Impact of the leaf width on the prostate VMAT about dosimeto-volmetric and delivering parameters

Purpose/Objective

While 4-5 mm thick MLC (L40 or L50) is usual for VMAT, we have been using 2.5 mm MLC (L25). So we compared dosimetric, volumetric and dose delivering parameters with them.

Materials and Methods

Twenty five cases were selected from our database. SmartArc system of Pinnacle3 was used with Novalis Tx (L25), Elekta synergy BM (L40) and Siemens® ARTISTE loaded on Varian Clinac-21EX virtually (L50). The same consolidations for optimization were used with single arc procedure.
Kruskal-Wallis Rank Sum Test (k-test) was applied to find the difference in D98, D95, D50 and D02 of PTV, rectal V40, irradiation time and MU, then Wilcoxon signed rank test examined the order along the leaf width.
To analyze relationships between these values and rectal volume overlapping PTV (ROV) grouped by L25, L40 or L50, linear regression model was employed with an analysis of covariance. We defined significant p level as <0.001.

Results

Kruskal-Wallis Rank Sum Test (k-test) of all variants

#Data
L25vs40vs50 <- read.csv("L25vs40vs50.csv")
#
with(L25vs40vs50,kruskal.test(D98~L25vs40vs50))$p.value
## [1] 0.9428088
with(L25vs40vs50,kruskal.test(D95~L25vs40vs50))$p.value
## [1] 0.3351991
with(L25vs40vs50,kruskal.test(D50~L25vs40vs50))$p.value
## [1] 0.006568651
with(L25vs40vs50,kruskal.test(D02~L25vs40vs50))$p.value 
## [1] 8.632889e-12
with(L25vs40vs50,kruskal.test(V40~L25vs40vs50))$p.value 
## [1] 0.1255103
with(L25vs40vs50,kruskal.test(Time~L25vs40vs50))$p.value
## [1] 3.286513e-14
with(L25vs40vs50,kruskal.test(MU~L25vs40vs50))$p.value  
## [1] 0.2729107

D02 and Time shoewd significant difference.

Wilcoxon signed rank test of D02 and Time

library("coin")
## Loading required package: survival
L25vs40<-L25vs40vs50[(L25vs40vs50$L25vs40vs50=="L25" | L25vs40vs50$L25vs40vs50=="L40"), ]
L25vs50<-L25vs40vs50[(L25vs40vs50$L25vs40vs50=="L25" | L25vs40vs50$L25vs40vs50=="L50"), ]
L40vs50<-L25vs40vs50[(L25vs40vs50$L25vs40vs50=="L40" | L25vs40vs50$L25vs40vs50=="L50"), ]
#D02
wilcox_test(D02~L25vs40vs50,data=L25vs40,paired=T)
## 
##  Asymptotic Wilcoxon-Mann-Whitney Test
## 
## data:  D02 by L25vs40vs50 (L25, L40)
## Z = -5.8308, p-value = 5.515e-09
## alternative hypothesis: true mu is not equal to 0
wilcox_test(D02~L25vs40vs50,data=L25vs50,paired=T)
## 
##  Asymptotic Wilcoxon-Mann-Whitney Test
## 
## data:  D02 by L25vs40vs50 (L25, L50)
## Z = 2.8139, p-value = 0.004895
## alternative hypothesis: true mu is not equal to 0
#Time
wilcox_test(Time~L25vs40vs50,data=L25vs40,paired=T)
## 
##  Asymptotic Wilcoxon-Mann-Whitney Test
## 
## data:  Time by L25vs40vs50 (L25, L40)
## Z = -5.5874, p-value = 2.304e-08
## alternative hypothesis: true mu is not equal to 0
wilcox_test(Time~L25vs40vs50,data=L40vs50,paired=T)
## 
##  Asymptotic Wilcoxon-Mann-Whitney Test
## 
## data:  Time by L25vs40vs50 (L40, L50)
## Z = -5.8462, p-value = 5.031e-09
## alternative hypothesis: true mu is not equal to 0

Linear Regression Analysis of D98 to D02

#lib
library("ggplot2")
#for ggplot facet use pile up the data to sequatial one
nm<-names(L25vs40vs50)
nn<-length(L25vs40vs50[,1])
l25<-subset(L25vs40vs50,select=L25vs40vs50) #factor 
ROV<-subset(L25vs40vs50,select=ROV)         #x of lm
DD<-NULL #pile up data
for(i in 2:(length(L25vs40vs50)-1)){
        dd<-L25vs40vs50[,i]      #data
        dn<-rep(nm[i],nn)        #name as group
        dc<-cbind(l25,dd,dn,ROV) #fator,y,group,x
        DD<-rbind(DD,dc)
}
DD.dose<-DD[1:(75*4),]  # D98, D95, D50, D02
DD.time<-DD[(75*4+1):(75*7),] #V40,Time,MU 
DD.dosetime<-rbind(DD.dose,DD.time)
#plot
g<-ggplot(DD.dose,aes(x=ROV,y=dd))
g<- g+ geom_point(aes(shape=factor(L25vs40vs50),col=factor(L25vs40vs50)),size=1)
g<- g+ facet_grid(.~dn)
g<-g+geom_smooth(method="lm",aes(color=factor(L25vs40vs50)),cex=1 )
g<-g+theme(text=element_text(size=15))
g<-g+ylab("Dose")
g<-g+ggtitle("Dose distribution vs. ROV")
g<-g+theme(axis.title=element_text(size=7),axis.text=element_text(size=5))
g<-g + theme(legend.title = element_text(size=7),legend.text = element_text(size=5)) 
g

D98

with(L25vs40vs50,(anova(lm(D98~ROV*L25vs40vs50))))
## Analysis of Variance Table
## 
## Response: D98
##                 Df  Sum Sq Mean Sq F value    Pr(>F)    
## ROV              1  26.056 26.0559 12.9940 0.0005858 ***
## L25vs40vs50      2   0.920  0.4599  0.2294 0.7956513    
## ROV:L25vs40vs50  2   0.836  0.4179  0.2084 0.8123922    
## Residuals       69 138.361  2.0052                      
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

D95

with(L25vs40vs50,anova(lm(D95~ROV*L25vs40vs50)))
## Analysis of Variance Table
## 
## Response: D95
##                 Df Sum Sq Mean Sq F value   Pr(>F)   
## ROV              1  7.116  7.1158  8.8992 0.003942 **
## L25vs40vs50      2  2.292  1.1458  1.4329 0.245620   
## ROV:L25vs40vs50  2  0.475  0.2375  0.2970 0.743968   
## Residuals       69 55.172  0.7996                    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

D50

with(L25vs40vs50,anova(lm(D50~ROV*L25vs40vs50)))
## Analysis of Variance Table
## 
## Response: D50
##                 Df  Sum Sq Mean Sq F value   Pr(>F)   
## ROV              1  0.9333 0.93334  2.6344 0.109134   
## L25vs40vs50      2  4.7803 2.39015  6.7463 0.002109 **
## ROV:L25vs40vs50  2  0.6996 0.34981  0.9873 0.377768   
## Residuals       69 24.4461 0.35429                    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

D02

with(L25vs40vs50,anova(lm(D02~ROV*L25vs40vs50)))
## Analysis of Variance Table
## 
## Response: D02
##                 Df Sum Sq Mean Sq F value Pr(>F)    
## ROV              1  0.501   0.501  1.0856 0.3011    
## L25vs40vs50      2 77.772  38.886 84.2750 <2e-16 ***
## ROV:L25vs40vs50  2  0.347   0.173  0.3755 0.6883    
## Residuals       69 31.838   0.461                   
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Linear Regression Analysis of Rectal V40, Time and MU

g<-ggplot(DD.time,aes(x=ROV,y=dd))
g<- g+ geom_point(aes(shape=factor(L25vs40vs50),col=factor(L25vs40vs50)),size=1)
g<- g+ facet_wrap(~dn,scales = "free")
g<-g+geom_smooth(method="lm",aes(color=factor(L25vs40vs50)),cex=1 )
g<-g+theme(text=element_text(size=15))
g<-g+ylab("Rate/Sec/Unit")
g<-g+ggtitle("V40, Time and MU vs. ROV")
g<-g+theme(axis.title=element_text(size=7),axis.text=element_text(size=5))
g<-g + theme(legend.title = element_text(size=7),legend.text = element_text(size=5)) 
g

V40

with(L25vs40vs50,anova(lm(V40~ROV*L25vs40vs50)))
## Analysis of Variance Table
## 
## Response: V40
##                 Df    Sum Sq    Mean Sq F value    Pr(>F)    
## ROV              1 0.0017671 0.00176705 17.4560 8.466e-05 ***
## L25vs40vs50      2 0.0005547 0.00027733  2.7397   0.07162 .  
## ROV:L25vs40vs50  2 0.0000881 0.00004407  0.4353   0.64881    
## Residuals       69 0.0069848 0.00010123                      
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Time

with(L25vs40vs50,anova(lm(Time~ROV*L25vs40vs50)))
## Analysis of Variance Table
## 
## Response: Time
##                 Df Sum Sq Mean Sq F value    Pr(>F)    
## ROV              1   6010  6009.5 16.5323 0.0001250 ***
## L25vs40vs50      2  60460 30230.0 83.1633 < 2.2e-16 ***
## ROV:L25vs40vs50  2   6239  3119.5  8.5818 0.0004695 ***
## Residuals       69  25082   363.5                      
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

MU

with(L25vs40vs50,anova(lm(MU~ROV*L25vs40vs50)))
## Analysis of Variance Table
## 
## Response: MU
##                 Df  Sum Sq Mean Sq F value    Pr(>F)    
## ROV              1  290572  290572 16.2207 0.0001428 ***
## L25vs40vs50      2  121939   60969  3.4035 0.0389335 *  
## ROV:L25vs40vs50  2   73938   36969  2.0637 0.1347454    
## Residuals       69 1236046   17914                      
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Only those mean values of D02 and Time were significantly different by k-test. Order along the leaf width about D02 (L50 = L25 < L40) and Time (L25 < L40 < L50) were significant. D98, Time and MU depended on ROV. Intercepts grouped by L25, L40 and L50 were very similar except D02 and Time.

Conclusions

L25, L40 and L50 plans were very similar from the dosimetric point of view (except D02) and from the volumetric (V40) point of view. In terms of dose delivery, Time had remarkable differences (L25 < L40 < L50) and large dependency on ROV.

DATA

L25vs40vs50,D98,D95,D50,D02,V40,Time,MU,ROV,PTV,Prostate,SV,No
L25,75.22,76.98,80.84,83.98,0.2,81,655,2.54,82.14,21.45,10.44,1
L40,74.72,76.6,81.21,85.22,0.21,96,642.2,2.54,82.14,21.45,10.44,1
L50,74.35,76.14,79.98,83.21,0.2,151,754.3,2.54,82.14,21.45,10.44,1
L25,75.34,76.57,80.83,84.99,0.2,84,666,3.04,117.56,38.37,7.24,2
L40,74.26,75.94,82.4,88.89,0.2,105,1038.4,3.04,117.56,38.37,7.24,2
L50,75.67,77.54,81.27,84.12,0.19,157,786.2,3.04,117.56,38.37,7.24,2
L25,76.68,78.07,81.26,84.6,0.2,77,495,2.05,121.65,33.19,10.41,3
L40,76.24,77.55,81.41,86.16,0.19,89,517.2,2.05,121.65,33.19,10.41,3
L50,76.87,78.05,82.04,83.81,0.18,97,452.3,2.05,121.65,33.19,10.41,3
L25,76.83,78.12,80.79,83.76,0.21,88,728,3.45,119.56,24.18,19.07,4
L40,75.26,76.76,81.03,85.06,0.21,102,819.6,3.45,119.56,24.18,19.07,4
L50,75.84,77.3,81.2,83.66,0.21,180,898.2,3.45,119.56,24.18,19.07,4
L25,75.01,77.69,81.77,84.38,0.21,88,686,5.45,128.37,28.91,16.35,5
L40,74.82,76.82,81.19,86.59,0.21,101,767.6,5.45,128.37,28.91,16.35,5
L50,74.54,76.52,80.49,83.09,0.2,204,1020.3,5.45,128.37,28.91,16.35,5
L25,77.03,78.28,81.72,84.77,0.18,81,599,3.74,91.82,25.98,0,6
L40,77.34,78.55,82.26,85.41,0.19,92,523.4,3.74,91.82,25.98,0,6
L50,76.99,78.31,81.89,83.82,0.19,124,587.7,3.74,91.82,25.98,0,6
L25,76.69,77.74,80.89,83.99,0.2,79,553,8.99,145.03,21.82,36.64,7
L40,76.32,77.84,81.14,84.74,0.2,90,511.1,8.99,145.03,21.82,36.64,7
L50,77.1,78.26,82.03,83.82,0.19,119,527.2,8.99,145.03,21.82,36.64,7
L25,77.11,78.23,81.74,83.97,0.19,79,528,3.64,197.91,70.14,28.86,8
L40,77.73,78.6,82.42,86.11,0.2,85,530.2,3.64,197.91,70.14,28.86,8
L50,77.16,78.38,82.08,84.01,0.19,111,524.9,3.64,197.91,70.14,28.86,8
L25,77.23,78.31,80.54,83.67,0.21,80,620,0.03,84.32,23.74,21.43,9
L40,77.46,76.71,81.08,85.1,0.2,90,616,0.03,84.32,23.74,21.43,9
L50,77.68,78.51,81.77,83.56,0.2,127,635.1,0.03,84.32,23.74,21.43,9
L25,75.9,77.6,81.67,84.89,0.19,86,705,4.74,199.32,81.02,6.94,10
L40,76.92,78.32,82.14,85.58,0.21,94,585.2,4.74,199.32,81.02,6.94,10
L50,77.62,78.84,82.04,84.04,0.2,124,621.1,4.74,199.32,81.02,6.94,10
L25,75.79,77.28,81.23,83.96,0.21,84,642,5.23,109.79,20.24,7.48,11
L40,75.14,76.44,81.05,86.12,0.21,92,592.3,5.23,109.79,20.24,7.48,11
L50,75.18,76.66,80.87,83.74,0.2,147,735,5.23,109.79,20.24,7.48,11
L25,74.84,77.32,80.96,83.44,0.2,78,611,1.79,92.99,17.77,5.77,12
L40,74.38,76.48,81.39,86.94,0.19,97,575.9,1.79,92.99,17.77,5.77,12
L50,74.22,77.03,80.86,83.58,0.18,133,665.3,1.79,92.99,17.77,5.77,12
L25,75.16,77.06,81.01,84.56,0.21,85,660,5.24,189.41,56.5,28.1,13
L40,74,75.7,80.93,85.79,0.21,94,639.3,5.24,189.41,56.5,28.1,13
L50,74.9,76.51,80.37,83.49,0.2,185,925,5.24,189.41,56.5,28.1,13
L25,76.95,77.83,81.12,84.36,0.2,79,563,2.22,64.85,15.14,7.25,14
L40,77.09,78.11,81.78,85.62,0.19,95,582.7,2.22,64.85,15.14,7.25,14
L50,77.16,78.17,81.44,83.91,0.19,111,552.6,2.22,64.85,15.14,7.25,14
L25,73.5,77.2,81.58,84.14,0.22,95,862,9.39,105.09,14.5,12.36,15
L40,72.66,75.66,81.09,86.73,0.22,114,881.2,9.39,105.09,14.5,12.36,15
L50,70.49,74.87,79.65,83.49,0.21,229,1146.2,9.39,105.09,14.5,12.36,15
L25,73.69,75.68,81.05,85.23,0.22,83,767,5.21,110.79,22.95,22.1,16
L40,73.94,76.81,81.36,86.26,0.22,104,799.6,5.21,110.79,22.95,22.1,16
L50,72.6,76.21,80.32,84.28,0.21,224,1118.6,5.21,110.79,22.95,22.1,16
L25,75.84,77.88,81.9,84.71,0.19,78,604,2.48,115.6,30.55,11.55,17
L40,75.7,77.32,81.29,86.27,0.2,90,543.2,2.48,115.6,30.55,11.55,17
L50,75.51,77.47,80.69,83.91,0.2,111,537.1,2.48,115.6,30.55,11.55,17
L25,73.86,76.69,81.77,84.35,0.22,91,802,7.46,133.49,26.98,10.91,18
L40,72.44,74.81,80.99,86.96,0.22,103,867.5,7.46,133.49,26.98,10.91,18
L50,73.47,76.04,80.48,84.91,0.21,203,1012.7,7.46,133.49,26.98,10.91,18
L25,76.39,77.89,81.32,83.79,0.21,87,691,4.84,98.99,12.66,14.09,19
L40,74.84,76.26,81.48,87.6,0.21,95,678.9,4.84,98.99,12.66,14.09,19
L50,74.64,76.35,80.34,83.87,0.2,170,851.9,4.84,98.99,12.66,14.09,19
L25,73.82,75.55,79.57,84.05,0.19,84,703,6.51,168.3,29.04,24.33,20
L40,75.85,77,81.79,87.24,0.21,91,650.3,6.51,168.3,29.04,24.33,20
L50,75.26,76.85,80.54,83.97,0.2,185,927,6.51,168.3,29.04,24.33,20
L25,77.03,77.96,81.53,83.95,0.2,76,554,4.84,79.87,17.27,6.86,21
L40,77.68,78.06,82.25,86.6,0.2,92,539.7,4.84,79.87,17.27,6.86,21
L50,75.84,76.99,80.09,83.17,0.19,113,561.2,4.84,79.87,17.27,6.86,21
L25,77.82,78.78,82.22,83.78,0.19,81,586,4.49,114.02,28.19,14.64,22
L40,77.57,78.51,81.62,85.63,0.18,95,515.9,4.49,114.02,28.19,14.64,22
L50,77.65,78.42,81.27,84.43,0.17,126,627.8,4.49,114.02,28.19,14.64,22
L25,76.05,77.66,81.14,83.67,0.2,82,647,3.69,159.26,42.35,17.81,23
L40,76.33,77.61,81.74,85.54,0.2,92,617.9,3.69,159.26,42.35,17.81,23
L50,75.49,77.04,80.89,83.81,0.2,135,675.8,3.69,159.26,42.35,17.81,23
L25,76.09,78.07,81.4,84.61,0.2,79,591,3.39,78.08,14.33,5.41,24
L40,77,78.39,82.45,87.1,0.19,95,649.7,3.39,78.08,14.33,5.41,24
L50,77.21,77.76,80.16,83.43,0.2,132,653.9,3.39,78.08,14.33,5.41,24
L25,75.05,77.78,81.77,83.95,0.17,83,617,0.86,124.26,41.83,2.71,25
L40,74.68,76.84,81.91,84.64,0.19,92,576,0.86,124.26,41.83,2.71,25
L50,74.85,77.07,81.18,84.49,0.19,109,536.2,0.86,124.26,41.83,2.71,25