Matt

=======================================

PATIENTS CHARACTERISTICS

=======================================

	Overall
n	131
Age (mean (SD))	62.70 (11.20)
AgeG = <45 (%)	7 ( 5.3)
Male1 = Yes (%)	101 (77.1)
Nephrectomy1 = Yes (%)	74 (57.4)
Sarcomatoid1 = Yes (%)	16 (13.8)
risk (%)
favourable	15 (11.5)
intermediate	74 (56.9)
poor	41 (31.5)
Line1 (%)
IO	9 ( 6.9)
IO+V	44 (33.6)
IO+IO	53 (40.5)
IO+IO+V	25 (19.1)
BoneMet1 = Yes (%)	37 (28.2)
LiverMet1 = Yes (%)	17 (13.0)
BrainMet1 = Yes (%)	11 ( 8.4)
iraeL1 = Yes (%)	28 (21.4)
Line2 = 1 (%)	76 (58.0)
Line3 = 1 (%)	35 (26.7)
Line4 = 1 (%)	10 ( 7.6)
Line5 = 1 (%)	1 ( 0.8)

=======================================

RESPONSE

=======================================

Baseline outcomes

	Overall
n	131
Best.Response (%)
CR	5 ( 3.9)
PD	30 (23.3)
PR	54 (41.9)
SD	40 (31.0)
ORR_12w = Yes (%)	40 (31.2)
CRR_12w = Yes (%)	99 (77.3)
OSL1fu2 (%)
OS <2y	66 (50.4)
OS 2-5y	47 (35.9)
OS ≥5y	18 (13.7)
PFSL1fu2 (%)
PFS <2y	96 (74.4)
PFS 2-5y	27 (20.9)
PFS ≥5y	6 ( 4.7)

ORR No vs Yes at week 12 and all variables

Here I try to take a first look at the relationship between ORR and other variables.

As you see, the only statistically significant variables are those at the 12 week.

I’ve created a variable called “VarHb.cut” to measure Hb_12w/Hb_0w.

	No	Yes	p
n	88	40
PltHigh_0w = Yes (%)	18 (20.5)	9 (22.5)	0.977
HbLow_0w = Yes (%)	48 (54.5)	22 (55.0)	1.000
NLR_0w.cut = ≥3 (%)	48 (55.2)	25 (62.5)	0.560
PLR_0w.cut = ≥150 (%)	61 (71.8)	28 (71.8)	1.000
PltHigh_12w = Yes (%)	16 (18.4)	1 ( 2.6)	0.034
HbLow_12w = Yes (%)	53 (60.9)	15 (37.5)	0.023
NLR_12w.cut = ≥3 (%)	47 (54.0)	12 (30.0)	0.020
PLR_12w.cut = ≥150 (%)	47 (58.8)	21 (52.5)	0.648
BoneMet1 = Yes (%)	29 (33.0)	8 (20.0)	0.198
LiverMet1 = Yes (%)	9 (10.2)	6 (15.0)	0.630
iraeL1 = Yes (%)	20 (22.7)	8 (20.0)	0.908
VarHb.cut = >1 (%)	44 (50.6)	28 (70.0)	0.063
VarNLR.cut = >1 (%)	45 (51.7)	14 (35.0)	0.118

CRR at week 12 and all variables

If we analyse CRR, it’s quite similar to the previous analysis with ORR.

Measurements at week 12 are significant. Here “VarHb.cut” is significant.

That means that if patients increase their Hb at week 12 it is related with CRR.

	No	Yes	p
n	29	99
PltHigh_0w = Yes (%)	7 (24.1)	20 (20.2)	0.843
HbLow_0w = Yes (%)	18 (62.1)	52 (52.5)	0.486
NLR_0w.cut = ≥3 (%)	18 (64.3)	55 (55.6)	0.543
PLR_0w.cut = ≥150 (%)	24 (85.7)	65 (67.7)	0.104
PltHigh_12w = Yes (%)	9 (32.1)	8 ( 8.2)	0.003
HbLow_12w = Yes (%)	22 (78.6)	46 (46.5)	0.005
NLR_12w.cut = ≥3 (%)	20 (71.4)	39 (39.4)	0.005
PLR_12w.cut = ≥150 (%)	18 (75.0)	50 (52.1)	0.072
BoneMet1 = Yes (%)	12 (41.4)	25 (25.3)	0.147
LiverMet1 = Yes (%)	6 (20.7)	9 ( 9.1)	0.168
iraeL1 = Yes (%)	4 (13.8)	24 (24.2)	0.346
VarHb.cut = >1 (%)	10 (35.7)	62 (62.6)	0.020
VarNLR.cut = >1 (%)	16 (57.1)	43 (43.4)	0.285

=======================================

PREDICTORS OF RESPONSE

=======================================

Here I’ve made a logistic regression to try to measure the relationships found in the previous analysis.

I’ve taken a look one by one and adjusted by IMDC risk.

As we are looking for “risk factors” (means Odds Ratio), the analysis is made to predict “no ORR”.

We can change the language latter.

=======================================

Predictor of not ORR adjusted by IMDC risk

=======================================

Plt High 12w

Plt high at week 12 have a ORR 8.83 (IC 95%: 1.59-166.09)

##      (Intercept) riskintermediate         riskpoor   PltHigh_12wYes 
##         1.634134         1.164024         1.100805         8.832934

## Waiting for profiling to be done...

##                      2.5 %     97.5 %
## (Intercept)      0.5517073   5.382049
## riskintermediate 0.3234541   3.842000
## riskpoor         0.2701761   4.228448
## PltHigh_12wYes   1.5992462 166.093805

Hb Low 12w

Hb low at week 12 have a ORR 2.72 (IC 95%: 1.19-6.45)

##      (Intercept) riskintermediate         riskpoor     HbLow_12wYes 
##        1.3994977        0.9350421        0.9132770        2.7188537

## Waiting for profiling to be done...

##                      2.5 %   97.5 %
## (Intercept)      0.4619766 4.678153
## riskintermediate 0.2560538 3.104009
## riskpoor         0.2104736 3.670853
## HbLow_12wYes     1.1910374 6.458384

NLR 12w cut ≥3

NLR ≥3 at week 12 have a ORR 2.65 (IC 95%: 1.16-6.30)

##      (Intercept) riskintermediate         riskpoor    NLR_12w.cut≥3 
##        1.5014965        0.9269957        1.0028330        2.6494640

## Waiting for profiling to be done...

##                      2.5 %   97.5 %
## (Intercept)      0.5034946 4.973435
## riskintermediate 0.2551503 3.062470
## riskpoor         0.2388274 3.930954
## NLR_12w.cut≥3    1.1680137 6.301835

PLR 12w cut ≥150

PLR ≥150 at week 12 have a ORR 1.24 (IC 95%: 0.55-2.77)

This value is not significant if we adjust by IMDC risk factor.

##      (Intercept) riskintermediate         riskpoor  PLR_12w.cut≥150 
##        1.7211722        0.9936375        1.0900417        1.2391278

## Waiting for profiling to be done...

##                      2.5 %   97.5 %
## (Intercept)      0.5860465 5.662498
## riskintermediate 0.2724033 3.300772
## riskpoor         0.2579696 4.346954
## PLR_12w.cut≥150  0.5516499 2.777934

VarHb = Hb_12w/Hb_0w > 1 = Increase of Hb at week 12

The ORR 0.42 (IC 95%: 0.18-0.93)

Means that this variable is a protective factor against not ORR.

##      (Intercept) riskintermediate         riskpoor      VarHb.cut>1 
##        2.6581732        1.2333207        1.8806848        0.4261822

## Waiting for profiling to be done...

##                      2.5 %    97.5 %
## (Intercept)      0.8496418 9.4978898
## riskintermediate 0.3394838 4.1292098
## riskpoor         0.4675347 7.2537447
## VarHb.cut>1      0.1842366 0.9399897

=======================================

Predictor of not CRR = PD adjusted by IMDC risk

=======================================

Plt High 12w

Plt high at week 12 have a ORR 3.71 (IC 95%: I don’t know why we have those values)

##      (Intercept) riskintermediate         riskpoor   PltHigh_12wYes 
##     7.528612e-09     2.604953e+07     5.575146e+07     3.717741e+00

## Waiting for profiling to be done...

##                         2.5 %        97.5 %
## (Intercept)                NA  4.965154e+41
## riskintermediate 4.156419e-32 3.362366e+181
## riskpoor         2.364020e-32 2.431154e+142
## PltHigh_12wYes   1.140829e+00  1.246222e+01

Hb Low 12w

Hb low at week 12 have a ORR 2.81 (IC 95%: 1.02-8.64)

##      (Intercept) riskintermediate         riskpoor     HbLow_12wYes 
##     5.962611e-09     2.065766e+07     4.588359e+07     2.818321e+00

## Waiting for profiling to be done...

##                         2.5 %        97.5 %
## (Intercept)                NA  2.123545e+41
## riskintermediate 4.571128e-32 1.447992e+180
## riskpoor         8.620805e-26 1.547314e+245
## HbLow_12wYes     1.021876e+00  8.646204e+00

NLR 12w cut ≥3

NLR ≥3 at week 12 have a ORR 2.85 (IC 95%: 1.11-7.77)

##      (Intercept) riskintermediate         riskpoor    NLR_12w.cut≥3 
##     6.470073e-09     2.005895e+07     4.771346e+07     2.850190e+00

## Waiting for profiling to be done...

##                         2.5 %        97.5 %
## (Intercept)                NA  2.965743e+41
## riskintermediate 6.501707e-26 5.080974e+246
## riskpoor         9.464987e-32 1.517139e+181
## NLR_12w.cut≥3    1.114811e+00  7.770895e+00

PLR 12w cut ≥150

PLR ≥150 at week 12 have a ORR 1.92 (IC 95%: 1.00-5.88)

##      (Intercept) riskintermediate         riskpoor  PLR_12w.cut≥150 
##     7.333909e-09     1.810423e+07     4.954075e+07     1.921884e+00

## Waiting for profiling to be done...

##                         2.5 %       97.5 %
## (Intercept)                NA 1.694152e+42
## riskintermediate 3.421541e-35           NA
## riskpoor         9.362757e-35           NA
## PLR_12w.cut≥150  6.946169e-01 5.887283e+00

VarHb = Hb_12w/Hb_0w > 1 = Increase of Hb at week 12

Means that this variable is a protective factor against not CRR.

However, I can’t get the exact value of ORR.

##      (Intercept) riskintermediate         riskpoor      VarHb.cut>1 
##     1.429299e-08     2.904180e+07     1.118738e+08     2.267677e-01

## Waiting for profiling to be done...

##                         2.5 %        97.5 %
## (Intercept)                NA  1.641666e+40
## riskintermediate 2.540193e-24 1.465387e+243
## riskpoor         7.586493e-40            NA
## VarHb.cut>1      8.259633e-02  5.763779e-01

VarNLR = NLR_12w/NLR_0w > 1 = Increase of NLR at week 12

Means that this variable is a protective factor against not CRR.

However, I can’t get the exact value of ORR.

##      (Intercept) riskintermediate         riskpoor     VarNLR.cut>1 
##     4.812088e+07     1.192060e-07     3.944923e-08     6.718712e-01

## Waiting for profiling to be done...

##                          2.5 %       97.5 %
## (Intercept)       1.052605e-23           NA
## riskintermediate 4.037019e-144 1.682019e+15
## riskpoor                    NA 2.425301e+22
## VarNLR.cut>1      2.721161e-01 1.626771e+00

=======================================

OVERALL SURVIVAL FIRST-LINE

=======================================

Differences in OS by ORR at week 12

Although there is a differences is median OS, the K-M curve is not significant.

## Call: survfit(formula = Surv(OSL1m, Ddeathcnsr) ~ ORR_12w, data = db, 
##     type = "kaplan-meier")
## 
##    3 observations deleted due to missingness 
##              n events median 0.95LCL 0.95UCL
## ORR_12w=No  88     46   40.3    17.5      NA
## ORR_12w=Yes 40     19   34.8    29.2      NA

Differences in OS by CRR at week 12

This is quire obvious. Patients with PD at 1L have worst survival.

## Call: survfit(formula = Surv(OSL1m, Ddeathcnsr) ~ CRR_12w, data = db, 
##     type = "kaplan-meier")
## 
##    3 observations deleted due to missingness 
##              n events median 0.95LCL 0.95UCL
## CRR_12w=No  29     22   11.2    8.64    18.6
## CRR_12w=Yes 99     43   51.6   34.76      NA

Differences in OS by Plt high at week 12

Plt high at week 12 seems to be associated with worst median OS (43.8 vs 14.3 months)

However, if we adjusted by IMDC risk, the HR is not significant (IC 95% 0.82-3.42).

## Call: survfit(formula = Surv(OSL1m, Ddeathcnsr) ~ PltHigh_12w, data = db, 
##     type = "kaplan-meier")
## 
##    5 observations deleted due to missingness 
##                   n events median 0.95LCL 0.95UCL
## PltHigh_12w=No  109     53   43.8   31.54    81.3
## PltHigh_12w=Yes  17     11   14.3    8.25      NA

## Call:
## coxph(formula = Surv(OSL1m, Ddeathcnsr) ~ risk + PltHigh_12w, 
##     data = db)
## 
##   n= 125, number of events= 64 
##    (6 observations deleted due to missingness)
## 
##                    coef exp(coef) se(coef)     z Pr(>|z|)   
## riskintermediate 0.7219    2.0584   0.5296 1.363  0.17281   
## riskpoor         1.6774    5.3516   0.5613 2.988  0.00281 **
## PltHigh_12wYes   0.5179    1.6786   0.3638 1.424  0.15451   
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
##                  exp(coef) exp(-coef) lower .95 upper .95
## riskintermediate     2.058     0.4858    0.7291     5.811
## riskpoor             5.352     0.1869    1.7811    16.080
## PltHigh_12wYes       1.679     0.5957    0.8228     3.424
## 
## Concordance= 0.671  (se = 0.031 )
## Likelihood ratio test= 21.03  on 3 df,   p=1e-04
## Wald test            = 22.07  on 3 df,   p=6e-05
## Score (logrank) test = 24.92  on 3 df,   p=2e-05

Differences in OS by Hb low at week 12

Hb low at week 12 seems to be associated with worst median OS (51.6 vs 20.3 months)

However, if we adjusted by IMDC risk,the HR is not significant (IC 95% 0.84-2.44).

## Call: survfit(formula = Surv(OSL1m, Ddeathcnsr) ~ HbLow_12w, data = db, 
##     type = "kaplan-meier")
## 
##    4 observations deleted due to missingness 
##                n events median 0.95LCL 0.95UCL
## HbLow_12w=No  59     27   51.6    41.3      NA
## HbLow_12w=Yes 68     37   20.3    14.6      NA

## Call:
## coxph(formula = Surv(OSL1m, Ddeathcnsr) ~ risk + HbLow_12w, data = db)
## 
##   n= 126, number of events= 64 
##    (5 observations deleted due to missingness)
## 
##                    coef exp(coef) se(coef)     z Pr(>|z|)   
## riskintermediate 0.6876    1.9890   0.5304 1.296  0.19483   
## riskpoor         1.6509    5.2118   0.5622 2.937  0.00332 **
## HbLow_12wYes     0.3610    1.4347   0.2704 1.335  0.18190   
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
##                  exp(coef) exp(-coef) lower .95 upper .95
## riskintermediate     1.989     0.5028    0.7033     5.625
## riskpoor             5.212     0.1919    1.7316    15.687
## HbLow_12wYes         1.435     0.6970    0.8445     2.438
## 
## Concordance= 0.689  (se = 0.03 )
## Likelihood ratio test= 21.06  on 3 df,   p=1e-04
## Wald test            = 21.46  on 3 df,   p=8e-05
## Score (logrank) test = 23.96  on 3 df,   p=3e-05

Differences in OS by NLR at week 12

NLR ≥ 3 at week 12 is associated with worst median OS (46.7 vs 17.5 months)

HR 2.11 (IC 95% 1.27-3.48), adjusted by IMDC risk.

## Call: survfit(formula = Surv(OSL1m, Ddeathcnsr) ~ NLR_12w.cut, data = db, 
##     type = "kaplan-meier")
## 
##    4 observations deleted due to missingness 
##                 n events median 0.95LCL 0.95UCL
## NLR_12w.cut=<3 68     29   46.7    40.3      NA
## NLR_12w.cut=≥3 59     35   17.5    12.6    53.6

## Call:
## coxph(formula = Surv(OSL1m, Ddeathcnsr) ~ risk + NLR_12w.cut, 
##     data = db)
## 
##   n= 126, number of events= 64 
##    (5 observations deleted due to missingness)
## 
##                    coef exp(coef) se(coef)     z Pr(>|z|)   
## riskintermediate 0.6423    1.9008   0.5314 1.209  0.22681   
## riskpoor         1.6654    5.2881   0.5578 2.986  0.00283 **
## NLR_12w.cut≥3    0.7462    2.1089   0.2567 2.906  0.00366 **
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
##                  exp(coef) exp(-coef) lower .95 upper .95
## riskintermediate     1.901     0.5261    0.6708     5.386
## riskpoor             5.288     0.1891    1.7721    15.779
## NLR_12w.cut≥3        2.109     0.4742    1.2750     3.488
## 
## Concordance= 0.698  (se = 0.031 )
## Likelihood ratio test= 27.71  on 3 df,   p=4e-06
## Wald test            = 27.48  on 3 df,   p=5e-06
## Score (logrank) test = 30.82  on 3 df,   p=9e-07

Differences in OS by NLR at week 0 + week 12

This is also interesting.

Patients who achieve a NLR < 3 at week 12 have better prognosis.

## Call: survfit(formula = Surv(OSL1m, Ddeathcnsr) ~ NLR_0w.cut + NLR_12w.cut, 
##     data = db, type = "kaplan-meier")
## 
##    4 observations deleted due to missingness 
##                                n events median 0.95LCL 0.95UCL
## NLR_0w.cut=<3, NLR_12w.cut=<3 41     15   85.0    41.3      NA
## NLR_0w.cut=<3, NLR_12w.cut=≥3 13      7   26.8    17.5      NA
## NLR_0w.cut=≥3, NLR_12w.cut=<3 27     14   40.3    24.0      NA
## NLR_0w.cut=≥3, NLR_12w.cut=≥3 46     28   14.7    10.3    62.6

Differences in OS by NLR at week 12 and risk

The importance of NLR at week 12 remains in all IMDC risk categories.

## Call: survfit(formula = Surv(OSL1m, Ddeathcnsr) ~ risk + NLR_12w.cut, 
##     data = db, type = "kaplan-meier")
## 
##    5 observations deleted due to missingness 
##                                    n events median 0.95LCL 0.95UCL
## risk=favourable, NLR_12w.cut=<3   11      3     NA   29.21      NA
## risk=favourable, NLR_12w.cut=≥3    3      1   62.6   62.55      NA
## risk=intermediate, NLR_12w.cut=<3 45     20   46.7   40.31      NA
## risk=intermediate, NLR_12w.cut=≥3 29     15   31.5   17.54      NA
## risk=poor, NLR_12w.cut=<3         12      6   41.3   14.75      NA
## risk=poor, NLR_12w.cut=≥3         26     19   10.3    7.85    18.6

Differences in OS by PLR at week 12

There is no differences in OS with PLR at week 12

## Call: survfit(formula = Surv(OSL1m, Ddeathcnsr) ~ PLR_12w.cut, data = db, 
##     type = "kaplan-meier")
## 
##    11 observations deleted due to missingness 
##                   n events median 0.95LCL 0.95UCL
## PLR_12w.cut=<150 52     23   43.8    31.6      NA
## PLR_12w.cut=≥150 68     36   33.4    20.3      NA

Matt

TapiaJC

2023-05-03

=======================================

=======================================

PATIENTS CHARACTERISTICS

=======================================

=======================================

=======================================

=======================================

RESPONSE

=======================================

=======================================

Baseline outcomes

ORR No vs Yes at week 12 and all variables

Here I try to take a first look at the relationship between ORR and other variables.

As you see, the only statistically significant variables are those at the 12 week.

I’ve created a variable called “VarHb.cut” to measure Hb_12w/Hb_0w.

CRR at week 12 and all variables

If we analyse CRR, it’s quite similar to the previous analysis with ORR.

Measurements at week 12 are significant. Here “VarHb.cut” is significant.

That means that if patients increase their Hb at week 12 it is related with CRR.

=======================================

=======================================

PREDICTORS OF RESPONSE

=======================================

=======================================

Here I’ve made a logistic regression to try to measure the relationships found in the previous analysis.

I’ve taken a look one by one and adjusted by IMDC risk.

As we are looking for “risk factors” (means Odds Ratio), the analysis is made to predict “no ORR”.

We can change the language latter.

=======================================

Predictor of not ORR adjusted by IMDC risk

=======================================

Plt High 12w

Plt high at week 12 have a ORR 8.83 (IC 95%: 1.59-166.09)

Hb Low 12w

Hb low at week 12 have a ORR 2.72 (IC 95%: 1.19-6.45)

NLR 12w cut ≥3

NLR ≥3 at week 12 have a ORR 2.65 (IC 95%: 1.16-6.30)

PLR 12w cut ≥150

PLR ≥150 at week 12 have a ORR 1.24 (IC 95%: 0.55-2.77)

This value is not significant if we adjust by IMDC risk factor.

VarHb = Hb_12w/Hb_0w > 1 = Increase of Hb at week 12

The ORR 0.42 (IC 95%: 0.18-0.93)

Means that this variable is a protective factor against not ORR.

=======================================

Predictor of not CRR = PD adjusted by IMDC risk

=======================================

Plt High 12w

Plt high at week 12 have a ORR 3.71 (IC 95%: I don’t know why we have those values)

Hb Low 12w

Hb low at week 12 have a ORR 2.81 (IC 95%: 1.02-8.64)

NLR 12w cut ≥3

NLR ≥3 at week 12 have a ORR 2.85 (IC 95%: 1.11-7.77)

PLR 12w cut ≥150

PLR ≥150 at week 12 have a ORR 1.92 (IC 95%: 1.00-5.88)

VarHb = Hb_12w/Hb_0w > 1 = Increase of Hb at week 12

Means that this variable is a protective factor against not CRR.

However, I can’t get the exact value of ORR.

VarNLR = NLR_12w/NLR_0w > 1 = Increase of NLR at week 12

Means that this variable is a protective factor against not CRR.

However, I can’t get the exact value of ORR.

=======================================

=======================================

OVERALL SURVIVAL FIRST-LINE

=======================================

=======================================

Differences in OS by ORR at week 12

Although there is a differences is median OS, the K-M curve is not significant.

Differences in OS by CRR at week 12

This is quire obvious. Patients with PD at 1L have worst survival.

Differences in OS by Plt high at week 12

Plt high at week 12 seems to be associated with worst median OS (43.8 vs 14.3 months)

However, if we adjusted by IMDC risk, the HR is not significant (IC 95% 0.82-3.42).

Differences in OS by Hb low at week 12

Hb low at week 12 seems to be associated with worst median OS (51.6 vs 20.3 months)

However, if we adjusted by IMDC risk,the HR is not significant (IC 95% 0.84-2.44).

Differences in OS by NLR at week 12

NLR ≥ 3 at week 12 is associated with worst median OS (46.7 vs 17.5 months)