Bicuspid Expansion v5

library(readxl)
library(ggplot2)
library(CGPfunctions)
library(plotly)
library(lmtest)

#Definir directorio
setwd("G:/TRABAJO/CONSULTORIAS/TRABAJOS VARIOS/JORGE CHAVARRIA/analisis3")

data1 = read_excel("Bicuspid3.xlsx")
head(data1,5)

ANALISIS RELATIVE STENT EXPANSION

#Relacion entre Maximum Sinus Diameter Indexed vs Relative Stent Expansion %
g1=ggplot(data=data1,mapping=
            aes(x=MaxSinAnnDcalc,y=RelativeStentExpansion,))+geom_point()+theme_bw()+
            geom_smooth(method = "lm") 

g1 +  labs(title = "Relative Stent Expansion % vs Maximum Sinus Diameter Indexed",
           x = "Maximum Sinus Diameter Indexed",
           y= "Relative Stent Expansion %") 

`geom_smooth()` using formula 'y ~ x'

correlation_test1 <- cor.test(data1$MaxSinAnnDcalc, data1$RelativeStentExpansion)
print(correlation_test1)

    Pearson's product-moment correlation

data:  data1$MaxSinAnnDcalc and data1$RelativeStentExpansion
t = 0.89077, df = 97, p-value = 0.3753
alternative hypothesis: true correlation is not equal to 0
95 percent confidence interval:
 -0.109279  0.282465
sample estimates:
       cor 
0.09007603 

Podemos observar que no existe una correlacion significativa entre MaxSinAnnDcalc y RelativeStentExpansion

#Relacion entre Minimun Sinus Diameter Indexed vs Relative Stent Expansion %
g2=ggplot(data=data1,mapping=
            aes(x=MinSinusAnnDcalc,y=RelativeStentExpansion,))+geom_point()+theme_bw()+
            geom_smooth(method = "lm") 

g2 +  labs(title = "Relative Stent Expansion % vs Minimun Sinus Diameter Indexed",
           x = "Minimun Sinus Diameter Indexed",
           y= "Relative Stent Expansion %") 

`geom_smooth()` using formula 'y ~ x'

correlation_test2 <- cor.test(data1$MinSinusAnnDcalc, data1$RelativeStentExpansion)
print(correlation_test2)

    Pearson's product-moment correlation

data:  data1$MinSinusAnnDcalc and data1$RelativeStentExpansion
t = 1.4692, df = 97, p-value = 0.145
alternative hypothesis: true correlation is not equal to 0
95 percent confidence interval:
 -0.05136679  0.33519012
sample estimates:
      cor 
0.1475412 

#Relacion entre Postdilatation vs Relative Stent Expansion %
boxplot(data1$RelativeStentExpansion~data1$Postdilation,
        xlab = 'Postdilation',
        ylab = 'Relative Stent Expansion %',
        title=('Postdilatation vs Relative Stent Expansion %'),
          col= 'bisque')

t_test1 <- t.test(data1$RelativeStentExpansion ~ as.factor(data1$Postdilation))
print(t_test1)

    Welch Two Sample t-test

data:  data1$RelativeStentExpansion by as.factor(data1$Postdilation)
t = -4.3252, df = 43.026, p-value = 8.89e-05
alternative hypothesis: true difference in means between group 0 and group 1 is not equal to 0
95 percent confidence interval:
 -0.04056690 -0.01476687
sample estimates:
mean in group 0 mean in group 1 
      0.8991558       0.9268227 

Se evidencia una diferencia significativa de RelativeStentExpansion en funcion de la Postdilatacion, si postdilata aumenta la expacion relativa.

ggplot(data = data1) + geom_density(aes(x=RelativeStentExpansion,fill=factor(Postdilation)),
                                    bins=10, position = "identity",alpha = 0.5)

Warning: Ignoring unknown parameters: bins

#Relacion entre Predilatation vs Relative Stent Expansion %
boxplot(data1$RelativeStentExpansion~data1$Predilatation,
        xlab = 'Predilatation',
        ylab = 'Relative Stent Expansion %',
        title=('Predilatation vs Relative Stent Expansion %'),
          col= 'bisque')

ggplot(data = data1) + geom_density(aes(x=RelativeStentExpansion,fill=factor(Predilatation)),
                                    bins=10, position = "identity",alpha = 0.5)

Warning: Ignoring unknown parameters: bins

t_test2 <- t.test(data1$RelativeStentExpansion ~ as.factor(data1$Predilatation))
print(t_test2)

    Welch Two Sample t-test

data:  data1$RelativeStentExpansion by as.factor(data1$Predilatation)
t = -0.73773, df = 71.657, p-value = 0.4631
alternative hypothesis: true difference in means between group 0 and group 1 is not equal to 0
95 percent confidence interval:
 -0.018353678  0.008439107
sample estimates:
mean in group 0 mean in group 1 
      0.9036015       0.9085588 

No existe una diferencia significativa de RelativeStentExpansion en función de la Predilatation

#Relacion entre Raphe Calcification vs Relative Stent Expansion %
boxplot(data1$RelativeStentExpansion~data1$Raphaecalcification,
        xlab = 'Raphae Calcification',
        ylab = 'Relative Stent Expansion %',
        title=('Raphae Calcification vs Relative Stent Expansion %'),
          col= 'ivory')

ggplot(data = data1) + geom_density(aes(x=RelativeStentExpansion,fill=factor(Raphaecalcification)),
                                    bins=10, position = "identity",alpha = 0.5)

Warning: Ignoring unknown parameters: bins

t_test3 <- t.test(data1$RelativeStentExpansion ~ as.factor(data1$Raphaecalcification))
print(t_test3)

    Welch Two Sample t-test

data:  data1$RelativeStentExpansion by as.factor(data1$Raphaecalcification)
t = 0.5185, df = 85.377, p-value = 0.6055
alternative hypothesis: true difference in means between group 0 and group 1 is not equal to 0
95 percent confidence interval:
 -0.009547276  0.016283981
sample estimates:
mean in group 0 mean in group 1 
      0.9074135       0.9040452 

No existe diferencia significativa de RelativeStentExpansion en función del RelativeStentExpansion

#Relacion entre AVAi vs Relative Stent Expansion %
g4=ggplot(data=data1,mapping=
            aes(x=AVAi,y=RelativeStentExpansion,))+geom_point()+theme_bw()+
            geom_smooth(method = "lm") 

g4 +  labs(title = "Relative Stent Expansion % vs AVAi",
           x = "AVAi",
           y= "Relative Stent Expansion %") 

`geom_smooth()` using formula 'y ~ x'
Warning: Removed 24 rows containing non-finite values (stat_smooth).
Warning: Removed 24 rows containing missing values (geom_point).

correlation_test3 <- cor.test(data1$AVAi, data1$RelativeStentExpansion)
print(correlation_test3)

    Pearson's product-moment correlation

data:  data1$AVAi and data1$RelativeStentExpansion
t = 1.9905, df = 73, p-value = 0.05029
alternative hypothesis: true correlation is not equal to 0
95 percent confidence interval:
 -7.606249e-05  4.316249e-01
sample estimates:
      cor 
0.2268897 

Existe correlacion positiva significativa entre AVAi y RelativeStentExpansion

#Relacion entre Mean Gradient (mmHg) vs Relative Stent Expansion %
g5=ggplot(data=data1,mapping=
            aes(x=MeanGradientmmHg,y=RelativeStentExpansion,))+geom_point()+theme_bw()+
            geom_smooth(method = "lm") 

g5 +  labs(title = "Relative Stent Expansion % vs Mean Gradient (mmHg)",
           x = "Mean Gradient (mmHg)",
           y= "Relative Stent Expansion %") 

`geom_smooth()` using formula 'y ~ x'

correlation_test4 <- cor.test(data1$MeanGradientmmHg, data1$RelativeStentExpansion)
print(correlation_test4)

    Pearson's product-moment correlation

data:  data1$MeanGradientmmHg and data1$RelativeStentExpansion
t = -3.3768, df = 97, p-value = 0.001056
alternative hypothesis: true correlation is not equal to 0
95 percent confidence interval:
 -0.4903450 -0.1355981
sample estimates:
       cor 
-0.3243281 

la correlacion entre MeanGradientmmHg y RelativeStentExpansion es negativa y significativa

#Relacion entre Relative Stent Expansion % (1;0) vs MeanGradientmmHg
boxplot(data1$MeanGradientmmHg~data1$RSEcalc,
        xlab = 'RSE Calc',
        ylab = 'MeanGradientmmHg',
        title=('Relative Stent Expansion % (1;0) vs MeanGradientmmHg'),
          col= 'ivory')

t_test4 <- t.test(data1$MeanGradientmmHg ~ as.factor(data1$RSEcalc))
print(t_test4)

    Welch Two Sample t-test

data:  data1$MeanGradientmmHg by as.factor(data1$RSEcalc)
t = 1.4598, df = 49.241, p-value = 0.1507
alternative hypothesis: true difference in means between group 0 and group 1 is not equal to 0
95 percent confidence interval:
 -0.6662386  4.2060294
sample estimates:
mean in group 0 mean in group 1 
      10.900541        9.130645 

#Relacion entre ICD4mm_calc(IDC4mm/Area Derived Diameter) vs Relative Stent Expansion %
g6=ggplot(data=data1,mapping=
            aes(x=ICD4mm_calc,y=RelativeStentExpansion,))+geom_point()+theme_bw()+
            geom_smooth(method = "lm") 

g6 +  labs(title = "Relative Stent Expansion % vs Intercomisural Diameter at 4 mm Indexed",
           x = "Intercomisural Diameter at 4 mm Indexed",
           y= "Relative Stent Expansion %") 

`geom_smooth()` using formula 'y ~ x'

correlation_test5 <- cor.test(data1$ICD4mm_calc, data1$RelativeStentExpansion)
print(correlation_test5)

    Pearson's product-moment correlation

data:  data1$ICD4mm_calc and data1$RelativeStentExpansion
t = 0.10038, df = 97, p-value = 0.9202
alternative hypothesis: true correlation is not equal to 0
95 percent confidence interval:
 -0.1875974  0.2071868
sample estimates:
       cor 
0.01019186 

La correlacion es nula para estas variables

#Relacion entre CCV vs Relative Stent Expansion %
g7=ggplot(data=data1,mapping=
            aes(x=CCV,y=RelativeStentExpansion,))+geom_point()+theme_bw()+
            geom_smooth(method = "lm") 

g7 +  labs(title = "Relative Stent Expansion % vs Calcium contrast volume",
           x = "Calcium contrast volume",
           y= "Relative Stent Expansion %") 

`geom_smooth()` using formula 'y ~ x'

correlation_test6 <- cor.test(data1$CCV, data1$RelativeStentExpansion)
print(correlation_test6)

    Pearson's product-moment correlation

data:  data1$CCV and data1$RelativeStentExpansion
t = 0.21294, df = 97, p-value = 0.8318
alternative hypothesis: true correlation is not equal to 0
95 percent confidence interval:
 -0.1765496  0.2180967
sample estimates:
       cor 
0.02161556 

La correlacion entre CCV y RelativeStentExpansion es nula.

#Relacion entre SVDmax vs Relative Stent Expansion %
g8=ggplot(data=data1,mapping=
            aes(x=SVDmax,y=RelativeStentExpansion,))+geom_point()+theme_bw()+
            geom_smooth(method = "lm") 

g8 +  labs(title = "Relative Stent Expansion % vs SVDmax",
           x = "SVDmax",
           y= "Relative Stent Expansion %") 

`geom_smooth()` using formula 'y ~ x'

correlation_test7 <- cor.test(data1$SVDmax, data1$RelativeStentExpansion)
print(correlation_test7)

    Pearson's product-moment correlation

data:  data1$SVDmax and data1$RelativeStentExpansion
t = 2.0889, df = 97, p-value = 0.03933
alternative hypothesis: true correlation is not equal to 0
95 percent confidence interval:
 0.01049871 0.38896086
sample estimates:
      cor 
0.2074805 

la correlacion entre SVDmax y RelativeSten expasion es significativa

#Relacion entre SVDmax vs Relative Stent Expansion %
g9=ggplot(data=data1,mapping=
            aes(x=SVDmin,y=RelativeStentExpansion,))+geom_point()+theme_bw()+
            geom_smooth(method = "lm") 

g9 +  labs(title = "Relative Stent Expansion % vs SVDmin",
           x = "SVDmin",
           y= "Relative Stent Expansion %") 

`geom_smooth()` using formula 'y ~ x'

correlation_test8 <- cor.test(data1$SVDmin, data1$RelativeStentExpansion)
print(correlation_test8)

    Pearson's product-moment correlation

data:  data1$SVDmin and data1$RelativeStentExpansion
t = 2.5185, df = 97, p-value = 0.01342
alternative hypothesis: true correlation is not equal to 0
95 percent confidence interval:
 0.05291916 0.42439847
sample estimates:
      cor 
0.2477429 

la correlacion entre SVDmin y RelativeSten expasion es significativa

#Relacion entre CCV vs RSEcalc
boxplot(data1$CCV~data1$RSEcalc,
        xlab = 'RSEcalc',
        ylab = 'CCV',
        title=('RSEcalc vs CCV'),
          col= 'ivory')

REGRESION LINEAL EN FUNCION DE RELATIVE STENT EXPANSION

#Regresion inicial
mod1=lm(RelativeStentExpansion ~ Predilatation + Postdilation +
                      CCV + Raphaecalcification +
                      ICD4mm + SVDmax
        , 
        data = data1)
summary(mod1)

Call:
lm(formula = RelativeStentExpansion ~ Predilatation + Postdilation + 
    CCV + Raphaecalcification + ICD4mm + SVDmax, data = data1)

Residuals:
      Min        1Q    Median        3Q       Max 
-0.095076 -0.009110  0.006183  0.015923  0.057555 

Coefficients:
                      Estimate Std. Error t value Pr(>|t|)    
(Intercept)          8.184e-01  3.273e-02  25.002  < 2e-16 ***
Predilatation        3.359e-03  6.708e-03   0.501 0.617715    
Postdilation         2.917e-02  7.373e-03   3.957 0.000149 ***
CCV                 -6.865e-06  6.108e-06  -1.124 0.263972    
Raphaecalcification -1.790e-03  6.630e-03  -0.270 0.787762    
ICD4mm               1.466e-03  1.618e-03   0.906 0.367438    
SVDmax               1.232e-03  9.888e-04   1.246 0.215808    
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 0.03005 on 92 degrees of freedom
Multiple R-squared:  0.1939,    Adjusted R-squared:  0.1413 
F-statistic: 3.688 on 6 and 92 DF,  p-value: 0.002502

Calculos Intervalo de confianza coeficientes Regresion Lineal

# Obtener el resumen del modelo
summary_mod1 <- summary(mod1)

# Extraer los coeficientes
coefficients <- coef(summary_mod1)[, 1]

# Extraer los intervalos de confianza
conf_intervals <- confint(mod1)

# Extraer los p-values
p_values <- summary_mod1$coefficients[, 4]

# Crear el data frame consolidado
consolidado <- data.frame(Coefficient = coefficients,
                          Lower_CI = conf_intervals[, 1],
                          Upper_CI = conf_intervals[, 2],
                          p_value = p_values)

# Mostrar el data frame consolidado
print(consolidado)

NA

# Aplicar el procedimiento stepwise
mod1_1 <- step(mod1)

Start:  AIC=-687.22
RelativeStentExpansion ~ Predilatation + Postdilation + CCV + 
    Raphaecalcification + ICD4mm + SVDmax

                      Df Sum of Sq      RSS     AIC
- Raphaecalcification  1 0.0000658 0.083152 -689.14
- Predilatation        1 0.0002265 0.083312 -688.95
- ICD4mm               1 0.0007409 0.083827 -688.34
- CCV                  1 0.0011408 0.084227 -687.87
- SVDmax               1 0.0014028 0.084489 -687.56
<none>                             0.083086 -687.22
- Postdilation         1 0.0141404 0.097226 -673.66

Step:  AIC=-689.14
RelativeStentExpansion ~ Predilatation + Postdilation + CCV + 
    ICD4mm + SVDmax

                Df Sum of Sq      RSS     AIC
- Predilatation  1 0.0002199 0.083372 -690.88
- ICD4mm         1 0.0008000 0.083952 -690.19
- SVDmax         1 0.0013865 0.084538 -689.50
- CCV            1 0.0014501 0.084602 -689.43
<none>                       0.083152 -689.14
- Postdilation   1 0.0142979 0.097450 -675.43

Step:  AIC=-690.88
RelativeStentExpansion ~ Postdilation + CCV + ICD4mm + SVDmax

               Df Sum of Sq      RSS     AIC
- ICD4mm        1 0.0007859 0.084158 -691.95
- CCV           1 0.0012633 0.084635 -691.39
- SVDmax        1 0.0014588 0.084830 -691.16
<none>                      0.083372 -690.88
- Postdilation  1 0.0144030 0.097775 -677.10

Step:  AIC=-691.95
RelativeStentExpansion ~ Postdilation + CCV + SVDmax

               Df Sum of Sq      RSS     AIC
- CCV           1 0.0008092 0.084967 -693.00
<none>                      0.084158 -691.95
- SVDmax        1 0.0057460 0.089903 -687.41
- Postdilation  1 0.0137097 0.097867 -679.01

Step:  AIC=-693
RelativeStentExpansion ~ Postdilation + SVDmax

               Df Sum of Sq      RSS     AIC
<none>                      0.084967 -693.00
- SVDmax        1 0.0050034 0.089970 -689.34
- Postdilation  1 0.0136643 0.098631 -680.24

summary(mod1_1)

Call:
lm(formula = RelativeStentExpansion ~ Postdilation + SVDmax, 
    data = data1)

Residuals:
     Min       1Q   Median       3Q      Max 
-0.09468 -0.01011  0.00591  0.01579  0.06160 

Coefficients:
              Estimate Std. Error t value Pr(>|t|)    
(Intercept)  0.8410974  0.0246528  34.118  < 2e-16 ***
Postdilation 0.0282769  0.0071966   3.929 0.000161 ***
SVDmax       0.0015349  0.0006456   2.378 0.019406 *  
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 0.02975 on 96 degrees of freedom
Multiple R-squared:  0.1756,    Adjusted R-squared:  0.1584 
F-statistic: 10.23 on 2 and 96 DF,  p-value: 9.419e-05

REGRESION LOGISTICA EN FUNCION DE RSEcalc(1;0)

#Modelo binario
mod2 <- glm(RSEcalc ~ Predilatation + Postdilation +
                      CCV + Raphaecalcification +
                      ICD4mm + SVDmax 
        , 
                data = data1, family = "binomial")
summary(mod2)

Call:
glm(formula = RSEcalc ~ Predilatation + Postdilation + CCV + 
    Raphaecalcification + ICD4mm + SVDmax, family = "binomial", 
    data = data1)

Deviance Residuals: 
    Min       1Q   Median       3Q      Max  
-1.8558  -0.9999   0.4974   0.9399   1.5382  

Coefficients:
                      Estimate Std. Error z value Pr(>|z|)   
(Intercept)         -7.5848759  2.6526375  -2.859  0.00424 **
Predilatation        0.1432666  0.5154728   0.278  0.78106   
Postdilation         1.9870174  0.7142392   2.782  0.00540 **
CCV                 -0.0010839  0.0004925  -2.201  0.02775 * 
Raphaecalcification  0.3676468  0.5113617   0.719  0.47217   
ICD4mm               0.1931991  0.1264376   1.528  0.12651   
SVDmax               0.0862144  0.0796681   1.082  0.27918   
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 130.86  on 98  degrees of freedom
Residual deviance: 112.14  on 92  degrees of freedom
AIC: 126.14

Number of Fisher Scoring iterations: 4

# Aplicar el procedimiento stepwise
mod2_1 <- step(mod2)

Start:  AIC=126.14
RSEcalc ~ Predilatation + Postdilation + CCV + Raphaecalcification + 
    ICD4mm + SVDmax

                      Df Deviance    AIC
- Predilatation        1   112.22 124.22
- Raphaecalcification  1   112.66 124.66
- SVDmax               1   113.36 125.36
<none>                     112.14 126.14
- ICD4mm               1   114.58 126.58
- CCV                  1   117.42 129.42
- Postdilation         1   122.38 134.38

Step:  AIC=124.22
RSEcalc ~ Postdilation + CCV + Raphaecalcification + ICD4mm + 
    SVDmax

                      Df Deviance    AIC
- Raphaecalcification  1   112.72 122.72
- SVDmax               1   113.50 123.50
<none>                     112.22 124.22
- ICD4mm               1   114.63 124.63
- CCV                  1   117.48 127.48
- Postdilation         1   122.51 132.51

Step:  AIC=122.73
RSEcalc ~ Postdilation + CCV + ICD4mm + SVDmax

               Df Deviance    AIC
- SVDmax        1   113.97 121.97
<none>              112.72 122.72
- ICD4mm        1   115.06 123.06
- CCV           1   117.48 125.48
- Postdilation  1   122.75 130.75

Step:  AIC=121.96
RSEcalc ~ Postdilation + CCV + ICD4mm

               Df Deviance    AIC
<none>              113.97 121.97
- CCV           1   117.98 123.98
- ICD4mm        1   122.98 128.98
- Postdilation  1   124.44 130.44

summary(mod2_1)

Call:
glm(formula = RSEcalc ~ Postdilation + CCV + ICD4mm, family = "binomial", 
    data = data1)

Deviance Residuals: 
    Min       1Q   Median       3Q      Max  
-2.0277  -1.0723   0.5231   0.9297   1.5930  

Coefficients:
               Estimate Std. Error z value Pr(>|z|)   
(Intercept)  -6.7111946  2.5587175  -2.623  0.00872 **
Postdilation  1.9734008  0.7017581   2.812  0.00492 **
CCV          -0.0008416  0.0004300  -1.958  0.05029 . 
ICD4mm        0.2784577  0.0992875   2.805  0.00504 **
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 130.86  on 98  degrees of freedom
Residual deviance: 113.96  on 95  degrees of freedom
AIC: 121.96

Number of Fisher Scoring iterations: 4

#Modelo binario RSE (1;0) final
mod2_final <- glm(formula = RSEcalc ~ Postdilation + CCV + ICD4mm, 
    family = "binomial", data = data1)

# Obtener el resumen del modelo
summary_mod2 <- summary(mod2_final)
summary_mod2

Call:
glm(formula = RSEcalc ~ Postdilation + CCV + ICD4mm, family = "binomial", 
    data = data1)

Deviance Residuals: 
    Min       1Q   Median       3Q      Max  
-2.0277  -1.0723   0.5231   0.9297   1.5930  

Coefficients:
               Estimate Std. Error z value Pr(>|z|)   
(Intercept)  -6.7111946  2.5587175  -2.623  0.00872 **
Postdilation  1.9734008  0.7017581   2.812  0.00492 **
CCV          -0.0008416  0.0004300  -1.958  0.05029 . 
ICD4mm        0.2784577  0.0992875   2.805  0.00504 **
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 130.86  on 98  degrees of freedom
Residual deviance: 113.96  on 95  degrees of freedom
AIC: 121.96

Number of Fisher Scoring iterations: 4

Calculos de Odd Ratio

# Extraer los odds ratios
odds_ratios <- exp(summary_mod2$coefficients[, 1])

# Extraer los intervalos de confianza
conf_intervals <- exp(confint(mod2_final))

Waiting for profiling to be done...

# Extraer los p-values
p_values <- summary_mod2$coefficients[, 4]

# Crear el data frame consolidado
consolidado <- data.frame(Odds_Ratio = odds_ratios,
                          Lower_CI = conf_intervals[, 1],
                          Upper_CI = conf_intervals[, 2],
                          p_value = p_values)

# Mostrar el data frame consolidado
print(consolidado)

NA