SEER and A-Bomb Data Analysis Tools

Description

Creates SEER (Surveillance, Epidemiology and End Results) and Japanese A-bomb survivor data binaries from ASCII sources and provides tools for estimating SEER second cancer risks. Depicted below, mkSEER merges cancers and populations of all three of the SEER databases into single cancer and population data frames.

SEER data field positions and names change over the years and the original purpose of SEERaBomb was to buffer/protect R scripts from such changes. A second purpose was to speed up SEER data computations by reducing the data [via pickFields()] to only fields of interest. SEERaBomb now has an additional purpose: estimating relative risks of SEER second cancers after diagnoses of first cancers, using all three SEER databases. Note: SEER no longer includes radiation therapy data by default. Users must thus obtain custom SEER treatment data https://seer.cancer.gov/data/treatment.html.

Details

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

References

Surveillance, Epidemiology, and End Results (SEER) Program (www.seer.cancer.gov) Research Data (1973-2015), National Cancer Institute, DCCPS, Surveillance Research Program, Surveillance Systems Branch, released April 2018, based on the November 2017 submission.

See Also

getFields,pickFields,mkSEER,mkSEERold,mkAbomb

Chromosome translocation versus age data

Description

This is chromosome translocation versus age data that is pooled across gender and race.

Usage

data(Sigurdson)

Format

A data frame named Sigurdson with the following columns.

age

Age of donor of lymphocytes.

tlcn

Total number of chromosomal translocations per 100 cell equivalents.

Details

The data were obtained using FISH, see reference below. This dataset is loaded automatically with library(SEERaBomb). As such, the function data() is not needed.

References

Sigurdson et al. Mutation Research 652 (2008) 112-121

Examples

library(SEERaBomb)
with(Sigurdson,plot(age,tlcn,cex=2,cex.axis=2,cex.lab=2,las=1,cex.main=2,
               ylab="",main="Translocations per 100 cells"))

Converts canc made by mkSEER into a data.frame with py at risk after firsts

Description

Makes a data frame in the exernal second cancer data format expected by esd.

Usage

canc2py(canc,firstS,secondS) 

Arguments

Details

Use is currently limited to 2nd cancers in SEER since 1973, e.g. AML, since esd assumes its py column is in synch with times since diagnosis. This is not the case when e.g. MDS or CMML are in secondS.

Value

data.frame with columns: yrdx, agedx, sex, py at risk (in years), cancer1, and cancer2. Cases not ending in a second cancer have cancer2 set to "none".

Note

This function was developed with support from the Cleveland Clinic Foundation.

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

See Also

SEERaBomb-package, esd,msd

Examples

## Not run: 
rm(list=ls()) 
library(SEERaBomb)
load("~/data/SEER/mrgd/cancDef.RData") 
secS=c("AML","MDS")
frstS=c("HL")
canc=canc%>%filter(cancer%in%union(frstS,secS))
load("~/data/SEER/mrgd/popsae.RData")
popsa=popsae%>%group_by(db,race,sex,age,year)%>%summarize(py=sum(py)) # sum on regs
pm=seerSet(canc,popsa,Sex="male",ageStart=0,ageEnd=100) 
pf=seerSet(canc,popsa,Sex="female",ageStart=0,ageEnd=100) 
pm=mk2D(pm,secondS=secS) 
pf=mk2D(pf,secondS=secS)
brks=c(0,1,5,10)
pm=csd(pm,brkst=brks,firstS=frstS) 
pf=csd(pf,brkst=brks,firstS=frstS)
S=rbind(pf$DF,pm$DF)%>%select(int,t,cancer1,cancer2,O,E) #merge sexes
S=S%>%group_by(int,cancer1,cancer2)%>%summarize(O=sum(O),E=sum(E),t=mean(t))
S=S%>%mutate(RR=O/E,rrL=qchisq(.025,2*O)/(2*E),rrU=qchisq(.975,2*O+2)/(2*E))
S$data="SEER"

ds=canc2py(canc,frstS,secS)
S2=esd(ds,pf$D,pm$D,brks)
S2=S2%>%group_by(int,cancer1,cancer2)%>%summarize(O=sum(O),E=sum(E),t=mean(t))
S2=S2%>%mutate(RR=O/E,rrL=qchisq(.025,2*O)/(2*E),rrU=qchisq(.975,2*O+2)/(2*E))
S2$data="SEER2"
d=rbind(S,S2)
d%>%filter(cancer1=="HL",cancer2=="AML")%>%arrange(int)
d%>%filter(cancer1=="HL",cancer2=="MDS")%>%arrange(int) #low because MDS not observable before 2001
yrcut=2001  # try to fix it like this
ds=ds%>%filter(yeaR+py>yrcut)
I=ds$yeaR<yrcut
ds$py[I]=(ds$py-(yrcut-ds$yeaR))[I]
S2=esd(ds,pf$D,pm$D,brks)
S2=S2%>%group_by(int,cancer1,cancer2)%>%summarize(O=sum(O),E=sum(E),t=mean(t))
S2=S2%>%mutate(RR=O/E,rrL=qchisq(.025,2*O)/(2*E),rrU=qchisq(.975,2*O+2)/(2*E))
S2$data="SEER2"
d=rbind(S,S2)
d%>%filter(cancer1=="HL",cancer2=="MDS")%>%arrange(int) 
#the problem is that esd assumes that py is synched with tsx. csd handles the delays correctly.
d%>%filter(cancer1=="HL",cancer2=="MDS")%>%group_by(data)%>%summarize(o=sum(O)) #cases were shifted

## End(Not run)

Cancer risk vs years Since Diagnosis of other cancer

Description

Computes relative risks (RR) of 2nd cancers over specified intervals of times since diagnoses of a 1st cancer. 2D spline fits are used to produce expected cases E controlling for background risk dedepence on age and calendar year. RR is then O/E where O is the number of observed cases.

Usage

csd(seerSet,brkst=c(0),brksy=c(1975),brksa=c(0),trts=NULL,
                    PYLong=FALSE,firstS="all",exclUnkSurv=FALSE) 

Arguments

Value

The input with an L component added to it or extended it if it already existed. Each component of L is a nested list of lists that can yield second cancer relative risks as a function of time since 1st cancer diagnosis. The most recent component of L is also provided as a data.frame seerSet$DF produced internally using getDF.

Note

This function was developed with support from the Cleveland Clinic Foundation.

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

See Also

SEERaBomb-package, mk2D,seerSet

Examples

## Not run: 
library(SEERaBomb)
pm=simSeerSet()
pm=mk2D(pm)
pm$canc
pm=csd(pm,brkst=c(0,5),brksy=c(1975,2000),brksa=c(0,50),trts=c("noRad","rad"))
pm
library(ggplot2)
theme_set(theme_gray(base_size = 16)) 
theme_update(legend.position = "top")
g=qplot(x=t,y=RR,data=subset(pm$DF,cancer1=="A"&cancer2=="B"),col=trt,geom=c("line","point"),
        xlab="Years Since First Cancer Diagnosis",ylab="Relative Risk")
g=g+facet_grid(yearG~ageG,scales="free")+geom_abline(intercept=1, slope=0) 
g+geom_errorbar(aes(ymin=rrL,ymax=rrU,width=.15))

## End(Not run)

Event vs years Since Diagnosis

Description

Computes relative risks (RR) of second cancers over specified years-since-diagnosis intervals. SEER incidence rates are used to compute background/expected numbers of cases E, sex, age, and calendar year specifically. RR = O/E where O and E are the numbers of observed and expected cases.

Usage

esd(d,srfF,srfM,brkst=c(0,2,5),brksy=NULL) 

Arguments

Value

data.frame with observed and expected cases, RR, and RR CI for each time since diagnosis interval.

Note

This function was developed with support from the Cleveland Clinic Foundation.

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

See Also

SEERaBomb-package, mk2D,msd

Fills age-year person year (PY) matrix

Description

This internal function converts a matrix with lots of individual PY contributions and starting ages and years as rows, into a population-level PY histrogram matrix of (1-year age)x(1-year year) resolution bins. The output matrix is more square-like (currently at 40 calendar years by 126 age years) than the tall dataframe-like input matrix.

Usage

fillPYM(PYin,PYM)

Arguments

Value

The second argument becomes the output. This matrix should be filled with zeros in R before calling this function. After the call, it will be filled; this function uses pointers in Rcpp C++ code.

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

Examples

  yrs=1975:1990
  ages=0.5:70.5
  PYM=matrix(0,ncol=length(yrs),nrow=length(ages))
  colnames(PYM)=yrs
  rownames(PYM)=ages
(PYin=structure(c(3.5, 11.25,5.2, 51.5, 58.5,0.75, 1976, 1977,1980),.Dim = c(3L,3L), 
                  .Dimnames=list(c("1","2","3"),c("py", "ageL", "year"))))
fillPYM(PYin, PYM)

Gets the lower and upper limit and index of a tsd bin

Description

Extract time since diagnosis (tsd) interval information in strings produced by cut.

Usage

getBinInfo(binLab, binS)

Arguments

Value

A numeric vector containting the lower limit (LL), upper limit (UL), and position (index) in the parent vector binS.

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

See Also

SEERaBomb-package

Examples

library(SEERaBomb)
brks=c(0,0.25,1,3,5)  
(binS=levels(cut(brks+0.1,breaks=c(brks,100)))) #make a vector of intervals 
getBinInfo(binS[4],binS) # test getBinInfo

Converts a seerSet$L series to a data.frame

Description

Creates a data.frame of observed and expected cases for each first and second cancer and treatment. csd() calls this internally for the most recent time series, so it may not need to be called directly.

Usage

getDF(seerSet,srs=NULL)

Arguments

Value

A data.frame in long format that can be used by ggplot.

Note

I envision getting away from saving multiseries seerSet objects and instead just saving several DF outputs of getDF. Besides smaller objects, a reason for this is that two L objects out of csd can now be confounded if they have the same time since diagnosis series but a different series for age and/or year of diagnosis.

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

See Also

SEERaBomb-package, csd

Examples

## Not run:  
library(SEERaBomb)
load("~/data/SEER/mrgd/cancDef.RData") #load in canc
load("~/data/SEER/mrgd/popsae.RData") # load in popsae
canc=canc%>%select(-reg,-recno,-agerec,-numprims,-COD,
          -age19,-age86,-radiatn,-ICD9,-db,-histo3) 
popsa=popsae%>%group_by(db,race,sex,age,year)%>%summarize(py=sum(py)) # sum on regs
pm=seerSet(canc,popsa,Sex="male",ageStart=0,ageEnd=100) #pooled (races) male seerSet
pm=mk2D(pm,secondS=c("AML","MDS"))
firstS=c("NHL","MM")
pm=csd(pm,brkst=c(0,1,5),trts=c("rad","noRad"),firstS=firstS)
pm$DF 
getDF(pm)

## End(Not run)

Get expected numbers of cases

Description

Converts list of matrices produced by post1PYO into a matrix that corresponds to the matrix of observed cases produced by post1PYO.

Usage

getE(LPYM,D,ageStart,ageEnd,yearEnd,firstS,secondS)

Arguments

Value

A matrix of expected cases with first cancers labeling the rows and second cancers labeling the columns.

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

See Also

SEERaBomb-package, seerSet

Examples

## Not run: 
library(SEERaBomb)
pm=simSeerSet()
pm=mk2D(pm,secondS="B") 
L=post1PYO(pm$canc,brks=c(0,2,5),binIndx=1,Trt="rad",yearEnd=2012 )
names(L)
names(pm)
with(pm,getE(L$LPYM,D,ageStart,ageEnd,yearEnd,cancerS,secondS))

## End(Not run)

Get fields from SEER SAS file

Description

Converts the SAS file in the SEER ‘incidence’ directory into a data frame in R.

Usage

getFields(seerHome="~/data/SEER")

Arguments

Details

SEER provides a SAS file for reading SEER ASCII data files into SAS. This file is parsed by getFields() to generate a data frame in R that contains all of the SEER fields. This data frame describes these fields in terms of their names (short and long forms), their starting points, and their widths.

Value

A data frame with one row for each field and columns that contain corresponding starting positions, widths, sas names, short names, and expansions thereof.

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

See Also

SEERaBomb-package, mkSEER, pickFields

Examples

## Not run: 
library(SEERaBomb)
(df=getFields())
head(df,20)

## End(Not run)

Get tsd interval PY

Description

Gets person year (PY) contributions to a particular time since diagnosis (tsd) interval from survival times.

Usage

getPY(surv, bin, binS, brks)

Arguments

Value

A vector as long as the survival vector input of PY at risk in a particular interval.

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

See Also

SEERaBomb-package, getBinInfo

Examples

library(SEERaBomb)
brks=c(0,0.25,1,3,6)  
(binS=levels(cut(brks+0.1,breaks=c(brks,100)))) #make a vector of intervals 
survTimes=c(8,16,1.5,3.7)
getPY(survTimes,binS[1],binS,brks)# all contribute 0.25 to first interval 
getPY(survTimes,binS[4],binS,brks)# 3rd and 4th survivals contribute 0 and 0.7 to (3,6]
getPY(survTimes,binS[5],binS,brks)# 1st and 2nd survival contribute 2 and 10 years to (6,100]

Computes A-bomb incidences

Description

Creates A-bomb survivor incidence rates and confidence intervals.

Usage

incidAbomb(d)

Arguments

Details

The columns DG and py must exist in d, in that order. Person-year weighted means will be formed of any columns between them. Its OK if none exist. Sums are formed on py and anything to their right. It is assumed that cancer types begin after py, with at most upy and/or subjects intervening them.

Value

A tibble data frame, summarized by groups, with cancers after py in a new cancers column, and new columns O (observed cases), incid and incid 95% CI limits LL and UL.

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

See Also

SEERaBomb-package

Computes SEER incidences

Description

Creates SEER incidence rates and confidence intervals.

Usage

incidSEER(canc,popsae,cancers)

Arguments

Details

This left joins popsae and cancers in canc.

Value

A data frame with observed cases (O), incid, and incid 95% CI limits LL and UL.

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

See Also

SEERaBomb-package

Map CODs to strings

Description

Maps integer cause of death (COD) codes in COD of a SEER cancer data frame to a factor CODS with recognizable levels. This is a bit slow, so it is called within mkSEER.

Usage

mapCODs(D)

Arguments

Value

The input data frame with an additional CODS column added on.

Note

Typing mapCODs dumps the function definition and thus the mapping used.

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

See Also

SEERaBomb-package, mapRegs

Examples

library(SEERaBomb)
mapCODs # shows default definitions

Map ICD9 and ICD-O3 codes to cancers

Description

Adds a factor cancer with easily recognizable levels to a SEER cancer data.frame.

Usage

mapCancs(D)

Arguments

Value

The input data frame with an additional cancer column added on.

Note

This is used by mkSEER() when it generates R binaries of the SEER data. Otherwise it provides current cancer definitions (seen by looking at the function definition).

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

See Also

SEERaBomb-package, mapRegs

Examples

library(SEERaBomb)
mapCancs # shows default definitions

Map registry codes to acronyms

Description

Maps codes for SEER registries to 2-letter acronyms and corresponding descriptions.

Usage

mapRegs(code=NA)

Arguments

Value

A dataframe of SEER registry symbols and descriptions with rownames such as "1501" for sf (san francisco) and "1520" for dM (detroit Michigan), or just the symbol if the rowname is given. Note that city characters are in lower case and state characters are in upper case.

Note

This function is used by mkSEER when it generates merged R binaries. It is exposed to provide quick access to registry acronym definitions.

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

See Also

SEERaBomb-package, mapCancs

Examples

library(SEERaBomb)
mapRegs(1501)
mapRegs()

Map treatment codes to factor

Description

Uses SEER codes in the SEER field radiatn to add a factor named trt with levels "noRad","rad", and "unk" to a cancer data frame.

Usage

mapTrts(D)

Arguments

Value

The input data frame with an additional trt column added to its end.

Note

This function is used by mkSEER when it generates merged R binaries. It is exposed to state the default definition of trt and, by way of example, to show how to override it.

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

See Also

SEERaBomb-package, mapTrts

Examples

library(SEERaBomb)
mapTrts # exposes default definition of trt

Make 2D-spline fits of incidences

Description

Produces two dimensional (2D) spline fits of cancer incidence versus age and calendar year, with interactions. In conjunction with person years (PY) at risk, this is used in csd() to produce expected numbers of cases under a null hypothesis that prior cancers do not impact subsequent cancer risks.

Usage

mk2D(seerSet, knots=5, write=FALSE, outDir="~/Results",txt=NULL,secondS=NULL)

Arguments

Value

The input seerSet with an additional data frame D added to this list. D holds background/expected incidences over a 1-year resolution age-year grid.

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

See Also

SEERaBomb-package, plot2D, seerSet

Examples

## Not run: 
library(SEERaBomb)
(pm=simSeerSet())
(pm=mk2D(pm))
names(pm)
head(pm$D)
tail(pm$D)

## End(Not run)

Make Abomb Binaries

Description

Converts Abomb files ‘lsshempy.csv’ and ‘lssinc07.csv’ into tibbles heme and solid in the file ‘abomb.RData’.

Usage

mkAbomb(AbombHome="~/data/abomb")

Arguments

Details

The files ‘lsshempy.csv’ and ‘lssinc07.csv’ can be found under The incidence of leukemia, lymphoma and multiple myeloma among atomic bomb survivors: 1950-2001 and Life Span Study Cancer Incidence Data, 1958-1998 of the Radiation Effects Research Foundation (RERF) website http://www.rerf.or.jp/.

Value

None. This function is called for its side-effect of producing ‘abomb.RData’.

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

See Also

SEERaBomb-package,mkSEER

Examples

## Not run: 
library(SEERaBomb)
mkAbomb()
load("~/data/abomb/abomb.RData")
View(heme)

## End(Not run)

Converts seerSet$L series to a data.frame

Description

Creates a data.frame of observed and expected cases for each first and second cancer and treatment. Use of this function is deprecated. Use getDF() instead, if needed: csd now calls getDF internally for the most recent time series, so even it may not need to be called directly.

Usage

mkDF(seerSet,srs)

Arguments

Value

A data.frame in long format that can be used by ggplot.

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

See Also

SEERaBomb-package, mk2D, tsd

Examples

## Not run:  
library(SEERaBomb)
load("~/data/SEER/mrgd/cancDef.RData") #load in canc
load("~/data/SEER/mrgd/popsae.RData") # load in popsae
canc=canc%>%select(-reg,-recno,-agerec,-numprims,-COD,
          -age19,-age86,-radiatn,-ICD9,-db,-histo3) 
popsa=popsae%>%group_by(db,race,sex,age,year)%>%summarize(py=sum(py)) # sum on regs
pm=seerSet(canc,popsa,Sex="male",ageStart=0,ageEnd=100) #pooled (races) male seerSet
pm=mk2D(pm,secondS=c("AML","MDS"))
brks=c(0,1,5)
firstS=c("NHL","MM")
pm=tsd(pm,brks=brks,trts=c("rad","noRad"),firstS=firstS)
mkDF(pm)

## End(Not run)

Make Demographics Tables

Description

Provides, in an Excel file, quartiles of age at diagnoses in one sheet and median overall survival times on a second. Many tables are placed in each sheet. One Excel file is produced per cancer type.

Usage

mkDemographics(canc,outDir="~/Results/SEERaBomb")

Arguments

Value

Returned invisibly is a list of data frames corresponding to tables of the Excel file(s).

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

See Also

SEERaBomb-package

Examples

## Not run: 
library(SEERaBomb)
rm(list=ls())
load("~/data/SEER/mrgd/cancDef.RData")
canc$year=cut(canc$yrdx,c(1973,2003,2009,2015),include.lowest = T,dig.lab=4)
canc$age=cut(canc$agedx,c(0,40,50,60,70,80,90,126),include.lowest = T)
canc=canc%>%filter(surv<9999)
canc=canc%>%select(-age86,-radiatn,-chemo,-db,-casenum,-modx,-seqnum,-yrbrth,-ICD9,-reg,-histo3)
canc=canc%>%filter(cancer%in%c("AML","MDS","MPN"))
head(canc,3)
mkDemographics(canc)

## End(Not run)

Make RR Excel file from csd output

Description

Provides relative risks (RR) organized by 1st and 2nd cancers, times since 1st cancer diagnoses, and 1st cancer treatment. RR = O/E where O = observed cases and E = cases expected under a null hypothesis that prior cancers do not impact subsequent risks. If flip = FALSE (default), sheets = 1st cancers and rows = 2nd cancers, else sheets = 2nd cancers and rows = 1st cancers; columns are always intervals of years since diagnosis, in 1st cancer treatment blocks. RR CI and observed numbers are included in each data cell.

Usage

mkExcelCsd(seerSet,tsdn,biny="[1975,2017)",bina="(0,126]",
                        outDir="~/Results",outName=NULL,flip=FALSE)

Arguments

Value

Returned invisibly, a list of data frames corresponding to sheets of the Excel file.

Note

Outputs are for a given sex. Races are typically pooled.

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

See Also

SEERaBomb-package, mk2D,seerSet

Examples

## Not run: 
library(SEERaBomb)
pm=simSeerSet()
pm=mk2D(pm) 
mybrks=c(0,1,5,10)
pm=csd(pm,brkst=mybrks,trts=c("noRad","rad")) 
(lab=paste0("b",paste(mybrks,collapse="_")))
(L=mkExcelCsd(pm,lab))
(L=mkExcelCsd(pm,lab,flip=TRUE))

## End(Not run)

Make RR Excel file from tsd function output

Description

Provides relative risks (RR) organized by 1st and 2nd cancers, times since 1st cancer diagnoses, and 1st cancer treatment. RR = O/E where O = observed cases and E = cases expected under a null hypothesis that prior cancers do not impact subsequent risks. If flip = FALSE (default), sheets = 1st cancers and rows = 2nd cancers, else sheets = 2nd cancers and rows = 1st cancers; columns are always intervals of years since diagnosis, in 1st cancer treatment blocks. RR CI and observed numbers are included in each data cell.

Usage

mkExcelTsd(seerSet,tsdn,outDir="~/Results",outName=NULL,flip=FALSE)

Arguments

Value

Returned invisibly, a list of data frames corresponding to sheets of the Excel file.

Note

Outputs are for a given sex. Races are typically pooled.

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

See Also

SEERaBomb-package, mk2D,seerSet

Examples

## Not run: 
library(SEERaBomb)
pm=simSeerSet()
pm=mk2D(pm) 
mybrks=c(0,1,5,10)
pm=tsd(pm,brkst=mybrks,trts=c("noRad","rad")) 
(lab=paste0("b",paste(mybrks,collapse="_")))
(L=mkExcelTsd(pm,lab))
(L=mkExcelTsd(pm,lab,flip=TRUE))

## End(Not run)

Make Life Tables

Description

Makes life tables from mortality data binaries produced by mkMrt() and places the files in the same folder.

Usage

 mkLT(mrtHome="~/data/usMort",input="mrt.RData",output="ltb.RData")

Arguments

Value

None. This function is called for its side-effect of producing male and female life tables in files.

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

See Also

SEERaBomb-package,mkSEER

Examples

## Not run: 
library(SEERaBomb)
mkLT()
load("~/data/usMort/ltb.RData")
tail(ltb$Female)

## End(Not run)

Make mortality binaries

Description

Gets mortality data from the Human Mortality Database http://www.mortality.org/ and puts it in the file ‘mrt.RData’.

Usage

 mkMrt(username,passwd,country="USA",mrtHome="~/data/usMort")

Arguments

Value

None. This function is called for its side-effect of producing ‘mrt.RData’.

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

References

Barbieri M, Wilmoth JR, Shkolnikov VM, et al. Data Resource Profile: The Human Mortality Database (HMD). Int J Epidemiol. 2015;44: 1549-1556.

See Also

SEERaBomb-package,mkSEER

Examples

## Not run: 
library(SEERaBomb)
mkMrt("username", "password")# sub in your personal account info
load("~/data/usMort/mrt.RData")
head(mrt$Female)

## End(Not run)

Make mortality binaries from local HMD data files

Description

Converts locally installed Human Mortality Data http://www.mortality.org/ into an R binary file ‘mrtCOUNTRY.RData’.

Usage

mkMrtLocal(country="USA",mrtHome="~/data/mrt",
                    mrtSrc1="~/data/hmd_countries",
                    mrtSrc2="~/data/hmd_statistics/death_rates/Mx_1x1"
                    )

Arguments

Value

None. This function is called for its side-effect of producing ‘mrt.RData’ from HMD files organized as all HMD countries or all HMD statistics on the HMD download page (you need at least one of these).

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

References

Barbieri M, Wilmoth JR, Shkolnikov VM, et al. Data Resource Profile: The Human Mortality Database (HMD). Int J Epidemiol. 2015;44: 1549-1556.

See Also

SEERaBomb-package,mkSEER

Examples

## Not run: 
library(SEERaBomb)
mkMrtLocal() 
load("~/data/mrt/mrtUSA.RData")
head(mrt$Female)

## End(Not run)

Make R binaries of SEER data.

Description

Converts SEER ASCII text files into large R binaries that include all cancer types and registries combined.

Usage

mkSEER(df,seerHome="~/data/SEER",outDir="mrgd",outFile="cancDef",
                  indices = list(c("sex","race"), c("histo3","seqnum"),  "ICD9"),
                  writePops=TRUE,writeRData=TRUE,writeDB=FALSE)

Arguments

Details

This function uses the R package LaF to access the fixed-width format data files of SEER. LaF is fast, but it requires knowledge of all the widths of columns wanted, as well as the the widths of unwanted stretches in between. This knowledge is produced by getFields() and pickFields() combined. It is passed to mkSEER() via the argument df.

Value

None, it produces R binary files of the SEER data.

Note

This takes a substantial amount of RAM (it works on a Mac with 16 GB of RAM) and time (~3 minutes using default fields).

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

See Also

SEERaBomb-package,getFields,pickFields

Examples

## Not run: 
library(SEERaBomb)
(df=getFields())
(df=pickFields(df))
# the following will take a several minutes, but may only need 
# to be done roughly once per year, with each release.
mkSEER(df)

## End(Not run)

Make SEER binaries as before

Description

Converts SEER ASCII text files into smaller R binaries. This is being maintained to avoid fixing old scripts. Please use mkSEER for new scripts.

Usage

mkSEERold(df,seerHome="~/data/SEER",
            dataset=c("00","73","92"),SQL=TRUE, mkDFs=FALSE)

Arguments

Details

See mkSEER.

Value

None. This function produces R binary data files.

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

See Also

SEERaBomb-package,mkSEER,getFields,pickFields

Examples

## Not run: 
library(SEERaBomb)
(df=getFields())
(df=pickFields(df))
for (i in c("73","92","00")) mkSEERold(df,dataset=i)

## End(Not run)

Mortality vs years Since Diagnosis

Description

Computes relative risks (RR) of death over specified years-since-diagnosis intervals. US mortality rates obtained via the R package demography are used to compute background death dedepence on age and calendar year. RR is then O/E where O and E are the number of observed and expected cases.

Usage

msd(canc,mrt,brkst=c(0,2,5),brksy=NULL) 

Arguments

Value

data.frame with observed and expected cases, RR, and RR CI for each tsd interval.

Note

This function was developed with support from the Cleveland Clinic Foundation.

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

See Also

SEERaBomb-package, mk2D,seerSet

Examples

## Not run: 
library(SEERaBomb)
load("~/data/SEER/mrgd/cancDef.RData") #loads in canc
lu=canc%>%filter(cancer=="lung")
lu=lu%>%mutate(status=as.numeric(COD>0))%>%select(yrdx,agedx,sex,surv,status) 
lu=lu%>%mutate(surv=round((surv+0.5)/12,3))#convert surv to years

# library(demography)
# d=hmd.mx("USA", "username", "password") #make an account and put your info in here
# mrt=d$rate
# save(mrt,file="~/data/usMort/mrt.RData")
load("~/data/usMort/mrt.RData"); object.size(mrt)# 250kb
brks=c(0,0.5,3,6,10,15,20,25)    
(dlu=msd(lu,mrt,brkst=brks))

## End(Not run)

National Vital Statistic Report (nvsr) Data

Description

US mortality rates (probability of death that year) in 2010; report published Nov. 2014.

Usage

nvsr

Format

A data frame with the following columns.

age

Single-year resolution ages up to 99.5.

p

Pooled sexes and races.

pm

Pooled races, males.

pf

Pooled races, females.

w

Whites, sexes pooled.

b

Blacks, sexes pooled.

o

Others, sexes pooled.

wm

White males.

wf

White females.

bm

Black males.

bf

Black females.

om

Other males.

of

Other females.

References

National Vital Statistics Reports, Vol. 63, No. 7, November 6, 2014 http://www.cdc.gov/nchs/data/nvsr/nvsr63/nvsr63_07.pdf

Examples

library(SEERaBomb)
head(nvsr)

National Vital Statistic Report (nvsr01) Data

Description

US mortality rates (probability of death that year) in 2001 (Report 52_14).

Usage

nvsr01

Format

A data frame with the following columns.

age

Single-year resolution ages up to 99.5.

p

Pooled sexes and races.

pm

Pooled races, males.

pf

Pooled races, females.

wm

White males.

wf

White females.

bm

Black males.

bf

Black females.

Details

This data is used to extrapolate PY at risk in SEER population files from 85+ to older ages.

References

ftp://ftp.cdc.gov/pub/Health_Statistics/NCHS/Publications/NVSR/52_14/

Examples

library(SEERaBomb)
head(nvsr01)

Primary to Secondary

Description

Using a SEER data frame, this function computes times between primary and secondary cancers. In the resulting data frame, surv and status can be analyzed at the individual level, e.g. using Cox regression.

Usage

p2s(canc,firstS,secondS,yrcut=2010) 

Arguments

Value

Data frame with a row for each primary (first cancer) diagnosed on or after yrcut. The surv column holds the time in months to last follow up or death (status=0), or to the time of diagnosis of the second cancer (status=1).

Author(s)

Remco J. Molenaar (r.j.molenaar@amc.uva.nl )

See Also

SEERaBomb-package, mk2D,seerSet

Examples

## Not run: 
# 

## End(Not run)

Pick SEER fields of interest

Description

Reduces the full set of SEER data fields to a smaller set of interest. SEER fields are rows of the input and output dataframes of this function. The output dataframe differs from the input dataframe not only in there being fewer rows but also in there being an additional column needed by mkSEER() downstream.

Usage

pickFields(sas,picks=c("casenum","reg","race","sex","agedx",
        "yrbrth","seqnum","modx","yrdx","histo3",
        "ICD9","COD","surv","radiatn","chemo"))

Arguments

Details

R binaries become too large if all of the fields are selected. SEERaBomb is faster than SEER*Stat because it tailors/streamlines the database to your interests. The default picks are a reasonable place to start; if you determine later that you need more fields, you can always rebuild the binaries. Grabbing all fields is discouraged, but if you want this anyway, note that you still need pickFields() to create a data type column, i.e. you cannot bypass pickFields() by sending the output of getFields() straight to mkSEER().

Value

The SAS-based input data frame sas, shortened to just the rows of picks, and expanded to include spacer rows of fields of no interest pooled into single strings: the width of such a spacer row is equal to the distance in bytes between the fields of interest above and below it. This data frame is then used by laf_open_fwf() of LaF in mkSEER() to read the SEER files. Proper use of this function, and of the SEER data in general, requires an understanding of the contents of ‘seerdic.pdf’ in the ‘incidence’ directory of seerHome.

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

See Also

SEERaBomb-package, getFields, pickFields, mkSEER

Examples

## Not run: 
library(SEERaBomb)
(df=getFields())
(df=pickFields(df))


## End(Not run)

Plot 2D cancer incidence splines

Description

Plots splines of incidence versus age and calendar year produced by mk2D.

Usage

plot2D(seerSet, write=TRUE,outDir="~/Results/plots",col="red")

Arguments

Details

A plot will be produced for each cancer fitted by mk2D. For the first of these, RGL will open a new X11 window. Adjustments of size and angle of this first plot will hold for all subsequent plots. After each plot, the user hits any key to write the plot to a file and advance through the list of cancers.

Value

None, results go to the screen and to png files.

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

See Also

SEERaBomb-package, mk2D,seerSet

Examples

## Not run: 
library(SEERaBomb)
n=simSeerSet()
n=mk2D(n) 
plot2D(n)

## End(Not run)

Get person-years at risk and observed cases after first cancers

Description

Converts a canc data.frame into a list of objects containing information regarding person years at risk for a second cancer after having a first cancer, and the number observed, in a defined time since exposure interval and after a defined first cancer therapy.

Usage

post1PYO(canc, brks=c(0,2,5),binIndx=1,Trt="rad",PYLong=FALSE,yearEnd,firstS,secondS)

Arguments

Value

A list where the first element is a list LPYM with as many PY matrices (PYM) as cancers in canc. The second element is a matrix of cases observed in this interval after this treatment, where row names are first cancers and column names are second cancers. The third element is a trivial scalar, the py-weighted midpoint of the time interval selected.

Note

After the SEER data is installed, see the script mkRRtsx.R in the examples folder.

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

See Also

SEERaBomb-package, getE,seerSet

Get person-years at risk and observed cases after first cancers

Description

Converts a canc data.frame into a list of objects containing information regarding person years at risk for a second cancer after having a first cancer, and the number observed, in defined intervals of time since diagnosis, calendar year, and age at diagnosis.

Usage

post1PYOc(canc, brkst=c(0),binIndxt=1,brksy=c(1973),binIndxy=1,brksa=c(0),binIndxa=1,
             Trt="rad",PYLong=FALSE,yearEnd,firstS,secondS)

Arguments

Value

A list where the first element is a list LPYM with as many PY matrices (PYM) as cancers in canc. The second element is a matrix of cases observed in this interval after this treatment, where row names are first cancers and column names are second cancers. And after a few other slots, the last element is a trivial scalar, the py-weighted midpoint of the time interval selected.

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

See Also

SEERaBomb-package, getE,seerSet

Get person-years at risk and observed deaths after cancer

Description

Converts data.frame D into a list of objects containing information regarding person years at risk of death after cancer and the number of deaths observed, in a defined time since exposure interval.

Usage

post1PYOm(D,brks=c(0,2,5),binIndx=1,yearEnd)

Arguments

Value

A list where the first element is a PY matrix (PYM), the second element is a vector of deaths observed in the intervals, and the third is a scalar, the py-weighted midpoint of the time intervals.

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

See Also

SEERaBomb-package, getE,post1PYO

Prints seerSet.summary objects

Description

Renders data.frame of cases and median ages and survival times with a title above and notes below it. Also qplots PY versus years.

Usage

## S3 method for class 'seerSet.summary'
print(x, ...)

Arguments

Value

None.

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

See Also

SEERaBomb-package, summary.seerSet, seerSet

Examples

## Not run:  
library(SEERaBomb)
load("~/data/SEER/mrgd/cancDef.RData") #load in canc
load("~/data/SEER/mrgd/popsae.RData") # load in popsae
canc=canc%>%select(casenum,race:histo3,surv,cancer,trt,id) 
popsa=popsae%>%group_by(db,race,sex,age,year)%>%summarize(py=sum(py)) # sum on regs
pm=seerSet(canc,popsa,Sex="male",ageStart=0,ageEnd=100) #pooled (races) male seerSet
pm # no print method for seerSet object, so we see the list
summary(pm) # print method for summary renders the summary and a plot of PY

## End(Not run)

Second cancer risk vs attained age after first cancer

Description

Computes absolute risk of 2nd cancers as a function of attained age after first cancer.

Usage

riskVsAge(canc,firstS=c("NHL","HL","MM"),
                secondS=c("AML","MDS"),brksa=c(0,30,50,60,70,80)) 

Arguments

Value

data.frame with incidence vs age.

Note

This function was developed with support from the Cleveland Clinic Foundation.

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

See Also

SEERaBomb-package, mkSEER

Join SEER cancers and PY

Description

Creates a sex-specific list of cancer and population person year (PY) data frames, possibly specific to a race and interval of ages at diagnosis.

Usage

seerSet(canc,popsa,Sex, Race="pooled",ageStart=15,ageEnd=85)

Arguments

Details

In the output: 1) 0.5 years is added to ages at diagnosis (agedx) to reverse SEER flooring to integers; 2) 0.5 months is added to survival months (again, to reverse flooring) before dividing by 12 to convert to years; 3) year of diagnosis integers are converted to reals by adding to them the month of diagnosis (modx) - 0.5 divided by 12 (note that a modx of 1 represents anytime in the month of January). If ageEnd>85, popsae (extended to ages up to 99) should be used as the input for popsa. If popsa is used, the age86 column of popsa will be replaced by an age column. The age86 and yrbrth columns of a canc are not used and will be removed if they happen to be present; users should manually remove any other columns not needed to minimize seerSet object sizes. Sex and race columns in inputs are removed from outputs as they are specified in other (scalar) seerSet elements. Also removed from canc are cancer factor levels not present for that sex.

Value

A list containing sex specific subsets of canc and popsa and information regarding how they were reduced.

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

See Also

SEERaBomb-package, mk2D, plot2D

Examples

## Not run:  
library(SEERaBomb)
simSeerSet() # without data, a simulated seerSet
#  else, with data ...
load("~/data/SEER/mrgd/cancDef.RData") #load in canc
load("~/data/SEER/mrgd/popsae.RData") # load in popsae
# trim columns
library(dplyr)  
canc=canc%>%select(-reg,-recno,-agerec,-numprims,-COD,
          -age19,-age86,-radiatn,-ICD9,-db,-histo3) 
popsae=popsae%>%select(-reg,-db)   
seerSet(canc,popsae,Sex="male",ageStart=0,ageEnd=100) 

## End(Not run)

Summarize SEER data.

Description

Shows numbers of cases per cancer in each SEER database and PY in each registry. Sexes are pooled.

Usage

seerStats(canc,popsa)

Arguments

Value

A list of 2 data.frames with sexes pooled, one of cases with cancer types as rows and as columns, databases, totals, cases >=100 years old or not, and numbers of first-, second-, third- and higher cancers. The second data.frame holds population PY, and PY-weighted ages, per registry.

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

See Also

SEERaBomb-package, mk2D, plot2D

Examples

## Not run:  
library(SEERaBomb)
load("~/data/SEER/mrgd/cancDef.RData") #load in canc
load("~/data/SEER/mrgd/popsae.RData") # load in popsae
seerStats(canc,popsae)

## End(Not run)

Simulate SEER cancers and population person years

Description

Simulates data for two cancers, A and B.

Usage

simSeerSet(N=2e9,yearEnd=2016,ka=1e-5,kb=0.04,Ab=1e-5,
                 tauA=10,tauB=1,delay=1,period=4)

Arguments

Value

A simulated seerSet object with popsa filled using US 2000 Std population proportions and canc with cancers A and B where the incidence of A increases linearly with age and B increase exponentially in age. Survival times are assumed to be exponentially distributed with means of tauA years for A and tauB years for B. Radiation therapy of A is assumed to increase RR of B to 5 uniformly for period years after delay years.

Note

Supported by the Cleveland Clinic Foundation.

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

See Also

SEERaBomb-package, seerSet,mk2D, plot2D

Examples

## Not run: 
library(SEERaBomb)
n=simSeerSet()
n=mk2D(n,secondS="B") 
mybrks=c(0,0.75,0.9,1.1,1.25,2,2.5,3,3.5,4,4.75,4.9,5.1,5.25,6)
n=tsd(n,brks=mybrks,trts=c("rad","noRad")) 
D=mkDF(n)%>%filter(cancer1=="A")%>%select(t,RR,L=rrL,U=rrU,trt)
head(D,2)
library(ggplot2)
theme_update(legend.position = c(.8, .815),
             axis.text=element_text(size=rel(1.2)),
             axis.title=element_text(size=rel(1.3)),
             legend.title=element_text(size=rel(1.2)),
             legend.text=element_text(size=rel(1.2)))
g=qplot(x=t,y=RR,col=trt,data=D,geom=c("line","point"),
        xlab="Years Since First Cancer Diagnosis",ylab="Relative Risk")
g+geom_abline(intercept=1, slope=0)+geom_errorbar(aes(ymin=L,ymax=U,width=.05))

## End(Not run)

Simulate Survival Times

Description

Uses background mortality rates to simulate background survival time for matching sex, age and year.

Usage

 simSurv(d,mrt,rep=1,ltb=NULL,unif=TRUE)

Arguments

Value

Input d with simulation rows added below it, identified as "simulated" in a new column called type.

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

See Also

SEERaBomb-package,mkSEER

Examples

## Not run: 
library(SEERaBomb)
mkLT()
load("~/data/usMort/ltb.RData")
tail(ltb$Female)

## End(Not run)

The standard population of the US in 2000

Description

The US population in 2000 for ages up to 100 years. Sexes are pooled.

Usage

stdUS

Format

A data frame with the following columns.

age

Sigle-year resolution ages up to 100.

pop

The population within each age group.

prop

Proportion of the total population within each age group.

Details

This population data can be used to map age specific incidence rate vectors into summarizing scalars. It allows cancer incidence rates across different SEER registries to be compared without concerns of differences in age distributions of the populations.

References

http://seer.cancer.gov/stdpopulations/

Examples

library(SEERaBomb)
with(stdUS,plot(age,pop/1e6,type="l",xlab="Age",
     ylab="People (Millions)",main="US Population in 2000"))
library(dplyr)
stdUS%>%filter(age>=85)%>%summarize(weighted.mean(age,w=pop)) 
### so ave age >=85.0 is 89.4

Summary of seerSet object

Description

Creates a data.frame of cases and median ages and survival times for each cancer and treatment type.

Usage

## S3 method for class 'seerSet'
summary(object, ...)

Arguments

Value

A list that includes: a data.frame of cases, median ages at diagnosis, and survival times, in years, for each cancer and treatment type; a data.frame of person-years by year; and smaller things such as a title, sex, race, and notes. The resulting list is set to class seerSet.summary which has a print method.

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

See Also

SEERaBomb-package, mk2D, plot2D

Examples

## Not run:  
library(SEERaBomb)
load("~/data/SEER/mrgd/cancDef.RData") #load in canc
load("~/data/SEER/mrgd/popsae.RData") # load in popsae
canc=canc%>%select(casenum,race:yrdx,surv,cancer,trt,id) 
popsa=popsae%>%group_by(db,race,sex,age,year)%>%summarize(py=sum(py)) # sum on regs
pm=seerSet(canc,popsa,Sex="male",ageStart=0,ageEnd=100) #pooled (races) male seerSet
pm # no print method for seerSet object, so we see the list
(x=summary(pm)) # print renders summary and plot of PY
class(x)<-NULL #if you want to see the list as is, kill its class.
x    # It then goes through the regular print method for lists.

## End(Not run)

Compute RR vs tsd

Description

Computes relative risks (RR) over specified times since diagnoses (tsd) of a 1st cancer. 2D spline fits are used to produce expected cases E controlling for background risk dedepence on age and calendar year. RR is then O/E where O is the number of observed cases. WARNING: Use of this function is deprecated, please use csd() instead.

Usage

tsd(seerSet,brks,trts=NULL,PYLong=FALSE,firstS="all") 

Arguments

Value

The input with an L component added to it or extended it if it already existed. Each component of L is a nested list of lists that can yield second cancer relative risks as a function of time since diagnosis and different first cancers and if they were irradiated or not.

Note

This function was developed with support from the Cleveland Clinic Foundation.

Author(s)

Tom Radivoyevitch (radivot@ccf.org)

See Also

SEERaBomb-package, mk2D,seerSet

Examples

## Not run: 
library(SEERaBomb)
pm=simSeerSet()
pm=mk2D(pm) 
mybrks=c(0,1,5,10)
pm=tsd(pm,brks=mybrks,trts=c("noRad","rad"),PYM=TRUE)
(lab=paste0("b",paste(mybrks,collapse="_")))
LM=pm$L[[lab]]$'rad'
names(LM)
LM$PyM
LM$Obs
LM$Exp
table(LM$PyM$`(0,1]`$cancer2)
table(LM$PyM$`(1,5]`$cancer2)
table(LM$PyM$`(5,10]`$cancer2)

## End(Not run)