International Journal of Scientific Research in Engineering and Management (IJSREM)

Volume: 08 Issue: 11 | Nov - 2024 SJIF 2023: 8.176 ISSN: 2582-3930

Enhancing Healthcare with Personalized Recommendations:

A Machine Learning Approach to Symptom-Driven Diagnosis
Sanjay K. S.1, Yashaswini Carel Rodrigues2, Yash Rodrigues3
1Third Year Student, Department of Information Science and Engineering, Global Academy of Technology
2Third Year Student, Department of Information Science and Engineering, SJB Institute of Technology
3Fourth Year MBBS Student, Bangalore Medical College and Research Institute

---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - The development of high-throughput, data- harmonized, and made available for analysis in order to
intensive biomedical research assays and technologies, such identify patterns that would normally not be identified if the
as DNA sequencing, imaging protocols, and wireless health different data points were analyzed independently; and 4.
monitoring devices, has created a need for researchers to Artificial Intelligence (AI; which we consider here to include
develop strategies for analyzing, integrating and interpreting algorithms based machine learning, deep learning, neural
the massive amounts of data they generate. Although a wide network constructs and a wide variety of related
variety of statistical methods have been designed to techniques[1]), which can be used to find relevant patterns in
accommodate the ‘big data’ produced by these assays, massive data sets.
experiences with the use of artificial intelligence (AI)
techniques suggest that they might be particularly appropriate.
In addition, the application of data-intensive biomedical
technologies in research studies has revealed that humans vary
widely at the genetic, biochemical, physiological, exposure
and behavioral levels, especially with respect to disease
processes and treatment responsiveness. This suggests that
there is often a need to tailor, or ‘personalize,’ medicines to
the nuanced and often unique features possessed by individual
patients. Given how important data-intensive assays are to
revealing appropriate intervention targets and strategies for
personalizing medicines, AI can play an important role in the
development of personalized medicines at all relevant phases
of the clinical development and implementation of new
personalized health products, from finding appropriate
intervention targets to testing them for their utility. We
describe many areas where AI can play a role in the Fig 1: Four emerging complementary themes in biomedical
development of personalized medicines, and argue that AI’s science: personalized medicine, emerging data-intensive
ability to advance personalized medicine will depend critically technologies, big data and information technologies (IT)
on the refinement of relevant assays and ways of storing, infrastructure and artificial intelligence (AI). These
aggregating, accessing and ultimately integrating the data they technologies can be fuel, and be fueled by, AI in all phases of
produce. We also point out the limitations of many AI the development (T0-T4) of personalized medicines.
techniques, as well as consider areas for further research. These four themes are highly interrelated in that, e.g.,
personalizing a medicine or tailoring an intervention to a
Key Words: personalized healthcare, machine learning, patient requires a very deep understanding of that patient’s
disease prediction, medical recommendation system, symptom condition and circumstances, and this requires the extensive
analysis, data privacy. use of sophisticated assays that generate massive amounts of
data, such as DNA sequencing or an imaging protocol.
Essentially, the data produced by these assays needs to be
1.INTRODUCTION organized so that analyses can be pursued to identify features
that the patient possesses that may indicate the optimal
Modern biomedical science is guided, if not intervention. Research associated with each of these themes is
dominated, by many interrelated themes. Four of the most often pursued independently of the others because of the very
prominent and important of these themes are (see Figure 1): 1. specialized expertise required. For example, there are scientific
Personalized medicine, or the belief that health interventions journals devoted to Personal Medicine (e.g., ‘Personalized
need to be tailored to the nuanced and often unique genetic, Medicine,’ ‘Journal of Personalized Medicine’), emerging
biochemical, physiological, exposure and behavioral features assays (e.g., ‘Nature Biotechnology,’ ‘Nature Digital
individuals possess; 2. The exploitation of emerging data- Medicine’), big data (e.g., ‘Big Data,’ ‘Journal of Big Data,’
intensive assays, such as DNA sequencing, proteomics, ‘Gigascience’), and artificial intelligence (e.g., ‘IEEE
imaging protocols, and wireless health monitoring devices; 3. Transactions on Neural Networks and Learning Systems,’
‘Big data’ research paradigms in which massive amounts of ‘IEEE Transactions on Pattern Analysis and Machine
data, say of the type generated from emerging data-intensive Intelligence’) that publish very focused studies. However, the
biomedical assays, are aggregated from different sources, integrated use of the insights, information and strategies

© 2023, IJSREM | www.ijsrem.com | Page 1

 International Journal of Scientific Research in Engineering and Management (IJSREM)
Volume: 08 Issue: 11 | Nov - 2024 SJIF 2023: 8.176 ISSN: 2582-3930

obtained from research associated with each of these themes is The CTSA program focuses on all aspects of biomedical
necessary for creating personalized health products, such as science associated with attempts to translate very basic
drugs and interventions. biomedical insights concerning, e.g., pathogenic processes
contributing to diseases, into clinically-useful products like
Bringing together research activities associated with these four drugs or interventions for those diseases. The CTSA scheme
emerging themes is not trivial, as it will require does, however, incorporate elements of the Phase 0-Phase IV
communication and participation from researchers and clinical trials workflow or transition scheme for developing
practitioners with a wide variety of skills and expertise, drugs or health products to treat or manage diseases in the
including molecular biology, genetics, pathology, informatics, population at large. Thus, in accordance with the CTSA
computer science, statistics, clinical science, and medicine. AI scheme (left panel of Figure 2), T0 science involves very basic
will have a special role to play in this integration process if the research focusing on that could lead to the identification of a
goal is to advance personalized medicine, since it is unclear drug or intervention target and then crafting an appropriate
how relevant clinically-meaningful insights can be drawn from drug or intervention that modulates that target; T1 science
big data-generating assays that would complement or build off focuses on testing an intervention or health device in a small
the insights from experts in different domains. In this light, clinical studies to determine if it is safe and at what dosages it
there are a number of phases in the development of medicines, should be used; T2 science involves vetting the drug or device
general interventions, and other products, such as diagnostics, in a large number of individuals in a well-designed study to
prognostics, decision support tools, etc., where AI could have a assess its efficacy in the population at large; T3 science
significant impact. These different phases are emphasized in focuses on the implementation of the drug or device for use in
subsections of this chapter that describe and comment on the population, including adapting existing workflows (e.g.,
recent studies leveraging AI. This chapter does not provide an custody chains for tests or physician-patient interaction points);
exhaustive literature review of AI in medicine, however, as and T4 science involves a re-evaluation or assessment of the
there are some excellent reviews for this[2–4], but rather utility of the drug or device post-deployment and
considers the potential that AI has in developing new implementation. Each of these phases can leverage AI
medicines, health devices and products. In particular, a focus techniques if the right data and motivation is present.
on the need for greater integration across the various phases of
the development of health interventions and products could
result in very radical yet positive changes in the way medicine
is practiced. In this sense, this chapter is as much a summary of
the ways in which AI can be exploited in modern medicine as
it is a vision of the future.

2. Body of Paper
THE TRANSLATIONAL WORKFLOW
As noted, the development of interventions and health
products, as with the diagnosis and treatment of a patients,
proceeds in different phases. There are various ways of
defining and referring to these phases, however, and all of
them point to opportunities for AI to have a substantial impact
if leveraged appropriately. For example, in the context of
clinical trials to vet a new drug or intervention for treating a
disease like cancer, a common progression or workflow runs
from Phase 0 trials, which involve characterizing the
pharmacokinetic properties of a drug using what may turn out
to be non-physiologic (i.e., does not have an appreciable effect
on the body) and physiologic (i.e., does have an appreciable
effect on the body) doses of a drug in a very small number of
individuals, to Phase I trials, which involve establishing safe
and effective doses of a drug in small number of individuals, to Fig 2: A representation of the stages in the ‘translation’ of
Phase II trials, which seek to establish whether a drug is likely basic insights into clinical useful products considered in
to be efficacious in a moderately sized group of individuals, to initiatives such as the Clinical and Translational Science
Phase III trials, which attempt to establish the utility of a drug Award (CTSA) initiative overseen by the National Center on
in the population at large by studying a very large number of Advancing Translational Science of the United States National
individuals, to Phase IV trials, which evaluate the adoption, Institutes of Health (NCATS; left panel;[5]). An analogous
uptake and acceptance, as well as any evidence for adverse representation of the stages in the diagnosis and treatment of
consequences, associated with the use of the drug in the an individual patient are provided in the right hand panel.
population at large.
The actual practice of personalized medicine can be
We take a broader view of the workflows behind developing seen as involving an analogous process to the develop of drugs
new products than that reflected in the traditional clinical trial and health devices (right panel of Figure 2; see also Schork
progression. This broader view is consistent with the scheme and Nazor [6]). Thus, P0 activity involves making a diagnosis
used by, for example, the National Institutes of Health (NIH)’s or determining an individual’s risk of diseases; P1 activity
Clinical and Translational Science Award (CTSA) program.[5] involves identifying the key pathophysiologic processes, if not

© 2023, IJSREM | www.ijsrem.com | Page 2

 International Journal of Scientific Research in Engineering and Management (IJSREM)
Volume: 08 Issue: 11 | Nov - 2024 SJIF 2023: 8.176 ISSN: 2582-3930

known, that are causing (or likely to cause) a disease that pharmacological modulation is the design of appropriate
might be amenable to modulation by an appropriate studies. For example, if the relationships between different
intervention; P2 considers the identification of an appropriate potential drug targets and their effect on a molecular system or
intervention given what was identified in the P0 and P1 stages; pathway is not known, researchers may have to systematically
P3 involves testing the intervention on the relevant individual perturb each element in such a system and examine what the
undergoing the diagnosis and pathobiology assessment; and P4 effect on the system these perturbations have.[14] As one can
involves warehousing the result in appropriate databases so imagine, such studies can be tedious and laborious. However,
that the insights and information obtained on a patient can be recent studies suggest that one can use robotics and AI to
exploited in examinations of further patients or used in broader conduct such experiments and, in fact, anticipate further
data mining initiatives to find further clinically-meaningful experiments that might be called for based on the results of
patterns. initial experiments.[15] Experimental infrastructure for
pursuing AI-based experiments has also been developed, but
A couple of items about these workflows are worth noting. only for a few select settings.[16, 17]
First, as noted, although the science associated with each
component involves unique insight and expertise and provide a If an extant drug or compound is not found that could
fertile ground for collaborations with AI tools and scientists appropriately modulate a target, then creating a novel
independently of the other components, the transmission of pharmacotherapeutic (i.e., drug) often requires insights
information from one component to another – or the transitions obtained from materials science in order to make sure the
from one component to another – is of crucial importance (e.g., molecular structure of the drug produced has favorable
consider that a diagnostic would not be particularly useful if it properties inside the body. Very recent work suggests that AI
did not help a physician choose an appropriate course of can be leveraged to identify materials that may not be easily
action). Second, a goal of personalized medicine, and the identified with traditional brute-force approaches.[18, 19] In
improvement of health care generally, is to make workflows addition, AI has been used to design new structures, which
like those in Figure 2 more efficient and reliable and AI can may be relevant to crafting better interventions, whether a drug
have a substantial role to play in this broader goal. These two or mechanical device, as well as aid in the selection of
notes are emphasized in the subsection on ‘Integration and the appropriate chemical syntheses.[20–22] AI has also been used
Personalized Medicine Workflow‘ below. in studies of very basic phenomena, such as particle physics, to
probe how materials interact.[23] Interestingly, the design of
PRECLINICAL (T0) AND DIAGNOSTIC (P0) RESEARCH new drugs, for example in the context of refining the structure
The identification of targets for therapies (T0 research from the of a therapeutic protein or molecule, could also be greatly
left hand side of Figure 2) can be greatly aided by AI in quite a facilitated by AI. It has been shown that in many design
few contexts. For example, many assays used to uncover contexts in which optimization of materials and the way they
potential therapeutic targets generate massive amounts of data are put together, the use of AI can identify superior designs to
and as such often require sophisticated statistical methods to those based on legacy strategies.[20] This theme of harnessing
identify meaningful patterns. AI has been used to identify such AI to identify ways of optimizing the assembly of materials, or
patterns from, e.g., DNA sequence data and molecular the manner in which an objective function of whatever sort is
pathology imaging protocols.[7–10] For individual patients optimized given some starting materials and appropriate yet
whose comprehensive diagnosis may lay the foundation for basic principles for assembling them, was on display in the
crafting a personalized treatment or intervention plan and recent description of the system for playing the age-old game
exploit data-intensive assays (i.e., P0 in the right hand figure of of GO developed by Google’s DeepMind group.[24]
Figure 2), many AI-based tools can be leveraged. For example, Essentially, the system developed at DeepMind was not only
mining DNA sequence information obtained from an able to easily beat all human experts as well as other GO-
individual in order to make a genetic diagnosis of a disease can playing systems, but was able to do this by identifying
be greatly aided by AI-based analyses.[7] In addition, strategies and moves that were completely beyond those which
facilitating, e.g., cancer diagnoses with AI-based analyses of humans had used to play and try to master the game for
blood analytes has been shown to have great potential.[11] centuries.

If a particular therapeutic target has been identified that is If a therapeutic target has been identified, then one could
consistent with the molecular pathology underlying an potentially pursue compound screening studies to identify
individual patient’s disease, then AI-based strategies for compounds that modulate the chosen target. Such studies often
analyzing drug screening data collected to determine if any of require a particular output or phenotype (e.g., the expression
a large number of extant drugs and compounds have activity level of a target gene). In the absence of such a phenotype,
against that target have been shown to be very reliable.[12] In high-content screens, in which many different phenotypes are
addition, AI-based studies have been shown to reveal a great evaluated to see if any of the compounds, often numbering in
many insights into how drugs and compounds may impact the thousands or tens of thousands, affects any one or some
various structural and functional features of a cell.[13] Finally, subset of these phenotypes as a sign of its activity. AI
web sites with large databases like DeepChem techniques have been used to identify potential compounds
(https://deepchem.io/about.html), which leverage AI in impacting a target in this setting.[25–27]
chemistry settings, can be used specifically to identify
properties of drugs and compounds. Once a therapeutic concept has been defined, relevant drugs or
an intervention apparatus embodying the concept must be
An interesting research area of relevance to the identification manufactured at scale for distribution. The manufacture of
of pathologies underlying diseases that might be amenable to drugs and interventions of all sorts have been greatly facilitated

© 2023, IJSREM | www.ijsrem.com | Page 3

 International Journal of Scientific Research in Engineering and Management (IJSREM)
Volume: 08 Issue: 11 | Nov - 2024 SJIF 2023: 8.176 ISSN: 2582-3930

by robotics and AI.[28] In the context of personalized

Study
medicines, it may be that the manufacture of drugs and Authors Ref. Comments
Elements
interventions will require nuanced features based on patients’
profiles and therefore have to be designed and crafted in real- n et al. 287817 treatment treatments for
time as opposed to be created at scale, stored and distributed (2016)[95 44 study compromised sleep
when needed – a topic to be discussed in a later section ]
focusing on ‘Integration and the Personalized Medicine
Workflow.[29, 30]’ Zeevi et
PMID: Glycemic 800 people studied to
al.
265904 responses to develop personalized
FIRST-IN-HUMAN (T1) AND PATHOBIOLOGY (P1) (2015)[96
18 diet diets
STUDIES ]
Once a drug had been created, it must be shown to safe
through Phase I clinical trials, or studies that are often referred Smarr et
PMID: Gut microbiome
to as ‘first-in-human-studies’ (T1 in the left panel of Figure 2). al. Colonoscopy
295829 changes post
Such studies focus on safety and are typically pursued on a (2017)[97 effects
16 colonoscopy
small number of individuals in case there are side-effects. To ]
minimize the risk of exposing individuals to a new drug that Tremmell
might cause them harm, insights into the likelihood that PMID: Single individual
et al. NAD dose
individuals with a certain profile will have an adverse response 277214 dosing study + study
(2016)[98 response
are required. Studies that consider genetic factors that 79 of others
]
predispose to responses to drugs (good or bad) have revealed
many compelling and clinically-useful connections between Forsdyke PMID: Response to Decade of seasonal
genetic variants and drug responses. Discovering such (2015)[99 260551 antihypertensi variation in response
‘pharmacogenetic’ insights have been greatly enhanced ] 03 ves to Losartan
through the use of AI tools applied to very large databases with
relevant genetic and drug response information.[31, 32] O’Rawe
PMID: Deep brain 2 year study of a man
et al.
241095 stimulation with OCD treated with
The design of phase I clinical trials is an ongoing area of (2013)[10
60 (DBS) DBS
research. The fact that only a few individuals are enrolled in 0]
such trials, and a great sensitivity to the detection of the effects Li et al. PMID: Single individual
of the proposed drug or intervention are of focus, suggests that Parkinson cell
(2016)[10 271406 traced for 24 years
careful monitoring of the subjects enrolled in the trial is replacement
1] 03 post cell transplant
required. Extensions of N-of-1 and aggregated N-of-1 trial
designs could be appropriate for Phase I trials.[33] Although Bloss et a. PMID: Idiopathic Severely disabled
discussed in greater detail in the context of vetting the efficacy (2016)[10 257901 neurologic individual treated for
of a personalized medicine in the section on the 2] 60 disease sleep tremors
‘Implementation (T3) and Clinical Assessment (P3) Studies‘
below, such studies can leverage massive amounts of data and Piening et
PMID: Multiomics
AI techniques to identify patterns in a patient’s data that might al. 23 individuals
293614 weight loss
be indicative of response to the intervention (see, e.g., Serhani (2018)[10 observed over time
66 study
et al.[34]). Table 1 lists examples of studies that have focused 3]
on monitoring a single individual over time to explore how
Zalusky
they responded to a particular intervention, or how their health PMID: Folate Study of an individual
et al.
status may have changed over that time, using various data 140097 supplementati with severe dietary
(1961)[10
collection schemes. 35 on constraints
4]
Table 1.
Single patient-oriented studies leveraging intensive monitoring V.
to identify either health status changes or the effects of an PMID: Folate Study on the
Herbert
intervention. 139539 supplementati researcher himself;
(1962)[10
04 on dietary restriction
Study 5]
Authors Ref. Comments
Elements
P. H.
PMID: Folate Study on the
David et Golding
PMID: 2 individuals fecal 253328 supplementati researcher himself;
al. Diet and (2014)[10
251463 microbiome tracked 50 on dietary restriction
(2014)[93 microbiome 6]
75 for a year
]
In the context of personalized medicine studies, once an
Chen et
PMID: Individual tracked individual is found to possess a certain pathology, a need to
al. Multiomics
224242 over a year (the identify how that pathology can be corrected arises (P1 in the
(2012)[94 profiling
36 ‘Snyderome’ study) right hand panel of Figure 2). For many common chronic
]
diseases this is obvious (e.g., for someone diagnosed with
Magnuso PMID: Sleep Patient with multiple high blood pressure, providing blood pressure lowering
medications makes sense). However, nuanced features of the

© 2023, IJSREM | www.ijsrem.com | Page 4

 International Journal of Scientific Research in Engineering and Management (IJSREM)
Volume: 08 Issue: 11 | Nov - 2024 SJIF 2023: 8.176 ISSN: 2582-3930

patient, and the optimal way to correct the pathology given exist then the choice could be based on simply matching the
those nuances, are not often clear. Similar AI-based strategies patient’s pathophysiologic profile to the mechanisms of action
for making diagnoses can be exploited to identify potentially of the drugs, consistent with the underlying theme governing
correctable pathologies (P0 in the right panel of Figure 2). For basket trials. If the choice is not obvious, then one could
example, the company Arterys recently announced the first leverage personalized drug screening strategies using biopsies
FDA-approved AI-based application to be used in facilitating or biomaterials obtained from the patient, as suggested by
clinical diagnoses. The Arterys system used deep-learning Kodack et al. in the context of cancer.[41] These, and more
applied to a medical imaging platform to help diagnose heart general, personalized drug screening strategies have been
problems.[35] Other systems have been developed that developed and could benefit from AI techniques to find
consider more comprehensive approaches to understanding a patterns of relevance in the data that could indicate which
patient’s profiles in a way that could facilitate the choice of an drug or intervention is the most optimal.[42, 43]
intervention.[36]
IMPLEMENTATION (T3) AND CLINICAL ASSESSMENT
There is growing sophistication in the way that biopsied (P3) STUDIES
material or patient biosamples can be studied in a laboratory If a drug or health product has been shown to benefit
to identify targets for intervention. For example, emerging individuals in the general population, then considerations
induced pluripotent stem cell (iPSC) and organoid about the routine implementation and/or use of the product
technologies have shown great promise in yielding insights arise (T3 in the left side of Figure 2). Implementation can
into patient-specific pathologies that could be overcome with come in many forms. For example, if a drug shows sufficient
specific interventions (see Table 2 of Schork and Nazor [6] as proof to be safe and efficacious it can be approved for use by
well as Rossi et al.[37]). When recently-developed single cell regulatory agencies such as the FDA and become embedded
assays are combined with the use of organoids even greater in clinical practice. Of greater relevance to AI is the
resolution concerning pathologies and drug targets can be implementation of insights that might benefit physicians with
revealed and AI-based analyses have been shown to have respect to intervention choices when confronted with patients
great promise in this area.[38] with unique profiles (e.g., implementing a treatment strategy
of the type tested in a basket trial). Implementing such
LATE PHASE HUMAN (T2) AND INTERVENTION insights require codifying and providing them to physicians
CHOICE (P2) STUDIES through electronic medical record (EMR) systems typically
If a drug or health product has been shown to be safe and used to convey patient information to physicians.[44]
likely to be efficacious in early phase trials, then it must be Implementation of AI-based insights is a major topic of
tested for its general utility in the population at large. AI can discussion among pathologists since they are typically
be exploited in relevant large scale population trials that seek responsible for evaluating evidence that a patient has
to minimize the deployment and use of inappropriate drugs or particular condition, as well as pointing out nuances
interventions studies for each participant in the trial, as associated with that condition that may require special
described recently by Yauney and Shah.[39] Of great interest attention when intervention decisions are made.[45] The
in this context are the design and conduct of bucket (or provision of ‘decision support’ information to physicians and
variations termed ‘basket’), umbrella and adaptive trial health care workers – especially that derived from AI-based
designs.[40] Although each has unique features to them, a analyses – opens up a number of thorny ethical issues,
description of basket trials provides the general strategy however, such as who to blame if the use of the decision
behind each and also points out where AI can be exploited. support leads to poorer outcomes (i.e., the algorithm and its
Essentially bucket trials enroll eligible patients, profile them developers or the users who may be using it
to identify nuanced pathophysiologic profiles they possess inappropriately).[46]
(e.g., sequencing their tumor DNA in the context of a cancer
clinical trial), and assign them one of possibly many One important element of the implementation of AI-based
treatments based on the mechanisms of action of the decision support tools in EMR systems is that as new patient
treatments (i.e., put them into one of many treatment data are collected, the prediction models they are based on can
‘buckets’). If the patients provided with treatments dictated by be improved. Thus, ‘Learning Systems’ can be created that
their pathophysiologic profiles (i.e., assigned to the different continually evolve and improve based on the accrual of more
baskets) have better outcomes than those provided treatments patient information and outcomes associated with the patients
without recourse to the profiling and treatment matching provided interventions based on the algorithms behind the
scheme, then one could infer that the strategy or ‘algorithm’ learning systems’ recommendations.[47–49] Sophisticated AI-
for matching the treatments to the patients has merit. AI could techniques can be used to enhance this learning, including
be of great use in not only identifying treatment targets in the aggregating data from multiple EMRs or sources.[50]
patient profiles, but also aid in determining the strategy for
matching the treatments to the patient profiles. This would To vet the utility of an intervention for an individual patient
especially be the case if one could envision the use of many (P3 in the right side of Figure 2), N-of-1 trials can be pursued
different treatment baskets (for example due to considering (and could, as noted previously, be exploited in drug
many treatment combinations or complicated temporal development phase I clinical trials)[33, 51]. AI techniques can
treatment schemes). be used to identify patterns in data collected on the patient –
say through wireless sensors – that might be indicative of that
In the context of choosing an intervention for a particular patient’s response (or lack thereof) to the intervention.[34]
individual via the personalized medicine paradigm (P2 in the The studies listed in Table 1 provide example published N-of-
right panel of Figure 1), if available drugs and interventions 1 studies focusing on an individual’s response to a treatment

© 2023, IJSREM | www.ijsrem.com | Page 5

 International Journal of Scientific Research in Engineering and Management (IJSREM)
Volume: 08 Issue: 11 | Nov - 2024 SJIF 2023: 8.176 ISSN: 2582-3930

or an individual undergoing monitoring for health status harboring those mutations. If such neo-antigens are found,
changes. then cells from either a donor (allogeneic transplantation) or
from the patient him or herself (autologous transplantation)
POST-DEPLOYMENT EVALUATION (T4) STUDIES are harvested and sensitized to recognize the neo-antigens.
AND WAREHOUSING (P4) The basic idea is that these cells will attract the host’s immune
After the implementation and adoption of a new drug, cells towards the tumor cells harboring the neo-antigens when
treatment intervention or health product, continuous introduced into the patient’s body.[62, 63] Since the creation
monitoring of that product must occur in order to determine if and manufacture of the cells cannot be pursued in advance of
either unanticipated side effects are occurring or the product knowing what neo-antigens are present in the patient’s tumor,
can be improved or replaced for various reasons (T4 in the left they must be created in near real time. The production of
side of Figure 2). AI-based learning systems of the type treatments in real-time based on the patient’s unique and
mentioned in the previous section provide an excellent immediate needs is termed the ‘magistral’ production of
foundation for such monitoring, and early experiences with treatments, as opposed to the traditional or ‘officinal’
such systems bear this out.[52–54] In addition to the creation production of treatments.[29, 30] Magistral production of
of learning systems, there are many initiatives to aggregate drugs is likely to be a reality for personalized medicine in
data on patients and patient materials to enable data mining many settings, even beyond cancer, since it would be too
and AI-based analyses, for example in cancer contexts,[55] difficult to anticipate all the treatments (e.g., cells sensitive to
but also for more general settings as well.[56, 57] every neo-antigen profile) and stockpile them for use in the
future, as is assumed in the case with the officinal production
The implementation of AI-based products, such as EMR of treatments.
decision support tools and learning systems, will also affect
doctor/patient relationships in profound ways. This is To advance and generalize this concept of the magistral
especially likely with respect to discussions around the production of personalized medicine treatments, one could
justification of intervention choices,[58] but also with respect imagine leveraging AI-powered robotics technologies to
to discussions about predictions concerning future health care enable the efficient and precise manufacture of relevant
issues.[59] Large, government-sponsored national initiatives treatments.[64] 3D printing of treatments also has the
are being pursued to identify patterns among individuals potential to facilitate the real-time production of treatments in
tracked for health care-related phenomena that might be near real time, as the first US FDA-approval for a 3D printed
useful in clinical and public health practices in the future, such drug was made in 2015.[65] One could also envision the
as the UK Biobank initiative in the United Kingdom and the immediate conduct of N-of-1 trials involving AI-based pattern
‘All-of-Us’ study in the United States.[60, 61] These studies discovery with sophisticated treatment outcome monitoring
raise important questions about the ethics, legal and social devices after a treatment has been crafted to assess its impact
implications of aggregating data on individual patients for the on the patient.[33, 51, 66] Further, one could potentially
purposes of benefitting individuals who may need focused exploit AI-based simulation studies to anticipate directions
care going forward.[61] that a treatment strategy might take.[67]

INTEGRATION AND THE PERSONALIZED MEDICINE LIMITATIONS OF AI IN ADVANCING PERSONALIZED

WORKFLOW MEDICINE
Implicit in the P0-P4 personalized medicine translational There are many limitations to the use of AI in the
workflow that has common elements with the T0-T4 drug and development of personalized medicines. We briefly discuss
health product development workflow (Figure 2), is the some of the more salient issues. First, there is an argument
suggestion that what is important for these workflows to that many big data analyses that combine information on
function properly is to make sure the transitions from each many individuals to identify patterns that reflect population-
stage of the workflow are eased and enabled. This can be level relationships between data points do not get at important
difficult since the expertise and technologies needed at each individual-level relationships.[68] This potential lack of
stage are very different, often leading to their independent ‘ergodicity’ could result in models that are not useful for
pursuit. However, emerging strategies and concepts in the making individual treatment decisions. For example, in terms
way personalized medicine is practiced, coupled with the use of identifying trends in a target individual’s health data that
of AI techniques, could lead to more holistic and efficient could indicate a health status change for that target individual
ways of treating individual patients that run through the entire based on data collected on a large number of individuals, as
P0-P4 workflow.[6] more data points are collected on each individual, any
predictions of the target individual’s heath trajectory should
Thus, the ideal setting for personalized medicine and health rely more on the legacy data points on that target individual
care is one in which the diagnoses, treatment and follow-up and less on the population level data.[69] Sensitizing AI
monitoring of individual patients is streamlined into a single techniques to this fact is crucial for advancing personalized
process with very smooth and coordinated transitions from medicine.
one relevant activity or sub-process to another. A good
paradigm for this involves the creation of cell replacement Second, there is a need to vet or test the utility of health care
therapies for a wide variety of conditions.[62, 63] Thus, for products rooted in AI. This is motivated by the inconsistent
example, in certain immunotherapeutic-oriented cell results observed with the use of some AI or big data based
replacement therapies for cancer, a patient’s tumor is profiled health care products, such as IBM’s Watson treatment
for the existence of unique ‘neo-antigens’ or mutations that decision support system. [70, 71] Testing such systems via
might attract the host’s own immune system to attack cells traditional randomized clinical trials has been discussed in the

© 2023, IJSREM | www.ijsrem.com | Page 6

 International Journal of Scientific Research in Engineering and Management (IJSREM)
Volume: 08 Issue: 11 | Nov - 2024 SJIF 2023: 8.176 ISSN: 2582-3930

literature, and some AI-based decision support tools have in seed and key off for the development of more reliable
fact been shown to pass muster in bona fide clinical trials.[72] predictions.[83]
A potential need for vetting, for example, AI-based decision
support products, like IBM’s Watson, is that if the underlying In addition to exploiting greater speed and computational
system’s decision making capability is trained on an efficiency, AI-based health products and tools will likely
incomplete or biased data set, then the recommendations or incorporate greater understanding of biology in their
predictions it provides are likely to be unreliable. A rather formulations in the future. Thus, the discovery of simple
infamous case of this involves Google’s system for prediction input/output relationships among data points that has been
flu outbreaks.[73] In addition, in the context of basket trials, focus of a great deal of AI, machine learning and statistical
in which the underlying scheme for matching drugs to patient analysis research, could be pursued with constraints that are
profiles is being tested, if the scheme is shown not work better known to govern phenomena of relevance (e.g., known
than standard of care or an alternative way of matching drugs biophysical constraints involving the production of
to patient profiles, then a couple of questions could be raised. metabolites in a biochemical pathway, first principles having
It could be that the drugs are ineffective, or some subset are to do with Watson-Crick base pairing, etc.).[84, 85]
ineffective, essentially negatively impacting the overall
performance of the matching scheme. Alternatively, it could Finally, as a closing note, much of the use of AI in the
be that the drugs work, but simply are not matched properly to development of personalized medicines is focused on the
the patient profiles; i.e., the matching algorithm or scheme is treatment of individuals with overt disease: identifying the
simply wrong. These questions were raised in the context of underlying pathology, determining which interventions might
the SHIVA trial – a bucket trail in which the drug matching make most sense to provide given what is known about that
scheme was shown not to benefit the patients any more than pathology and the mechanism of action of the intervention,
legacy ways of treating patients.[74] and testing to see if the intervention works. Thus, the vast
majority of AI-based products and tools used in advancing
Third, it may be that the best way to vet at least decision personalized medicine focus on the diagnosis, prognosis and
support tools, is not to test them in randomized clinical trials, treatment of individuals. This makes sense as there is a great
but rather to implement them in learning systems in which the need for advances and efficiency gains in treating patients
decision support rules or algorithms are continuously given the costs of current treatments, especially in the context
updated.[66, 75–78] However, this would not only require a of cancer. However, the application of AI to disease
lack of bias in the initial data sets used to seed the learning prevention is gaining a great deal of attention and traction. For
system in order to ensure generalizable results, but could also example, AI and machine learning techniques have been
take a lengthy time period for the system to evolve into a shown to be useful in the development of ‘polygenic risk
system with accurate and reliable decision making. scores’ that can be used to identify individuals with an
elevated genetic risk for disease that could be monitored more
Fourth, many AI-based decision support products leverage closely.[86–88] In addition, by combining insights into
deep learning and neural network-based algorithms. Such genetic predisposition to disease with continuous monitoring
algorithms can produce very reliable predictions if a large to identify early signs of disease, one could potentially stop
enough training set is used, but the connections between the diseases in their tracks before complicated treatments are
inputs (i.e., data) and the outputs (predictions) can be very needed for fulminant forms of the disease manifest.[89, 90]
hard to understand. Thus, the ‘Black Box’ problem associated Such monitoring could be greatly enhanced by applying AI
with many AI-based tools can be problematic and lead to a techniques to novel sensors.[91, 92]
lack of confidence or sense of trepidation about relying on the
predictions in the real world where real lives are at stake.[79] Ultimately, enthusiasm for leveraging AI techniques is not
In addition, not all AI techniques are designed to identify likely to slow down any time soon. AI is likely to impact
causal relationships between various input and outputs, but virtually every industry, from manufacturing, to sales and
rather mere associations or predictions (i.e., focus on marketing, to banking, to transportation. All of these
correlation and not causation).[80] This may suffice if the industries can obviously be improved with AI playing an
goal is to develop accurate predictions, but if the goal is to, important role in the needed innovations. The health care
e.g., identify a drug target that, when modulated, leads to a industry is no less likely to benefit from AI, as this chapter has
desired effect, then identifying causal relationships is crucial. made clear, as long as appropriate integration and vetting
occurs.
3. FUTURE DIRECTIONS AND CONCLUDING REFERENCES
REMARKS
[1] Russell S and Norvig P, Artificial Intelligence: A Modern
The future contributions of AI in advancing personalized Approach (3rd Edition). 3rd ed 2009, Carmel, Indiana:
medicine are likely to be very pronounced, as this chapter Pearson. [Google Scholar]
makes clear. Not only will there be greater adoption of AI- [2] Mahmud M, et al. , Applications of Deep Learning and
based health products but such products could be developed Reinforcement Learning to Biological Data. IEEE Trans
and exploit emerging computing capabilities such as quantum Neural Netw Learn Syst, 2018. 29(6): p. 2063–2079. [DOI]
computing[81, 82] to achieve increased speed and an ability to [PubMed] [Google Scholar]
[3] Webb S, Deep learning for biology. Nature, 2018.
handle larger and larger data sets. These larger data sets are
554(7693): p. 555–557. [DOI] [PubMed] [Google Scholar]
likely to derive from better and more sophisticated monitoring [4] Fleming N, How artificial intelligence is changing drug
health monitoring devices which can be used to gather data to discovery. Nature, 2018. 557(7707): p. S55–S57. [DOI]
[PubMed] [Google Scholar]

© 2023, IJSREM | www.ijsrem.com | Page 7

 International Journal of Scientific Research in Engineering and Management (IJSREM)
Volume: 08 Issue: 11 | Nov - 2024 SJIF 2023: 8.176 ISSN: 2582-3930

[5] Committee to Review the Clinical and Translational [23] Radovic A, et al. , Machine learning at the energy and
Science Awards Program at the National Center for intensity frontiers of particle physics. Nature, 2018.
Advancing Translational Sciences; Board on Health 560(7716): p. 41–48. [DOI] [PubMed] [Google Scholar]
Sciences Policy; Institute of Medicine; Leshner AI TS, ed. [24] Silver D, et al. , Mastering the game of Go without human
The CTSA Program at NIH: Opportunities for Advancing knowledge. Nature, 2017. 550(7676): p. 354–359. [DOI]
Clinical and Translational Research. The National [PubMed] [Google Scholar]
Academies Collection: Reports funded by National [25] Madhukar NS, et al. , A New Big-Data Paradigm For
Institutes of Health. 2013, National Academies Press: Target Identification And Drug Discovery. BioRxiv (
Washington, DC. [PubMed] [Google Scholar] 10.1101/134973), 2018. [DOI] [Google Scholar]
[6] Schork NJ and Nazor K, Integrated Genomic Medicine: A [26] Chen B and Butte AJ, Leveraging big data to transform
Paradigm for Rare Diseases and Beyond. Adv Genet, 2017. target selection and drug discovery. Clin Pharmacol Ther,
97: p. 81–113. [DOI] [PMC free article] [PubMed] 2016. 99(3): p. 285–97. [DOI] [PMC free article]
[Google Scholar] [PubMed] [Google Scholar]
[7] Telenti A, et al. , Deep learning of genomic variation and [27] Hu Y and Bajorath J, Entering the ‘big data’ era in
regulatory network data. Hum Mol Genet, 2018. 27(R1): p. medicinal chemistry: molecular promiscuity analysis
R63–R71. [DOI] [PMC free article] [PubMed] [Google revisited. Future Sci OA, 2017. 3(2): p. FSO179. [DOI]
Scholar] [PMC free article] [PubMed] [Google Scholar]
[8] Esteva A, et al. , Dermatologist-level classification of skin [28] Hernandez D, How Robots Are Making Better Drugs,
cancer with deep neural networks. Nature, 2017. Faster, in Wall Street Journal. 2018, Dow Jones &
542(7639): p. 115–118. [DOI] [PMC free article] Company New York, New York. [Google Scholar]
[PubMed] [Google Scholar] [29] Patient-centered drug manufacture. Nat Biotechnol, 2017.
[9] Gerstung M, et al. , Precision oncology for acute myeloid 35(6): p. 485. [DOI] [PubMed] [Google Scholar]
leukemia using a knowledge bank approach. Nat Genet, [30] Schellekens H, et al. , Making individualized drugs a
2017. 49(3): p. 332–340. [DOI] [PMC free article] reality. Nat Biotechnol, 2017. 35(6): p. 507–513. [DOI]
[PubMed] [Google Scholar] [PubMed] [Google Scholar]
[10] Gulshan V, et al. , Development and Validation of a Deep [31] Lavertu A, et al. , Pharmacogenomics and big genomic
Learning Algorithm for Detection of Diabetic Retinopathy data: from lab to clinic and back again. Hum Mol Genet,
in Retinal Fundus Photographs. JAMA, 2016. 316(22): p. 2018. 27(R1): p. R72–R78. [DOI] [PMC free article]
2402–2410. [DOI] [PubMed] [Google Scholar] [PubMed] [Google Scholar]
[11] Cohen JD, et al. , Detection and localization of surgically [32] Kalinin AA, et al. , Deep learning in pharmacogenomics:
resectable cancers with a multi-analyte blood test. Science, from gene regulation to patient stratification.
2018. 359(6378): p. 926–930. [DOI] [PMC free article] Pharmacogenomics, 2018. 19(7): p. 629–650. [DOI] [PMC
[PubMed] [Google Scholar] free article] [PubMed] [Google Scholar]
[12] Bray MA, et al. , A dataset of images and morphological [33] Schork NJ, Personalized medicine: Time for one-person
profiles of 30 000 small-molecule treatments using the Cell trials. Nature, 2015. 520(7549): p. 609–11. [DOI]
Painting assay. Gigascience, 2017. 6(12): p. 1–5. [DOI] [PubMed] [Google Scholar]
[PMC free article] [PubMed] [Google Scholar] [34] Serhani MA, et al. , New algorithms for processing time-
[13] Ma J, et al. , Using deep learning to model the hierarchical series big EEG data within mobile health monitoring
structure and function of a cell. Nat Methods, 2018. 15(4): systems. Comput Methods Programs Biomed, 2017. 149:
p. 290–298. [DOI] [PMC free article] [PubMed] [Google p. 79–94. [DOI] [PubMed] [Google Scholar]
Scholar] [35] Marr B, First FDA Approval For Clinical Cloud-Based
[14] Ideker T, et al. , Integrated genomic and proteomic Deep Learning In Healthcare, in Forbes. 2017, Forbes
analyses of a systematically perturbed metabolic network. Publishing Company: New York City. [Google Scholar]
Science, 2001. 292(5518): p. 929–34. [DOI] [PubMed] [36] Miotto R, et al. , Deep Patient: An Unsupervised
[Google Scholar] Representation to Predict the Future of Patients from the
[15] Bohannon J, The cyberscientist. Science, 2017. 357(6346): Electronic Health Records. Sci Rep, 2016. 6: p.
p. 18–21. [DOI] [PubMed] [Google Scholar] 26094. [DOI] [PMC free article] [PubMed] [Google
[16] King RD, et al. , The automation of science. Science, 2009. Scholar]
324(5923): p. 85–9. [DOI] [PubMed] [Google Scholar] [37] Rossi G, Manfrin A, and Lutolf MP, Progress and potential
[17] Sparkes A and Clare A, AutoLabDB: a substantial open in organoid research. Nat Rev Genet, 2018. [DOI]
source database schema to support a high-throughput [PubMed] [Google Scholar]
automated laboratory. Bioinformatics, 2012. 28(10): p. [38] Deng Y, et al. , Massive single-cell RNA-seq analysis and
1390–7. [DOI] [PubMed] [Google Scholar] imputation via deep learning. BioRxiv
[18] Butler KT, et al. , Machine learning for molecular and 2018. https://t.co/EGBwlYFLLK. [Google Scholar]
materials science. Nature, 2018. 559(7715): p. 547– [39] Yauney G and Shah P, Reinforcement Learning with
555. [DOI] [PubMed] [Google Scholar] Action-Derived Rewards for Chemotherapy and Clinical
[19] Sanchez-Lengeling B and Aspuru-Guzik A, Inverse Trial Dosing Regimen Selection. Proceedings of Machine
molecular design using machine learning: Generative Learning Research, 2018. 85. [Google Scholar]
models for matter engineering. Science, 2018. 361(6400): [40] Biankin AV, Piantadosi S, and Hollingsworth SJ, Patient-
p. 360–365. [DOI] [PubMed] [Google Scholar] centric trials for therapeutic development in precision
[20] Aage N, et al. , Giga-voxel computational morphogenesis oncology. Nature, 2015. 526(7573): p. 361–70. [DOI]
for structural design. Nature, 2017. 550(7674): p. 84– [PubMed] [Google Scholar]
86. [DOI] [PubMed] [Google Scholar] [41] Kodack DP, et al. , Primary Patient-Derived Cancer Cells
[21] Segler MHS, Preuss M, and Waller MP, Planning chemical and Their Potential for Personalized Cancer Patient Care.
syntheses with deep neural networks and symbolic AI. Cell Rep, 2017. 21(11): p. 3298–3309. [DOI] [PMC free
Nature, 2018. 555(7698): p. 604–610. [DOI] [PubMed] article] [PubMed] [Google Scholar]
[Google Scholar] [42] Gorshkov K, et al. , Advancing precision medicine with
[22] Ahneman DT, et al. , Predicting reaction performance in C- personalized drug screening. Drug Discov Today,
N cross-coupling using machine learning. Science, 2018. 2018. [DOI] [PMC free article] [PubMed] [Google
360(6385): p. 186–190. [DOI] [PubMed] [Google Scholar] Scholar]

© 2023, IJSREM | www.ijsrem.com | Page 8

 International Journal of Scientific Research in Engineering and Management (IJSREM)
Volume: 08 Issue: 11 | Nov - 2024 SJIF 2023: 8.176 ISSN: 2582-3930

[43] Miranda CC, et al. , Towards Multi-Organoid Systems for 2017. 19(7): p. 743–750. [DOI] [PubMed] [Google
Drug Screening Applications. Bioengineering (Basel), Scholar]
2018. 5(3). [DOI] [PMC free article] [PubMed] [Google [62] Li C, et al. , Application of induced pluripotent stem cell
Scholar] transplants: Autologous or allogeneic? Life Sci,
[44] Scott IA, et al. , Using EMR-enabled computerized 2018. [DOI] [PubMed] [Google Scholar]
decision support systems to reduce prescribing of [63] Graham C, et al. , Allogeneic CAR-T Cells: More than
potentially inappropriate medications: a narrative review. Ease of Access? Cells, 2018. 7(10). [DOI] [PMC free
Ther Adv Drug Saf, 2018. 9(9): p. 559–573. [DOI] [PMC article] [PubMed] [Google Scholar]
free article] [PubMed] [Google Scholar] [64] Tan R, Yang X, and Shen Y, Robot-aided electrospinning
[45] Dreyer KJ and Geis JR, When Machines Think: toward intelligent biomedical engineering. Robotics
Radiology’s Next Frontier. Radiology, 2017. 285(3): p. Biomim, 2017. 4(1): p. 17. [DOI] [PMC free article]
713–718. [DOI] [PubMed] [Google Scholar] [PubMed] [Google Scholar]
[46] Nabi J, How Bioethics Can Shape Artificial Intelligence [65] Osouli-Bostanabad K and Adibkia K, Made-on-demand,
and Machine Learning. Hastings Cent Rep, 2018. 48(5): p. complex and personalized 3D-printed drug products.
10–13. [DOI] [PubMed] [Google Scholar] Bioimpacts, 2018. 8(2): p. 77–79. [DOI] [PMC free article]
[47] Etheredge LM, A rapid-learning health system. Health Aff [PubMed] [Google Scholar]
(Millwood), 2007. 26(2): p. w107–18. [DOI] [PubMed] [66] Schork NJ, Randomized clinical trials and personalized
[Google Scholar] medicine: A commentary on deaton and cartwright. Soc
[48] Shrager J and Tenenbaum JM, Rapid learning for precision Sci Med, 2018. 210: p. 71–73. [DOI] [PMC free article]
oncology. Nat Rev Clin Oncol, 2014. 11(2): p. 109– [PubMed] [Google Scholar]
18. [DOI] [PubMed] [Google Scholar] [67] Shamsuddin R, et al. Virtual patient model: An approach
[49] Shah A, et al. , Building a Rapid Learning Health Care for generating synthetic healthcare time series data in IEEE
System for Oncology: Why CancerLinQ Collects Interantional COnference on Healthcare Informatics. 2018.
Identifiable Health Information to Achieve Its Vision. J IEEE Computer Society. [Google Scholar]
Clin Oncol, 2016. 34(7): p. 756–63. [DOI] [PubMed] [68] Fisher AJ, Medaglia JD, and Jeronimus BF, Lack of group-
[Google Scholar] to-individual generalizability is a threat to human subjects
[50] Hastie T, Tibshirani R, and Friedman J, The Elements of research. Proc Natl Acad Sci U S A, 2018. 115(27): p.
Statistical Learning: Data Mining, Inference, and E6106–E6115. [DOI] [PMC free article] [PubMed]
Prediction. 2nd ed 2009, New York, New York: [Google Scholar]
Springer. [Google Scholar] [69] Drescher CW, et al. , Longitudinal screening algorithm that
[51] Schork NJ and Goetz LH, Single-Subject Studies in incorporates change over time in CA125 levels identifies
Translational Nutrition Research. Annu Rev Nutr, 2017. ovarian cancer earlier than a single-threshold rule. J Clin
37: p. 395–422. [DOI] [PMC free article] [PubMed] Oncol, 2013. 31(3): p. 387–92. [DOI] [PMC free article]
[Google Scholar] [PubMed] [Google Scholar]
[52] Agarwala V, et al. , Real-World Evidence In Support Of [70] Zhou N, et al. , Concordance Study Between IBM Watson
Precision Medicine: Clinico-Genomic Cancer Data As A for Oncology and Clinical Practice for Patients with
Case Study. Health Aff (Millwood), 2018. 37(5): p. 765– Cancer in China. Oncologist, 2018. [DOI] [PMC free
772. [DOI] [PubMed] [Google Scholar] article] [PubMed] [Google Scholar]
[53] Williams MS, et al. , Patient-Centered Precision Health In [71] Schmidt C, Anderson MD Breaks With IBM Watson,
A Learning Health Care System: Geisinger’s Genomic Raising Questions About Artificial Intelligence in
Medicine Experience. Health Aff (Millwood), 2018. 37(5): Oncology. J Natl Cancer Inst, 2017. 109(5): p. 4–5. [DOI]
p. 757–764. [DOI] [PubMed] [Google Scholar] [PubMed] [Google Scholar]
[54] Rajkomar A, et al. , Scalable and accurate deep learning [72] Abramoff MD, et al. , Pivotal trial of an autonomous AI-
with electronic health records. NPJ Digital Medicine, 2018. based diagnostic system for detection of diabetic
1(18). [DOI] [PMC free article] [PubMed] [Google retinopathy in primary care offices. NPJ Digital Medicine,
Scholar] 2018. 1(39). [DOI] [PMC free article] [PubMed] [Google
[55] Ali M and Aittokallio T, Machine learning and feature Scholar]
selection for drug response prediction in precision [73] Lazer D, et al. , Big data. The parable of Google Flu: traps
oncology applications. Biophys Rev, 2018. [DOI] [PMC in big data analysis. Science, 2014. 343(6176): p. 1203–
free article] [PubMed] [Google Scholar] 5. [DOI] [PubMed] [Google Scholar]
[56] Mathe E, et al. , The Omics Revolution Continues: The [74] Le Tourneau C, et al. , Molecularly targeted therapy based
Maturation of High-Throughput Biological Data Sources. on tumour molecular profiling versus conventional therapy
Yearb Med Inform, 2018. 27(1): p. 211–222. [DOI] [PMC for advanced cancer (SHIVA): a multicentre, open-label,
free article] [PubMed] [Google Scholar] proof-of-concept, randomised, controlled phase 2 trial.
[57] Varghese J, et al. , CDEGenerator: an online platform to Lancet Oncol, 2015. 16(13): p. 1324–34. [DOI] [PubMed]
learn from existing data models to build model registries. [Google Scholar]
Clin Epidemiol, 2018. 10: p. 961–970. [DOI] [PMC free [75] Ioannidis JPA and Khoury MJ, Evidence-based medicine
article] [PubMed] [Google Scholar] and big genomic data. Hum Mol Genet, 2018. 27(R1): p.
[58] Lerner I, et al. , Revolution in Health Care: How Will Data R2–R7. [DOI] [PMC free article] [PubMed] [Google
Science Impact Doctor-Patient Relationships? Front Public Scholar]
Health, 2018. 6: p. 99. [DOI] [PMC free article] [PubMed] [76] AI diagnostics need attention. Nature, 2018. 555(7696): p.
[Google Scholar] 285. [DOI] [PubMed] [Google Scholar]
[59] Savage N, Machine learning: Calculating disease. Nature, [77] Frieden TR, Evidence for Health Decision Making -
2017. 550(7676): p. S115–S117. [DOI] [PubMed] [Google Beyond Randomized, Controlled Trials. N Engl J Med,
Scholar] 2017. 377(5): p. 465–475. [DOI] [PubMed] [Google
[60] Bycroft C, et al. , The UK Biobank resource with deep Scholar]
phenotyping and genomic data. Nature, 2018. 562: p. 203– [78] Abernethy A and Khozin S, Clinical Drug Trials May Be
209. [DOI] [PMC free article] [PubMed] [Google Scholar] Coming to Your Doctor’s Office, in Wall Street Journal.
[61] Sankar PL and Parker LS, The Precision Medicine 2017, Dow Jones & Company: New York, New
Initiative’s All of Us Research Program: an agenda for York. [Google Scholar]
research on its ethical, legal, and social issues. Genet Med,

© 2023, IJSREM | www.ijsrem.com | Page 9

 International Journal of Scientific Research in Engineering and Management (IJSREM)
Volume: 08 Issue: 11 | Nov - 2024 SJIF 2023: 8.176 ISSN: 2582-3930

[79] Voosen P, The AI detectives. Science, 2017. 357(6346): p. 2016. 7: p. 12948. [DOI] [PMC free article] [PubMed]
22–27. [DOI] [PubMed] [Google Scholar] [Google Scholar]
[80] Marwala T, Causality, Correlation and Artificial [99] Forsdyke DR, Summertime dosage-dependent
Intelligence for Rational Decision Making. 2015, New hypersensitivity to an angiotensin II receptor blocker.
Jersey: World Scientific. [Google Scholar] BMC Res Notes, 2015. 8: p. 227. [DOI] [PMC free article]
[81] Ciliberto C, et al. , Quantum machine learning: a classical [PubMed] [Google Scholar]
perspective. Proc Math Phys Eng Sci, 2018. 474(2209): p. [100] O’Rawe JA, et al. , Integrating precision medicine in the
20170551. [DOI] [PMC free article] [PubMed] [Google study and clinical treatment of a severely mentally ill
Scholar] person. PeerJ, 2013. 1: p. e177. [DOI] [PMC free article]
[82] Li RY, et al. , Quantum annealing versus classical machine [PubMed] [Google Scholar]
learning applied to a simplified computational biology [101] Li W, et al. , Extensive graft-derived dopaminergic
problem. npj Quantum Inf, 2018. 4. [DOI] [PMC free innervation is maintained 24 years after transplantation in
article] [PubMed] [Google Scholar] the degenerating parkinsonian brain. Proc Natl Acad Sci U
[83] Vashistha R, et al. , Futuristic biosensors for cardiac health S A, 2016. 113(23): p. 6544–9. [DOI] [PMC free article]
care: an artificial intelligence approach. 3 Biotech, 2018. [PubMed] [Google Scholar]
8(8): p. 358. [DOI] [PMC free article] [PubMed] [Google [102] Bloss CS, et al. , A genome sequencing program for novel
Scholar] undiagnosed diseases. Genet Med, 2015. 17(12): p. 995–
[84] Palsson B, Systems Biology: Constraint-based 1001. [DOI] [PMC free article] [PubMed] [Google
Reconstruction and Analysis. 2nd ed 2015, Boston, MA: Scholar]
Cambridge University Press. [Google Scholar] [103] Piening BD, et al. , Integrative Personal Omics Profiles
[85] Ideker T, Dutkowski J, and Hood L, Boosting signal-to- during Periods of Weight Gain and Loss. Cell Syst, 2018.
noise in complex biology: prior knowledge is power. Cell, 6(2): p. 157–170 e8. [DOI] [PMC free article] [PubMed]
2011. 144(6): p. 860–3. [DOI] [PMC free article] [Google Scholar]
[PubMed] [Google Scholar] [104] Zalusky R and Herbert V, Megaloblastic anemia in scurvy
[86] Khera AV, et al. , Genome-wide polygenic scores for with response to 50 microgm. of folic acid daily. N Engl J
common diseases identify individuals with risk equivalent Med, 1961. 265: p. 1033–8. [DOI] [PubMed] [Google
to monogenic mutations. Nat Genet, 2018. 50(9): p. 1219– Scholar]
1224. [DOI] [PMC free article] [PubMed] [Google [105] Herbert V, Experimental nutritional folate deficiency in
Scholar] man. Trans Assoc Am Physicians, 1962. 75: p. 307–
[87] Warren M, The approach to predictive medicine that is 20. [PubMed] [Google Scholar]
taking genomics research by storm. Nature, 2018. [106] Golding PH, Severe experimental folate deficiency in a
562(7726): p. 181–183. [DOI] [PubMed] [Google Scholar] human subject - a longitudinal study of biochemical and
[88] Schork AJ, Schork MA, and Schork NJ, Genetic risks and haematological responses as megaloblastic anaemia
clinical rewards. Nat Genet, 2018. 50(9): p. 1210– develops. Springerplus, 2014. 3: p. 442. [DOI] [PMC free
1211. [DOI] [PMC free article] [PubMed] [Google article] [PubMed] [Google Scholar]
Scholar]
[89] Patel CJ, et al. , Whole genome sequencing in support of
wellness and health maintenance. Genome Med, 2013.
5(6): p. 58. [DOI] [PMC free article] [PubMed] [Google
Scholar]
[90] Schork NJ, Genetic parts to a preventive medicine whole.
Genome Med, 2013. 5(6): p. 54. [DOI] [PMC free article]
[PubMed] [Google Scholar]
[91] Mapara SS and Patravale VB, Medical capsule robots: A
renaissance for diagnostics, drug delivery and surgical
treatment. J Control Release, 2017. 261: p. 337–351. [DOI]
[PubMed] [Google Scholar]
[92] Topol EJ, Deep Medicine: How Artificial Intelligence Can
Make Healthcare Human Again. 2019, New York: Basic
Books. [Google Scholar]
[93] David LA, et al. , Host lifestyle affects human microbiota
on daily timescales. Genome Biol, 2014. 15(7): p.
R89. [DOI] [PMC free article] [PubMed] [Google Scholar]
[94] Chen R, et al. , Personal omics profiling reveals dynamic
molecular and medical phenotypes. Cell, 2012. 148(6): p.
1293–307. [DOI] [PMC free article] [PubMed] [Google
Scholar]
[95] Magnuson V, Wang Y, and Schork N, Normalizing sleep
quality disturbed by psychiatric polypharmacy: a single
patient open trial (SPOT). F1000Res, 2016. 5: p.
132. [DOI] [PMC free article] [PubMed] [Google Scholar]
[96] Zeevi D, et al. , Personalized Nutrition by Prediction of
Glycemic Responses. Cell, 2015. 163(5): p. 1079–
1094. [DOI] [PubMed] [Google Scholar]
[97] Smarr L, et al. , Tracking Human Gut Microbiome
Changes Resulting from a Colonoscopy. Methods Inf Med,
2017. 56(6): p. 442–447. [DOI] [PubMed] [Google
Scholar]
[98] Trammell SA, et al. , Nicotinamide riboside is uniquely
and orally bioavailable in mice and humans. Nat Commun,

© 2023, IJSREM | www.ijsrem.com | Page 10