[EN] - Why cutting-edge AI research doesn't scale by itself ?

WEBVTT

1
00:00:06.273 --> 00:00:08.374
Artificial intelligence in oncology is booming.

2
00:00:08.594 --> 00:00:09.855
We have public programs,

3
00:00:09.935 --> 00:00:10.955
private partnerships,

4
00:00:11.756 --> 00:00:13.076
and massive volumes of data.

5
00:00:14.017 --> 00:00:16.098
China is closely observing European research.

6
00:00:16.598 --> 00:00:18.439
Natalie Lasso tells us more about it.

7
00:00:18.739 --> 00:00:23.161
She says she is particularly impressed by the commitment of the Gustav Rusi Institute,

8
00:00:23.421 --> 00:00:26.382
but cutting-edge research does not necessarily guarantee patient access.

9
00:00:26.582 --> 00:00:30.244
The patient caught between academic excellence and industrial deployment.

10
00:00:30.484 --> 00:00:31.985
remains in a fragile position.

11
00:00:32.085 --> 00:00:33.606
The situation is still precarious.

12
00:00:35.447 --> 00:00:37.548
This is the gap I invite you to look at today.

13
00:00:39.389 --> 00:00:40.449
If these topics interest you,

14
00:00:40.610 --> 00:00:43.151
feel free to subscribe to follow the entire series.

15
00:00:44.051 --> 00:00:46.773
So we no longer work in silos when it comes to imaging.

16
00:00:47.373 --> 00:00:49.494
We do liquid biopsies and imaging.

17
00:00:50.095 --> 00:00:52.236
It's also useful for all pharmaceutical companies.

18
00:00:52.296 --> 00:00:53.557
Now everyone is joining in.

19
00:00:54.157 --> 00:00:58.099
How did you first get started in the fields of liquid biopsy and AI?

20
00:00:58.779 --> 00:00:59.580
Is there a formula?

21
00:00:59.960 --> 00:01:01.161
It's true that at this year's

22
00:01:01.601 --> 00:01:02.361
Chicago conference,

23
00:01:02.381 --> 00:01:04.642
we proved we're actually the first in the world to do this.

24
00:01:04.722 --> 00:01:05.582
We're real pioneers.

25
00:01:06.002 --> 00:01:09.924
We presented a study of 1,000 patients and 55,000 metastases.

26
00:01:10.264 --> 00:01:11.585
A Chinese man in the room said,

27
00:01:11.885 --> 00:01:12.705
that's impossible.

28
00:01:12.725 --> 00:01:13.426
You didn't do that.

29
00:01:13.506 --> 00:01:14.726
It's impossible to do that.

30
00:01:15.006 --> 00:01:17.107
How did you manage to get French people to work together?

31
00:01:17.487 --> 00:01:20.148
It was an artificial intelligence lab at Central Suppilic,

32
00:01:20.368 --> 00:01:21.049
the CVN,

33
00:01:21.269 --> 00:01:22.029
and also Okin,

34
00:01:22.289 --> 00:01:22.929
with whom I worked.

35
00:01:23.129 --> 00:01:23.750
Then I told them,

36
00:01:23.970 --> 00:01:24.790
we're going to do a study,

37
00:01:24.870 --> 00:01:26.811
and I would like to use each of your technologies,

38
00:01:26.851 --> 00:01:27.551
which were different.

39
00:01:27.931 --> 00:01:29.252
One of them was working on machine.

40
00:01:29.472 --> 00:01:29.712
learning.

41
00:01:29.713 --> 00:01:30.493
So in deep learning,

42
00:01:30.793 --> 00:01:31.993
I told them it's a bet.

43
00:01:32.534 --> 00:01:33.294
May the best one win.

44
00:01:33.554 --> 00:01:35.355
We'll take the best technology and use it.

45
00:01:36.256 --> 00:01:36.476
Hello,

46
00:01:36.796 --> 00:01:37.316
everyone.

47
00:01:38.077 --> 00:01:38.317
Today,

48
00:01:38.337 --> 00:01:40.838
I am joined here by Professor Nathalie Lassau.

49
00:01:41.579 --> 00:01:49.203
She is a highly distinguished expert in the specialized care of cancer patients and a radiologist currently practicing at the Institute.

50
00:01:49.443 --> 00:01:50.343
Gustave Roussy,

51
00:01:50.443 --> 00:01:51.184
a professor,

52
00:01:51.624 --> 00:01:55.646
PUPH at the University of Saclay and also a researcher.

53
00:01:56.006 --> 00:01:58.127
So I am delighted to be with you today

54
00:01:58.528 --> 00:01:58.968
Nathalie

55
00:01:59.168 --> 00:01:59.408
Hello.

56
00:01:59.909 --> 00:02:00.109
Hello.

57
00:02:00.289 --> 00:02:01.210
It is nice to meet you.

58
00:02:01.290 --> 00:02:02.351
Thank you for the warm welcome.

59
00:02:02.711 --> 00:02:09.336
The reason I came to the Gustave Roussy Institute today to have this conversation is because you have a fascinating background,

60
00:02:09.737 --> 00:02:10.918
an incredible career,

61
00:02:11.378 --> 00:02:13.700
but also a truly unique personality.

62
00:02:14.200 --> 00:02:16.322
And we're really going to try to understand that.

63
00:02:16.582 --> 00:02:18.264
Maybe we can start by introducing this place,

64
00:02:18.265 --> 00:02:19.925
the Gustave Roussy Institute.

65
00:02:20.305 --> 00:02:23.048
So the Gustave Roussy Institute is a truly incredible place.

66
00:02:23.728 --> 00:02:24.369
And in the end,

67
00:02:25.690 --> 00:02:26.510
it draws people in.

68
00:02:27.031 --> 00:02:28.012
That's how I found myself.

69
00:02:28.612 --> 00:02:29.792
incredibly attracted to it.

70
00:02:29.852 --> 00:02:30.793
Once I fell into the pot,

71
00:02:30.873 --> 00:02:32.333
I just never managed to get back out.

72
00:02:32.853 --> 00:02:34.234
That's the story of Gustave Roussy.

73
00:02:34.434 --> 00:02:36.054
It's a place of genuine passion,

74
00:02:36.094 --> 00:02:36.774
of vocation,

75
00:02:36.954 --> 00:02:40.475
of wonderful exchanges with patients that makes you never want to leave.

76
00:02:40.555 --> 00:02:41.556
Especially in oncology,

77
00:02:41.896 --> 00:02:43.556
you form a very strong bond.

78
00:02:43.896 --> 00:02:44.737
And that's why I love it.

79
00:02:45.017 --> 00:02:45.457
Ultrasound,

80
00:02:45.517 --> 00:02:47.217
I'm not from the CT or MRI side.

81
00:02:47.637 --> 00:02:50.458
That's how I was also hired at Gustave as an ultrasound specialist.

82
00:02:50.718 --> 00:02:51.899
I was really immersed in it.

83
00:02:52.339 --> 00:02:52.779
And for me,

84
00:02:52.899 --> 00:02:54.519
it was through research that I managed.

85
00:02:54.539 --> 00:02:55.039
I knew it.

86
00:02:55.380 --> 00:02:56.760
So I had to innovate.

87
00:02:56.920 --> 00:02:57.480
As a result,

88
00:02:57.520 --> 00:02:58.801
I invented some patents.

89
00:02:59.802 --> 00:03:01.303
These were bought by Toshiba,

90
00:03:01.383 --> 00:03:01.963
for example,

91
00:03:02.023 --> 00:03:05.985
which made me known worldwide and established my reputation.

92
00:03:06.505 --> 00:03:08.487
I love being a pioneer in what I do.

93
00:03:08.747 --> 00:03:10.788
We've just filed another patent with Gerbit.

94
00:03:11.348 --> 00:03:12.669
I really like being a pioneer,

95
00:03:12.769 --> 00:03:15.971
being the first demonstrating things and then passing them on,

96
00:03:16.031 --> 00:03:18.372
spreading them so that they benefit patients.

97
00:03:18.892 --> 00:03:22.615
And it is definitely true that I have this drive to discover new things,

98
00:03:22.995 --> 00:03:24.996
to prove things that we might have suspected.

99
00:03:25.708 --> 00:03:28.709
but that hadn't yet been demonstrated and I absolutely love that.

100
00:03:29.570 --> 00:03:32.371
If I look back at the stories about Doppler ultrasound,

101
00:03:32.951 --> 00:03:38.113
there's still this fundamental aspect taking on topics no matter what they are.

102
00:03:38.393 --> 00:03:39.233
It's about cancer,

103
00:03:39.634 --> 00:03:42.635
ultimately looking for things and staying curious,

104
00:03:43.395 --> 00:03:47.697
but as a result not necessarily doing what everyone else does or what's most trendy right now.

105
00:03:48.197 --> 00:03:48.377
Yes,

106
00:03:48.417 --> 00:03:49.758
it's about moving into innovation,

107
00:03:50.178 --> 00:03:50.638
that's for sure.

108
00:03:50.738 --> 00:03:53.239
It's true that I'm I'm vice president for innovation and all that.

109
00:03:54.139 --> 00:03:55.240
It's really about innovating.

110
00:03:55.688 --> 00:03:57.870
innovating in imaging or innovating for the patient,

111
00:03:58.230 --> 00:03:59.551
innovating with oncologists,

112
00:03:59.611 --> 00:04:00.352
with new drugs.

113
00:04:00.672 --> 00:04:02.554
I had worked a lot with pharmaceutical companies.

114
00:04:03.494 --> 00:04:09.439
So it's about seeking different topics and combining them because medicine is very compartmentalized and siloed.

115
00:04:09.900 --> 00:04:11.581
You have radiologists on one side,

116
00:04:12.001 --> 00:04:13.162
oncologists on the other.

117
00:04:14.023 --> 00:04:14.403
Early on,

118
00:04:14.503 --> 00:04:16.485
I worked closely with oncologists,

119
00:04:16.585 --> 00:04:18.346
sometimes even more than with radiologists,

120
00:04:18.507 --> 00:04:19.888
because you had to understand their issues.

121
00:04:20.248 --> 00:04:20.568
Obviously,

122
00:04:20.628 --> 00:04:21.889
what did they need for their patients?

123
00:04:22.210 --> 00:04:25.032
A tool that would allow in early stage trials.

124
00:04:25.352 --> 00:04:27.073
to know whether a drug was working or not.

125
00:04:27.533 --> 00:04:27.913
So I thought,

126
00:04:27.953 --> 00:04:28.193
okay,

127
00:04:28.273 --> 00:04:33.335
I'm going to develop my technique that will help the oncologist determine when a pharmaceutical company proposes a new drug,

128
00:04:33.415 --> 00:04:34.996
whether we can quickly know if it works or not.

129
00:04:35.776 --> 00:04:36.196
That was it.

130
00:04:36.736 --> 00:04:38.817
It's a very maybe practical approach as well.

131
00:04:39.257 --> 00:04:39.677
Pragmatic,

132
00:04:39.717 --> 00:04:40.738
you've got it exactly right.

133
00:04:41.138 --> 00:04:44.219
It is not something that is driven by the expectations of other people.

134
00:04:44.959 --> 00:04:46.940
It is really just more about being basic.

135
00:04:47.997 --> 00:04:48.717
That's exactly it.

136
00:04:49.138 --> 00:04:50.858
Every time I talk to my young colleagues,

137
00:04:50.898 --> 00:04:51.679
I always tell them,

138
00:04:52.199 --> 00:04:52.439
okay,

139
00:04:52.539 --> 00:04:53.980
what is the question being asked?

140
00:04:54.340 --> 00:04:55.800
We do a study at the end,

141
00:04:55.820 --> 00:05:00.202
we publish and we deliver a result that allows us to better monitor patients,

142
00:05:00.262 --> 00:05:02.323
better select patients or better assist them.

143
00:05:02.823 --> 00:05:04.464
We need something concrete for the patients.

144
00:05:04.824 --> 00:05:05.324
That is to say,

145
00:05:05.365 --> 00:05:08.426
there are so many beautiful publications with little significant p-values,

146
00:05:08.786 --> 00:05:10.587
but the oncologist can't do anything with them.

147
00:05:11.727 --> 00:05:13.168
What I want every time is to say,

148
00:05:13.248 --> 00:05:13.528
okay,

149
00:05:13.628 --> 00:05:14.608
we're doing research,

150
00:05:14.669 --> 00:05:15.589
but afterwards,

151
00:05:16.029 --> 00:05:18.711
It has to be usable in routine practice for patients.

152
00:05:19.171 --> 00:05:19.571
That's it.

153
00:05:19.931 --> 00:05:21.852
And that the end of the road is always complicated.

154
00:05:22.292 --> 00:05:23.533
We realized that in research,

155
00:05:23.793 --> 00:05:25.034
a lot of things are developed,

156
00:05:25.174 --> 00:05:27.475
but they remain pending and are never applied.

157
00:05:28.136 --> 00:05:28.556
And the end,

158
00:05:28.596 --> 00:05:29.416
which is always hard,

159
00:05:29.516 --> 00:05:30.117
ultimately to,

160
00:05:30.617 --> 00:05:31.157
can we use it?

161
00:05:31.257 --> 00:05:33.018
Can we implement it routinely for patients?

162
00:05:33.198 --> 00:05:34.119
It's always hard at the end.

163
00:05:35.179 --> 00:05:35.920
It's complicated.

164
00:05:36.360 --> 00:05:37.361
And that's the hardest part.

165
00:05:38.161 --> 00:05:41.383
And that's the strength and value of what you do and what is recognized.

166
00:05:41.803 --> 00:05:43.344
Why does it come to an end too soon?

167
00:05:43.784 --> 00:05:45.305
Because it's hard to implement at the end.

168
00:05:45.633 --> 00:05:47.094
People stop right after they finish.

169
00:05:47.334 --> 00:05:47.455
No,

170
00:05:47.595 --> 00:05:48.916
you must see things through to the end.

171
00:05:49.176 --> 00:05:50.877
You have to go all the way for the patient.

172
00:05:51.318 --> 00:05:55.201
You have to put yourself in the clinician's or the patient's shoes and say,

173
00:05:55.301 --> 00:05:55.521
okay,

174
00:05:55.561 --> 00:05:58.704
what I'm developing could change the doctor's life or the patient's life,

175
00:05:59.044 --> 00:06:00.145
and how can I change it?

176
00:06:00.525 --> 00:06:02.467
It's in their hands or all the doctors now.

177
00:06:02.827 --> 00:06:03.167
Exactly,

178
00:06:03.227 --> 00:06:03.568
exactly,

179
00:06:03.648 --> 00:06:03.968
exactly.

180
00:06:04.549 --> 00:06:08.872
And we realize that there are a lot of things that have been published that never made it to the end.

181
00:06:09.252 --> 00:06:14.637
So maybe it's because people don't have that tenacity to see things through to the end in some way.

182
00:06:14.837 --> 00:06:15.878
Do you have concrete examples?

183
00:06:16.118 --> 00:06:16.458
I don't know.

184
00:06:16.699 --> 00:06:17.859
A publication where you say,

185
00:06:17.980 --> 00:06:18.300
actually,

186
00:06:18.660 --> 00:06:19.601
I could have stopped there,

187
00:06:19.661 --> 00:06:19.941
but no,

188
00:06:19.961 --> 00:06:20.522
I kept going.

189
00:06:20.702 --> 00:06:20.842
Yes,

190
00:06:20.862 --> 00:06:21.422
for example,

191
00:06:21.423 --> 00:06:22.543
the technique I invented.

192
00:06:23.004 --> 00:06:23.484
You could say,

193
00:06:23.644 --> 00:06:23.864
okay,

194
00:06:23.925 --> 00:06:24.165
great,

195
00:06:24.225 --> 00:06:25.106
I filed a patent,

196
00:06:25.126 --> 00:06:25.686
it got bought.

197
00:06:26.427 --> 00:06:26.547
No,

198
00:06:26.687 --> 00:06:28.528
the idea was to say it has to be used,

199
00:06:28.529 --> 00:06:30.190
it has to be included in the guidelines.

200
00:06:30.470 --> 00:06:31.090
The next step,

201
00:06:31.491 --> 00:06:32.331
after I had done,

202
00:06:32.392 --> 00:06:32.732
I don't know,

203
00:06:32.812 --> 00:06:33.773
four or five studies,

204
00:06:33.853 --> 00:06:34.914
go right to the guideline.

205
00:06:35.274 --> 00:06:35.454
Yes,

206
00:06:35.514 --> 00:06:35.814
that's it,

207
00:06:35.815 --> 00:06:36.475
the recognition.

208
00:06:36.495 --> 00:06:36.755
You say,

209
00:06:36.835 --> 00:06:37.015
okay,

210
00:06:37.055 --> 00:06:39.297
this is the global reference used in various contexts.

211
00:06:39.698 --> 00:06:42.520
I had done several single center studies and could have said that's enough.

212
00:06:42.580 --> 00:06:43.661
They're in my publications.

213
00:06:44.129 --> 00:06:44.449
But no,

214
00:06:44.609 --> 00:06:45.290
I told myself,

215
00:06:45.710 --> 00:06:46.930
for it to be truly recognized,

216
00:06:47.250 --> 00:06:49.031
it's essential to have a multi-center study.

217
00:06:49.411 --> 00:06:52.613
I went to see INCA when there were major calls for proposals.

218
00:06:53.073 --> 00:06:55.914
I responded to a major call for 1.5 million euros.

219
00:06:56.375 --> 00:06:59.496
I brought together 19 teams in France from cancer centers,

220
00:06:59.936 --> 00:07:01.037
university hospitals,

221
00:07:01.237 --> 00:07:01.797
and so on.

222
00:07:02.197 --> 00:07:03.818
I told them I'd invented a technique,

223
00:07:03.898 --> 00:07:06.499
but we had to prove it works beyond Gustave Roussy,

224
00:07:06.879 --> 00:07:08.380
across France with residents,

225
00:07:08.480 --> 00:07:09.020
professors,

226
00:07:09.040 --> 00:07:09.901
and department heads.

227
00:07:10.337 --> 00:07:12.818
We conducted a large study on 500 patients,

228
00:07:12.898 --> 00:07:13.878
followed for six years,

229
00:07:14.758 --> 00:07:20.160
and we demonstrated the relevance of the tool that it helped to predict the effectiveness of drugs that destroy blood vessels.

230
00:07:20.200 --> 00:07:23.601
We validated a biomarker because we correlated it with survival.

231
00:07:24.021 --> 00:07:26.802
And afterwards it was referenced in a nature review clinical oncology,

232
00:07:27.542 --> 00:07:30.043
and it's included in the international ultrasound guidelines.

233
00:07:30.723 --> 00:07:31.483
So it was really,

234
00:07:31.983 --> 00:07:32.603
this is truly,

235
00:07:32.964 --> 00:07:33.364
I search,

236
00:07:33.424 --> 00:07:33.844
I find,

237
00:07:33.864 --> 00:07:34.404
and I apply.

238
00:07:35.144 --> 00:07:37.665
So it took quite a long time to develop.

239
00:07:37.805 --> 00:07:43.190
because the original patent dates all the way back to 2005 and the multicenter publications,

240
00:07:43.490 --> 00:07:45.051
all of that came out in 2015.

241
00:07:45.472 --> 00:07:46.472
So it took 10 years.

242
00:07:46.953 --> 00:07:47.733
That's when it gets tough.

243
00:07:47.974 --> 00:07:48.794
So at that point,

244
00:07:48.834 --> 00:07:49.435
you can't give up.

245
00:07:50.176 --> 00:07:50.476
First,

246
00:07:50.496 --> 00:07:52.638
you have to demonstrate proof of concept on a phantom,

247
00:07:52.718 --> 00:07:54.199
then do small studies on mice,

248
00:07:54.619 --> 00:07:56.081
monocentric studies on patients,

249
00:07:56.101 --> 00:07:57.021
the multicenter study,

250
00:07:57.402 --> 00:07:58.543
and then convince Toshiba.

251
00:07:59.183 --> 00:08:01.285
And now on ultrasound machines all over the world,

252
00:08:01.365 --> 00:08:03.287
my technique is implemented on the devices.

253
00:08:03.407 --> 00:08:03.827
All right.

254
00:08:04.027 --> 00:08:04.388
There you go.

255
00:08:04.428 --> 00:08:05.008
You can use it.

256
00:08:05.549 --> 00:08:05.849
All right.

257
00:08:06.473 --> 00:08:07.254
And the manufacturer,

258
00:08:07.314 --> 00:08:08.354
when did they come into the picture?

259
00:08:08.774 --> 00:08:09.315
At what point,

260
00:08:09.515 --> 00:08:10.475
right from the very beginning,

261
00:08:10.656 --> 00:08:10.956
Canon?

262
00:08:11.356 --> 00:08:13.077
Canon got involved pretty quickly.

263
00:08:13.517 --> 00:08:13.637
Well,

264
00:08:13.657 --> 00:08:15.318
it was Toshiba and now it's Canon.

265
00:08:15.698 --> 00:08:17.099
But back then it was Toshiba.

266
00:08:17.319 --> 00:08:18.620
So I was with Pierre Perronneau,

267
00:08:18.680 --> 00:08:19.641
the inventor of Doppler,

268
00:08:19.781 --> 00:08:21.362
and I had told him my idea and everything.

269
00:08:21.522 --> 00:08:23.303
And I had gone to present my first work.

270
00:08:23.683 --> 00:08:24.264
And at the time,

271
00:08:24.304 --> 00:08:24.544
actually,

272
00:08:24.584 --> 00:08:26.605
we were doing quantification with Photoshop.

273
00:08:26.805 --> 00:08:28.266
I don't know if you ever experienced that.

274
00:08:28.886 --> 00:08:29.807
I think you're not too young,

275
00:08:29.847 --> 00:08:31.148
but it was,

276
00:08:31.428 --> 00:08:31.668
you know,

277
00:08:31.768 --> 00:08:33.369
quantifying images with Photoshop.

278
00:08:33.869 --> 00:08:35.550
I presented it in the United States.

279
00:08:35.551 --> 00:08:37.871
It was a great study I had done on sarcomas,

280
00:08:38.912 --> 00:08:40.192
evaluating glyphosate,

281
00:08:40.232 --> 00:08:43.074
which had just come out from Superdrug Revolutionary and all that.

282
00:08:43.954 --> 00:08:44.275
And so

283
00:08:44.975 --> 00:08:47.276
I was quantifying my contrast echoes with Photoshop.

284
00:08:47.336 --> 00:08:50.738
I presented this at RSNA and there I got myself into a mess.

285
00:08:50.898 --> 00:08:54.060
It was a radiologist from MD Anderson back in Houston,

286
00:08:54.720 --> 00:08:56.101
someone who really tore me apart.

287
00:08:56.681 --> 00:08:57.262
He said to me,

288
00:08:57.582 --> 00:08:59.843
how can you propose quantification with Photoshop?

289
00:09:00.223 --> 00:09:01.024
I was completely,

290
00:09:01.704 --> 00:09:02.565
I wanted to cry,

291
00:09:02.725 --> 00:09:03.205
basically.

292
00:09:04.053 --> 00:09:06.154
It was done on 50 patients with sarcomas.

293
00:09:06.714 --> 00:09:07.054
I thought,

294
00:09:07.554 --> 00:09:08.815
I might as well just go home.

295
00:09:09.575 --> 00:09:10.875
I went to see Pierre Perronon,

296
00:09:11.075 --> 00:09:11.555
and he said,

297
00:09:12.116 --> 00:09:12.236
OK,

298
00:09:12.396 --> 00:09:12.796
Nathalie,

299
00:09:13.296 --> 00:09:16.237
we need the raw linear data from the manufacturers.

300
00:09:16.777 --> 00:09:21.338
So then we went to see every ultrasound manufacturer like GE and Shimatsu.

301
00:09:22.259 --> 00:09:22.979
And there we said,

302
00:09:23.019 --> 00:09:23.159
look,

303
00:09:23.199 --> 00:09:26.260
we want the raw linear data because we need to do real quantification.

304
00:09:26.980 --> 00:09:28.281
We can't just use Photoshop.

305
00:09:28.741 --> 00:09:29.081
So then...

306
00:09:29.709 --> 00:09:30.850
Most of the manufacturers said,

307
00:09:30.870 --> 00:09:30.990
no,

308
00:09:31.030 --> 00:09:31.150
no,

309
00:09:31.230 --> 00:09:31.930
that's classified.

310
00:09:31.931 --> 00:09:33.151
We don't give out the raw data.

311
00:09:33.251 --> 00:09:33.871
It's secret,

312
00:09:34.471 --> 00:09:36.112
except for the Japanese Toshiba,

313
00:09:36.272 --> 00:09:36.852
who agreed.

314
00:09:37.433 --> 00:09:38.033
It was a gamble,

315
00:09:38.113 --> 00:09:39.234
so they brought me to Japan.

316
00:09:39.354 --> 00:09:40.654
I explained everything and they said,

317
00:09:40.674 --> 00:09:40.794
OK,

318
00:09:40.994 --> 00:09:41.895
let's sign the contracts.

319
00:09:42.515 --> 00:09:42.795
All right,

320
00:09:42.835 --> 00:09:43.856
we'll give you the raw data.

321
00:09:44.156 --> 00:09:44.716
And from that,

322
00:09:44.796 --> 00:09:46.257
we were able to create the patent.

323
00:09:46.840 --> 00:09:47.581
So they followed us.

324
00:09:47.901 --> 00:09:49.282
They followed me all around the world.

325
00:09:49.382 --> 00:09:50.242
They gave me access.

326
00:09:50.903 --> 00:09:51.423
And to tell you,

327
00:09:51.443 --> 00:09:52.244
it took a long time.

328
00:09:52.564 --> 00:09:53.124
For example,

329
00:09:53.225 --> 00:09:53.965
just five years ago,

330
00:09:54.525 --> 00:09:55.106
all of a sudden,

331
00:09:55.446 --> 00:09:57.467
they announced at the big Congress in Chicago,

332
00:09:58.028 --> 00:09:58.408
that's it,

333
00:09:58.728 --> 00:10:00.630
we are giving access to the raw linear data.

334
00:10:00.910 --> 00:10:02.811
That was 15 years after what we had requested.

335
00:10:03.492 --> 00:10:06.434
But so we had managed to convince the Japanese.

336
00:10:06.554 --> 00:10:06.954
And for me,

337
00:10:07.014 --> 00:10:07.675
the Japanese,

338
00:10:07.815 --> 00:10:09.336
everyone says it's hard to work with them.

339
00:10:10.016 --> 00:10:11.017
It's hard to convince them.

340
00:10:11.277 --> 00:10:12.198
But the day it's okay.

341
00:10:12.778 --> 00:10:13.579
When you sign the deal,

342
00:10:13.639 --> 00:10:15.040
they're really funny and they stick with you.

343
00:10:15.416 --> 00:10:17.237
Their loyalty is absolutely unwavering.

344
00:10:17.577 --> 00:10:18.638
So it was hard to convince them.

345
00:10:18.698 --> 00:10:19.218
But after that,

346
00:10:19.278 --> 00:10:20.199
they've always been there.

347
00:10:20.639 --> 00:10:21.799
And they're still with us today.

348
00:10:22.260 --> 00:10:22.680
I've always,

349
00:10:22.940 --> 00:10:23.260
right now,

350
00:10:23.300 --> 00:10:24.601
I'm still a reference in ultrasound.

351
00:10:24.661 --> 00:10:25.522
And so is Gustav.

352
00:10:26.162 --> 00:10:27.102
I have machines on loan.

353
00:10:27.162 --> 00:10:29.704
I have a full-time engineer who's been paid for 10 years.

354
00:10:30.044 --> 00:10:30.984
And I keep publishing a lot.

355
00:10:31.004 --> 00:10:31.605
I work a lot.

356
00:10:31.965 --> 00:10:33.046
So there you have it.

357
00:10:33.126 --> 00:10:34.586
It's a partnership that's lasted.

358
00:10:34.686 --> 00:10:35.327
It's incredible.

359
00:10:35.687 --> 00:10:40.249
It's a really beautiful story of a partnership with industrialists who are straightforward,

360
00:10:40.309 --> 00:10:40.710
reliable,

361
00:10:40.750 --> 00:10:41.170
and loyal.

362
00:10:41.890 --> 00:10:42.511
The Japanese,

363
00:10:42.671 --> 00:10:44.172
it's true that they don't play tricks.

364
00:10:44.612 --> 00:10:44.832
That's...

365
00:10:45.404 --> 00:10:46.904
That's not the case with everyone.

366
00:10:47.045 --> 00:10:52.526
I now work with other industrial partners and it's true that at first I was criticized in radiology because they didn't understand.

367
00:10:52.546 --> 00:10:53.826
They said you have to work with everyone.

368
00:10:54.647 --> 00:10:55.027
I said no,

369
00:10:55.047 --> 00:11:00.988
if you're doing research very early stage and pioneering and if you really want to do R&D,

370
00:11:01.248 --> 00:11:03.069
you can't work with all the industrial partners.

371
00:11:03.369 --> 00:11:04.389
You have to make a choice.

372
00:11:04.609 --> 00:11:06.470
Afterwards you can't pick the wrong horse.

373
00:11:06.471 --> 00:11:09.151
You have to tell yourself it's the right industrial partner and all that.

374
00:11:09.551 --> 00:11:10.651
But in this case you can't do...

375
00:11:10.971 --> 00:11:14.112
Now I do more things with other industrial partners but on different topics.

376
00:11:14.564 --> 00:11:19.427
I always make sure to have a well-defined scope so there's no overlap because otherwise you ruin everything.

377
00:11:20.147 --> 00:11:22.348
And that's also the advantage of working at Gustav,

378
00:11:22.849 --> 00:11:27.732
having the possibility to work with all the industrial partners because it's also a research accelerator.

379
00:11:28.272 --> 00:11:28.432
Yes,

380
00:11:28.452 --> 00:11:31.254
it took 10 years but without them you would have done nothing.

381
00:11:31.714 --> 00:11:32.214
Well you...

382
00:11:33.055 --> 00:11:33.255
Yes,

383
00:11:33.335 --> 00:11:36.236
I wouldn't have filed my model after that and so what,

384
00:11:36.316 --> 00:11:36.857
nothing really.

385
00:11:37.317 --> 00:11:38.498
I'd just be planting cabbages.

386
00:11:38.918 --> 00:11:39.698
So indeed it's...

387
00:11:40.059 --> 00:11:40.179
No,

388
00:11:40.339 --> 00:11:40.879
but that's it.

389
00:11:41.059 --> 00:11:42.520
So in the end in PharmaMind...

390
00:11:42.700 --> 00:11:45.222
it's essentially personalities who make up the pharma world.

391
00:11:45.482 --> 00:11:51.146
So it's inevitably people who actually see the final stages of innovations when they arrive,

392
00:11:51.266 --> 00:11:53.307
either on the market or when they're about to arrive.

393
00:11:53.948 --> 00:11:54.828
But as a result,

394
00:11:54.848 --> 00:11:57.991
it's interesting to have something that's really upstream and to say to yourself,

395
00:11:58.051 --> 00:11:58.471
sometimes,

396
00:11:58.531 --> 00:11:58.731
yes,

397
00:11:58.911 --> 00:11:59.512
it's a gamble.

398
00:11:59.912 --> 00:12:00.392
It's a gamble.

399
00:12:00.813 --> 00:12:01.313
Afterwards,

400
00:12:01.373 --> 00:12:03.855
I've had one or two things that didn't work out or that I stopped.

401
00:12:04.195 --> 00:12:04.375
Yes,

402
00:12:04.435 --> 00:12:04.855
there's that,

403
00:12:04.975 --> 00:12:06.096
which I just picked up again.

404
00:12:06.677 --> 00:12:07.117
For example,

405
00:12:07.137 --> 00:12:08.938
there's molecular imaging with ultrasound.

406
00:12:09.659 --> 00:12:11.340
I filed a patent with some Americans.

407
00:12:12.132 --> 00:12:13.552
That was in 2003.

408
00:12:13.553 --> 00:12:14.833
I realized I wouldn't be able to do it.

409
00:12:15.413 --> 00:12:16.293
It cost too much money.

410
00:12:16.433 --> 00:12:17.073
My boss told me,

411
00:12:17.474 --> 00:12:18.454
we won't be able to make it.

412
00:12:18.734 --> 00:12:20.995
Yet I had raised the money and everything from the region,

413
00:12:21.095 --> 00:12:22.115
but I just couldn't do it.

414
00:12:22.315 --> 00:12:23.955
I stopped for a moment and put it aside,

415
00:12:24.356 --> 00:12:26.396
but it was still on my mind.

416
00:12:26.576 --> 00:12:28.077
It's molecular imaging.

417
00:12:29.117 --> 00:12:33.698
And then all of a sudden we managed to convince an industry partner because I'm more visible now.

418
00:12:34.218 --> 00:12:37.619
And so now we're going to do a phase one trial with patients.

419
00:12:38.640 --> 00:12:38.980
There you go.

420
00:12:39.080 --> 00:12:39.720
It's incredible.

421
00:12:40.388 --> 00:12:43.910
with a molecular ultrasound contrast agent that targets receptors.

422
00:12:44.450 --> 00:12:45.690
But you have to know when to stop.

423
00:12:45.810 --> 00:12:46.691
My boss always told me,

424
00:12:47.051 --> 00:12:47.351
Natalie,

425
00:12:47.411 --> 00:12:49.232
because he saw that I was still pretty stubborn.

426
00:12:49.692 --> 00:12:50.192
He said to me,

427
00:12:50.212 --> 00:12:50.332
OK,

428
00:12:50.452 --> 00:12:50.872
that's good.

429
00:12:51.312 --> 00:12:52.633
But sometimes you have to tell yourself,

430
00:12:52.753 --> 00:12:52.873
OK,

431
00:12:53.113 --> 00:12:54.134
now you need to stop.

432
00:12:54.154 --> 00:12:55.094
You have to cut the branch.

433
00:12:55.294 --> 00:12:55.634
Stop.

434
00:12:55.714 --> 00:12:56.735
You're wearing yourself out.

435
00:12:56.835 --> 00:12:58.295
And then you move on to other things.

436
00:12:58.435 --> 00:12:59.956
And maybe you'll come back to it later.

437
00:13:00.676 --> 00:13:01.377
And that project,

438
00:13:01.457 --> 00:13:01.617
yeah,

439
00:13:02.057 --> 00:13:02.777
it was picked up again.

440
00:13:02.837 --> 00:13:04.898
It started up again because there were problems.

441
00:13:05.398 --> 00:13:07.039
With gadolinium contrast,

442
00:13:07.339 --> 00:13:08.679
scanners used radiation.

443
00:13:08.940 --> 00:13:09.860
Now we have ultrasound.

444
00:13:10.080 --> 00:13:10.980
It's radiation free,

445
00:13:11.340 --> 00:13:13.481
infinitely repeatable and cheaper than a PET scan.

446
00:13:13.801 --> 00:13:15.161
We finally have the technology,

447
00:13:15.421 --> 00:13:17.442
so suddenly it's the right time.

448
00:13:17.522 --> 00:13:18.222
It's always like that,

449
00:13:18.242 --> 00:13:19.383
the stars must align.

450
00:13:20.023 --> 00:13:22.083
And now everything is lining up perfectly.

451
00:13:22.283 --> 00:13:24.404
So we're just starting again to see what happens.

452
00:13:25.124 --> 00:13:26.164
It's a massive gamble.

453
00:13:26.985 --> 00:13:27.705
It has to work.

454
00:13:28.745 --> 00:13:36.107
I would be truly happy to have conducted this molecular imaging study using ultrasound because it was something I really cared about at the time,

455
00:13:36.147 --> 00:13:37.747
but it was far too innovative,

456
00:13:37.808 --> 00:13:38.248
too early,

457
00:13:38.308 --> 00:13:39.228
I didn't have enough money.

458
00:13:39.668 --> 00:13:40.749
It just wasn't possible.

459
00:13:41.710 --> 00:13:43.571
And when the industry did not follow along,

460
00:13:43.591 --> 00:13:44.652
you still need the industry.

461
00:13:44.813 --> 00:13:46.214
We can do academic research,

462
00:13:46.454 --> 00:13:47.535
but in this day and age,

463
00:13:48.236 --> 00:13:51.378
you still need the powerful drive from industry and funding from industry.

464
00:13:51.538 --> 00:13:53.100
We need all three of these elements.

465
00:13:53.480 --> 00:13:54.701
You need power behind it.

466
00:13:55.102 --> 00:13:58.825
Something must be pushing from behind or we'll be stranded in the middle of nowhere.

467
00:13:59.125 --> 00:14:00.266
That is actually interesting.

468
00:14:00.486 --> 00:14:01.647
It's about knowing how to wait.

469
00:14:01.727 --> 00:14:03.049
That's something my bosses taught me.

470
00:14:03.469 --> 00:14:04.029
Bide your time,

471
00:14:04.109 --> 00:14:04.750
know how to wait.

472
00:14:05.090 --> 00:14:06.852
Since you have young people doing research with you.

473
00:14:07.112 --> 00:14:09.373
You shouldn't set them up for failure on those topics in the end.

474
00:14:09.693 --> 00:14:10.373
You think about that?

475
00:14:10.753 --> 00:14:10.933
Well,

476
00:14:11.033 --> 00:14:12.593
I'm still responsible for my little ones.

477
00:14:13.054 --> 00:14:13.914
We do great research,

478
00:14:13.954 --> 00:14:16.415
but then they need to find jobs and have a life of passion.

479
00:14:17.015 --> 00:14:18.355
And so at that point,

480
00:14:18.755 --> 00:14:20.436
I thought I couldn't take them down that path.

481
00:14:20.476 --> 00:14:21.576
It was kind of a dead end.

482
00:14:22.576 --> 00:14:22.956
So yeah,

483
00:14:23.776 --> 00:14:23.897
well,

484
00:14:23.957 --> 00:14:24.077
yeah,

485
00:14:24.117 --> 00:14:25.737
we're still responsible for the young people.

486
00:14:26.437 --> 00:14:27.738
You can't control everything.

487
00:14:27.758 --> 00:14:28.898
You can't do it all either.

488
00:14:29.478 --> 00:14:30.658
Even if with experience,

489
00:14:30.678 --> 00:14:31.819
you can know what's going to happen.

490
00:14:32.059 --> 00:14:33.159
It has to be at the right time.

491
00:14:33.919 --> 00:14:35.060
It has to be the right people.

492
00:14:35.220 --> 00:14:36.660
You have to analyze the ecosystem.

493
00:14:36.920 --> 00:14:38.002
You need to have a kind of vision.

494
00:14:38.363 --> 00:14:40.427
We can't control everything,

495
00:14:40.587 --> 00:14:41.288
never in life.

496
00:14:41.348 --> 00:14:41.789
But still.

497
00:14:42.310 --> 00:14:43.112
What does that really mean?

498
00:14:43.886 --> 00:14:44.306
Mentoring,

499
00:14:44.366 --> 00:14:46.207
guiding and leading a research group or students?

500
00:14:46.687 --> 00:14:46.807
Well,

501
00:14:46.847 --> 00:14:47.127
for me,

502
00:14:47.548 --> 00:14:49.408
guiding always means leading a research group.

503
00:14:49.548 --> 00:14:49.668
OK,

504
00:14:49.909 --> 00:14:50.849
we have our topics.

505
00:14:51.269 --> 00:14:54.250
And right now we're working on biomarkers in oncology,

506
00:14:55.031 --> 00:14:56.151
predictive prognosis.

507
00:14:56.631 --> 00:14:58.972
So I know what could be promising for patients,

508
00:14:59.012 --> 00:14:59.713
for oncology,

509
00:14:59.793 --> 00:15:02.414
for pharmaceutical companies and for imaging in general.

510
00:15:02.914 --> 00:15:04.695
Five or even 10 years gets more complicated.

511
00:15:05.135 --> 00:15:06.996
People who project 10 years ahead,

512
00:15:07.436 --> 00:15:08.036
I think that's,

513
00:15:08.736 --> 00:15:10.757
I like to have something concrete and pragmatic.

514
00:15:10.777 --> 00:15:12.198
So 10 years is too far for me.

515
00:15:13.006 --> 00:15:15.927
While research or ministers might plan that far ahead,

516
00:15:16.107 --> 00:15:17.327
that's not how I operate.

517
00:15:18.047 --> 00:15:20.888
Since I like to achieve concrete results beyond five years,

518
00:15:20.988 --> 00:15:22.449
it's hard to see anything tangible.

519
00:15:23.489 --> 00:15:24.929
I always tell young people

520
00:15:25.349 --> 00:15:27.570
I prefer to build my castle with small bricks first.

521
00:15:28.410 --> 00:15:32.491
Then I expand and make my first room rather than trying to build the big castle first.

522
00:15:33.052 --> 00:15:34.012
You have to be realistic.

523
00:15:34.432 --> 00:15:36.012
Even if you're visionary and a pioneer,

524
00:15:36.052 --> 00:15:37.073
you have to be realistic.

525
00:15:37.533 --> 00:15:38.853
And you always have to in the end,

526
00:15:38.933 --> 00:15:39.793
because for young people,

527
00:15:39.853 --> 00:15:41.614
we're here to help them fill out their CVs.

528
00:15:42.078 --> 00:15:43.279
That's how they're going to find a job.

529
00:15:43.399 --> 00:15:45.199
What remains in the end are the publications,

530
00:15:45.499 --> 00:15:48.461
patents or publications that catch someone's eye.

531
00:15:49.241 --> 00:15:49.621
Basically,

532
00:15:49.681 --> 00:15:50.021
that's it.

533
00:15:50.341 --> 00:15:50.522
Okay,

534
00:15:50.542 --> 00:15:51.282
you have the degrees.

535
00:15:51.702 --> 00:15:52.342
But when I look,

536
00:15:52.642 --> 00:15:54.463
I pay attention to which internships they've done,

537
00:15:54.823 --> 00:15:56.644
where they've been and which team they've worked with.

538
00:15:56.964 --> 00:15:58.225
Was the person able to publish?

539
00:15:58.465 --> 00:15:59.665
Is their name listed first?

540
00:16:00.046 --> 00:16:02.026
Did they keep a good pace in a fairly short time?

541
00:16:02.126 --> 00:16:03.047
That's what I focus on.

542
00:16:04.087 --> 00:16:05.008
Then after that,

543
00:16:05.148 --> 00:16:06.228
the desire to go further,

544
00:16:06.288 --> 00:16:07.309
the passion and all that.

545
00:16:07.709 --> 00:16:09.309
So as a result of this,

546
00:16:09.389 --> 00:16:10.490
the projects I start

547
00:16:10.922 --> 00:16:12.823
are things that span over five years.

548
00:16:13.544 --> 00:16:15.926
I need to achieve a solid result within five years,

549
00:16:16.246 --> 00:16:17.086
or in one year,

550
00:16:17.206 --> 00:16:17.827
or in two years,

551
00:16:17.947 --> 00:16:18.647
or in three years,

552
00:16:19.548 --> 00:16:21.069
but you have to achieve concrete results.

553
00:16:21.669 --> 00:16:23.271
So tell me what are the topics?

554
00:16:24.331 --> 00:16:25.372
So the topics right now,

555
00:16:25.412 --> 00:16:29.015
the themes I've chosen include molecular imaging and molecular ultrasound.

556
00:16:29.475 --> 00:16:31.196
That's really the ultrasound part,

557
00:16:31.556 --> 00:16:33.277
which is my little baby until the end,

558
00:16:33.638 --> 00:16:34.458
and I'll see it through.

559
00:16:35.359 --> 00:16:38.701
And then it's everything related to biomarkers,

560
00:16:38.761 --> 00:16:39.662
it's tumor burden.

561
00:16:40.470 --> 00:16:41.851
So now we're getting into a topic.

562
00:16:42.271 --> 00:16:45.372
For more than 20 years we've been explaining to radiologists,

563
00:16:45.692 --> 00:16:46.593
so internationally,

564
00:16:47.213 --> 00:16:51.135
we have a single rule all together to make sure we're talking about the same thing.

565
00:16:51.335 --> 00:16:52.015
That is to say,

566
00:16:52.415 --> 00:16:54.917
when we evaluate a patient who has several metastases,

567
00:16:54.997 --> 00:16:56.337
we say here is the rule,

568
00:16:56.797 --> 00:17:00.679
you will only measure five lesions and that way from one scan to another,

569
00:17:00.959 --> 00:17:02.320
no matter where the scan is done,

570
00:17:02.580 --> 00:17:03.761
we will measure in the same way.

571
00:17:04.121 --> 00:17:04.881
Except that you say,

572
00:17:04.981 --> 00:17:05.181
well,

573
00:17:05.481 --> 00:17:06.562
it's extremely restrictive.

574
00:17:06.978 --> 00:17:07.478
For example,

575
00:17:07.479 --> 00:17:10.199
the patient had multiple metastases but you only measure five of them.

576
00:17:10.519 --> 00:17:11.119
Up until now,

577
00:17:11.419 --> 00:17:12.840
that hasn't been a problem for anyone.

578
00:17:13.280 --> 00:17:13.800
Except now,

579
00:17:14.040 --> 00:17:15.400
liquid biopsy has arrived.

580
00:17:15.800 --> 00:17:16.561
Liquid biopsy,

581
00:17:16.781 --> 00:17:18.281
I've been talking about it non-stop,

582
00:17:18.881 --> 00:17:21.842
but it's a new tool that's going to revolutionize oncology.

583
00:17:22.342 --> 00:17:23.062
Up until now,

584
00:17:23.122 --> 00:17:24.203
it's been 10 years,

585
00:17:24.283 --> 00:17:25.443
I think even 15 years,

586
00:17:25.483 --> 00:17:27.404
that we've been talking about precision medicine.

587
00:17:28.124 --> 00:17:30.024
Millions and millions have been invested.

588
00:17:30.865 --> 00:17:31.425
Previously,

589
00:17:32.145 --> 00:17:34.186
we'd biopsy a patient's lesion,

590
00:17:34.726 --> 00:17:35.506
take a sample,

591
00:17:36.122 --> 00:17:38.063
examine the cells and say,

592
00:17:38.623 --> 00:17:41.203
it's this type of cancer or abnormality.

593
00:17:41.984 --> 00:17:43.544
Now we've broken things down quite a bit.

594
00:17:44.044 --> 00:17:44.584
For example,

595
00:17:44.664 --> 00:17:45.725
lung and breast cancers,

596
00:17:46.025 --> 00:17:48.285
all categorized by the genome,

597
00:17:48.586 --> 00:17:50.006
except there's a gap in the system.

598
00:17:50.606 --> 00:17:54.267
We realize that if you do a biopsy somewhere else on another lesion,

599
00:17:54.847 --> 00:17:56.568
you don't necessarily get a good result.

600
00:17:56.988 --> 00:17:57.768
You think to yourself,

601
00:17:57.888 --> 00:17:58.688
all that for this?

602
00:17:58.968 --> 00:18:02.389
So we don't have a comprehensive analysis of the patient.

603
00:18:03.170 --> 00:18:04.090
What is a liquid biopsy?

604
00:18:04.862 --> 00:18:06.783
You take a blood sample directly from a patient.

605
00:18:07.224 --> 00:18:11.026
You detect in the blood the DNA from tumors that are circulating in the bloodstream.

606
00:18:11.427 --> 00:18:14.269
So the DNA from all of the patient's metastases,

607
00:18:14.729 --> 00:18:15.490
and you collect it,

608
00:18:15.770 --> 00:18:17.591
and then you analyze the modifications.

609
00:18:18.052 --> 00:18:19.693
So you are much more holistic,

610
00:18:19.793 --> 00:18:20.393
much broader.

611
00:18:20.874 --> 00:18:21.034
Well,

612
00:18:21.035 --> 00:18:21.494
you might say,

613
00:18:21.934 --> 00:18:23.055
a CT scan is the same.

614
00:18:23.115 --> 00:18:25.297
I'm not going to measure all five of my metastases.

615
00:18:25.477 --> 00:18:25.757
Actually,

616
00:18:25.817 --> 00:18:26.718
I need to measure everything,

617
00:18:27.258 --> 00:18:28.279
what we call the tumor burden.

618
00:18:28.759 --> 00:18:29.160
Except here,

619
00:18:29.200 --> 00:18:29.780
there's a problem.

620
00:18:30.240 --> 00:18:32.482
You've heard about the demographics of radiologists.

621
00:18:32.874 --> 00:18:34.795
They say there are fewer and fewer radiologists.

622
00:18:35.175 --> 00:18:35.976
A CT scan,

623
00:18:36.016 --> 00:18:36.976
when you evaluate a scan,

624
00:18:37.016 --> 00:18:38.977
it's one every 10 or 15 minutes at most.

625
00:18:40.017 --> 00:18:41.498
Each time that's 2,000 images.

626
00:18:41.978 --> 00:18:45.860
That means every 10 or 15 minutes you have 2,000 images and a report to write.

627
00:18:46.140 --> 00:18:48.482
By the end of the day your eyes hurt and your head is spinning.

628
00:18:49.362 --> 00:18:52.924
And if on top of that you ask radiologists to measure every single lesion,

629
00:18:53.484 --> 00:18:54.044
for example,

630
00:18:54.104 --> 00:18:57.626
if there are 100 lesions it's impossible to do in 10 minutes.

631
00:18:58.134 --> 00:19:00.035
So at that point you realize,

632
00:19:00.075 --> 00:19:00.375
okay,

633
00:19:00.615 --> 00:19:02.436
if we want to assess this tumor burden,

634
00:19:02.556 --> 00:19:03.756
we need to develop AI,

635
00:19:03.976 --> 00:19:05.577
artificial intelligence tools.

636
00:19:05.977 --> 00:19:06.958
That's just mathematics,

637
00:19:07.098 --> 00:19:07.998
so it's being developed.

638
00:19:08.058 --> 00:19:11.780
It's a project we've just submitted with Gerbit and we'll get the answer at the end of the week.

639
00:19:12.240 --> 00:19:13.460
Probably BPI France,

640
00:19:13.500 --> 00:19:14.941
a major source of funding.

641
00:19:15.081 --> 00:19:16.102
So what's our vision?

642
00:19:16.402 --> 00:19:25.165
So within five years we should be able to produce a platform where you upload a patient scan and automatically an algorithm detects all the patients metastases,

643
00:19:25.586 --> 00:19:26.486
segments all of them.

644
00:19:26.962 --> 00:19:30.703
and says for example this patient has 200 millimeters of tumor,

645
00:19:31.243 --> 00:19:38.545
200 cubic millimeters or 300 cubic millimeters and we know that this tumor burden before starting treatment is highly prognostic.

646
00:19:38.765 --> 00:19:42.766
You can stratify and just now we published an article in the European Journal of Cancer.

647
00:19:43.206 --> 00:19:47.267
These are patients with metastatic colon cancer who respond very well to immunotherapy.

648
00:19:47.728 --> 00:19:48.708
It's a particular status.

649
00:19:48.948 --> 00:19:52.169
These are MSI positive patients so it's a molecular abnormality.

650
00:19:52.629 --> 00:19:56.290
Half of patients respond to immunotherapy so double immunotherapy is great.

651
00:19:56.810 --> 00:19:58.451
But there's a public health issue.

652
00:19:58.792 --> 00:20:00.393
Immunotherapy is expensive,

653
00:20:00.573 --> 00:20:02.495
costing 70,000 euros each,

654
00:20:02.915 --> 00:20:05.838
so double immunotherapy is 140,000 euros.

655
00:20:06.335 --> 00:20:08.176
unlike chemotherapy at 300 euros.

656
00:20:08.536 --> 00:20:08.676
Well,

657
00:20:08.716 --> 00:20:10.837
now by looking at this significant tumor burden,

658
00:20:11.538 --> 00:20:13.259
the number of metastases and everything,

659
00:20:13.419 --> 00:20:17.621
we've shown that we are able to stratify to know for whom we should immediately give double immunotherapy.

660
00:20:18.081 --> 00:20:23.444
Those with a very poor prognosis versus those for whom a single immunotherapy was enough for them to have a...

661
00:20:24.185 --> 00:20:25.365
So that's something concrete.

662
00:20:26.026 --> 00:20:26.946
It's for the oncologist.

663
00:20:27.086 --> 00:20:27.827
It's a scoring system.

664
00:20:28.107 --> 00:20:29.828
They count the number of cubic millimeters,

665
00:20:29.868 --> 00:20:31.008
the number of metastases.

666
00:20:31.148 --> 00:20:34.250
This small scoring system assesses four parameters for stratification.

667
00:20:34.590 --> 00:20:36.011
We did this with an oncologist.

668
00:20:36.295 --> 00:20:38.196
a world-leading expert in colon cancer,

669
00:20:38.356 --> 00:20:40.516
Professor Thierry André at Saint-Antoine.

670
00:20:41.176 --> 00:20:42.577
And we conducted this study together,

671
00:20:42.757 --> 00:20:43.677
so it's really concrete.

672
00:20:43.877 --> 00:20:44.597
It's tumor research,

673
00:20:44.877 --> 00:20:45.598
and I work on this.

674
00:20:45.778 --> 00:20:46.718
These are AI tools.

675
00:20:47.058 --> 00:20:47.378
Right now,

676
00:20:47.398 --> 00:20:49.459
my radiologists are segmenting all the tumors.

677
00:20:49.819 --> 00:20:51.019
It takes time for research,

678
00:20:51.079 --> 00:20:53.460
but the goal is to have a fully automated platform.

679
00:20:53.660 --> 00:20:54.340
Once that's done,

680
00:20:54.560 --> 00:20:56.241
and it will be ready before I retire,

681
00:20:56.461 --> 00:20:57.761
I am confident it'll work.

682
00:20:58.561 --> 00:20:58.741
Okay,

683
00:20:58.901 --> 00:20:59.182
good news.

684
00:21:00.162 --> 00:21:01.182
And if it happens before,

685
00:21:01.322 --> 00:21:01.842
I'll retire.

686
00:21:02.282 --> 00:21:02.903
It's exhausting,

687
00:21:02.923 --> 00:21:04.163
but truly great research.

688
00:21:04.519 --> 00:21:06.781
We are correlating liquid biopsy with imaging,

689
00:21:07.041 --> 00:21:09.783
so we're no longer working in silos with imaging alone.

690
00:21:10.403 --> 00:21:12.024
We do liquid biopsy and imaging.

691
00:21:12.885 --> 00:21:15.107
This will benefit all pharmaceutical companies.

692
00:21:15.447 --> 00:21:16.648
Now everyone is joining.

693
00:21:17.188 --> 00:21:20.270
So how do you actually bridge the gap between liquid biopsy and AI?

694
00:21:20.871 --> 00:21:22.392
How do you integrate these fields together?

695
00:21:22.792 --> 00:21:23.713
How do you mix everything?

696
00:21:23.853 --> 00:21:24.433
At what point?

697
00:21:24.593 --> 00:21:25.394
How do you make it happen?

698
00:21:25.794 --> 00:21:29.097
It's true that this year when we presented it at the restaurant in Chicago,

699
00:21:29.297 --> 00:21:31.118
actually we are the first in the world to do this.

700
00:21:31.158 --> 00:21:32.219
We are real pioneers.

701
00:21:32.603 --> 00:21:34.725
Because we presented a study of 1,000 patients,

702
00:21:34.905 --> 00:21:37.027
we tagged 55,000 metastases.

703
00:21:37.447 --> 00:21:38.708
There was a Chinese man in the room.

704
00:21:38.909 --> 00:21:39.709
He stood up and said,

705
00:21:40.710 --> 00:21:41.431
that's impossible.

706
00:21:41.451 --> 00:21:42.151
You didn't do that.

707
00:21:42.211 --> 00:21:43.352
It's impossible to do that.

708
00:21:43.713 --> 00:21:47.196
How did you manage to get French people to tag 55,000 metastases?

709
00:21:48.117 --> 00:21:52.480
I told them it was extremely important and that I had taken young people for a year of research to do it.

710
00:21:52.640 --> 00:21:53.581
At noon after that,

711
00:21:53.641 --> 00:21:55.603
it's because for the past 25 to 30 years,

712
00:21:56.023 --> 00:21:59.867
I've always had a very close relationship with the oncologists here who are really at the forefront.

713
00:22:00.635 --> 00:22:02.596
and they educate me about what's new and everything.

714
00:22:02.996 --> 00:22:04.817
They'd alerted me about liquid biopsy.

715
00:22:05.057 --> 00:22:05.998
Benjamin Besser said,

716
00:22:06.318 --> 00:22:08.439
you'll see it's going to replace the CT scan.

717
00:22:08.639 --> 00:22:08.879
I said,

718
00:22:08.999 --> 00:22:09.139
what?

719
00:22:09.199 --> 00:22:09.919
That's not possible.

720
00:22:10.380 --> 00:22:11.360
It will be complimentary,

721
00:22:11.400 --> 00:22:12.761
but you're not going to replace the scan.

722
00:22:13.321 --> 00:22:14.201
And that's how it started,

723
00:22:14.321 --> 00:22:14.962
just by talking.

724
00:22:15.722 --> 00:22:17.883
And it's always about having those kinds of exchanges.

725
00:22:18.303 --> 00:22:18.563
Actually,

726
00:22:18.644 --> 00:22:19.444
that's what's needed,

727
00:22:19.544 --> 00:22:20.284
not staying stuck.

728
00:22:21.065 --> 00:22:22.325
Maybe it's because very early on,

729
00:22:22.345 --> 00:22:23.466
I was involved in research.

730
00:22:23.846 --> 00:22:25.107
So as a result from early on,

731
00:22:25.147 --> 00:22:25.867
I didn't stay put.

732
00:22:26.319 --> 00:22:27.640
I was a radiology intern,

733
00:22:27.720 --> 00:22:28.720
so that was my whole thing.

734
00:22:28.760 --> 00:22:30.141
And so I had to open up.

735
00:22:30.161 --> 00:22:31.941
I think it's because of that research training.

736
00:22:32.622 --> 00:22:33.062
So for me,

737
00:22:33.122 --> 00:22:34.262
whatever topic I'm given,

738
00:22:34.382 --> 00:22:34.542
well,

739
00:22:34.622 --> 00:22:34.782
yes,

740
00:22:34.822 --> 00:22:35.363
I go for it.

741
00:22:35.364 --> 00:22:35.863
I dive in.

742
00:22:36.283 --> 00:22:36.463
Okay,

743
00:22:36.543 --> 00:22:37.123
so what is it?

744
00:22:37.203 --> 00:22:37.604
The topic?

745
00:22:37.724 --> 00:22:38.264
What's the problem?

746
00:22:38.664 --> 00:22:39.985
Should we involve the stakeholders?

747
00:22:40.045 --> 00:22:40.925
Who should we include?

748
00:22:41.085 --> 00:22:45.807
It takes mental flexibility and energy to get people working together who have different professions,

749
00:22:46.127 --> 00:22:46.487
skills,

750
00:22:46.607 --> 00:22:47.307
and objectives.

751
00:22:47.508 --> 00:22:48.668
We don't have the same goals,

752
00:22:48.928 --> 00:22:54.530
so it requires a certain level of positive energy by saying you have to get people on board,

753
00:22:54.630 --> 00:22:55.811
get them on the fast train.

754
00:22:56.211 --> 00:22:58.472
But the most beautiful projects right now are those,

755
00:22:58.912 --> 00:23:02.954
the ones that break down silos with people who really have different visions,

756
00:23:03.014 --> 00:23:03.674
different minds.

757
00:23:04.074 --> 00:23:05.615
And it's always people from the hospital,

758
00:23:05.855 --> 00:23:08.356
people from research and people from industry.

759
00:23:08.876 --> 00:23:11.717
All my projects now always mix these three worlds together.

760
00:23:12.078 --> 00:23:13.818
So there's something you truly like,

761
00:23:14.299 --> 00:23:16.740
which is bringing together an industry professional,

762
00:23:16.840 --> 00:23:18.960
the hospital and research at the same time.

763
00:23:20.181 --> 00:23:24.943
Is there a specific project that you are especially proud to have successfully carried out using this exact methodology?

764
00:23:25.959 --> 00:23:26.079
Well,

765
00:23:26.119 --> 00:23:31.924
I think it will be my best publication and my greatest experience of collective intelligence,

766
00:23:31.984 --> 00:23:34.206
even though it was an artificial intelligence project.

767
00:23:34.546 --> 00:23:35.547
It was during COVID.

768
00:23:35.907 --> 00:23:40.291
You know that at that time all research labs were asked to close except those that were going to work on COVID.

769
00:23:40.691 --> 00:23:41.992
I was still a researcher at heart,

770
00:23:42.452 --> 00:23:43.353
along with another colleague.

771
00:23:44.234 --> 00:23:48.317
Here we were also busy with Gustave because we were receiving all the patients who had COVID.

772
00:23:48.477 --> 00:23:49.278
And even afterwards,

773
00:23:49.538 --> 00:23:52.521
we were asked to take in patients since the APHP was completely...

774
00:23:52.981 --> 00:23:53.141
well,

775
00:23:53.161 --> 00:23:53.802
there was a wave,

776
00:23:53.822 --> 00:23:54.062
you see,

777
00:23:54.082 --> 00:23:55.383
they no longer had enough bids

778
00:23:55.975 --> 00:23:58.356
So we were asked to open Gustav as well,

779
00:23:58.436 --> 00:24:00.297
to take in non-cancer patients who had COVID.

780
00:24:00.677 --> 00:24:01.677
We were overwhelmed,

781
00:24:01.778 --> 00:24:03.218
but I still wanted to do research.

782
00:24:03.718 --> 00:24:05.999
So a colleague and I decided to study COVID.

783
00:24:07.020 --> 00:24:09.741
This allowed us to keep working since it was essential research.

784
00:24:10.761 --> 00:24:11.822
Outside regular hours,

785
00:24:11.882 --> 00:24:14.863
since we spent our days on Doppler ultrasounds and working in the ICU,

786
00:24:15.443 --> 00:24:16.484
then we said to ourselves,

787
00:24:16.524 --> 00:24:16.644
OK,

788
00:24:16.744 --> 00:24:17.684
what do our patients need?

789
00:24:17.984 --> 00:24:18.725
Always the same thing.

790
00:24:18.805 --> 00:24:23.807
So we talked with the intensivists and the issue is a patient arrives in the emergency room.

791
00:24:24.247 --> 00:24:24.567
In fact,

792
00:24:24.587 --> 00:24:30.769
we don't know whether to let them go home or keep them if in 12 hours they'll be in the ICU dead or back at home.

793
00:24:31.710 --> 00:24:33.010
So he said,

794
00:24:33.170 --> 00:24:34.851
if you can find us a scoring system,

795
00:24:35.211 --> 00:24:39.092
something that can tell us whether we should at least keep the patient or send them home.

796
00:24:39.272 --> 00:24:39.392
OK,

797
00:24:40.512 --> 00:24:41.213
we said to ourselves,

798
00:24:41.253 --> 00:24:41.373
OK,

799
00:24:41.593 --> 00:24:43.633
since we've started working with artificial intelligence,

800
00:24:44.334 --> 00:24:45.054
we said to ourselves,

801
00:24:45.094 --> 00:24:48.095
let's use imaging because there was a lot of talk about CT scans.

802
00:24:48.506 --> 00:24:52.508
We could see the shading on the scan with the ground glass opacity and all that.

803
00:24:52.688 --> 00:24:54.810
We decided to work on AI using the CT scans.

804
00:24:55.330 --> 00:24:57.151
We decided not to stay isolated.

805
00:24:57.251 --> 00:25:02.254
We reviewed early reports from the Chinese because they published one or two months before us.

806
00:25:02.514 --> 00:25:04.475
We decided to collect all clinical,

807
00:25:04.575 --> 00:25:06.356
biological and medical history data.

808
00:25:06.536 --> 00:25:08.517
We'll collect everything to build a scoring system.

809
00:25:08.797 --> 00:25:09.298
At the time,

810
00:25:09.318 --> 00:25:12.199
we didn't have a CT scanner running constantly for COVID.

811
00:25:12.940 --> 00:25:14.801
But at the APHP Kremlin Bicêtre,

812
00:25:15.301 --> 00:25:17.022
my colleagues in the research lab you

813
00:25:17.502 --> 00:25:23.667
Franz Belen and the others actually had an entire dedicated CT scanner that was only doing COVID cases.

814
00:25:23.908 --> 00:25:24.468
So I told them,

815
00:25:24.848 --> 00:25:25.008
well,

816
00:25:25.189 --> 00:25:27.771
I would like to do a study with our patients and your patients.

817
00:25:28.231 --> 00:25:34.856
That will provide a validation cohort to try to see how we can predict when the patient arrives at the emergency room based on all these parameters,

818
00:25:35.197 --> 00:25:36.658
whether they need to be admitted or not.

819
00:25:37.038 --> 00:25:37.579
They said to me,

820
00:25:37.839 --> 00:25:37.959
OK.

821
00:25:38.800 --> 00:25:39.140
Then I said,

822
00:25:39.160 --> 00:25:39.340
well,

823
00:25:39.380 --> 00:25:40.721
I need to find the best people in AI.

824
00:25:41.722 --> 00:25:46.466
So then I went looking of the two teams I worked with.

825
00:25:47.010 --> 00:25:47.971
One was from INRIA,

826
00:25:48.411 --> 00:25:50.073
an AI lab at Centralis Superlec,

827
00:25:50.113 --> 00:25:50.733
the CVN,

828
00:25:51.133 --> 00:25:51.774
and then Kin,

829
00:25:51.994 --> 00:25:53.035
who I also worked with.

830
00:25:53.515 --> 00:25:54.256
And I told them,

831
00:25:54.736 --> 00:25:55.177
basically,

832
00:25:55.237 --> 00:25:55.737
I told them,

833
00:25:55.817 --> 00:25:55.977
well,

834
00:25:56.037 --> 00:25:56.618
here's the plan.

835
00:25:57.078 --> 00:25:59.720
We'll do a study and I'd like to use both your technologies,

836
00:25:59.901 --> 00:26:00.681
which were different.

837
00:26:01.522 --> 00:26:03.323
One was in machine learning,

838
00:26:03.523 --> 00:26:04.264
so in deep learning.

839
00:26:04.384 --> 00:26:05.205
I told them after that,

840
00:26:05.305 --> 00:26:05.705
it's a bet.

841
00:26:05.865 --> 00:26:06.586
May the best one win.

842
00:26:06.706 --> 00:26:08.367
We'll take the best technology and use it.

843
00:26:08.948 --> 00:26:09.068
OK,

844
00:26:09.668 --> 00:26:12.931
that's a challenge between getting the contract signed with APHP,

845
00:26:13.351 --> 00:26:14.112
retrieving the data,

846
00:26:14.192 --> 00:26:15.173
which was no small matter,

847
00:26:15.533 --> 00:26:16.274
and everything else.

848
00:26:16.966 --> 00:26:17.106
Well,

849
00:26:17.146 --> 00:26:17.646
let me tell you,

850
00:26:17.647 --> 00:26:18.367
in just 10 days,

851
00:26:18.427 --> 00:26:19.247
which is unimaginable,

852
00:26:19.307 --> 00:26:19.708
we did it.

853
00:26:20.028 --> 00:26:21.309
We signed the contracts with APHP,

854
00:26:21.569 --> 00:26:22.129
with Aukin,

855
00:26:22.209 --> 00:26:22.649
with INRIA,

856
00:26:22.750 --> 00:26:23.190
and so on.

857
00:26:23.690 --> 00:26:24.571
And I got everyone working.

858
00:26:24.591 --> 00:26:25.451
We were 45 people.

859
00:26:25.871 --> 00:26:27.252
They put half their lab on it for me.

860
00:26:27.592 --> 00:26:28.733
You can check in the publication,

861
00:26:28.793 --> 00:26:29.914
which was published in Nature.

862
00:26:30.354 --> 00:26:30.934
It's my longest,

863
00:26:30.974 --> 00:26:31.595
most major work.

864
00:26:32.115 --> 00:26:32.996
Half Aukin on it,

865
00:26:33.236 --> 00:26:34.036
45 total.

866
00:26:34.637 --> 00:26:36.278
There are 20 co-authors on my paper.

867
00:26:36.458 --> 00:26:37.618
And then everyone got involved.

868
00:26:37.718 --> 00:26:38.199
In the end,

869
00:26:38.219 --> 00:26:40.680
I told them I need a school for intensive care.

870
00:26:41.200 --> 00:26:43.662
So we included 1,000 patients in one month.

871
00:26:44.306 --> 00:26:45.727
Every evening at 9 p.m.

872
00:26:45.847 --> 00:26:49.348
we went to the Ambicet ward to record all patient data by hand.

873
00:26:49.788 --> 00:26:50.229
Honestly,

874
00:26:50.409 --> 00:26:51.049
it was true,

875
00:26:51.129 --> 00:26:52.450
we did everything ourselves.

876
00:26:53.030 --> 00:26:56.091
People were stressed by what was happening in the hospital and the study,

877
00:26:56.471 --> 00:26:57.852
but we finished the publication.

878
00:26:58.272 --> 00:26:59.293
We sent it to the study,

879
00:26:59.353 --> 00:27:02.614
which challenged us to compare our results to 15 others.

880
00:27:02.974 --> 00:27:05.716
And there was an amazing methodologist who is the last author.

881
00:27:06.636 --> 00:27:07.156
Michael Bloom,

882
00:27:07.157 --> 00:27:07.837
you should look him up,

883
00:27:07.877 --> 00:27:08.997
he was an incredible guy.

884
00:27:09.657 --> 00:27:10.818
And we published a great paper.

885
00:27:11.262 --> 00:27:11.823
And in the end,

886
00:27:11.824 --> 00:27:14.044
you have a kind of scoring system with a color code.

887
00:27:14.465 --> 00:27:15.566
When the patient arrives,

888
00:27:15.746 --> 00:27:16.006
you know,

889
00:27:16.046 --> 00:27:16.547
with AI,

890
00:27:16.587 --> 00:27:17.467
the scan results.

891
00:27:18.328 --> 00:27:20.710
And then it was the oxygen saturation level,

892
00:27:20.750 --> 00:27:21.771
the platelet count,

893
00:27:21.851 --> 00:27:23.933
the duration rate that no one expected,

894
00:27:24.373 --> 00:27:25.454
and then age and sex.

895
00:27:26.115 --> 00:27:26.635
And with that,

896
00:27:26.675 --> 00:27:27.035
you knew,

897
00:27:27.175 --> 00:27:29.077
depending on the color and the involvement.

898
00:27:29.738 --> 00:27:32.420
So it was a paper with explainability in AI,

899
00:27:32.520 --> 00:27:33.621
which is very important.

900
00:27:34.422 --> 00:27:35.542
This allowed a doctor to say,

901
00:27:35.683 --> 00:27:35.803
OK,

902
00:27:35.943 --> 00:27:37.024
this patient is orange-red.

903
00:27:37.824 --> 00:27:38.545
I need to keep them,

904
00:27:38.565 --> 00:27:39.586
but this one is blue-green.

905
00:27:40.394 --> 00:27:41.415
so I don't need to keep them.

906
00:27:41.895 --> 00:27:42.775
And that was incredible.

907
00:27:43.336 --> 00:27:44.616
And we did that in three months.

908
00:27:45.057 --> 00:27:46.497
But it's surreal when you say that.

909
00:27:47.038 --> 00:27:48.859
Any industry professional or any laboratory,

910
00:27:48.919 --> 00:27:49.859
like if you explain this,

911
00:27:49.899 --> 00:27:50.239
they'll say,

912
00:27:50.620 --> 00:27:51.340
how did she do that?

913
00:27:52.261 --> 00:27:52.841
I told myself,

914
00:27:52.901 --> 00:27:53.061
yeah,

915
00:27:53.481 --> 00:27:59.705
never again will I do a study with 100 meters between the hospital industry and academia each time.

916
00:27:59.885 --> 00:28:00.545
It was possible.

917
00:28:00.765 --> 00:28:01.966
It was a bit tough at the beginning,

918
00:28:02.026 --> 00:28:02.446
but well,

919
00:28:02.626 --> 00:28:03.047
in the end,

920
00:28:03.067 --> 00:28:05.548
it produces something magnificent.

921
00:28:06.468 --> 00:28:06.989
Magnificent,

922
00:28:07.049 --> 00:28:07.269
really.

923
00:28:07.689 --> 00:28:09.310
It's the best paper of my entire career.

924
00:28:09.470 --> 00:28:12.032
And it was truly an adventure in collective intelligence,

925
00:28:12.092 --> 00:28:13.513
more than artificial intelligence.

926
00:28:13.873 --> 00:28:14.974
Even though we used AI,

927
00:28:14.994 --> 00:28:16.755
it was really collective intelligence.

928
00:28:17.295 --> 00:28:18.936
Every time we made a video with everyone,

929
00:28:19.017 --> 00:28:20.037
with all the stakeholders,

930
00:28:20.097 --> 00:28:20.598
to be sure.

931
00:28:20.938 --> 00:28:21.238
All right,

932
00:28:21.278 --> 00:28:22.399
so what do you have the data,

933
00:28:22.739 --> 00:28:23.159
the stuff?

934
00:28:23.459 --> 00:28:24.040
Every evening,

935
00:28:24.041 --> 00:28:25.521
we would have a video call with everyone,

936
00:28:25.801 --> 00:28:27.702
because I had to keep motivating everyone,

937
00:28:27.703 --> 00:28:28.363
the whole team.

938
00:28:29.323 --> 00:28:30.604
Those who were locked up at home,

939
00:28:30.624 --> 00:28:31.585
but working like crazy,

940
00:28:31.765 --> 00:28:32.986
analyzing the scans as well.

941
00:28:33.806 --> 00:28:35.007
Those who were collecting the data,

942
00:28:35.107 --> 00:28:36.048
those who were on site.

943
00:28:36.428 --> 00:28:36.909
It was funny.

944
00:28:37.329 --> 00:28:37.929
It was amazing.

945
00:28:38.189 --> 00:28:38.830
Very rewarding.

946
00:28:39.202 --> 00:28:40.183
a very tough period.

947
00:28:40.523 --> 00:28:40.783
For me,

948
00:28:40.983 --> 00:28:45.767
I think it was the hardest period of my career as a doctor to lose our patients like that,

949
00:28:46.467 --> 00:28:47.948
but at the same time very rewarding.

950
00:28:48.709 --> 00:28:54.913
I think it was also a way for us to keep our heads above water and not let ourselves be overwhelmed by what was happening.

951
00:28:55.454 --> 00:28:57.035
It was really hard.

952
00:28:57.315 --> 00:28:58.556
I was not really afraid to read.

953
00:28:59.196 --> 00:28:59.316
Me,

954
00:28:59.356 --> 00:28:59.937
I was afraid.

955
00:29:00.317 --> 00:29:01.378
I was 55 years old.

956
00:29:01.498 --> 00:29:03.680
I was in the age group of people who could die quickly.

957
00:29:04.500 --> 00:29:05.501
I could catch it and die.

958
00:29:05.841 --> 00:29:06.902
In a week I could be dead.

959
00:29:07.342 --> 00:29:09.023
Those are still things that cross your mind.

960
00:29:09.783 --> 00:29:10.903
Even if we had the equipment,

961
00:29:11.083 --> 00:29:11.803
we were protected,

962
00:29:11.903 --> 00:29:13.304
dressed like little astronauts all day.

963
00:29:13.924 --> 00:29:15.985
But those are still things that cross your mind.

964
00:29:16.105 --> 00:29:16.225
Me,

965
00:29:16.505 --> 00:29:17.105
I said it.

966
00:29:17.265 --> 00:29:18.005
I was scared.

967
00:29:18.205 --> 00:29:18.665
It's true.

968
00:29:19.466 --> 00:29:19.826
Thank you.

969
00:29:20.106 --> 00:29:23.387
It was very interesting to see the innovations coming out of the research centers.

970
00:29:23.507 --> 00:29:24.627
Thank you for this exchange.

971
00:29:24.907 --> 00:29:26.147
It was a pleasure meeting you.

972
00:29:26.508 --> 00:29:26.848
Goodbye.

973
00:29:27.068 --> 00:29:28.608
If Europe excels in AI research,

974
00:29:28.708 --> 00:29:30.149
why does access remain unequal?

975
00:29:30.629 --> 00:29:31.929
Is it because of regulation,

976
00:29:32.189 --> 00:29:33.109
industrial caution,

977
00:29:33.870 --> 00:29:34.970
fragmentation of systems?

978
00:29:36.834 --> 00:29:37.716
In the next episode,

979
00:29:37.736 --> 00:29:39.380
we will address another structural gap.

980
00:29:40.142 --> 00:29:42.909
Why are there 100 new drugs already available?

981
00:29:43.370 --> 00:29:44.212
In the United States,

982
00:29:44.252 --> 00:29:45.034
but not in Europe?

983
00:29:47.581 --> 00:29:48.762
And to follow up on this video,

984
00:29:48.862 --> 00:29:50.044
I'll be waiting for your comments.

985
00:29:50.604 --> 00:29:51.205
In your country,

986
00:29:51.285 --> 00:29:53.407
what is the biggest barrier to accessing AI?

987
00:29:53.627 --> 00:29:54.769
Is it funding,

988
00:29:55.229 --> 00:29:55.790
regulation,

989
00:29:56.431 --> 00:29:58.513
or collaboration between industry and academia?