KM Plot (PFS) — Figure 14.2.x.x

Program: f_14_2_x_x_pfs_km_plot_saffl.sas
Study: MYFAKESTUDY
Output: Figure 14.2.x.x Kaplan–Meier Plot of Progression-Free Survival by Treatment Group
Population: Safety Analysis Set (SAFFL='Y')
14_2_x_x

---

Purpose

Create the Kaplan–Meier PFS figure per SAP, including:

• Survival S(t) or failure 1–S(t) curve (per SAP)
• Censored observations (markers)
• Numbers at risk at selected time points

---

Inputs

• ADaM.ADSL — treatment, flags/denominators
• ADaM.ADTTE — time-to-event records (PARAMCD='PFS' or per SAP)

---

Outputs

• /outputs/rtf/f_14_2_x_x_pfs_km_plot_saffl.rtf
• /outputs/datasets/f_14_2_x_x_kmplotdata.sas7bdat

---

Key Variables

• ADSL: USUBJID, TRT01A / TRT01AN, SAFFL
• ADTTE: USUBJID, PARAMCD, AVAL (time), CNSR

---

Methods

KM estimation via PROC LIFETEST (or SGRENDER from KM outputs):

• Curve: S(t) or 1–S(t) (per SAP)
• Censoring markers from censored records
• At-risk counts via TIMELIST= and (template) BLOCKPLOT in INNERMARGIN
• Group styling via attribute map (color/pattern/symbol), if required

---

Author: Alpha Traore
Date: <YYYY-MM-DD>
QC Program: qc_f_14_2_x_x_pfs_km_plot_saffl.sas

---

Change Log

Version	Date	By	Description
v1.0	YYYY-MM-DD	<Name>	Initial version
v1.1	YYYY-MM-DD	<Name>	<What changed>

Main program steps (outline)

1. Build analysis sets: merge ADSL + ADTTE(PFS), apply SAFFL and PARAMCD filter
2. Derive censoring counts/percents by treatment
3. KM: percentiles + survival rates at Day 0/50/100/150 (per SAP)
4. Cox model: HR (95% CI) for treatment comparison (and p-value if required)
5. Assemble final shell order + labels
6. Produce RTF/PDF output with titles/footnotes

options mprint mlogic symbolgen; 
options validvarname=upcase;

227  ods listing close;ods html5 (id=saspy_internal) options(bitmap_mode='inline') device=svg style=HTMLBlue; ods graphics on /
227! outputfmt=png;
NOTE: Writing HTML5(SASPY_INTERNAL) Body file: sashtml13.htm
228  
229  options mprint mlogic symbolgen;
230  options validvarname=upcase;
231  ods html5 (id=saspy_internal) close;ods listing;
232  

Read the ADaM Datasets Required for Table Generation: ADSL and ADTTE

libname aa "/export/viya/homes/alpha@alphatraore.com/toydata";

/* proc datasets lib=aa; quit;*/   


proc sort data=aa.adsl  out = adsl  (obs=5);  by usubjid; run;
proc sort data=aa.adtte out = adtte (obs=5); by usubjid; run;

data subj;
   set adsl;
   where saffl="Y";
   trt=trt01an;
run;
 
data tte;
   set adtte;
   trt=trtan;
run;
proc sort data=subj; by usubjid; run;
proc sort data=tte; by usubjid; run;

data temp1;
    merge subj tte;
    by usubjid;
run;
proc print data=temp1(obs=5); run;

Obs	USUBJID	TRT01AN	SAFFL	TRT	TRTAN	PARAM	PARAMCD	AVAL	CNSR	PARAMN
1	01-701-1015	1	Y	1	1	Progression Free Survival	PFS	2	0	1
2	01-701-1023	1	Y	1	1	Progression Free Survival	PFS	3	0	1
3	01-701-1028	2	Y	2	2	Progression Free Survival	PFS	3	0	1
4	01-701-1033	3	Y	3	3	Progression Free Survival	PFS	28	1	1
5	01-701-1034	2	Y	2	2	Progression Free Survival	PFS	58	0	1

Kaplan–Meier Cumulative Incidence (Failure) Estimate

• alphaqt=: significance level for quartile (e.g., median) confidence limits
• method= / conftype=: KM method and CI type
• plots=: request failure plot (1–S(t)) with numbers at risk
• timelist=: specify reporting/display time points and limit ProductLimitEstimates rows

Output: Time points and numbers at risk

Produce the KM failure plot with time-point markers on the x-axis and numbers at risk shown below the x-axis at the specified timelist= time points.

proc lifetest data=temp1
   alphaqt=0.05
   method=km
   plots=survival(failure atrisk=0 to 200 by 25 )
   timelist=(0 to 200 by 25)
   conftype=linear ;
 
   time aval*cnsr(1);
   strata trtan;
   ods output FailurePlot  = sur_fail ProductLimitEstimates=estimates;
run;

The LIFETEST Procedure

 

Stratum 1: Actual Treatment (N) = 1

Product-Limit Survival Estimates
Timelist	AVAL		Survival	Failure	Survival Standard Error	Number Failed	Number Left
0.000	0.000		1.0000	0	0	0	86
25.000	24.000		0.8811	0.1189	0.0353	10	72
50.000	48.000		0.7812	0.2188	0.0457	18	61
75.000	70.000		0.7414	0.2586	0.0488	21	53
100.000	97.000		0.6575	0.3425	0.0540	27	47
125.000	110.000		0.6435	0.3565	0.0546	28	45
150.000	110.000		0.6435	0.3565	0.0546	28	40
175.000	110.000		0.6435	0.3565	0.0546	28	37
200.000	.		.	.	.	29	0

Summary Statistics for Time Variable AVAL

Quartile Estimates
Percent	Point Estimate	95% Confidence Interval
		Transform	[Lower	Upper)
75	.	LINEAR	.	.
50	.	LINEAR	.	.
25	70.000	LINEAR	35.000	177.000

Mean	Standard Error
129.866	7.534

Note:The mean survival time and its standard error were underestimated because the largest observation was censored and the estimation was restricted to the largest event time.

---

The LIFETEST Procedure

 

Stratum 2: Actual Treatment (N) = 2

Product-Limit Survival Estimates
Timelist	AVAL		Survival	Failure	Survival Standard Error	Number Failed	Number Left
0.000	0.000		1.0000	0	0	0	84
25.000	25.000		0.5883	0.4117	0.0566	32	41
50.000	50.000		0.3417	0.6583	0.0576	48	21
75.000	71.000		0.1609	0.8391	0.0490	58	7
100.000	96.000		0.0919	0.9081	0.0411	61	4
125.000	96.000		0.0919	0.9081	0.0411	61	4
150.000	96.000		0.0919	0.9081	0.0411	61	4
175.000	96.000		0.0919	0.9081	0.0411	61	3
200.000	.		.	.	.	61	0

Summary Statistics for Time Variable AVAL

Quartile Estimates
Percent	Point Estimate	95% Confidence Interval
		Transform	[Lower	Upper)
75	58.000	LINEAR	47.000	89.000
50	36.000	LINEAR	24.000	46.000
25	14.000	LINEAR	5.000	22.000

Mean	Standard Error
40.075	3.689

Note:The mean survival time and its standard error were underestimated because the largest observation was censored and the estimation was restricted to the largest event time.

---

The LIFETEST Procedure

 

Stratum 3: Actual Treatment (N) = 3

Product-Limit Survival Estimates
Timelist	AVAL		Survival	Failure	Survival Standard Error	Number Failed	Number Left
0.000	0.000		1.0000	0	0	0	84
25.000	25.000		0.6390	0.3610	0.0538	29	49
50.000	48.000		0.3761	0.6239	0.0565	48	23
75.000	71.000		0.2751	0.7249	0.0545	54	15
100.000	100.000		0.2018	0.7982	0.0508	58	11
125.000	119.000		0.1467	0.8533	0.0458	61	7
150.000	126.000		0.1258	0.8742	0.0438	62	6
175.000	126.000		0.1258	0.8742	0.0438	62	5
200.000	.		.	.	.	62	0

Summary Statistics for Time Variable AVAL

Quartile Estimates
Percent	Point Estimate	95% Confidence Interval
		Transform	[Lower	Upper)
75	80.000	LINEAR	51.000	119.000
50	33.000	LINEAR	27.000	48.000
25	19.000	LINEAR	15.000	25.000

Mean	Standard Error
51.706	4.941

Note:The mean survival time and its standard error were underestimated because the largest observation was censored and the estimation was restricted to the largest event time.

Summary of the Number of Censored and Uncensored Values
Stratum	TRTAN	Total	Failed	Censored	Percent Censored
1	1	86	29	57	66.28
2	2	84	61	23	27.38
3	3	84	62	22	26.19
Total		254	152	102	40.16

---

The LIFETEST Procedure

Testing Homogeneity of Survival Curves for AVAL over Strata

Rank Statistics
TRTAN	Log-Rank	Wilcoxon
1	-45.268	-6456.0
2	25.627	3984.0
3	19.641	2472.0

Covariance Matrix for the Log-Rank Statistics
TRTAN	1	2	3
1	34.8938	-15.5885	-19.3053
2	-15.5885	25.8988	-10.3102
3	-19.3053	-10.3102	29.6155

Covariance Matrix for the Wilcoxon Statistics
TRTAN	1	2	3
1	984202	-458582	-525619
2	-458582	828329	-369746
3	-525619	-369746	895366

Test of Equality over Strata
Test	Chi-Square	DF	Pr > Chi-Square
Log-Rank	60.2696	2	<.0001
Wilcoxon	43.8983	2	<.0001
-2Log(LR)	100.9378	2	<.0001

title "FailurePlot";
proc print data=sur_fail (obs=5); run;

FailurePlot

Obs	TIME	SURVIVAL	ATRISK	EVENT	CENSORED	TATRISK	STRATUM	STRATUMNUM	_1_SURVIVAL_	_1_CENSORED_
1	0	1.00000	86	0	.	.	1	1	0.000000	.
2	0	.	86	.	.	0	1	1	.	.
3	1	0.98837	86	1	.	.	1	1	0.011628	.
4	2	0.97674	85	1	.	.	1	1	0.023256	.
5	3	0.95349	84	2	.	.	1	1	0.046512	.

title "ProductLimitEstimates";
proc print data=estimates (obs=5); run;

ProductLimitEstimates

Obs	STRATUM	TRTAN	TIMELIST	AVAL	CENSOR	SURVIVAL	FAILURE	STDERR	FAILED	LEFT
1	1	1	0.000	0.000	0	1.0000	0	0	0	86
2	1	1	25.000	24.000	0	0.8811	0.1189	0.0353	10	72
3	1	1	50.000	48.000	0	0.7812	0.2188	0.0457	18	61
4	1	1	75.000	70.000	0	0.7414	0.2586	0.0488	21	53
5	1	1	100.000	97.000	0	0.6575	0.3425	0.0540	27	47

Select the number of subjects at risk

Definition: Number at risk = count of subjects who have not experienced the event and have not been censored immediately prior to a given time point.

Programming note: Display the number at risk at selected time points on the Kaplan–Meier (failure/cumulative incidence) plot.

data est(rename=(timelist=time stratum=stratumnum));
   set estimates(keep=stratum timelist left );
run;

proc print data=est (obs=8); run;

ProductLimitEstimates

Obs	STRATUMNUM	TIME	LEFT
1	1	0.000	86
2	1	25.000	72
3	1	50.000	61
4	1	75.000	53
5	1	100.000	47
6	1	125.000	45
7	1	150.000	40
8	1	175.000	37

Output: Time-point block plot and numbers at risk

Generate a Kaplan–Meier figure that includes:

• Time-point block plot (tick/marker display):
  Marks the requested time points along the x-axis for consistent visual reference.

• Numbers at risk:
  Displays the number of subjects at risk immediately prior to each selected time point beneath the x-axis.

Implementation notes

• Use PLOTS= to request the failure plot and the at-risk table.
• Use TIMELIST= to align the displayed/reporting time points across the plot and output datasets.

proc sort data=sur_fail;
   by stratumnum time;
run;
 
data new_survival (drop=j);
   merge est sur_fail ;
   length param_t $100;
   by stratumnum time;
      do j = 0 to 200 by 25;
          if time = j  then do;
            blkrsk = put(left,3.);
            blkx   = time;
         end;
      end;
    trtan = stratumnum;
    param_t="parameter name";
    paramn=1;
run;
 
proc sort data=new_survival;
   by trtan time;
run;

title "Time-point block plot and numbers at risk";
proc print data=new_survival (obs=10); run;

Time-point block plot and numbers at risk

Obs	STRATUMNUM	TIME	LEFT	SURVIVAL	ATRISK	EVENT	CENSORED	TATRISK	STRATUM	_1_SURVIVAL_	_1_CENSORED_	PARAM_T	BLKRSK	BLKX	TRTAN	PARAMN
1	1	0.000	86	1.00000	86	0	.	.	1	0.000000	.	parameter name	86	0	1	1
2	1	0.000	86	.	86	.	.	0	1	.	.	parameter name	86	0	1	1
3	1	1.000	.	0.98837	86	1	.	.	1	0.011628	.	parameter name		.	1	1
4	1	2.000	.	0.97674	85	1	.	.	1	0.023256	.	parameter name		.	1	1
5	1	3.000	.	0.95349	84	2	.	.	1	0.046512	.	parameter name		.	1	1
6	1	7.000	.	0.94186	82	1	.	.	1	0.058140	.	parameter name		.	1	1
7	1	8.000	.	.	81	0	0.94186	.	1	.	0.058140	parameter name		.	1	1
8	1	9.000	.	0.93009	80	1	.	.	1	0.069913	.	parameter name		.	1	1
9	1	12.000	.	.	79	0	0.93009	.	1	.	0.069913	parameter name		.	1	1
10	1	13.000	.	.	78	0	0.93009	.	1	.	0.069913	parameter name		.	1	1

Build PROC TEMPLATE for the Kaplan–Meier Graph

• Use dynamic statements to parameterize titles, labels, and display options.
• Use ENTRYTITLE to define the graph title (and optional subtitle).
• Apply attribute mapping for consistent group styling:
  • Line: color and pattern
  • Marker: color and symbol
• Display numbers at risk in the inner margin, printed with BLOCKPLOT at selected time points.
• Customize axes (line, ticks, tick values, labels) and set the appropriate axis type with its type-specific options.
• In the KM plot section, layer two plot types:
  • STEPPLOT for the KM curve
  • SCATTERPLOT for censored observations (and/or time-point markers)

proc template;
   define statgraph kmplot;
   dynamic x_var y_var1 y_var2 ;
    begingraph;
         *Graph title;
          entrytitle "Category = Time to Event"  /
          textattrs=(size=9pt ) pad=(bottom=20px);
 
         *Colour map;
         discreteattrmap name='colors' / ignorecase=true;
            value 'Placebo'  / lineattrs=(color=blue pattern=shortdash) markerattrs=(color=blue symbol=trianglefilled);
            value 'Low Dose'     / lineattrs=(color=red  pattern=shortdash) markerattrs=(color=red symbol=circlefilled);
            value 'High Dose'     / lineattrs=(color=green  pattern=shortdash) markerattrs=(color=green symbol=squarefilled);
         enddiscreteattrmap;
 
         discreteattrvar attrvar=gmarker var=trtan attrmap='colors';
 
            %*Start KM plot*;
            layout overlay /
               xaxisopts=(Label="Time at Risk (days)"
               display=(tickvalues line label ticks )
               type=linear
               linearopts=(tickvaluesequence=(start=0 end=200 increment=25)
                    viewmin=0 viewmax=200))
               yaxisopts=( Label="Cumulative Incidence of Subjects with Event"
               type=linear  linearopts= (viewmin=0 viewmax=1) );
 
               StepPlot X=x_var Y=y_var1 / primary=true Group=gmarker
                  LegendLabel="Cumulative Incidence of Subjects with Event" NAME="STEP";
 
               %*Censored observations are suppressed but can be added here*;
               scatterPlot X=x_var Y=y_var2 / Group=gmarker markerattrs=(symbol=plus)
                    LegendLabel="Censored" NAME="SCATTER";
 
               Mergedlegend "STEP" "SCATTER"/
                  location=inside halign=left valign=top across=1 valueattrs=(family="Arial" size=8pt);
               referenceline y=0;
 
            %*Start at risk value plot;
            innermargin;
               axistable x=blkx value=blkrsk / class=gmarker colorgroup=gmarker
                  display=(label)
                  labelattrs=(family="Arial" size=8pt)
                  valueattrs=(family="Arial" size=8pt);
 
 
            endinnermargin;
            endlayout;
      endgraph;
   end;
run;

328  ods listing close;ods html5 (id=saspy_internal) options(bitmap_mode='inline') device=svg style=HTMLBlue; ods graphics on /
328! outputfmt=png;
NOTE: Writing HTML5(SASPY_INTERNAL) Body file: sashtml20.htm
329  
330  proc template;
331     define statgraph kmplot;
332     dynamic x_var y_var1 y_var2 ;
333      begingraph;
334           *Graph title;
335            entrytitle "Category = Time to Event"  /
336            textattrs=(size=9pt ) pad=(bottom=20px);
337  
338           *Colour map;
339           discreteattrmap name='colors' / ignorecase=true;
340              value 'Placebo'  / lineattrs=(color=blue pattern=shortdash) markerattrs=(color=blue symbol=trianglefilled);
341              value 'Low Dose'     / lineattrs=(color=red  pattern=shortdash) markerattrs=(color=red symbol=circlefilled);
342              value 'High Dose'     / lineattrs=(color=green  pattern=shortdash) markerattrs=(color=green symbol=squarefilled);
343           enddiscreteattrmap;
344  
345           discreteattrvar attrvar=gmarker var=trtan attrmap='colors';
346  
347              %*Start KM plot*;
348              layout overlay /
349                 xaxisopts=(Label="Time at Risk (days)"
350                 display=(tickvalues line label ticks )
351                 type=linear
352                 linearopts=(tickvaluesequence=(start=0 end=200 increment=25)
353                      viewmin=0 viewmax=200))
354                 yaxisopts=( Label="Cumulative Incidence of Subjects with Event"
355                 type=linear  linearopts= (viewmin=0 viewmax=1) );
356  
357                 StepPlot X=x_var Y=y_var1 / primary=true Group=gmarker
358                    LegendLabel="Cumulative Incidence of Subjects with Event" NAME="STEP";
359  
360                 %*Censored observations are suppressed but can be added here*;
361                 scatterPlot X=x_var Y=y_var2 / Group=gmarker markerattrs=(symbol=plus)
362                      LegendLabel="Censored" NAME="SCATTER";
363  
364                 Mergedlegend "STEP" "SCATTER"/
365                    location=inside halign=left valign=top across=1 valueattrs=(family="Arial" size=8pt);
366                 referenceline y=0;
367  
368              %*Start at risk value plot;
369              innermargin;
370                 axistable x=blkx value=blkrsk / class=gmarker colorgroup=gmarker
371                    display=(label)
372                    labelattrs=(family="Arial" size=8pt)
373                    valueattrs=(family="Arial" size=8pt);
374  
375  
376              endinnermargin;
377              endlayout;
378        endgraph;
379     end;
NOTE: Overwriting existing template/link: Kmplot
NOTE: STATGRAPH 'Kmplot' has been saved to: WORK.TEMPLAT
380  run;
NOTE: PROCEDURE TEMPLATE used (Total process time):
      real time           0.00 seconds
      cpu time            0.00 seconds
      

381  ods html5 (id=saspy_internal) close;ods listing;
382  

Set options for the graph

• Turn on ODS Graphics and set output quality (DPI), size (height/width), and border.
• Define titles and footnotes for the figure.
• Choose the output destination (inline for Jupyter, or file output such as PNG/RTF/PDF) and apply the desired style.
• Set image naming/path options when saving files.

Create a format for treatment display

• Create a user-defined format to present treatment groups consistently in outputs.
• Use the stratum (treatment) numeric values when building the format so labels align with the KM STRATA ordering and report presentation.

proc format;
   value trtanf
   1="Placebo"
   2="TRT A"
   3="TRT B"
   ;
run;

383  ods listing close;ods html5 (id=saspy_internal) options(bitmap_mode='inline') device=svg style=HTMLBlue; ods graphics on /
383! outputfmt=png;
NOTE: Writing HTML5(SASPY_INTERNAL) Body file: sashtml21.htm
384  
385  proc format;
386     value trtanf
387     1="Placebo"
388     2="TRT A"
389     3="TRT B"
390     ;
NOTE: Format TRTANF is already on the library WORK.FORMATS.
NOTE: Format TRTANF has been output.
391  run;

NOTE: PROCEDURE FORMAT used (Total process time):
      real time           0.00 seconds
      cpu time            0.00 seconds
      

392  ods html5 (id=saspy_internal) close;ods listing;
393  

ods listing image_dpi=300;

ods graphics on / reset=all border=off
  height=13cm width=14.72cm
  antialiasmax=20000;

title1 "Cumulative Incidence Plot";
title2 "Safety Analysis Set";


/* Generate the graph using sgrender procedure */
proc sgrender data=new_survival template=kmplot;
   dynamic x_var="time" y_var1="_1_survival_" y_var2="_1_censored_" ;
   format trtan trtanf.;
run;
ods graphics off;
ods rtf close;

title;
footnote;
ods graphics off;

Cumulative Incidence Plot

Safety Analysis Set