A retrospective cohort study was conducted on subjects with colorectal cancer over five years at a single center in Transylvania, Romania. Seven derived ratios were calculated based on laboratory data: neutrophil-to-lymphocyte (NLR), derived neutrophil-to-lymphocyte (dNLR), platelet-to-lymphocyte (PLR), lymphocyte-to-monocyte (LMR) and albumin-to-globulin (AGR) ratios, Systemic Immune Inflammation Index (SII) and Prognostic Nutritional Index (PNI). The utility of these ratios as predictors for seven outcomes was further evaluated in multivariable regression models. Results: Our study shows that the evaluated ratios exhibit specific performances for individual outcomes, proving a fair ability as screening tools (NLR and dNLR for survival, T stage and M stage; NLR and SII for T stage; and PLR for M stage). A dNLR over 3.1 (OR = 2.48, 95% CI (1.421 to 4.331)) shows predictive value for survival. A value of NLR over 3.10 (OR = 1.389, 95% CI (1.061 to 1.817)) is positively associated with an advanced T stage, while LMR is negatively related to the T stage (OR = 0.919, 95% CI (0.867 to 0.975)). NLR over 4.25 (OR = 2.647, 95% CI (2.128 to 3.360)) is positively associated with, while PNI is negatively related (OR = 0.970, 95% CI (0.947 to 0.993)) to, the M stage. Conclusion: Each of the evaluated ratios possesses prognostic value for certain outcomes considered, but the reported models need external validation to recommend their clinical practice utilization.