1
00:00:06,970 --> 00:00:09,930
In this episode of EHA 
Unplugged, we will revisit a 

2
00:00:09,930 --> 00:00:14,450
spotlight talk from EHA 2024. 
Doctor Yurian Firstlaus from the

3
00:00:14,450 --> 00:00:17,450
Erasmus Medical Center in 
Rotterdam discusses the need to 

4
00:00:17,450 --> 00:00:20,410
use innovative clinical trial 
designs in the era of 

5
00:00:20,410 --> 00:00:22,370
personalized medicine in 
hematology. 

6
00:00:22,490 --> 00:00:25,010
My name is Yurian Firstlaus. 
I'm a hematologist at the 

7
00:00:25,010 --> 00:00:30,290
Erasmus MCA in Rotterdam and I'm
also highly involved in trial 

8
00:00:30,290 --> 00:00:32,960
design and within the whole form
and. 

9
00:00:33,400 --> 00:00:38,600
Related to this topic, do some 
research so for those of you who

10
00:00:38,600 --> 00:00:40,320
are. 
Aware of clinical trial design. 

11
00:00:41,080 --> 00:00:43,440
There are a couple of stages to 
consider. 

12
00:00:43,920 --> 00:00:46,640
In the clinical trial 
development and the 1st. 

13
00:00:47,040 --> 00:00:48,280
Part of trial. 
Development. 

14
00:00:48,280 --> 00:00:50,840
Is the idea. 
Everything starts with an idea 

15
00:00:50,840 --> 00:00:53,480
and a drug and. 
Opportunity to investigate. 

16
00:00:55,160 --> 00:00:59,880
An expert do an experiment and 
usually use a control and when 

17
00:00:59,880 --> 00:01:01,920
there is a design you need to 
be. 

18
00:01:01,920 --> 00:01:03,680
Aware of the effect size of 
your. 

19
00:01:04,319 --> 00:01:07,120
Drug and discuss. 
That with your statistician. 

20
00:01:07,440 --> 00:01:09,080
And as a result of that. 
There usually. 

21
00:01:09,080 --> 00:01:12,680
Comes a protocol and. 
Eventually you'll start. 

22
00:01:13,120 --> 00:01:17,400
Recruiting patients and when you
recruit patients, there might be

23
00:01:18,440 --> 00:01:20,960
room for interim analysis. 
When you. 

24
00:01:20,960 --> 00:01:25,960
Have the time and the equipment 
to do that and at the end you do

25
00:01:25,960 --> 00:01:29,440
a final analysis which. 
Is usually set out at the 

26
00:01:29,440 --> 00:01:32,360
beginning. 
At the design of a trial and at 

27
00:01:32,360 --> 00:01:34,480
final analysis. 
You execute everything you. 

28
00:01:34,960 --> 00:01:36,720
You thought of when you designed
the trial? 

29
00:01:38,040 --> 00:01:40,080
This is quite a. 
Fixed and sometimes. 

30
00:01:40,080 --> 00:01:42,800
Rigid process. 
Where there's no room for 

31
00:01:42,800 --> 00:01:45,200
changes during the trial and 
these. 

32
00:01:45,200 --> 00:01:47,600
Are actually exactly. 
Those things that might be 

33
00:01:47,600 --> 00:01:49,800
innovative. 
Which I will talk about today. 

34
00:01:51,800 --> 00:01:54,120
So why do we need to innovate 
trial designs? 

35
00:01:54,960 --> 00:01:57,240
We are all aware. 
Especially in hematology. 

36
00:01:57,240 --> 00:01:59,840
Of the. 
ERA of personalized medicine 

37
00:02:00,200 --> 00:02:02,000
there are multiple drugs 
becoming. 

38
00:02:02,000 --> 00:02:06,200
Available specific to. 
Subtypes of patients who have. 

39
00:02:06,440 --> 00:02:10,080
A target. 
Mutation, for instance, for whom

40
00:02:10,600 --> 00:02:13,080
those drugs are available but in
the clinical trial. 

41
00:02:13,320 --> 00:02:15,280
It's usually difficult to. 
Enroll. 

42
00:02:15,280 --> 00:02:18,440
Large subsets of. 
Patients when there's only a 

43
00:02:18,440 --> 00:02:19,960
few. 
Patients who have that. 

44
00:02:19,960 --> 00:02:24,000
Actual mutation. 
So in that era, we need to think

45
00:02:24,000 --> 00:02:28,120
of other approaches to clinical 
trial design and also since 

46
00:02:28,160 --> 00:02:33,760
outcomes of our patients are 
getting so much better, studies 

47
00:02:33,760 --> 00:02:38,080
take a long time to enroll 
patients and as a result we need

48
00:02:38,080 --> 00:02:42,760
to wait a long time for the 
final analysis and studies 

49
00:02:42,760 --> 00:02:45,040
become expensive because of that
long follow up time. 

50
00:02:45,960 --> 00:02:49,880
So these are, I believe. 
Reasons to innovate or think of 

51
00:02:49,880 --> 00:02:54,120
innovation in clinical trial 
designs and there's also. 

52
00:02:54,120 --> 00:02:56,800
So I talked about the targeted. 
Treatments which are also 

53
00:02:56,800 --> 00:02:59,920
biomarker driven approaches and 
on the other hand, when talking 

54
00:02:59,920 --> 00:03:01,920
about time and costs. 
And we have. 

55
00:03:01,920 --> 00:03:05,160
The early endpoints which is a 
different. 

56
00:03:05,160 --> 00:03:07,800
Topic I believe. 
But surrogate endpoints who 

57
00:03:07,800 --> 00:03:11,760
might serve, for instance, for 
overall survival, which MRD is 

58
00:03:11,760 --> 00:03:15,560
for instance being used, those 
early endpoints might. 

59
00:03:15,560 --> 00:03:20,520
Actually facilitate a. 
Sooner end of your trial, but 

60
00:03:20,520 --> 00:03:23,400
that's a different topic to 
discuss and I'm not going to 

61
00:03:23,400 --> 00:03:26,240
focus on that today. 
There's also another methods 

62
00:03:26,600 --> 00:03:29,800
which is more important I 
believe is early intervention. 

63
00:03:30,160 --> 00:03:33,160
For a very efficacious or a 
futile trial. 

64
00:03:33,880 --> 00:03:35,880
And which? 
Is getting very. 

65
00:03:35,880 --> 00:03:38,040
Hot in the last. 
Years is the. 

66
00:03:38,040 --> 00:03:40,480
Use of external control data for
your clinical trial, which I 

67
00:03:40,480 --> 00:03:41,400
will. 
Also address. 

68
00:03:43,400 --> 00:03:45,520
So when we talk about personal, 
personalized. 

69
00:03:45,520 --> 00:03:47,280
Medicine in clinical trial 
designs there. 

70
00:03:47,280 --> 00:03:52,280
Are three key types of trial. 
Design one should consider and 

71
00:03:52,280 --> 00:03:53,760
the first. 
Is the basket. 

72
00:03:53,760 --> 00:03:55,120
Trial which? 
Is a type. 

73
00:03:55,120 --> 00:03:57,160
Of trial which? 
Usually. 

74
00:03:57,160 --> 00:04:00,960
Enrolls multiple diseases where 
there's one targeted 

75
00:04:00,960 --> 00:04:03,240
intervention. 
Specific for usually. 

76
00:04:03,240 --> 00:04:09,280
Mutation and that common target 
is is investigated in such a 

77
00:04:09,280 --> 00:04:10,320
trial but. 
Such a trial could. 

78
00:04:10,320 --> 00:04:14,280
Include solid oncology. 
Patients with also hematology 

79
00:04:14,280 --> 00:04:16,680
patients. 
And it's really a mixed bag of. 

80
00:04:16,680 --> 00:04:21,200
Different types of diseases. 
On the other hand, we have also 

81
00:04:21,519 --> 00:04:24,160
umbrella trials. 
Which is usually for a single 

82
00:04:24,160 --> 00:04:27,840
disease, but within. 
That single disease when there 

83
00:04:27,840 --> 00:04:30,840
are multiple targeted 
interventions available, those 

84
00:04:30,840 --> 00:04:34,360
targeted interventions are 
investigated in such a trial and

85
00:04:34,360 --> 00:04:35,760
in in these. 
Types of trials. 

86
00:04:35,760 --> 00:04:37,480
There could even be a 
randomization. 

87
00:04:38,720 --> 00:04:41,760
Both of these trials basket. 
Trials and umbrella trials. 

88
00:04:42,080 --> 00:04:45,640
Are early phase trials usually? 
Phase one or phase two clinical 

89
00:04:45,640 --> 00:04:47,520
trials? 
Mostly phase two clinical trials

90
00:04:49,800 --> 00:04:53,640
during the COVID-19 pandemic, 
platform trials have become 

91
00:04:53,640 --> 00:04:55,640
very. 
Popular where there's. 

92
00:04:55,880 --> 00:04:59,400
A continuously running standard 
of care arm where there. 

93
00:04:59,400 --> 00:05:01,240
Are always. 
Patients receiving that standard

94
00:05:01,240 --> 00:05:04,200
of care and when there, when 
there are becoming 

95
00:05:04,720 --> 00:05:06,800
interventional treatments, 
alternative treatments 

96
00:05:06,800 --> 00:05:10,240
available, patients are at that 
point could be randomized to 

97
00:05:10,320 --> 00:05:14,520
that intervention and there are 
in the platform trial. 

98
00:05:14,520 --> 00:05:18,400
Usually a couple. 
Of interim analysis where at 

99
00:05:18,400 --> 00:05:22,000
that time an intervention is 
reassessed and when there's a 

100
00:05:22,000 --> 00:05:24,800
clear sign of. 
Efficacy or futility? 

101
00:05:25,480 --> 00:05:27,360
That drug, that intervention is 
dropped. 

102
00:05:29,240 --> 00:05:31,640
These trials are attractive 
because if you. 

103
00:05:31,840 --> 00:05:35,120
If you design it properly, the 
the standard of care arm could 

104
00:05:35,120 --> 00:05:36,400
run for. 
Multiple years. 

105
00:05:36,720 --> 00:05:39,160
And when available, different 
types of. 

106
00:05:39,160 --> 00:05:40,640
Drugs or. 
Experiments could be. 

107
00:05:40,640 --> 00:05:42,400
Added. 
During that trial, and this 

108
00:05:42,400 --> 00:05:44,200
serves as a sort of master. 
Protocol. 

109
00:05:45,160 --> 00:05:47,520
Which you could. 
Always rely. 

110
00:05:47,520 --> 00:05:49,280
On when adding the new 
treatments. 

111
00:05:50,720 --> 00:05:52,400
These types of trials are a bit 
more. 

112
00:05:52,400 --> 00:05:57,200
Difficult to design when you 
talk to your statistician, but I

113
00:05:57,520 --> 00:05:59,520
really encourage you if you. 
Have these type of. 

114
00:05:59,520 --> 00:06:01,320
Treatments within your disease. 
Type. 

115
00:06:01,600 --> 00:06:03,680
To consider these. 
Trials because the. 

116
00:06:04,000 --> 00:06:06,560
Standard of continuously 
running, standard of care arm is

117
00:06:06,560 --> 00:06:12,760
very attractive. 
The next thing I want to address

118
00:06:12,760 --> 00:06:15,800
is, like I mentioned, the early 
intervention for and very. 

119
00:06:15,920 --> 00:06:19,120
Efficacious or futile trial? 
No one wants to do a. 

120
00:06:19,120 --> 00:06:20,720
Trial who? 
At the end. 

121
00:06:20,720 --> 00:06:23,280
Is is either. 
Very negative. 

122
00:06:23,280 --> 00:06:26,600
Or. 
Very positive and always comes 

123
00:06:26,600 --> 00:06:30,280
the question if you could have 
seen that coming and these type 

124
00:06:30,280 --> 00:06:31,440
of. 
Trials are called. 

125
00:06:31,440 --> 00:06:36,320
Adaptive trial designs and the 
example I want to highlight here

126
00:06:36,320 --> 00:06:39,760
is based on the whole phone data
and this is an example in 

127
00:06:39,760 --> 00:06:42,720
multiple myeloma. 
Where the trial? 

128
00:06:42,720 --> 00:06:46,440
Randomized 2 two treatment arms,
a control treatment arm versus 

129
00:06:46,440 --> 00:06:49,800
an intervention arm, and at the 
design of the. 

130
00:06:49,800 --> 00:06:54,000
Study it was. 
Estimated that 668 patients 

131
00:06:54,000 --> 00:06:55,000
needed. 
To be enrolled. 

132
00:06:56,560 --> 00:06:58,920
And at final analysis. 
The result? 

133
00:06:58,920 --> 00:07:03,000
Was quite disappointing because 
there was no difference between 

134
00:07:03,000 --> 00:07:05,120
the experimental and the control
arm. 

135
00:07:05,120 --> 00:07:07,960
We questioned whether we could 
have seen this coming and 

136
00:07:07,960 --> 00:07:09,040
whether. 
If we could. 

137
00:07:09,560 --> 00:07:11,960
Introduce interim analysis. 
Maybe have? 

138
00:07:12,120 --> 00:07:15,000
Changed the course in this. 
Trial retrospectively I. 

139
00:07:15,000 --> 00:07:18,240
Will show you 2 methods now. 
The first one is the group 

140
00:07:18,240 --> 00:07:20,880
sequential design. 
Which actually allows you. 

141
00:07:20,880 --> 00:07:24,800
To introduce interim analysis 
and AT interim analysis. 

142
00:07:25,080 --> 00:07:26,080
You could. 
You could. 

143
00:07:26,080 --> 00:07:28,720
Decide the based on the the 
boundaries. 

144
00:07:28,720 --> 00:07:31,960
I will show you whether a trial.
Should be continued. 

145
00:07:33,240 --> 00:07:36,120
And in this example, we 
introduced 2 interim analysis. 

146
00:07:36,520 --> 00:07:39,920
The first interim analysis was 
when there were still patients 

147
00:07:39,920 --> 00:07:43,080
being enrolled and at the second
interim analysis. 

148
00:07:43,080 --> 00:07:44,920
We observed. 
That all patients were already 

149
00:07:45,200 --> 00:07:46,880
included. 
But at 2:00. 

150
00:07:46,880 --> 00:07:51,280
Thirds of events we had still 
some follow up time to monitor 

151
00:07:52,240 --> 00:07:54,520
and if you for instance could 
consider at the first interim 

152
00:07:54,520 --> 00:07:56,920
analysis to stop the trial. 
You would actually. 

153
00:07:57,240 --> 00:08:00,280
Save patients from a potential 
futile or. 

154
00:08:00,400 --> 00:08:03,880
Efficacious treatment and. 
At the second interim analysis, 

155
00:08:04,240 --> 00:08:06,280
you could primarily say follow 
up time. 

156
00:08:07,880 --> 00:08:10,000
So if we. 
Again retrospectively. 

157
00:08:10,000 --> 00:08:13,120
Apply this to the. 
To this hopeful multiple myeloma

158
00:08:13,120 --> 00:08:16,760
trial, you'll see interim 
analysis one, interim analysis 

159
00:08:16,760 --> 00:08:19,520
two, we were not able. 
To firmly say that this. 

160
00:08:19,520 --> 00:08:22,160
Trial was going to be a negative
trial and we should stop the 

161
00:08:22,160 --> 00:08:24,200
trial. 
And a key concept of the group 

162
00:08:24,200 --> 00:08:27,600
sequential design is that when 
more data are being observed, 

163
00:08:28,440 --> 00:08:31,080
the bounds of efficacy and 
futility. 

164
00:08:31,280 --> 00:08:33,440
Are getting. 
Closer, In other words, your 

165
00:08:33,440 --> 00:08:37,600
estimate of of the data is 
getting more precise and 

166
00:08:37,600 --> 00:08:39,000
therefore you're more. 
Precisely to say. 

167
00:08:39,000 --> 00:08:42,240
If your trial is going to be 
still promising or whether your 

168
00:08:42,240 --> 00:08:44,039
trial is going to be a negative 
trial. 

169
00:08:44,080 --> 00:08:47,480
At the final analysis, we did 
not observe any difference. 

170
00:08:47,480 --> 00:08:51,000
Another approach we introduced. 
Here is a sample size. 3 

171
00:08:51,000 --> 00:08:55,160
estimation, design and what? 
Is really attractive here is 

172
00:08:55,160 --> 00:08:57,560
that when you. 
Are conducting your. 

173
00:08:57,560 --> 00:09:00,440
Trial. 
You're reassessing your power, 

174
00:09:01,280 --> 00:09:04,160
and the power is important when 
designing a trial because you 

175
00:09:04,160 --> 00:09:06,600
want to be sure you want, you 
want to have some. 

176
00:09:06,840 --> 00:09:09,000
Probability that your. 
Trial is actually going to be 

177
00:09:09,560 --> 00:09:12,080
showing a real effect and when 
the. 

178
00:09:12,080 --> 00:09:15,520
Power is very high. 
Above 90% usually. 

179
00:09:16,040 --> 00:09:17,600
That's a very favourable result 
and. 

180
00:09:17,600 --> 00:09:19,680
That's a scenario where you 
could consider. 

181
00:09:19,680 --> 00:09:21,320
Stopping your trial for 
efficacy. 

182
00:09:22,600 --> 00:09:24,600
On the other hand, when your. 
Trial is at. 

183
00:09:24,840 --> 00:09:28,160
Below 30%. 
Which is a probability of 30. 

184
00:09:28,160 --> 00:09:30,120
Percent that you're going to. 
Observe a real. 

185
00:09:30,160 --> 00:09:32,960
Difference. 
That's considered a futile 

186
00:09:32,960 --> 00:09:34,520
trial. 
And then usually a trial is 

187
00:09:34,520 --> 00:09:38,040
stopped and the promising zone 
is the interesting zone where 

188
00:09:38,040 --> 00:09:39,440
and that's. 
Usually set at. 

189
00:09:39,480 --> 00:09:42,160
Between 30 to 90%. 
Where the sample. 

190
00:09:42,160 --> 00:09:43,840
Size is being re. 
Estimated. 

191
00:09:44,280 --> 00:09:47,400
So during your trial, you. 
Have the feeling that this trial

192
00:09:47,800 --> 00:09:52,600
might have some interest, but at
that time you're not sure 

193
00:09:52,800 --> 00:09:53,760
definitely. 
Not sure whether. 

194
00:09:53,760 --> 00:09:56,400
Your trial is going to show that
eventually and then you could 

195
00:09:56,400 --> 00:09:59,360
enroll additional patients. 
You could re estimate your 

196
00:09:59,360 --> 00:10:03,640
sample size and in this scenario
that could have been resolved in

197
00:10:03,640 --> 00:10:06,040
1/3 of additional patients in 
your trial. 

198
00:10:07,760 --> 00:10:10,120
When we calculated the 
conditional power, and we did 

199
00:10:10,120 --> 00:10:13,400
that at the second interim 
analysis, we found that this. 

200
00:10:13,400 --> 00:10:18,360
Study actually had. 17% 
conditional power, meaning that 

201
00:10:18,360 --> 00:10:20,720
this trial was very unlikely. 
To show efficacy. 

202
00:10:21,160 --> 00:10:25,440
And actually was futile and 
therefore in this scenario 

203
00:10:25,440 --> 00:10:28,600
retrospectively again. 
We could have decided that this 

204
00:10:29,520 --> 00:10:32,560
trials should have stopped. 
Which in this case because it 

205
00:10:32,560 --> 00:10:37,680
was after already. 2/3 of. 
Events, meaning the trial was 

206
00:10:37,680 --> 00:10:40,960
already completed for inclusion.
But it could have saved. 

207
00:10:40,960 --> 00:10:44,560
Follow up time. 
These are the. 

208
00:10:44,560 --> 00:10:47,160
Adaptive trial designs. 
We also recently. 

209
00:10:48,000 --> 00:10:50,680
Used a Bayesian type of analysis
to. 

210
00:10:50,720 --> 00:10:53,520
Investigate whether we could. 
Improve decision making in a 

211
00:10:53,520 --> 00:10:57,320
clinical trial by using a 
Bayesian methodology. 

212
00:10:58,080 --> 00:11:03,640
The basin methodology. 
Is has three key concepts 

213
00:11:03,840 --> 00:11:07,280
consisting of a prior. 
Which is the belief what you 

214
00:11:07,280 --> 00:11:09,480
have before you start on a. 
Trial. 

215
00:11:09,640 --> 00:11:11,640
And this is, for instance, a 
complete remission rate you 

216
00:11:11,800 --> 00:11:15,160
expect for a leukemia patient. 
Which is usually. 

217
00:11:15,160 --> 00:11:18,000
After induction treatment, about
80 percent, 85%. 

218
00:11:18,720 --> 00:11:21,360
And when you observe additional.
Data in In a very positive 

219
00:11:21,360 --> 00:11:22,960
trial, you might. 
Observe a very. 

220
00:11:22,960 --> 00:11:24,480
High remission rate, for 
instance. 

221
00:11:24,760 --> 00:11:27,200
Updated likelihood. 
Your guess what you observe is 

222
00:11:27,200 --> 00:11:30,160
is better, so your updated. 
Belief is. 

223
00:11:30,200 --> 00:11:32,000
In between, that's the 
posterior. 

224
00:11:32,360 --> 00:11:36,360
And that's a key concept of 
Bayesian analysis, which. 

225
00:11:36,360 --> 00:11:38,280
I will show. 
You how that performed in our 

226
00:11:38,360 --> 00:11:43,360
AML analysis and this is again 
maybe a disappointing hopeful 

227
00:11:43,360 --> 00:11:45,960
example. 
But it helps to use these. 

228
00:11:45,960 --> 00:11:49,000
Examples. 
To investigate what what could 

229
00:11:49,000 --> 00:11:51,000
have changed? 
And if we looked in this. 

230
00:11:51,000 --> 00:11:54,480
Trial this was. 
The whole from 132 trial we 

231
00:11:54,480 --> 00:11:59,840
randomized almost 800 patients 
between a controlled treatment 

232
00:11:59,840 --> 00:12:03,200
and experimental treatments and 
we observed clearly no 

233
00:12:03,200 --> 00:12:07,280
difference, however, at the 
design of the study. 

234
00:12:07,720 --> 00:12:09,880
The expected hazard ratio for 
benefit. 

235
00:12:09,880 --> 00:12:12,360
Was .76. 
While the observed. 

236
00:12:12,800 --> 00:12:16,240
Hazard ratio was .99. 
And again, we questioned what 

237
00:12:16,240 --> 00:12:19,000
we've seen at this coming when 
we introduced interim analysis. 

238
00:12:19,760 --> 00:12:22,080
So that's what we did. 
We introduced 4 interim analysis

239
00:12:22,080 --> 00:12:24,040
and when using that group, 
sequential design. 

240
00:12:24,480 --> 00:12:28,120
We actually observed that. 
At the third interim analysis, 

241
00:12:28,480 --> 00:12:31,520
the fertility bound was crossed.
So at the third interim analysis

242
00:12:31,520 --> 00:12:35,960
after 450 patients we using. 
This type of. 

243
00:12:36,640 --> 00:12:39,880
Adaptive designs might have 
considered to stop the. 

244
00:12:39,880 --> 00:12:43,040
Trial because efficacy. 
Was never going to be shown. 

245
00:12:44,760 --> 00:12:48,160
However, if we. 
Use a Bayesian analysis and one 

246
00:12:48,160 --> 00:12:49,880
of the strengths of Bayesian 
analysis. 

247
00:12:50,120 --> 00:12:53,120
Is that you're also able. 
To reinforce your control arm by

248
00:12:53,680 --> 00:12:57,080
adding additional patients. 
And we use the previous. 

249
00:12:57,080 --> 00:13:00,120
Hopeful trial also including AML
patients. 

250
00:13:01,160 --> 00:13:06,200
We actually showed. 
That using that Bayesian 

251
00:13:06,200 --> 00:13:08,800
approach after the first interim
analysis. 

252
00:13:08,800 --> 00:13:12,800
So after 150 patients there was 
a 1.2%. 

253
00:13:12,800 --> 00:13:16,760
Probability of being. 
Successful of being able to show

254
00:13:16,760 --> 00:13:20,800
that .76 expected hazard. 
Ratio we had. 

255
00:13:20,800 --> 00:13:24,880
At the beginning of the trial. 
So using a Bayesian analysis and

256
00:13:24,880 --> 00:13:27,800
reinforcing your control arm, we
were. 

257
00:13:28,240 --> 00:13:30,320
Able to show that that trial was
already. 

258
00:13:30,320 --> 00:13:32,760
Very early, going to be a 
negative trial. 

259
00:13:34,640 --> 00:13:37,240
So these are. 
Examples of methods to introduce

260
00:13:37,240 --> 00:13:39,000
interim analysis which might 
really. 

261
00:13:39,000 --> 00:13:41,040
Help you. 
During a trial. 

262
00:13:42,480 --> 00:13:46,400
To to be sure that you're. 
Able to demonstrate the beliefs 

263
00:13:46,400 --> 00:13:47,920
you had when you designed the 
trial. 

264
00:13:50,800 --> 00:13:53,120
External control data, I 
mentioned that at the beginning,

265
00:13:53,120 --> 00:13:57,960
are increasingly being used, and
we also investigated thinking of

266
00:13:57,960 --> 00:13:59,720
the Bayesian analysis. 
Where you're able to. 

267
00:13:59,760 --> 00:14:02,440
Reinforce your control arm if 
you're. 

268
00:14:02,440 --> 00:14:04,200
Able to. 
Use external control. 

269
00:14:04,200 --> 00:14:06,920
Data whether that helps to 
compare to. 

270
00:14:06,920 --> 00:14:10,600
Reinforce your control arm. 
But the question is, are 

271
00:14:10,600 --> 00:14:12,920
external control? 
Data really helpful. 

272
00:14:13,200 --> 00:14:15,440
Are those really comparable? 
To your clinical trial. 

273
00:14:16,640 --> 00:14:20,320
So we again I used an approach. 
Where we. 

274
00:14:20,600 --> 00:14:23,480
Took the control arm of a 
hopeful trial and we. 

275
00:14:23,480 --> 00:14:26,120
Sought for. 
External control data to 

276
00:14:26,120 --> 00:14:30,440
reinforce the control arm and we
identified 2 potential 

277
00:14:30,440 --> 00:14:32,040
alternative. 
Data sources. 

278
00:14:32,440 --> 00:14:34,760
The 1st is the. 
Harmony and Alliance data. 

279
00:14:34,760 --> 00:14:36,320
Source. 
And the other is a. 

280
00:14:36,600 --> 00:14:40,440
Truly real world data set. 
Of patients gathered in the in 

281
00:14:40,440 --> 00:14:42,240
the Netherlands by the Dutch 
Cancer. 

282
00:14:42,440 --> 00:14:44,360
Society, who also received the 
same. 

283
00:14:44,680 --> 00:14:48,920
Intensive AML induction 
treatments and we questioned 

284
00:14:48,920 --> 00:14:51,280
whether the control patients of 
the clinical. 

285
00:14:51,280 --> 00:14:53,080
Trial were actually similar to 
the. 

286
00:14:53,480 --> 00:14:59,320
External real worlds control 
data arms and when we looked for

287
00:15:00,080 --> 00:15:03,000
baseline characteristics, we 
clearly, and that's not 

288
00:15:03,000 --> 00:15:05,560
surprising, observed 
differences. 

289
00:15:06,080 --> 00:15:08,520
Missing data was much more 
common in the real world data 

290
00:15:08,520 --> 00:15:10,160
source, but also within the 
harmony. 

291
00:15:10,800 --> 00:15:13,320
Data source and. 
When we performed the matching 

292
00:15:13,320 --> 00:15:15,640
methods, we. 
Observed this is overall 

293
00:15:15,640 --> 00:15:18,600
survival that. 
Outcome for the whole phone 

294
00:15:18,720 --> 00:15:20,480
control. 
Arm was actually. 

295
00:15:20,480 --> 00:15:23,000
Similar similar compared to the 
Harmony data. 

296
00:15:23,560 --> 00:15:26,080
But I have to acknowledge that 
the harmony data is basic. 

297
00:15:26,080 --> 00:15:28,080
Is basically a data set. 
Who consists? 

298
00:15:28,080 --> 00:15:30,120
Of the majority of clinical 
trial patients. 

299
00:15:30,400 --> 00:15:35,160
And in this case, in our 
analysis, we found that about 60

300
00:15:35,160 --> 00:15:37,880
to 70% were actually enrolled in
a clinical trial. 

301
00:15:39,000 --> 00:15:41,560
However, the real. 
Data from the Dutch Cancer. 

302
00:15:41,560 --> 00:15:44,120
Society. 
Those patients. 

303
00:15:44,400 --> 00:15:46,720
Clearly performed worse. 10% 
worse. 

304
00:15:46,720 --> 00:15:49,360
Survival Overall survival at two
years. 

305
00:15:50,040 --> 00:15:53,200
When compared to hope and 
controls and we really believe. 

306
00:15:53,200 --> 00:15:56,080
That this is due to. 
To various reasons, but for 

307
00:15:56,080 --> 00:15:58,360
instance the missing performance
status. 

308
00:15:58,360 --> 00:15:59,880
How? 
Fit is your patient, and is it 

309
00:15:59,880 --> 00:16:01,360
really a patient who could also 
be? 

310
00:16:02,000 --> 00:16:05,400
Included in a clinical trial? 
Is it comorbidity? 

311
00:16:05,400 --> 00:16:09,240
Data all all types of. 
Data that are not available in a

312
00:16:09,320 --> 00:16:11,720
real. 
World data set so we are really 

313
00:16:11,720 --> 00:16:15,800
hesitant to firmly say that real
real world data could actually. 

314
00:16:15,800 --> 00:16:18,680
Supplement the data in a 
clinical trial at this. 

315
00:16:18,680 --> 00:16:24,600
Stage so I hope to have show. 
You that innovative trial 

316
00:16:24,600 --> 00:16:26,320
design. 
Is the future. 

317
00:16:26,680 --> 00:16:30,040
And with the personalized 
medicine becoming available now 

318
00:16:30,320 --> 00:16:32,280
we need to. 
Adapt our trials according to 

319
00:16:32,280 --> 00:16:34,680
the. 
Needs to the accelerating needs 

320
00:16:34,680 --> 00:16:37,200
within the clinical trial 
development. 

321
00:16:37,960 --> 00:16:40,080
To deliver the right drug. 
For the right patient at the 

322
00:16:40,080 --> 00:16:45,000
right time, adaptive designs 
I've I've shown you could help 

323
00:16:45,000 --> 00:16:46,800
you. 
For potential and early trial 

324
00:16:46,800 --> 00:16:50,920
intervention. 
External data might reinforce 

325
00:16:50,920 --> 00:16:52,800
your control arm and important 
in that. 

326
00:16:52,800 --> 00:16:57,200
Perspective, but. 
You need to define and identify 

327
00:16:57,200 --> 00:16:59,000
the right and the proper 
external control. 

328
00:16:59,000 --> 00:17:01,800
Data set. 
And with all this, and 

329
00:17:01,800 --> 00:17:04,440
especially if you want to adapt 
your trial during the conduct of

330
00:17:04,440 --> 00:17:07,880
your trial, you need a solid 
trial infrastructure. 

331
00:17:08,280 --> 00:17:11,240
Where you have. 
Immediate data available and who

332
00:17:11,240 --> 00:17:15,480
can support the complex trial 
that you might design? 

333
00:17:17,119 --> 00:17:18,839
And with that, I would like to 
conclude. 

334
00:17:18,839 --> 00:17:19,319
Thank you.