Measurement of Radiation and Dosimetric Procedure.pptx
When digital medicine becomes the medicine (1/2)
1. Professor, SAHIST, Sungkyunkwan University
Director, Digital Healthcare Institute
Yoon Sup Choi, Ph.D.
디지털 의료가 ‘의료’가 될 때
When Digital Medicine Becomes the Medicine
6. 최윤섭 지음
의료인공지능
표지디자인•최승협
컴퓨터
털 헬
치를 만드는 것을 화두로
기업가, 엔젤투자가, 에반
의 대표적인 전문가로, 활
이 분야를 처음 소개한 장
포항공과대학교에서 컴
동 대학원 시스템생명공
취득하였다. 스탠퍼드대
조교수, KT 종합기술원 컨
구원 연구조교수 등을 거
저널에 10여 편의 논문을
국내 최초로 디지털 헬스
윤섭 디지털 헬스케어 연
국내 유일의 헬스케어 스
어 파트너스’의 공동 창업
스타트업을 의료 전문가
관대학교 디지털헬스학과
뷰노, 직토, 3billion, 서지
소울링, 메디히어, 모바일
자문을 맡아 한국에서도
고 있다. 국내 최초의 디
케어 이노베이션』에 활발
을 연재하고 있다. 저서로
와 『그렇게 나는 스스로
•블로그_ http://www
•페이스북_ https://w
•이메일_ yoonsup.c
최윤섭
의료 인공지능은 보수적인 의료 시스템을 재편할 혁신을 일으키고 있다. 의료 인공지능의 빠른 발전과
광범위한 영향은 전문화, 세분화되며 발전해 온 현대 의료 전문가들이 이해하기가 어려우며, 어디서부
터 공부해야 할지도 막연하다. 이런 상황에서 의료 인공지능의 개념과 적용, 그리고 의사와의 관계를 쉽
게 풀어내는 이 책은 좋은 길라잡이가 될 것이다. 특히 미래의 주역이 될 의학도와 젊은 의료인에게 유용
한 소개서이다.
━ 서준범, 서울아산병원 영상의학과 교수, 의료영상인공지능사업단장
인공지능이 의료의 패러다임을 크게 바꿀 것이라는 것에 동의하지 않는 사람은 거의 없다. 하지만 인공
지능이 처리해야 할 의료의 난제는 많으며 그 해결 방안도 천차만별이다. 흔히 생각하는 만병통치약 같
은 의료 인공지능은 존재하지 않는다. 이 책은 다양한 의료 인공지능의 개발, 활용 및 가능성을 균형 있
게 분석하고 있다. 인공지능을 도입하려는 의료인, 생소한 의료 영역에 도전할 인공지능 연구자 모두에
게 일독을 권한다.
━ 정지훈, 경희사이버대 미디어커뮤니케이션학과 선임강의교수, 의사
서울의대 기초의학교육을 책임지고 있는 교수의 입장에서, 산업화 이후 변하지 않은 현재의 의학 교육
으로는 격변하는 인공지능 시대에 의대생을 대비시키지 못한다는 한계를 절실히 느낀다. 저와 함께 의
대 인공지능 교육을 개척하고 있는 최윤섭 소장의 전문적 분석과 미래 지향적 안목이 담긴 책이다. 인공
지능이라는 미래를 대비할 의대생과 교수, 그리고 의대 진학을 고민하는 학생과 학부모에게 추천한다.
━ 최형진, 서울대학교 의과대학 해부학교실 교수, 내과 전문의
최근 의료 인공지능의 도입에 대해서 극단적인 시각과 태도가 공존하고 있다. 이 책은 다양한 사례와 깊
은 통찰을 통해 의료 인공지능의 현황과 미래에 대해 균형적인 시각을 제공하여, 인공지능이 의료에 본
격적으로 도입되기 위한 토론의 장을 마련한다. 의료 인공지능이 일상화된 10년 후 돌아보았을 때, 이 책
이 그런 시대를 이끄는 길라잡이 역할을 하였음을 확인할 수 있기를 기대한다.
━ 정규환, 뷰노 CTO
의료 인공지능은 다른 분야 인공지능보다 더 본질적인 이해가 필요하다. 단순히 인간의 일을 대신하는
수준을 넘어 의학의 패러다임을 데이터 기반으로 변화시키기 때문이다. 따라서 인공지능을 균형있게 이
해하고, 어떻게 의사와 환자에게 도움을 줄 수 있을지 깊은 고민이 필요하다. 세계적으로 일어나고 있는
이러한 노력의 결과물을 집대성한 이 책이 반가운 이유다.
━ 백승욱, 루닛 대표
의료 인공지능의 최신 동향뿐만 아니라, 의의와 한계, 전망, 그리고 다양한 생각거리까지 주는 책이다.
논쟁이 되는 여러 이슈에 대해서도 저자는 자신의 시각을 명확한 근거에 기반하여 설득력 있게 제시하
고 있다. 개인적으로는 이 책을 대학원 수업 교재로 활용하려 한다.
━ 신수용, 성균관대학교 디지털헬스학과 교수
최윤섭지음
의료인공지능
값 20,000원
ISBN 979-11-86269-99-2
최초의 책!
계 안팎에서 제기
고 있다. 현재 의
분 커버했다고 자
것인가, 어느 진료
제하고 효용과 안
누가 지는가, 의학
쉬운 언어로 깊이
들이 의료 인공지
적인 용어를 최대
서 다른 곳에서 접
를 접하게 될 것
너무나 빨리 발전
책에서 제시하는
술을 공부하며, 앞
란다.
의사 면허를 취득
저가 도움되면 좋
를 불러일으킬 것
화를 일으킬 수도
슈에 제대로 대응
분은 의학 교육의
예비 의사들은 샌
지능과 함께하는
레이닝 방식도 이
전에 진료실과 수
겠지만, 여러분들
도생하는 수밖에
미래의료학자 최윤섭 박사가 제시하는
의료 인공지능의 현재와 미래
의료 딥러닝과 IBM 왓슨의 현주소
인공지능은 의사를 대체하는가
값 20,000원
ISBN 979-11-86269-99-2
레이닝 방식도 이
전에 진료실과 수
겠지만, 여러분들
도생하는 수밖에
소울링, 메디히어, 모바일
자문을 맡아 한국에서도
고 있다. 국내 최초의 디
케어 이노베이션』에 활발
을 연재하고 있다. 저서로
와 『그렇게 나는 스스로
•블로그_ http://www
•페이스북_ https://w
•이메일_ yoonsup.c
14. • "2018년 3Q는 역대 최고로 투자 받기 좋은 시기였다”
• 2018년 3Q에서 이미 2017년 투자 규모를 능가
• 모든 라운드에서 더 높은 빈도로, 더 큰 금액이 투자되는 entrepreneurs’ market
15. 헬스케어넓은 의미의 건강 관리에는 해당되지만,
디지털 기술이 적용되지 않고, 전문 의료 영역도 아닌 것
예) 운동, 영양, 수면
디지털 헬스케어
건강 관리 중에 디지털 기술이 사용되는 것
예) 사물인터넷, 인공지능, 3D 프린터, VR/AR
모바일 헬스케어
디지털 헬스케어 중
모바일 기술이 사용되는 것
예) 스마트폰, 사물인터넷, SNS
개인 유전정보분석
예) 암유전체, 질병위험도,
보인자, 약물 민감도
예) 웰니스, 조상 분석
헬스케어 관련 분야 구성도 (ver 0.3)
의료
질병 예방, 치료, 처방, 관리
등 전문 의료 영역
원격의료
원격진료
32. Fig 1. What can consumer wearables do? Heart rate can be measured with an oximeter built into a ring [3], muscle activity with an electromyographi
sensor embedded into clothing [4], stress with an electodermal sensor incorporated into a wristband [5], and physical activity or sleep patterns via an
accelerometer in a watch [6,7]. In addition, a female’s most fertile period can be identified with detailed body temperature tracking [8], while levels of me
attention can be monitored with a small number of non-gelled electroencephalogram (EEG) electrodes [9]. Levels of social interaction (also known to a
PLOS Medicine 2016
42. Human genomes are being sequenced at an ever-increasing rate. The 1000 Genomes Project has
aggregated hundreds of genomes; The Cancer Genome Atlas (TGCA) has gathered several thousand; and
the Exome Aggregation Consortium (ExAC) has sequenced more than 60,000 exomes. Dotted lines show
three possible future growth curves.
DNA SEQUENCING SOARS
2001 2005 2010 2015 2020 2025
100
103
106
109
Human Genome Project
Cumulativenumberofhumangenomes
1000 Genomes
TCGA
ExAC
Current amount
1st personal genome
Recorded growth
Projection
Double every 7 months (historical growth rate)
Double every 12 months (Illumina estimate)
Double every 18 months (Moore's law)
Michael Einsetein, Nature, 2015
43. more rapid and accurate approaches to infectious diseases. The driver mutations and key biologic unde
Sequencing Applications in Medicine
from Prewomb to Tomb
Cell. 2014 Mar 27; 157(1): 241–253.
46. Digital Phenotype:
Your smartphone knows if you are depressed
J Med Internet Res. 2015 Jul 15;17(7):e175.
The correlation analysis between the features and the PHQ-9 scores revealed that 6 of the 10
features were significantly correlated to the scores:
• strong correlation: circadian movement, normalized entropy, location variance
• correlation: phone usage features, usage duration and usage frequency
48. ers, Jared B Hawkins & John S Brownstein
phenotypes captured to enhance health and wellness will extend to human interactions with
st Richard
pt of the
hat pheno-
biological
sis or tissue
effects that
or outside
m.Dawkins
phenotypes
can modify
difications
onsofone’s
ended phe-
cites damn
hebeaver’s
ncreasingly
there is an
heory—the
aspects of
ehowdiag-
Jan. 2013
0.000
0.002
0.004
Density
0.006
July 2013 Jan. 2014 July 2014
User 1
User 2
User 3
User 4
User 5
User 6
User 7
Date
Figure 1 Timeline of insomnia-related tweets from representative individuals. Density distributions
(probability density functions) are shown for seven individual users over a two-year period. Density on
the y axis highlights periods of relative activity for each user. A representative tweet from each user is
Your twitter knows if you cannot sleep
Timeline of insomnia-related tweets from representative individuals.
Nat. Biotech. 2015
50. Digital Phenotype:
Your Instagram knows if you are depressed
Rao (MVR) (24) .
Results
Both Alldata and Prediagnosis models were decisively superior to a null model
. Alldata predictors were significant with 99% probability.57.5;(KAll = 1 K 49.8) Pre = 1 7
Prediagnosis and Alldata confidence levels were largely identical, with two exceptions:
Prediagnosis Brightness decreased to 90% confidence, and Prediagnosis posting frequency
dropped to 30% confidence, suggesting a null predictive value in the latter case.
Increased hue, along with decreased brightness and saturation, predicted depression. This
means that photos posted by depressed individuals tended to be bluer, darker, and grayer (see
Fig. 2). The more comments Instagram posts received, the more likely they were posted by
depressed participants, but the opposite was true for likes received. In the Alldata model, higher
posting frequency was also associated with depression. Depressed participants were more likely
to post photos with faces, but had a lower average face count per photograph than healthy
participants. Finally, depressed participants were less likely to apply Instagram filters to their
posted photos.
Fig. 2. Magnitude and direction of regression coefficients in Alldata (N=24,713) and Prediagnosis (N=18,513)
models. Xaxis values represent the adjustment in odds of an observation belonging to depressed individuals, per
Reece & Danforth, “Instagram photos reveal predictive markers of depression” (2016)
Fig. 1. Comparison of HSV values. Right photograph has higher Hue (bluer), lower Saturation (grayer), and lower
Brightness (darker) than left photograph. Instagram photos posted by depressed individuals had HSV values
shifted towards those in the right photograph, compared with photos posted by healthy individuals.
Units of observation
In determining the best time span for this analysis, we encountered a difficult question:
When and for how long does depression occur? A diagnosis of depression does not indicate the
persistence of a depressive state for every moment of every day, and to conduct analysis using an
individual’s entire posting history as a single unit of observation is therefore rather specious. At
the other extreme, to take each individual photograph as units of observation runs the risk of
being too granular. DeChoudhury et al. (5) looked at all of a given user’s posts in a single day,
and aggregated those data into perperson, perday units of observation. We adopted this
precedent of “userdays” as a unit of analysis . 5
Statistical framework
We used Bayesian logistic regression with uninformative priors to determine the strength
of individual predictors. Two separate models were trained. The Alldata model used all
collected data to address Hypothesis 1. The Prediagnosis model used all data collected from
higher Hue (bluer)
lower Saturation (grayer)
lower Brightness (darker)
51. Digital Phenotype:
Your Instagram knows if you are depressed
Reece & Danforth, “Instagram photos reveal predictive markers of depression” (2016)
. In particular, depressedχ2 07.84, p .17e 64;( All = 9 = 9 − 1 13.80, p .87e 44)χ2Pre = 8 = 2 − 1
participants were less likely than healthy participants to use any filters at all. When depressed
participants did employ filters, they most disproportionately favored the “Inkwell” filter, which
converts color photographs to blackandwhite images. Conversely, healthy participants most
disproportionately favored the Valencia filter, which lightens the tint of photos. Examples of
filtered photographs are provided in SI Appendix VIII.
Fig. 3. Instagram filter usage among depressed and healthy participants. Bars indicate difference between observed
and expected usage frequencies, based on a Chisquared analysis of independence. Blue bars indicate
disproportionate use of a filter by depressed compared to healthy participants, orange bars indicate the reverse.
52. Digital Phenotype:
Your Instagram knows if you are depressed
Reece & Danforth, “Instagram photos reveal predictive markers of depression” (2016)
VIII. Instagram filter examples
Fig. S8. Examples of Inkwell and Valencia Instagram filters. Inkwell converts
color photos to blackandwhite, Valencia lightens tint. Depressed participants
most favored Inkwell compared to healthy participants, Healthy participants
54. Leading Edge
Review
Individualized Medicine
from Prewomb to Tomb
Eric J. Topol1 ,*
1The Scripps Translational Science Institute, The Scripps Research Institute and Scripps Health, La Jolla, CA 92037, USA
*Correspondence: etopol@scripps.edu
http://dx.doi.org/10.1016/j.cell.2014.02.012
That each of us is truly biologically unique, extending to even monozygotic, ‘‘identical’’ twins, is not
fully appreciated. Now that it is possible to perform a comprehensive ‘‘omic’’ assessment of an
individual, including one’s DNA and RNA sequence and at least some characterization of one’s
proteome, metabolome, microbiome, autoantibodies, and epigenome, it has become abundantly
clear that each of us has truly one-of-a-kind biological content. Well beyond the allure of the match-
less fingerprint or snowflake concept, these singular, individual data and information set up a
remarkable and unprecedented opportunity to improve medical treatment and develop preventive
strategies to preserve health.
From Digital to Biological to Individualized Medicine
In 2010, Eric Schmidt of Google said ‘‘The power of individual
targeting—the technology will be so good it will be very hard
for people to watch or consume something that has not in
some sense been tailored for them’’ (Jenkins, 2010). Although
referring to the capability of digital technology, we have now
reached a time of convergence of the digital and biologic do-
mains. It has been well established that 0 and 1 are interchange-
able with A, C, T, and G in books and Shakespeare sonnets and
that DNA may represent the ultimate data storage system
(Church et al., 2012; Goldman et al., 2013b). Biological transis-
tors, also known as genetic logic gates, have now been devel-
oped that make a computer from a living cell (Bonnet et al.,
2013). The convergence of biology and technology was further
captured by one of the protagonists of the digital era, Steve
Jobs, who said ‘‘I think the biggest innovations of the 21st
cen-
tury will be at the intersection of biology and technology. A
new era is beginning’’ (Issacson, 2011).
With whole-genome DNA sequencing and a variety of omic
technologies to define aspects of each individual’s biology at
many different levels, we have indeed embarked on a new era
of medicine. The term ‘‘personalized medicine’’ has been used
for many years but has engendered considerable confusion. A
recent survey indicated that only 4% of the public understand
what the term is intended to mean (Stanton, 2013), and the hack-
neyed, commercial use of ‘‘personalized’’ makes many people
think that this refers to a concierge service of medical care.
Whereas ‘‘person’’ refers to a human being, ‘‘personalized’’
can mean anything from having monogrammed stationary or
luggage to ascribing personal qualities. Therefore, it was not
surprising that a committee representing the National Academy
of Sciences proposed using the term ‘‘precision medicine’’ as
defined by ‘‘tailoring of medical treatment to the individual char-
acteristics of each patient’’ (National Research Council, 2011).
Although the term ‘‘precision’’ denotes the objective of exact-
ness, ironically, it too can be viewed as ambiguous in this context
because it does not capture the sense that the information is
derived from the individual. For example, many laboratory tests
could be made more precise by assay methodology, and treat-
ments could be made more precise by avoiding side effects—
without having anything to do with a specific individual. Other
terms that have been suggested include genomic, digital, and
stratified medicine, but all of these have a similar problem or
appear to be too narrowly focused.
The definition of individual is a single human being, derived
from the Latin word individu, or indivisible. I propose individual-
ized medicine as the preferred term because it has a useful
double entendre. It relates not only to medicine that is particular-
ized to a human being but also the future impact of digital
technology on individuals driving their health care. There will
increasingly be the flow of one’s biologic data and relevant
medical information directly to the individual. Be it a genome
sequence on a tablet or the results of a biosensor for blood pres-
sure or another physiologic metric displayed on a smartphone,
the digital convergence with biology will definitively anchor the
individual as a source of salient data, the conduit of information
flow, and a—if not the—principal driver of medicine in the future.
The Human GIS
Perhaps the most commonly used geographic information
systems (GIS) are Google maps, which provide a layered
approach to data visualization, such as viewing a location via
satellite overlaid with street names, landmarks, and real-time
traffic data. This GIS exemplifies the concept of gathering and
transforming large bodies of data to provide exquisite temporal
and location information. With the multiple virtual views, it gives
one the sense of physically being on site. Although Google has
digitized and thus created a GIS for the Earth, it is now possible
to digitize a human being. As shown in Figure 1, there are multi-
ple layers of data that can now be obtained for any individual.
This includes data from biosensors, scanners, electronic medi-
cal records, social media, and the various omics that include
Cell 157, March 27, 2014 ª2014 Elsevier Inc. 241
55. Leading Edge
Review
Individualized Medicine
from Prewomb to Tomb
Eric J. Topol1 ,*
1The Scripps Translational Science Institute, The Scripps Research Institute and Scripps Health, La Jolla, CA 92037, USA
*Correspondence: etopol@scripps.edu
http://dx.doi.org/10.1016/j.cell.2014.02.012
That each of us is truly biologically unique, extending to even monozygotic, ‘‘identical’’ twins, is not
fully appreciated. Now that it is possible to perform a comprehensive ‘‘omic’’ assessment of an
individual, including one’s DNA and RNA sequence and at least some characterization of one’s
proteome, metabolome, microbiome, autoantibodies, and epigenome, it has become abundantly
clear that each of us has truly one-of-a-kind biological content. Well beyond the allure of the match-
less fingerprint or snowflake concept, these singular, individual data and information set up a
remarkable and unprecedented opportunity to improve medical treatment and develop preventive
strategies to preserve health.
From Digital to Biological to Individualized Medicine
In 2010, Eric Schmidt of Google said ‘‘The power of individual
targeting—the technology will be so good it will be very hard
for people to watch or consume something that has not in
some sense been tailored for them’’ (Jenkins, 2010). Although
referring to the capability of digital technology, we have now
reached a time of convergence of the digital and biologic do-
mains. It has been well established that 0 and 1 are interchange-
able with A, C, T, and G in books and Shakespeare sonnets and
that DNA may represent the ultimate data storage system
(Church et al., 2012; Goldman et al., 2013b). Biological transis-
tors, also known as genetic logic gates, have now been devel-
oped that make a computer from a living cell (Bonnet et al.,
2013). The convergence of biology and technology was further
captured by one of the protagonists of the digital era, Steve
Jobs, who said ‘‘I think the biggest innovations of the 21st
cen-
tury will be at the intersection of biology and technology. A
new era is beginning’’ (Issacson, 2011).
With whole-genome DNA sequencing and a variety of omic
technologies to define aspects of each individual’s biology at
many different levels, we have indeed embarked on a new era
of medicine. The term ‘‘personalized medicine’’ has been used
for many years but has engendered considerable confusion. A
recent survey indicated that only 4% of the public understand
what the term is intended to mean (Stanton, 2013), and the hack-
neyed, commercial use of ‘‘personalized’’ makes many people
think that this refers to a concierge service of medical care.
Whereas ‘‘person’’ refers to a human being, ‘‘personalized’’
can mean anything from having monogrammed stationary or
luggage to ascribing personal qualities. Therefore, it was not
surprising that a committee representing the National Academy
of Sciences proposed using the term ‘‘precision medicine’’ as
defined by ‘‘tailoring of medical treatment to the individual char-
acteristics of each patient’’ (National Research Council, 2011).
Although the term ‘‘precision’’ denotes the objective of exact-
ness, ironically, it too can be viewed as ambiguous in this context
because it does not capture the sense that the information is
derived from the individual. For example, many laboratory tests
could be made more precise by assay methodology, and treat-
ments could be made more precise by avoiding side effects—
without having anything to do with a specific individual. Other
terms that have been suggested include genomic, digital, and
stratified medicine, but all of these have a similar problem or
appear to be too narrowly focused.
The definition of individual is a single human being, derived
from the Latin word individu, or indivisible. I propose individual-
ized medicine as the preferred term because it has a useful
double entendre. It relates not only to medicine that is particular-
ized to a human being but also the future impact of digital
technology on individuals driving their health care. There will
increasingly be the flow of one’s biologic data and relevant
medical information directly to the individual. Be it a genome
sequence on a tablet or the results of a biosensor for blood pres-
sure or another physiologic metric displayed on a smartphone,
the digital convergence with biology will definitively anchor the
individual as a source of salient data, the conduit of information
flow, and a—if not the—principal driver of medicine in the future.
The Human GIS
Perhaps the most commonly used geographic information
systems (GIS) are Google maps, which provide a layered
approach to data visualization, such as viewing a location via
satellite overlaid with street names, landmarks, and real-time
traffic data. This GIS exemplifies the concept of gathering and
transforming large bodies of data to provide exquisite temporal
and location information. With the multiple virtual views, it gives
one the sense of physically being on site. Although Google has
digitized and thus created a GIS for the Earth, it is now possible
to digitize a human being. As shown in Figure 1, there are multi-
ple layers of data that can now be obtained for any individual.
This includes data from biosensors, scanners, electronic medi-
cal records, social media, and the various omics that include
Cell 157, March 27, 2014 ª2014 Elsevier Inc. 241
countless hours of
context to the digit
DNA sequence, 2 T
transcriptome, and
first human GIS ca
feat and yielded k
individual. But, it ca
at this juncture. With
drop substantially,
automating the anal
ogy can readily be
providing meaningfu
The Omic Tools
Whole-Genome an
Perhaps the greates
domain has been t
sequence a human g
the pace of Moore’sFigure 1. Geographic Information System of a Human Being
70. • inter-omics correlation network 의 분석을 통해서 환자들을 몇가지 cluster로 분류
• 가장 큰 cluster (246 Vertices, 1645 Edges): Cardiometaboic Health
• four most connected clinical analyses: C-peptide, insulin, MOMA-IR, triglycerides
• four most-connected proteins: leptin, C-reactive protein, FGF21, INHBC
atureAmerica,Inc.,partofSpringerNature.Allrightsreserved.
A RT I C L E S
The largest community (246 V; 1,645 E) contains many clinical
analytes associated with cardiometabolic health, such as C-peptide,
triglycerides, insulin, homeostatic risk assessment–insulin resistance
(HOMA-IR), fasting glucose, high-density lipid (HDL) cholesterol,
and small low-density lipid (LDL) particle number (Fig. 3). The four
most-connected clinical analytes by degree (the number of edges
connecting a particular analyte) were C-peptide (degree 99), insulin
(88), HOMA-IR (88), and triglycerides (75). The four most-connected
proteins measured using targeted (i.e., selected reaction monitoring
analysis) mass spectrometry or Olink proximity extension assays
by degree are leptin (18), C-reactive protein (15), fibroblast growth
factor 21 (FGF21) (14), and inhibin beta C chain (INHBC) (10).
Leptin and C-reactive protein are indicators for cardiovascular
risk14,15. FGF21 is positively correlated with the clinical analytes
( = −0.41; padj = 2.1 × 10−3). Hypothyroidism has long been recog-
nized clinically as a cause of elevated cholesterol values19.
A community formed around plasma serotonin (18 V; 25 E) contain-
ing 12 proteins listed in Supplementary Table 6, for which the most
significant enrichment identified in a STRING ontology analysis20 was
platelet activation (padj = 1.7 × 10−3) (Fig. 4b). Serotonin is known to
induce platelet aggregation21; accordingly, selective serotonin reuptake
inhibitors (SSRIs) may protect against myocardial infarction22.
We identified several communities containing microbiome taxa,
suggesting that there are specific microbiome–analyte relationships.
Hydrocinnamate, l-urobilin, and 5-hydroxyhexanoate clustered with
the bacterial class Mollicutes and family Christensenellaceae (8 V;
8 E). Another community emerged around the Verrucomicrobiaceae
and Desulfovibrionaceae families and p-cresol-sulfate (7 V; 6 E). The
a
c
d
b
e
Figure 4 Cholesterol, serotonin, -diversity, IBD, and bladder cancer communities. (a) Cholesterol community. (b) Serotonin community. (c) -diversity
community. (d) The polygenic score for inflammatory bowel disease is negatively correlated with cystine. (e) The polygenic score for bladder cancer is
positively correlated with 5-acetylamino-6-formylamino-3-methyluracil (AFMU).
Cholesterol, serotonin, diversity, IBD, and bladder cancer communities. (a) Cholesterol community. (b) Serotonin community. (c)
-diversity community. (d) The polygenic score for inflammatory bowel disease is negatively correlated with cystine. (e) The
polygenic score for bladder cancer is positively correlated with 5-acetylamino-6-formylamino-3-methyluracil (AFMU).
71. 017NatureAmerica,Inc.,partofSpringerNature.Allrightsreserved.
identified with elevated fasting glucose or HbA1c at baseline (pre-
diabetes), the coach made recommendations based on the Diabetes
Prevention Program36, customized for each person’s lifestyle. These
individual recommendations typically fell into one of several major
factors (fasting insulin and HOMA-IR), and inflammation (IL-8 and
TNF-alpha). Lipoprotein fractionation, performed by both labora-
tory companies, produced significant but discordant results for LDL
particle number. We observed significant improvements in fasting
Table 1 Longitudinal analysis of clinical changes by round
Clinical laboratory test Changes in labs in participants out-of-range at baseline
Health area Name N per round P-value
Nutrition Vitamin D 95 +7.2 ng/mL/round 7.1 × 10−25
Nutrition Mercury 81 −0.002 mcg/g/round 8.9 × 10−9
Diabetes HbA1c 52 −0.085%/round 9.2 × 10−6
Cardiovascular LDL particle number (Quest) 30 +130 nmol/L/round 9.3 × 10−5
Nutrition Methylmalonic acid (Genova) 3 −0.49 mmol/mol creatinine/round 2.1 × 10−4
Cardiovascular LDL pattern (A or B) 28 −0.16 /round 4.8 × 10−4
Inflammation Interleukin-8 10 −6.1 pg/mL/round 5.9 × 10−4
Cardiovascular Total cholesterol (Quest) 48 −6.4 mg/dL/round 7.2 × 10−4
Cardiovascular LDL cholesterol 57 −4.8 mg/dL/round 8.8 × 10−4
Cardiovascular LDL particle number (Genova) 70 −69 nmol/L/round 1.2 × 10−3
Cardiovascular Small LDL particle number (Genova) 73 −56 nmol/L/round 3.5 × 10−3
Diabetes Fasting glucose (Quest) 45 −1.9 mg/dL/round 8.2 × 10−3
Cardiovascular Total cholesterol (Genova) 43 −5.4 mg/dL/round 1.2 × 10−2
Diabetes Insulin 16 −2.3 IU/mL/round 1.5 × 10−2
Inflammation TNF-alpha 4 −6.6 pg/mL/round 1.8 × 10−2
Diabetes HOMA-IR 19 −0.56 /round 2.0 × 10−2
Cardiovascular HDL cholesterol 5 +4.5 mg/dL/round 2.2 × 10−2
Nutrition Methylmalonic acid (Quest) 7 −42 nmol/L/round 5.2 × 10−2
Cardiovascular Triglycerides (Genova) 14 −18 mg/dL/round 1.4 × 10−1
Diabetes Fasting glucose (Genova) 47 −0.98 mg/dL/round 1.5 × 10−1
Nutrition Arachidonic acid 35 +0.24 wt%/round 1.9 × 10−1
Inflammation hs-CRP 51 −0.47 mcg/mL/round 2.1 × 10−1
Cardiovascular Triglycerides (Quest) 17 −14 mg/dL/round 2.4 × 10−1
Nutrition Glutathione 6 +11 micromol/L/round 2.5 × 10−1
Nutrition Zinc 4 −0.82 mcg/g/round 3.0 × 10−1
Nutrition Ferritin 10 −14 ng/mL/round 3.1 × 10−1
Inflammation Interleukin-6 4 −1.1 pg/mL/round 3.8 × 10−1
Cardiovascular HDL large particle number 8 +210 nmol/L/round 4.9 × 10−1
Nutrition Copper 10 +0.006 mcg/g/round 6.0 × 10−1
Nutrition Selenium 6 +0.035 mcg/g/round 6.2 × 10−1
Cardiovascular Medium LDL particle number 20 +2.8 nmol/L/round 8.5 × 10−1
Cardiovascular Small LDL particle number (Quest) 14 −2.3 nmol/L/round 8.8 × 10−1
Nutrition Manganese 0 N/A N/A
Nutrition EPA 0 N/A N/A
Nutrition DHA 0 N/A N/A
Generalized estimating equations (GEE) were used to calculate average changes in clinical laboratory tests over time, for those analytes that were actively coached on. The ‘ per
round’ column is the average change in the population for that analyte by round adjusted for age, sex, and self-reported ancestry. ‘Out-of-range at baseline’ indicates the average
change using only those participants who were out-of-range for that analyte at the beginning of the study. Rows in boldface indicate statistically significant improvement, while
the italicized row indicates statistically significant worsening. N/A values are present where no participants were out-of-range at baseline. For example, the average improvement
in vitamin D for the 95 participants that began the study out-of-range was +7.2 ng/mL per round. Several analytes are measured by both Quest and Genova; with the exception of
LDL particle number, the direction of effect for significantly changed analytes was concordant across the two laboratories. An independence working correlation structure was used
in the GEE. See Supplementary Table 10 for the complete results.
• 수치가 정상 범위를 벗어나면 코치가 개입하여, 해당 수치를 개선할 수 있는 라이프스타일의 변화 유도
• 예를 들어, 공복혈당 혹은 HbA1c 의 증가: 코치가 당뇨 예방 프로그램(DPP)을 권고
• 몇개의 major category로 나눠짐
• diet, exercise, stress management, dietary supplements, physician referral
• 이를 통해서 가장 크게 개선 효과가 있었던 수치들
• vitamin D, mercury, HbA1c
• 전반적으로 콜레스테롤 관련 수치나, 당뇨 위험 관련 수치, 염증 수치 등에 개선이 있었음
72. • 버릴리(구글)의 베이스라인 프로젝트
• 건강과 질병을 새롭게 정의하기 위한 프로젝트
• 4년 동안 만 명의 개인의 건강 상태를 면밀하게 추적하여 데이터를 축적
• 심박수와 수면패턴 및 유전 정보, 감정 상태, 진료기록, 가족력, 소변/타액/혈액 검사 등
73. • 버릴리(구글)의 베이스라인 프로젝트
• 건강과 질병을 새롭게 정의하기 위한 프로젝트
• 4년 동안 만 명의 개인의 건강 상태를 면밀하게 추적하여 데이터를 축적
• 심박수와 수면패턴 및 유전 정보, 감정 상태, 진료기록, 가족력, 소변/타액/혈액 검사 등
74. • 버릴리의 ‘Study Watch’
• 2017년 4월 공개한 베이스라인 스터디 용 스마트워치
• 심전도, 심박수, EDA(Electrodermal Activity), 관성움직임(inertial movement) 등 측정
• 장기간 추적연구를 위한 특징들: 배터리 수명(일주일), 데이터 저장 공간, 동기화 (일주일 한 번)
75.
76. • Linda Avey의 Precise.ly
• 23andMe의 공동창업자였던 Linda Avey가 2009년 회사를 떠난 이후, 2011년 창업
• ‘We Are Curious’ 라는 이름에서 최근에 Precise.ly로 회사 이름 변경
77.
78. • Linda Avey의 Precise.ly
• Genotype + Phenotype + Microbiome + environment 모두 결합하여 의학적인 insight
• Genotype: Helix의 플랫폼에서 WES 을 통하여 분석
• Phenotype: 웨어러블, IoT 기기를 이용
79. • ‘Modern diseases’를 주로 타게팅 하겠다고 언급하고 있음
• 예를 들어, autism spectrum syndrome을 다차원적 데이터를 기반으로 분류할 수 있을까?
• Helix 플랫폼을 통해서 먼저 Chronic Fatigue 에 대한 앱을 먼저 출시하고,
• 향후 autism, PD 등에 대한 앱을 출시할 예정이라고 함.
80. iCarbonX
•중국 BGI의 대표였던 준왕이 창업
•'모든 데이터를 측정'하고 이를 정밀 의료에 활용할 계획
•데이터를 측정할 수 있는 역량을 가진 회사에 투자 및 인수
•SomaLogic, HealthTell, PatientsLikMe
•향후 5년 동안 100만명-1000만 명의 데이터 모을 계획
•이 데이터의 분석은 인공지능으로
81. 현재 Arivale, Baseline Project,
Precisely, iCarbon-X 가
모두 잘 되고 있지는 않으나,
이러한 변화의 초창기 시도 정도로 해석 가능
93. ORIGINAL ARTICLE
Watson for Oncology and breast cancer treatment
recommendations: agreement with an expert
multidisciplinary tumor board
S. P. Somashekhar1*, M.-J. Sepu´lveda2
, S. Puglielli3
, A. D. Norden3
, E. H. Shortliffe4
, C. Rohit Kumar1
,
A. Rauthan1
, N. Arun Kumar1
, P. Patil1
, K. Rhee3
& Y. Ramya1
1
Manipal Comprehensive Cancer Centre, Manipal Hospital, Bangalore, India; 2
IBM Research (Retired), Yorktown Heights; 3
Watson Health, IBM Corporation,
Cambridge; 4
Department of Surgical Oncology, College of Health Solutions, Arizona State University, Phoenix, USA
*Correspondence to: Prof. Sampige Prasannakumar Somashekhar, Manipal Comprehensive Cancer Centre, Manipal Hospital, Old Airport Road, Bangalore 560017, Karnataka,
India. Tel: þ91-9845712012; Fax: þ91-80-2502-3759; E-mail: somashekhar.sp@manipalhospitals.com
Background: Breast cancer oncologists are challenged to personalize care with rapidly changing scientific evidence, drug
approvals, and treatment guidelines. Artificial intelligence (AI) clinical decision-support systems (CDSSs) have the potential to
help address this challenge. We report here the results of examining the level of agreement (concordance) between treatment
recommendations made by the AI CDSS Watson for Oncology (WFO) and a multidisciplinary tumor board for breast cancer.
Patients and methods: Treatment recommendations were provided for 638 breast cancers between 2014 and 2016 at the
Manipal Comprehensive Cancer Center, Bengaluru, India. WFO provided treatment recommendations for the identical cases in
2016. A blinded second review was carried out by the center’s tumor board in 2016 for all cases in which there was not
agreement, to account for treatments and guidelines not available before 2016. Treatment recommendations were considered
concordant if the tumor board recommendations were designated ‘recommended’ or ‘for consideration’ by WFO.
Results: Treatment concordance between WFO and the multidisciplinary tumor board occurred in 93% of breast cancer cases.
Subgroup analysis found that patients with stage I or IV disease were less likely to be concordant than patients with stage II or III
disease. Increasing age was found to have a major impact on concordance. Concordance declined significantly (P 0.02;
P < 0.001) in all age groups compared with patients <45 years of age, except for the age group 55–64 years. Receptor status
was not found to affect concordance.
Conclusion: Treatment recommendations made by WFO and the tumor board were highly concordant for breast cancer cases
examined. Breast cancer stage and patient age had significant influence on concordance, while receptor status alone did not.
This study demonstrates that the AI clinical decision-support system WFO may be a helpful tool for breast cancer treatment
decision making, especially at centers where expert breast cancer resources are limited.
Key words: Watson for Oncology, artificial intelligence, cognitive clinical decision-support systems, breast cancer,
concordance, multidisciplinary tumor board
Introduction
Oncologists who treat breast cancer are challenged by a large and
rapidly expanding knowledge base [1, 2]. As of October 2017, for
example, there were 69 FDA-approved drugs for the treatment of
breast cancer, not including combination treatment regimens
[3]. The growth of massive genetic and clinical databases, along
with computing systems to exploit them, will accelerate the speed
of breast cancer treatment advances and shorten the cycle time
for changes to breast cancer treatment guidelines [4, 5]. In add-
ition, these information management challenges in cancer care
are occurring in a practice environment where there is little time
available for tracking and accessing relevant information at the
point of care [6]. For example, a study that surveyed 1117 oncolo-
gists reported that on average 4.6 h per week were spent keeping
VC The Author(s) 2018. Published by Oxford University Press on behalf of the European Society for Medical Oncology.
All rights reserved. For permissions, please email: journals.permissions@oup.com.
Annals of Oncology 29: 418–423, 2018
doi:10.1093/annonc/mdx781
Published online 9 January 2018
Downloaded from https://academic.oup.com/annonc/article-abstract/29/2/418/4781689
by guest
WFO는 현재 정확성, 효용성에 대한
근거가 부족하지만,
10년 뒤에도 그러할까?
94. ORIGINAL ARTICLE
Watson for Oncology and breast cancer treatment
recommendations: agreement with an expert
multidisciplinary tumor board
S. P. Somashekhar1*, M.-J. Sepu´lveda2
, S. Puglielli3
, A. D. Norden3
, E. H. Shortliffe4
, C. Rohit Kumar1
,
A. Rauthan1
, N. Arun Kumar1
, P. Patil1
, K. Rhee3
& Y. Ramya1
1
Manipal Comprehensive Cancer Centre, Manipal Hospital, Bangalore, India; 2
IBM Research (Retired), Yorktown Heights; 3
Watson Health, IBM Corporation,
Cambridge; 4
Department of Surgical Oncology, College of Health Solutions, Arizona State University, Phoenix, USA
*Correspondence to: Prof. Sampige Prasannakumar Somashekhar, Manipal Comprehensive Cancer Centre, Manipal Hospital, Old Airport Road, Bangalore 560017, Karnataka,
India. Tel: þ91-9845712012; Fax: þ91-80-2502-3759; E-mail: somashekhar.sp@manipalhospitals.com
Background: Breast cancer oncologists are challenged to personalize care with rapidly changing scientific evidence, drug
approvals, and treatment guidelines. Artificial intelligence (AI) clinical decision-support systems (CDSSs) have the potential to
help address this challenge. We report here the results of examining the level of agreement (concordance) between treatment
recommendations made by the AI CDSS Watson for Oncology (WFO) and a multidisciplinary tumor board for breast cancer.
Patients and methods: Treatment recommendations were provided for 638 breast cancers between 2014 and 2016 at the
Manipal Comprehensive Cancer Center, Bengaluru, India. WFO provided treatment recommendations for the identical cases in
2016. A blinded second review was carried out by the center’s tumor board in 2016 for all cases in which there was not
agreement, to account for treatments and guidelines not available before 2016. Treatment recommendations were considered
concordant if the tumor board recommendations were designated ‘recommended’ or ‘for consideration’ by WFO.
Results: Treatment concordance between WFO and the multidisciplinary tumor board occurred in 93% of breast cancer cases.
Subgroup analysis found that patients with stage I or IV disease were less likely to be concordant than patients with stage II or III
disease. Increasing age was found to have a major impact on concordance. Concordance declined significantly (P 0.02;
P < 0.001) in all age groups compared with patients <45 years of age, except for the age group 55–64 years. Receptor status
was not found to affect concordance.
Conclusion: Treatment recommendations made by WFO and the tumor board were highly concordant for breast cancer cases
examined. Breast cancer stage and patient age had significant influence on concordance, while receptor status alone did not.
This study demonstrates that the AI clinical decision-support system WFO may be a helpful tool for breast cancer treatment
decision making, especially at centers where expert breast cancer resources are limited.
Key words: Watson for Oncology, artificial intelligence, cognitive clinical decision-support systems, breast cancer,
concordance, multidisciplinary tumor board
Introduction
Oncologists who treat breast cancer are challenged by a large and
rapidly expanding knowledge base [1, 2]. As of October 2017, for
example, there were 69 FDA-approved drugs for the treatment of
breast cancer, not including combination treatment regimens
[3]. The growth of massive genetic and clinical databases, along
with computing systems to exploit them, will accelerate the speed
of breast cancer treatment advances and shorten the cycle time
for changes to breast cancer treatment guidelines [4, 5]. In add-
ition, these information management challenges in cancer care
are occurring in a practice environment where there is little time
available for tracking and accessing relevant information at the
point of care [6]. For example, a study that surveyed 1117 oncolo-
gists reported that on average 4.6 h per week were spent keeping
VC The Author(s) 2018. Published by Oxford University Press on behalf of the European Society for Medical Oncology.
All rights reserved. For permissions, please email: journals.permissions@oup.com.
Annals of Oncology 29: 418–423, 2018
doi:10.1093/annonc/mdx781
Published online 9 January 2018
Downloaded from https://academic.oup.com/annonc/article-abstract/29/2/418/4781689
by guest
Table 2. MMDT and WFO recommendations after the initial and blinded second reviews
Review of breast cancer cases (N 5 638) Concordant cases, n (%) Non-concordant cases, n (%)
Recommended For consideration Total Not recommended Not available Total
Initial review (T1MMDT versus T2WFO) 296 (46) 167 (26) 463 (73) 137 (21) 38 (6) 175 (27)
Second review (T2MMDT versus T2WFO) 397 (62) 194 (30) 591 (93) 36 (5) 11 (2) 47 (7)
T1MMDT, original MMDT recommendation from 2014 to 2016; T2WFO, WFO advisor treatment recommendation in 2016; T2MMDT, MMDT treatment recom-
mendation in 2016; MMDT, Manipal multidisciplinary tumor board; WFO, Watson for Oncology.
31%
18%
1% 2% 33%
5% 31%
6%
0% 10% 20%
Not available Not recommended RecommendedFor consideration
30% 40% 50% 60% 70% 80% 90% 100%
8% 25% 61%
64%
64%
29% 51%
62%
Concordance, 93%
Concordance, 80%
Concordance, 97%
Concordance, 95%
Concordance, 86%
2%
2%
Overall
(n=638)
Stage I
(n=61)
Stage II
(n=262)
Stage III
(n=191)
Stage IV
(n=124)
5%
Figure 1. Treatment concordance between WFO and the MMDT overall and by stage. MMDT, Manipal multidisciplinary tumor board; WFO,
Watson for Oncology.
5%Non-metastatic
HR(+)HER2/neu(+)Triple(–)
Metastatic
Non-metastatic
Metastatic
Non-metastatic
Metastatic
10%
1%
2%
1% 5% 20%
20%10%
0%
Not applicable Not recommended For consideration Recommended
20% 40% 60% 80% 100%
5%
74%
65%
34% 64%
5% 38% 56%
15% 20% 55%
36% 59%
Concordance, 95%
Concordance, 75%
Concordance, 94%
Concordance, 98%
Concordance, 94%
Concordance, 85%
Figure 2. Treatment concordance between WFO and the MMDT by stage and receptor status. HER2/neu, human epidermal growth factor
receptor 2; HR, hormone receptor; MMDT, Manipal multidisciplinary tumor board; WFO, Watson for Oncology.
Annals of Oncology Original article
WFO는 현재 정확성, 효용성에 대한
근거가 부족하지만,
10년 뒤에도 그러할까?
96. •“향후 10년 동안 첫번째 cardiovascular event 가 올 것인가” 예측
•전향적 코호트 스터디: 영국 환자 378,256 명
•일상적 의료 데이터를 바탕으로 기계학습으로 질병을 예측하는 첫번째 대규모 스터디
•기존의 ACC/AHA 가이드라인과 4가지 기계학습 알고리즘의 정확도를 비교
•Random forest; Logistic regression; Gradient bossting; Neural network
97. •2018년 1월 구글이 전자의무기록(EMR)을 분석하여, 환자 치료 결과를 예측하는 인공지능 발표
•환자가 입원 중에 사망할 것인지
•장기간 입원할 것인지
•퇴원 후에 30일 내에 재입원할 것인지
•퇴원 시의 진단명
•이번 연구의 특징: 확장성
•과거 다른 연구와 달리 EMR의 일부 데이터를 pre-processing 하지 않고,
•전체 EMR 를 통채로 모두 분석하였음: UCSF, UCM (시카고 대학병원)
•특히, 비정형 데이터인 의사의 진료 노트도 분석
98. ARTICLE OPEN
Scalable and accurate deep learning with electronic health
records
Alvin Rajkomar 1,2
, Eyal Oren1
, Kai Chen1
, Andrew M. Dai1
, Nissan Hajaj1
, Michaela Hardt1
, Peter J. Liu1
, Xiaobing Liu1
, Jake Marcus1
,
Mimi Sun1
, Patrik Sundberg1
, Hector Yee1
, Kun Zhang1
, Yi Zhang1
, Gerardo Flores1
, Gavin E. Duggan1
, Jamie Irvine1
, Quoc Le1
,
Kurt Litsch1
, Alexander Mossin1
, Justin Tansuwan1
, De Wang1
, James Wexler1
, Jimbo Wilson1
, Dana Ludwig2
, Samuel L. Volchenboum3
,
Katherine Chou1
, Michael Pearson1
, Srinivasan Madabushi1
, Nigam H. Shah4
, Atul J. Butte2
, Michael D. Howell1
, Claire Cui1
,
Greg S. Corrado1
and Jeffrey Dean1
Predictive modeling with electronic health record (EHR) data is anticipated to drive personalized medicine and improve healthcare
quality. Constructing predictive statistical models typically requires extraction of curated predictor variables from normalized EHR
data, a labor-intensive process that discards the vast majority of information in each patient’s record. We propose a representation
of patients’ entire raw EHR records based on the Fast Healthcare Interoperability Resources (FHIR) format. We demonstrate that
deep learning methods using this representation are capable of accurately predicting multiple medical events from multiple
centers without site-specific data harmonization. We validated our approach using de-identified EHR data from two US academic
medical centers with 216,221 adult patients hospitalized for at least 24 h. In the sequential format we propose, this volume of EHR
data unrolled into a total of 46,864,534,945 data points, including clinical notes. Deep learning models achieved high accuracy for
tasks such as predicting: in-hospital mortality (area under the receiver operator curve [AUROC] across sites 0.93–0.94), 30-day
unplanned readmission (AUROC 0.75–0.76), prolonged length of stay (AUROC 0.85–0.86), and all of a patient’s final discharge
diagnoses (frequency-weighted AUROC 0.90). These models outperformed traditional, clinically-used predictive models in all cases.
We believe that this approach can be used to create accurate and scalable predictions for a variety of clinical scenarios. In a case
study of a particular prediction, we demonstrate that neural networks can be used to identify relevant information from the
patient’s chart.
npj Digital Medicine (2018)1:18 ; doi:10.1038/s41746-018-0029-1
INTRODUCTION
The promise of digital medicine stems in part from the hope that,
by digitizing health data, we might more easily leverage computer
information systems to understand and improve care. In fact,
routinely collected patient healthcare data are now approaching
the genomic scale in volume and complexity.1
Unfortunately,
most of this information is not yet used in the sorts of predictive
statistical models clinicians might use to improve care delivery. It
is widely suspected that use of such efforts, if successful, could
provide major benefits not only for patient safety and quality but
also in reducing healthcare costs.2–6
In spite of the richness and potential of available data, scaling
the development of predictive models is difficult because, for
traditional predictive modeling techniques, each outcome to be
predicted requires the creation of a custom dataset with specific
variables.7
It is widely held that 80% of the effort in an analytic
model is preprocessing, merging, customizing, and cleaning
nurses, and other providers are included. Traditional modeling
approaches have dealt with this complexity simply by choosing a
very limited number of commonly collected variables to consider.7
This is problematic because the resulting models may produce
imprecise predictions: false-positive predictions can overwhelm
physicians, nurses, and other providers with false alarms and
concomitant alert fatigue,10
which the Joint Commission identified
as a national patient safety priority in 2014.11
False-negative
predictions can miss significant numbers of clinically important
events, leading to poor clinical outcomes.11,12
Incorporating the
entire EHR, including clinicians’ free-text notes, offers some hope
of overcoming these shortcomings but is unwieldy for most
predictive modeling techniques.
Recent developments in deep learning and artificial neural
networks may allow us to address many of these challenges and
unlock the information in the EHR. Deep learning emerged as the
preferred machine learning approach in machine perception
www.nature.com/npjdigitalmed
•2018년 1월 구글이 전자의무기록(EMR)을 분석하여, 환자 치료 결과를 예측하는 인공지능 발표
•환자가 입원 중에 사망할 것인지
•장기간 입원할 것인지
•퇴원 후에 30일 내에 재입원할 것인지
•퇴원 시의 진단명
•이번 연구의 특징: 확장성
•과거 다른 연구와 달리 EMR의 일부 데이터를 pre-processing 하지 않고,
•전체 EMR 를 통채로 모두 분석하였음: UCSF, UCM (시카고 대학병원)
•특히, 비정형 데이터인 의사의 진료 노트도 분석
99. • 복잡한 의료 데이터의 분석 및 insight 도출
• 영상 의료/병리 데이터의 분석/판독
• 연속 데이터의 모니터링 및 예방/예측
의료 인공지능의 세 유형
108. Fig 1. What can consumer wearables do? Heart rate can be measured with an oximeter built into a ring [3], muscle activity with an electromyographi
sensor embedded into clothing [4], stress with an electodermal sensor incorporated into a wristband [5], and physical activity or sleep patterns via an
accelerometer in a watch [6,7]. In addition, a female’s most fertile period can be identified with detailed body temperature tracking [8], while levels of me
attention can be monitored with a small number of non-gelled electroencephalogram (EEG) electrodes [9]. Levels of social interaction (also known to a
PLOS Medicine 2016
109. • 복잡한 의료 데이터의 분석 및 insight 도출
• 영상 의료/병리 데이터의 분석/판독
• 연속 데이터의 모니터링 및 예방/예측
인공지능의 의료 활용
112. Sugar.IQ
사용자의 음식 섭취와 그에 따른 혈당 변화,
인슐린 주입 등의 과거 기록 기반
식후 사용자의 혈당이 어떻게 변화할지
Watson 이 예측
113.
114. An Algorithm Based on Deep Learning for Predicting In-Hospital
Cardiac Arrest
Joon-myoung Kwon, MD;* Youngnam Lee, MS;* Yeha Lee, PhD; Seungwoo Lee, BS; Jinsik Park, MD, PhD
Background-—In-hospital cardiac arrest is a major burden to public health, which affects patient safety. Although traditional track-
and-trigger systems are used to predict cardiac arrest early, they have limitations, with low sensitivity and high false-alarm rates.
We propose a deep learning–based early warning system that shows higher performance than the existing track-and-trigger
systems.
Methods and Results-—This retrospective cohort study reviewed patients who were admitted to 2 hospitals from June 2010 to July
2017. A total of 52 131 patients were included. Specifically, a recurrent neural network was trained using data from June 2010 to
January 2017. The result was tested using the data from February to July 2017. The primary outcome was cardiac arrest, and the
secondary outcome was death without attempted resuscitation. As comparative measures, we used the area under the receiver
operating characteristic curve (AUROC), the area under the precision–recall curve (AUPRC), and the net reclassification index.
Furthermore, we evaluated sensitivity while varying the number of alarms. The deep learning–based early warning system (AUROC:
0.850; AUPRC: 0.044) significantly outperformed a modified early warning score (AUROC: 0.603; AUPRC: 0.003), a random forest
algorithm (AUROC: 0.780; AUPRC: 0.014), and logistic regression (AUROC: 0.613; AUPRC: 0.007). Furthermore, the deep learning–
based early warning system reduced the number of alarms by 82.2%, 13.5%, and 42.1% compared with the modified early warning
system, random forest, and logistic regression, respectively, at the same sensitivity.
Conclusions-—An algorithm based on deep learning had high sensitivity and a low false-alarm rate for detection of patients with
cardiac arrest in the multicenter study. (J Am Heart Assoc. 2018;7:e008678. DOI: 10.1161/JAHA.118.008678.)
Key Words: artificial intelligence • cardiac arrest • deep learning • machine learning • rapid response system • resuscitation
In-hospital cardiac arrest is a major burden to public health,
which affects patient safety.1–3
More than a half of cardiac
arrests result from respiratory failure or hypovolemic shock,
and 80% of patients with cardiac arrest show signs of
deterioration in the 8 hours before cardiac arrest.4–9
However,
209 000 in-hospital cardiac arrests occur in the United States
each year, and the survival discharge rate for patients with
cardiac arrest is <20% worldwide.10,11
Rapid response systems
(RRSs) have been introduced in many hospitals to detect
cardiac arrest using the track-and-trigger system (TTS).12,13
Two types of TTS are used in RRSs. For the single-parameter
TTS (SPTTS), cardiac arrest is predicted if any single vital sign
(eg, heart rate [HR], blood pressure) is out of the normal
range.14
The aggregated weighted TTS calculates a weighted
score for each vital sign and then finds patients with cardiac
arrest based on the sum of these scores.15
The modified early
warning score (MEWS) is one of the most widely used
approaches among all aggregated weighted TTSs (Table 1)16
;
however, traditional TTSs including MEWS have limitations, with
low sensitivity or high false-alarm rates.14,15,17
Sensitivity and
false-alarm rate interact: Increased sensitivity creates higher
false-alarm rates and vice versa.
Current RRSs suffer from low sensitivity or a high false-
alarm rate. An RRS was used for only 30% of patients before
unplanned intensive care unit admission and was not used for
22.8% of patients, even if they met the criteria.18,19
From the Departments of Emergency Medicine (J.-m.K.) and Cardiology (J.P.), Mediplex Sejong Hospital, Incheon, Korea; VUNO, Seoul, Korea (Youngnam L., Yeha L.,
S.L.).
*Dr Kwon and Mr Youngnam Lee contributed equally to this study.
Correspondence to: Joon-myoung Kwon, MD, Department of Emergency medicine, Mediplex Sejong Hospital, 20, Gyeyangmunhwa-ro, Gyeyang-gu, Incheon 21080,
Korea. E-mail: kwonjm@sejongh.co.kr
Received January 18, 2018; accepted May 31, 2018.
ª 2018 The Authors. Published on behalf of the American Heart Association, Inc., by Wiley. This is an open access article under the terms of the Creative Commons
Attribution-NonCommercial License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited and is not used for
commercial purposes.
DOI: 10.1161/JAHA.118.008678 Journal of the American Heart Association 1
ORIGINAL RESEARCH
byguestonJune28,2018http://jaha.ahajournals.org/Downloadedfrom
116. •대학병원 신속 대응팀에서 처리 가능한 알림 수 (A, B 지점) 에서 더 큰 정확도 차이를 보임
•A: DEWS 33.0%, MEWS 0.3%
•B: DEWS 42.7%, MEWS 4.0%
(source: VUNO)
APPH(Alarms Per Patients Per Hour)
(source: VUNO)
Less False Alarm