A Conceptual Model-Based Application for the Treatment and

Management of Data in Pediatric Oncology:

The Neuroblastoma Use Case

Jes

us Carre

no-Bolufer

1 a

, Jos

e Fabi

an Reyes Rom

1 b

, Sergio P

erez Andr

1 c

esir

ee Ramal Pons

, V

ıctor Ju

arez Vidal

, Adela Ca

nete Nieto

2 d

and

Oscar Pastor

1 e

Valencian Research Institute for Artiﬁcial Intelligence (VRAIN), Universitat Polit

ecnica de Val

encia, Valencia, Spain

Grupo de Investigaci

on Cl

ınica y Traslacional en C

ancer, Instituto de Investigaci

on Sanitaria La Fe (IIS La Fe),

Valencia, Spain

Keywords:

ClinGenNBL, Conceptual Modeling, Neuroblastoma, Pediatric Oncology, Data Management, CMN.

Abstract:

Neuroblastoma is one of the leading causes of death in childhood oncology. Current treatments for these

patients are general and not targeted, including radiotherapy, chemotherapy, and surgery. There is a need

for more efﬁcient methods. Precision Medicine (PM) can help to overcome this challenge. PM incorporates

clinical, lifestyle, and genomic data, among others, into a standardized process to provide individualized

treatment. However, a large amount of data is needed to achieve PM, and the heterogeneity present in the

case of neuroblastoma poses a challenge for integration and, consequently, for knowledge generation. We

need a solid domain deﬁnition that provides a foundation for experts to work on, which implies generating

a conceptual model. Based on this model, any Information System (IS) can be developed. ISs play a vital

role in managing clinical data efﬁciently. Much of the clinical data has been captured and managed over the

years with inefﬁcient tools such as spreadsheets. In this work, we ﬁrst present the new Conceptual Model

of Neuroblastoma (CMN), with a special focus on genomics, and second, ClinGenNBL, a conceptual model-

based web application that implement the CMN with the goal of assisting clinicians in managing patients with

neuroblastoma through a user-friendly interface.

In Memoriam and in honor of the beloved Victoria

Castel S

anchez, who passed away during the research

and publication of this research work.

1 INTRODUCTION

The most frequent tumors in Spain (0-14 years, be-

tween 2010-2023) are leukemias 28.1%; lymphomas

12.1% and central nervous system tumors 24.8%

(Ca

nete Nieto et al., 2023). The Neuroblastoma

(NBL), or secondary nervous system tumor, with

7.7% of cases, comes in fourth place in terms of inci-

dence. NBL is also the most frequent solid extracra-

nial tumor in childhood (Castleberry, 1997). In the

https://orcid.org/0009-0006-4720-4037

https://orcid.org/0000-0002-9598-1301

https://orcid.org/0000-0002-9316-210X

https://orcid.org/0000-0002-5669-5097

https://orcid.org/0000-0002-1320-8471

United States, it represents 6% of all pediatric can-

cers, with a survival rate of 82% for ages from birth

to 14 years old (Siegel et al., 2024).

In light of this problem, the Spanish NBL Group

of SEHOP

was created in 1987, led by the Pediatric

Oncology Unit of Hospital Universitari i Polit

ecnic

La Fe (HUP/IIS LaFe) (HUPLF). Since then, the

group has developed and ﬁne-tuned therapeutic and

diagnostic protocols for NBLs and participated in in-

ternational groups with SIOPEN

and ANRA

. They

have also produced a wide variety of reports and dis-

seminated study results (Ca

nete Nieto et al., 2023)

We encounter massive amounts of data from

which knowledge could be generated to advance Pre-

cision Medicine (PM) (McCabe et al., 2024) (Ca-

Sociedad Espa

nola de Hematolog

ıa y Oncolog

ıa

Pedi

atricas. https://www.sehop.org/

International Society of Paediatric Oncology European

Neuroblastoma. https://www.siopen-r-net.org/

Advances in Neuroblastom Research Association.

https://www.anrmeeting.org/

Carreño-Bolufer, J., Reyes Román, J. F., Pérez Andrés, S., Ramal Pons, D., Juárez Vidal, V., Cañete Nieto, A. and Pastor, Ó.

A Conceptual Model-Based Application for the Treatment and Management of Data in Pediatric Oncology: The Neuroblastoma Use Case.

DOI: 10.5220/0013202600003928

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 20th International Conference on Evaluation of Novel Approaches to Software Engineering (ENASE 2025), pages 15-26

ISBN: 978-989-758-742-9; ISSN: 2184-4895

haney et al., 2022). This data can help us understand

patients who suffer from NBL. PM aims to provide

personalized treatment for each patient. Currently,

implementing PM in clinical data management re-

quires a high investment of resources, mainly due to

the complexity of the data and the lack of efﬁcient In-

formation Systems (IS). Each disease is unique and

requires specialized tools.

Fortunately, technological advances in Software

Engineering (SE) in recent years have created many

ISs for better information management in different

contexts (Evans, 2016). In particular, in the context

of PM, these ISs play a fundamental role in manag-

ing patient follow-up and conducting studies on data

collected in the clinical context over the years.

However, when applied to NBL, one of the main

challenges in building an IS is the data heterogeneity

in this ﬁeld. Consequently, there is a lack of inte-

grated information from a holistic perspective, which

hinders the development of automated pipelines for

identifying and analyzing relevant statistical data

(Digital Transition, ). A standard conceptual model

would facilitate the integration process from diverse

sources (Trujillo et al., 2018)(Ca

nete et al., 2022).

In summary, data management in this context

presents the following problems: high heterogeneity

and dispersion of data and lack of efﬁcient mecha-

nisms. To overcome these challenges, the following

goals are speciﬁed:

1. We aim to generate a Conceptual Model (CM) to

provide a common ontological framework for dis-

cussions with experts, increasing our understand-

ing of the neuroblastoma (NBL) domain. This

model will also serve as a foundation for de-

veloping an Information System (IS) to improve

data management and integration, encompassing

not only clinical data but also genomic data. To

achieve this, we establish a shared view of the

study domain, which must be supported by clin-

ical experts.

2. To design and develop an IS based on the de-

ﬁned CM. This IS will improve the integration and

management of the existing data in the NBL do-

main. It will also help clinical experts to improve

their decision-making.

To achieve these goals, this work contributes to

version 2.0 (V2) of the Conceptual Model of Neu-

roblastoma (CMN) upon which a software platform,

ClinGenNBL, is implemented. ClinGenNBL, a con-

ceptual model-based web application, serves as a

work tool for clinicians, combining the management

and analysis of clinical data.

The rest of the article is organized as follows: in

Section 2, we provide context for the project and re-

view how information systems (IS) have been applied

to improve clinical and genomic data management.

Section 3, we detail the CMN, the conceptual model

that deﬁnes the NBL domain as the result of meet-

ings with experts in the ﬁeld. Section 4 presents the

result of integrating the CMN with the Conceptual

Model of the Human Genome (CMHG) to expand the

genetic component and support the advancements in

the ﬁeld of the PM. In Section 5, we introduce Clin-

GenNBL, an IS for clinical data management based

on the CMN. Finally, Section 6 concludes the article

with the contribution of this work and outlines future

lines of research.

2 PREVIOUS AND RELATED

WORK

This section provides context for understanding the

work carried out by presenting the background on

which this work is based in Section 2.1 and a sum-

mary of related works in Section 2.2.

2.1 Previous Work

The project of this work started in 2018, with

the ﬁrst version (V1) of the CMN proposed by

Arevshatyan, Reyes Roman(co-author) and collabo-

rators (Arevshatyan et al., 2019), deﬁning the domain.

Their work highlighted the importance of building

a Genomic Information System (GeIS) upon a con-

ceptual model. Later, in 2020, the CMN was im-

proved (V1.5), and the ﬁrst version of the IS based on

this model was designed and developed (Arevshatyan

et al., 2020). It was adapted to the needs of the Pedi-

atric Oncology department of the HUPLF.

The clinical staff of the HUPLF used the prototype

to store patient data, diagnostic evaluations, and treat-

ments. However, this prototype has become obsolete

due to the lack of maintenance and the new needs of

the clinical staff. Most of the technologies used in

the prototype (jQuery, Bootstrap, Jasmine-Express)

are no longer maintained or have been replaced by

new ones. Moreover, the prototype does not have a

responsive design and has a lot of inconsistencies in

its user interfaces.

In recent years, new technologies have emerged

that facilitate development and maintenance. This

analysis, part of the work associated with this re-

search, culminated in a technical report that provides

a comprehensive analysis (Fern

andez Garc

ıa et al.,

2022). Rather than attempting to ﬁx all the proto-

type’s ﬂaws, a new version of the IS will be devel-

oped, incorporating best practices of modern tech-

ENASE 2025 - 20th International Conference on Evaluation of Novel Approaches to Software Engineering

nologies and based on an updated version of the

CMN.

Therefore, this work aims to generate the V2 of

the CMN using conceptual modeling techniques, as

they allow for an improved understanding of the do-

main and facilitate the treatment of complex data,

such as those about NBL. Based on this CMN, we

design and develop a new version of the IS, Clin-

GenNBL, that adapts to the new requirements de-

manded by the clinical staff. Finally, the resulting IS

must be implemented with novel but stable technolo-

gies to improve the development and maintenance

time.

2.2 Related Work

In order to develop a better solution, an analysis of

research related to this research work has been car-

ried out. This section presents several works that be-

long to the context of bioinformatics and/or aim to

design and develop an IS based on conceptual mod-

els. One of the areas where the use of conceptual

models has been most inﬂuential is in the human

genome. Since the contribution by Paton (Paton et al.,

2000), which introduced a collection of genomic (i.e.,

implementation-independent) conceptual data mod-

els for genomic data. These conceptual models are

amenable to (more or less direct) implementation on

different computing platforms.

Following the above work, numerous papers ap-

plying conceptual modeling techniques emerged. An

example of this is the work of Martin et. al., whose

goal is to design and develop an IS that integrates hu-

man genome variant data from different sources (Mar-

tin and Celma, 2011). To achieve this goal, the deﬁni-

tion and categorization of variations is uniﬁed through

conceptualization. Once the conceptual model was

established, a database (Human Genome Data Base,

HGDB) was implemented.

In the Ph.D. thesis of Burriel Coll (Burriel Coll,

2017), the aim is to solve the problem of heterogene-

ity and dispersion of genomic data on breast cancer.

Within this domain, the biggest challenge is the treat-

ment of a high volume of data, as it faces a disease

with high prevalence. It aims to analyze the informa-

tion available in the databases and integrate it into an

IS, thanks to creating a CM.

Focusing the context on genomics, Bernasconi

et. al. present an article that proposes a conceptual

model of genomic metadata, whose purpose is to fa-

vor queries to experimental databases. Their work

can be divided into three phases. The ﬁrst phase an-

alyzes the attributes of genomic metadata. The sec-

ond phase uses an up-bottom method to build the in-

tegrated schema. The last phase corresponds to vali-

dating the conceptual model with different databases

(Bernasconi et al., 2017).

The doctoral thesis of Reyes Rom

an (2018) de-

ﬁnes a framework focused on using CM. It proposes

to use an CMHG as a fundamental basis for generat-

ing Genomic Information Systems (GeIS), intending

to facilitate a conceptualization of the domain that al-

lows to reach accurate knowledge and to be able to

reach PM (Reyes Rom

an and Pastor, 2018). Within

the context of NBL is the work of Arevshatyan, which

analyzes and integrates clinical and genomic data. It

highlights that a Genomic IS based on a conceptual

model allows for improved adaptation to new domain

requirements. It also greatly simpliﬁes the integra-

tion and management of heterogeneous and homo-

geneous data (Arevshatyan et al., 2019). Within a

broader clinical context (not only NBL), Arevshatyan

et. al. present a practical experience of data analysis

and decision-making process where a CM is designed

to develop an IS to manage clinical, pathological, and

molecular data in an integrated manner in the oncol-

ogy department of two hospitals (Arevshatyan et al.,

2020).

As a result of the COVID-19 disease, numerous

databases were created to mitigate the effects of the

pandemic. Within this context, Bernasconi et. al. re-

views the data integration efforts needed to access and

search SARSCoV2 genome sequences and metadata

deposited in the most important viral sequence repos-

itories (Bernasconi et al., 2020). The paper applies

conceptual modeling techniques to structure and or-

ganize the information.

Finally, the work of Garcia et. al. faces the chal-

lenge of integrating a huge amount of existing omic

data (Garc

ıa Sim

on et al., 2021). To this end, concep-

tual modeling techniques are applied to facilitate un-

derstanding, communication, and problem-solving in

the domain while establishing a common ontological

framework to stimulate the communication and evolu-

tion of complex domain knowledge. The work’s main

contribution is the presentation of the Genome Con-

ceptual Schema, independent of the species, which al-

lows the representation of proteomic data and facili-

tates their integration.

3 CONCEPTUAL MODEL OF

NEUROBLASTOMA (CMN)

Conceptual Modeling (CM) is a technique used to un-

derstand a problem domain and communicate effec-

tively with users and subject matter experts. It has

been beneﬁcial in homogenizing data in the clinical

A Conceptual Model-Based Application for the Treatment and Management of Data in Pediatric Oncology: The Neuroblastoma Use Case

context, as discussed in Section 2.2. In this section,

we present the V2 of the Conceptual Model of Neu-

roblastoma (CMN).

Figure 1 shows the CMN resulting from the do-

main study and collaborative meetings with clinical

experts at HUPLF. We conducted a preliminary lit-

erature review and did not ﬁnd any other conceptual

models speciﬁcally addressing the neuroblastoma do-

main. Based on a previous CMN (Arevshatyan et al.,

2020), the analysis of each of the classes has been car-

ried out, structured in views, and extended to adapt it

to the new knowledge of the domain. Then, the model

is proposed to represent all the procedures performed

by the clinical staff throughout all the stages of patient

care.

The different views of the model shown in Figure

1 are discussed below:

• Patient View: It contains all demographic and

general information about the patient, including

their status throughout the course of the disease.

• Episode View: Encompasses all services provided

to the patient with the medical problem within

a speciﬁc period of time. This view includes

the protocol to which these services belong and

all information regarding symptoms, tumors, and

metastases.

• Treatment View: Represents all possible treat-

ments the patient can have. The subclasses rep-

resent the special types of treatments the patient

receives depending on the type of stage and how

it progresses.

• Test view: Represents the various diagnostic tests

physicians perform to detect and diagnose the dis-

ease.

• Genetic View: Contains all the genomic informa-

tion about the patient obtained through the perfor-

mance of genetic tests.

The Patient View is composed of all necessary pa-

tient data. It has been placed at the top as the rest of

the views depend on the existence of the patients. Be-

fore creating any patient, we must have at least one

”Hospital” in the system.

Once this requirement is fulﬁlled, it is possible

to start registering patients who belong to the class

”Patient”. This will occur on the ﬁrst visit to the

hospital center, where the basic patient data, such as

name, date of birth, gender, weight, height, etc., will

be taken.

In addition to the basic data, every patient should

have a history of conditions over time. This history

is represented by the class ”Status”. A patient may

be alive, either with a tumor, ”AliveWithTumor” or

without it, class ”AliveFreeDisease”, and have local

relapses or metastases in the course of its evolution,

in which case it will be necessary to indicate its type,

included in the class ”Relapse”, or mark the disease

progression as death cause, represented by the class

”Deceased”.

The Episode View is the central pillar of the model

and around which all other views revolve. In the hier-

archy of views presented in Figure 1, it is in the center,

not just by chance, as the patient’s evolution is related

to the episodes. This is the one in charge of keeping

the history of diagnostic tests, treatments, symptoms,

etc.

Each time the patient is evaluated or receives any

treatment, an episode is created in the system. This

episode can be of a special type called a diagnosis,

manifested in the diagram with the class ”Diagnosis”,

which represents the event for which the patient ﬁrst

goes to the hospital and the NBL diagnosis process

begins.

At the beginning of his assessment, the applica-

tion user will introduce the patient’s symptoms, repre-

sented in the CMN by the class ”Symptom”, through

a description and a date of the different symptoms of

the patient.

When diagnosing the patient, a primary tumor, de-

scribed by the class ”Tumor”, must be added. The ap-

plication user will specify the tumor’s type, location,

size, and date for each tumor. If a metastasis is de-

tected in a patient, it will be reﬂected through the en-

tity with that name and which contains the attributes

of type and date.

The episodes will be associated with a certain

phase, identiﬁed as ”Phase” in the model, correspond-

ing to a protocol called ”Protocol” in the model. A

medical protocol is a set of recommendations on the

diagnostic procedures to be used for any patient with

a given clinical condition. Or on the most appropriate

therapeutic approach to a clinical diagnosis or health

problem (UNITECO, 2019). Each patient receives

treatment based on a single protocol. These have a

duration and are renewed over the years.

Once the patient’s main data have been entered,

several diagnostic tests are performed to determine

the patient’s stage, making it possible to establish with

greater precision the treatment that the patient will re-

ceive.

Imaging methods, represented by the class ”Ra-

diological”, store all relevant data from tests such

as X-rays, ultrasound scans, and CT scans. Nuclear

medicine encompasses information on two tests in

each entity and is represented by the class ”Nuclear”.

The different laboratory tests that can collect infor-

mation on the blood or urine tests performed on the

ENASE 2025 - 20th International Conference on Evaluation of Novel Approaches to Software Engineering

Figure 1: Conceptual Model of NBL.

patient are represented in the class Biological. The

study of audiometry tests, represented by the class

with the same name ”Audiometries” and the assess-

ment of minimal residual disease (class ”Minimal-

Residual”), which stores the result of different tests

performed on the patient.

Histological tests correspond to those tests that

use a histological technique. This is the series of or-

A Conceptual Model-Based Application for the Treatment and Management of Data in Pediatric Oncology: The Neuroblastoma Use Case

dered steps that prepare the tissue for observation un-

der the microscope (Guerrero Alquicira et al., 2017).

These tests include:

• Bone marrow evaluation, represented by the class

”HistologicalMarrow”, a test that records the

method or technique used, its status, whether it

has been positive or not, and whether it has been

evaluable or not.

• Tumoral evaluation, represented by the class

”HistologicalTumor”, it stores genetic informa-

tion following a study on genetic information after

a study has been performed on a tumor sample.

Such a study can generate instances of the class

”GeneticTest”, which will be accompanied by a

result represented by the class ”GeneticResult”.

Finally, clinicians will rely on genetic testing,

which studies the different variants of the patient.

These types of tests are studies on genes and chromo-

somes, represented as ”Gene” and ”Chromosome”,

which are divided into various tests, such as the study

of the ampliﬁcations of a gene, represented by ”Am-

pliﬁcation”, its possible variants that appear in the

classes ”CNV” and ”Segmental” and, ﬁnally, the mu-

tations of the gene, represented by ”Mutation”.

Once the protocol to be followed for the patient

has been decided based on all the diagnostic tests, it

is the turn of the treatments, whose basic informa-

tion (date on which they were applied, the reason,

the protocol to which they belong, and whether or

not there were any complications, etc.) is available

in the class ”Treatments”. The treatments will always

be speciﬁc, i.e., they will always belong to surgery

(class ”Surgery”), chemotherapy (class ”Chemother-

apy”), autologous transplant (class ”TransplantAut”),

radiotherapy (class ”Radiotherapy”), or ”MIBG”.

Each treatment may cause different toxicity in

the patient (cardiac, nausea/vomiting, fever, infection,

etc.). For this reason, it is decided to represent the

different types of toxicities that may arise from each

treatment by the class ”Toxicity”, specifying the type

using its derivatives. In addition, toxicities can have a

degree and a speciﬁc description, represented by the

class ”CtcGrade”.

In summary, this section has described V2 of

CMN, supported by clinical experts at HUPLF. It

includes several views for patients, episodes, treat-

ments, tests, and genetics, providing a base to man-

age clinical data for neuroblastoma in common terms.

However, the genetic component can still be extended

with other conceptual models. This task is done in

Section 4. Numerous model-based applications can

be built upon the CMN, for example, ClinGenNBL

(Section 5).

4 INTEGRATION OF THE

CONCEPTUAL MODEL OF

NEUROBLASTOMA (CMN)

AND THE CONCEPTUAL

MODEL OF HUMAN GENOME

(CMHG)

As exposed in Section 1, PM is the pathway to effec-

tively and efﬁciently treat individuals suffering from

NBL. It consists of applying special treatments de-

pending on the characteristics of the patient. In this

context, DNA is what deﬁnes humans the most. To

this end, we focus on the genetic component of the

CMN by integrating the CMN with the Conceptual

Model of Human Genome (CMHG), which we detail

in this section.

The CMHG used is version 2 from the thesis

of Reyes Rom

an, which features a chromosome-

centered view (Reyes Rom

an and Pastor, 2018). Be-

low is a summary of the objectives of each view cor-

responding to the CMHG, which will be incorporated

into the CMN presented previously in this work (Sec-

tion 3) and that will give rise to the new model, CMN-

CMHG.

• Structural View: As its name indicates, it de-

scribes the structure of the genome.

• Transcript View: Shows the components and con-

cepts related to protein synthesis.

• Variation View: Models the knowledge related to

the differences found in the DNA sequences of

different individuals.

• Phenotype View: Represents the phenotypes as-

sociated with one or several DNA variations. The

phenotype is the set of physical, biochemical, and

behavioral characteristics that can be observed.

• Bibliographic View: View provides information

about the data sources from which the data to be

stored in the model has been extracted.

• Pathway View: Composition of processes com-

posed of chemical reactions that take place inside

a cell.

The CMN-CMHG integration can be found here

In the model presented, we ﬁnd three types of colors

that allow us to quickly and efﬁciently distinguish the

different parts where integration occurs. Blue marks

the existing classes in both models. Purple represents

the existing classes in the CMN that the inclusion of

the new model has modiﬁed. Finally, green represents

https://doi.org/10.5281/zenodo.13152470

ENASE 2025 - 20th International Conference on Evaluation of Novel Approaches to Software Engineering

the new connection classes, i.e. the ﬁrst point of con-

tact between the CMN and the CMHG.

The following subsections provide a detailed ex-

planation of how the integration took place. Some

views of the CMHG presented intersect with the Ge-

netic View in the CMN, leading to segregation of this

view (Subsection 4.1), while others introduce new

knowledge to the CMN (Subsection 4.2).

4.1 Extension of Knowledge:

Segregation of the Genetic View

All genetic information is grouped under the Genetic

View in the CMN. When this view is compared with

the CMHG for the ﬁrst time, it is divided into four

new views: the Variation View, Structural View, Tran-

scription View, and Bibliographic View, using the

same names as in the CMHG.

The ﬁrst new view comes from the variations, rep-

resented by the ”Variation” class in both models. This

class becomes the main class of the new Variation

View by adding the attributes provided by the CMHG

(clinical importance, private, and version of creation).

In addition to this class, the inheritance hierarchy of

the classes is maintained in both models.

The contribution on the Variation View with a

higher knowledge value is the extension of structural

polymorphic variants. Represented by the ”Struc-

tural” class in the CMN, is replaced by the class

”SNP” (Single Nucleotide Polymorphism) of the

CMHG. This renaming is closely related to data stor-

age, e.g. the use of dbSNP (Sherry et al., 2001).

Continuing with the segregation of the Genetic

View, the second new view is the Structural View, de-

rived from chromosomes. The ”Chromosome” class

is the core around which the CMHG revolves. The

”Chromosome” class now includes the sequence at-

tribute. Additionally, the ”Species” class has been in-

cluded to determine the family to which each chro-

mosome belongs; the points in the sequence where

recombinations occur, represented by the ”Hotspot”

class; and the cytogenetic bands, represented by the

”Cytoban” class, which provide characteristics of the

chromosome and are, in turn, related to the imprecise

subtype of the variant, in the ”Imprecise” class.

In the Structural View, special attention is given

to the elements of a chromosome. Where there was

previously a direct connection between chromosomes

and genes, the model is now enriched with transcrib-

able elements, conserved regions, and regulatory ele-

ments. All information related to the regions is con-

centrated in the Transcription View, the third new

view.

Of the three types of elements, special attention

is given to the transceptors, represented by the class

”TranscriptableElment” whose purpose is to represent

a region of DNA that can be transcribed. In turn, these

regions can be specialized into two types:

1. Gene: a concept that we ﬁnd in the CMN enriched

with new attributes, and more speciﬁcally, biotype

determines the specialization of the various types

of genes. of genes.

2. Exon: each of the transcribable elements that

form part of the gene.

The relationship between these two classes is made

through the transcripts, i.e. the ”Transcript” class rep-

resents the different transcripts present in a gene and

comprises a series of exons.

Finally, we ﬁnd the last new view, the Bib-

liographic View. Similarly to what occurred in

the CMN, this view provides information about the

source of the data, meaning where the information to

be stored in the model has been obtained. It retains

the data source, represented by the ”DataBank” class,

its version (”DataBankVersion” class), and the rela-

tionship between genes and the data source, shown

in the ”ElementDataBank” class. Notably, this ele-

ment’s relationship has changed from the Gene class

to its predecessor, ”ChromosomeElement”.

4.2 New Knowledge: Phenotype and

Pathway

During the integration process, entirely new views

were identiﬁed, meaning any element of the CMN

did not represent them. Therefore, the introduction of

these new views will provide previously unavailable

knowledge.

One of the most important contributions made in

the model’s V2 extension was the inclusion of the

Phenotype View. The phenotype results from genes

that can be expressed and the external factors that af-

fect their expression: environmental, nutritional, and

chemical factors.

This contribution, introduced in the context of

NBL, will endow the model with a combined

genotype-phenotype perspective within diagnostic

evaluations. This means that the model can represent

phenotypic expressions for a patient.

Just as in the CMN we ﬁnd parts unaffected by

integration, there is the list of metabolic pathways. A

metabolic pathway (”Pathway” class) is a succession

of chemical reactions that take place inside the cell.

In summary, in this section, we have presented

the integration of the CMN and the CMHG into a

uniﬁed framework (CMN-CMHG), with an expanded

genetic view that can serve as a common language

A Conceptual Model-Based Application for the Treatment and Management of Data in Pediatric Oncology: The Neuroblastoma Use Case

for communication in the domain of genetic Preci-

sion Medicine (PM) applied to neuroblastoma (NBL).

It includes information on chromosomes, variations,

phenotypes, bibliographic data, and pathways, among

other elements. This integration also facilitates inter-

operability between different PM and NBL tools, of-

fering deeper insights into the ﬁeld.

5 CLINGENNBL: A CMN-BASED

APPLICATION

This section introduces ClinGenNBL, the solution

generated based on the conceptual model deﬁned in

Section 3. On the one hand, this CM has helped im-

prove communication between the various multidisci-

plinary experts during the requirements deﬁnition. On

the other hand, it has served as ontological support in

creating software solutions developed with innovative

technologies and great support from the SE commu-

nity. All this is to provide an IS that responds to the

needs of clinical experts in NBL and has an appro-

priate degree of maturity for its application in medi-

cal practice. This section describes the solution de-

signed, developed, and validated. It also presents sev-

eral screenshots from extensive use involving more

than 800 real patients.

5.1 Requirements

The ﬁrst phase of the solution design consisted of

gathering the functional needs from the requirements

analysis sessions conducted with the HUPLF clinical

experts. The functional and non-functional require-

ments were analyzed and converted to small pieces of

information using the use case analysis. ClinGenNBL

allows doctors to manage and analyze the clinical data

associated with NBL with a user-friendly interface.

As introduced in this section, it is based on the CMN;

however, some concepts are not part of the require-

ments as they are not needed by the experts at the time

of the writing of this article.

Among its features, the data analytics section

should be noted, which includes the following ca-

pabilities: i) Generate a patient state report in CSV

format, ii) Search neuroblastoma hospitals by allow-

ing the user to search for hospitals, iii) Filter neurob-

lastoma protocols and hospitals by any property, iv)

Show the analysis of distributions in which patients

are found (INRG

or INSS

) and the number of pa-

tients associated with each status, v) Display the chro-

International Neuroblastoma Risk Group.

International Neuroblastoma Staging System.

mosome alteration analysis in a table composed of pa-

tient names, applied protocol, date of birth and diag-

nosis, INSS, as well as Del1p

, vi) Show gene ampli-

ﬁcation analysis composed of patient names, protocol

applied, date of birth and diagnosis, INSS, as well as

copies of the MYCN gene.

More information about the requirements

can be found in the following technical report

(Fern

andez Garc

ıa et al., 2022) (Spanish), which

includes comprehensive documentation covering all

the information, parameters, and options managed

by the developed system. During the design of an

application interface, mockups are common, which

allow us to present a possible approach to the ﬁnal

design.

5.2 Architecture

To ensure that an IS meets the needs of stakeholders

and has the necessary level of maturity for a medical

application, technologies that are widely used in the

software development industry and that receive the

support and backing of very important companies at a

professional level have been used.

ClinGenNBL consists of three components: fron-

tend, backend, and database. By decoupling the logic

and model (backend) from the view (frontend), we

achieve greater ﬂexibility and increased productivity

as we get a greater separation or composition of the

application. Figure 2 shows the architecture repre-

sented in a UML component diagram. Below, the

components are detailed:

1. Backend: NodeJS has been used as the JavaScript

execution platform. This level of abstraction aims

to provide an interface, usually an API, which al-

lows easier interaction with the database or other

external services.

2. Frontend: It consists of the user interface deﬁned

in ReactJS, coded in a proprietary language called

JSX, an extension of the JavaScript programming

language. Other libraries, such as i18n, have been

used as well. It was decided to use React as the

main tool for client-side development since it is

one of the most widely used web development li-

braries today. This component interacts with the

backend through a REST API, which uses the

HTTP protocol and exchanges messages in JSON

format.

3. Database: Object-Relational Mapping (ORM) has

been used, a technique that allows the manipu-

lation of database data using an object-oriented

Genetic deletion affecting the short arm of chromo-

some 1.

ENASE 2025 - 20th International Conference on Evaluation of Novel Approaches to Software Engineering

paradigm. In this case, it was decided to use

Sequelize

, a library written in TypeScript for

Node.js, which allows the deﬁnition of models

and relationships, the database query and the han-

dling of these models in the form of objects. In

addition, this library is compatible with many

DBMSs, including MySQL8

, which has been

chosen to store all the data related to the applica-

tion since it is a very mature system used through-

out the history of software application develop-

ment.

5.3 Validation

Validating the solution means justifying that it would

contribute to the stakeholders’ objectives if imple-

mented. In the Engineering Cycle (EC), validation is

performed before implementation. It involves investi-

gating the effects of the interaction between an artifact

prototype and the problem model, always compar-

ing them with the solution requirements (Wieringa,

2014).

Expert opinion is one of the simplest ways to val-

idate the solution (Wieringa, 2014). The design of

the solution is submitted to a committee of experts,

in this case, different physicians and doctors from the

Clinical and Translational Research Group in Cancer

(GICT-Cancer) of the HUPLF/IIS La Fe, who imag-

ine how it will interact with the context problem (NBL

data management) and try to predict the effects they

think the solution will have.

The main objectives of these meetings are: i) to

obtain the maximum knowledge of the domain and

the research problem and ii) to consolidate and vali-

date the progress on developing the solution. The ﬁrst

meetings held by the HUPLF team of experts focused

on deﬁning the requirements and studying the prob-

lems they were facing daily. Once analyzed, priori-

tized, and classiﬁed, they were shared along with the

conceptual model with the team to validate them. The

rest of the meetings that were agreed upon with the

group focused on reviewing the application’s progress

through demonstrations. Each of them had a deliver-

able associated with the progress of the functionalities

implemented since the previous meeting. Finally, to

test the ﬁnal developed software, the expert opinion

exercise was divided into two stages: i) User observa-

tion and ii) User interviews.

In the user observation phase, the experts of the

HUPLF were asked to perform a set of predeﬁned

tasks within the developed IS. The tasks that were

Sequelize. https://sequelize.org/

MySQL. https://www.mysql.com/

deﬁned following the application requirements in-

cluded.

• Hospital registration: The user must register a

new hospital, display the list of existing hospi-

tals, and search for the new hospital by two cri-

teria ”Name” and ”City”.

• Registration and protocol query: The user must

protocols, and add the different phases that make

up the protocol.

• Patient registration: The user must register a new

patient, display the list of existing patients, and

search for the new patient by MRN

• Symptom, tumor, metastasis registration: The

user must register a new symptom, tumor, and

metastasis on the previously created patient.

• Diagnostic evaluation registration: The user must

of diagnostic evaluations.

• Treatment Discharge: The user must discharge at

least once each treatment, including all types of

toxicity.

• Consultation of patient information: The user

must display the data entered in the previous

tasks.

• Correction of patient errors: The user will have to

edit one of the attributes of each entity and ﬁnally

delete them.

The objective of the user interview is to know the

user’s opinion regarding the use of the application af-

ter performing the previously deﬁned tasks. The com-

ments can be summarized in four relevant points:

• Improvement of daily work: Using the IS dur-

ing patient diagnosis and treatment allows them to

offer better care and achieve Precision Medicine

or Personalized Medicine. In addition, the IS al-

lows them to perform more efﬁcient and effective

queries on patient data, speeding up the data ex-

ploitation process.

• The tool is easy to use: Team members found the

application easy to use, as they could recognize

and locate most of the tasks they were asked to

perform without assistance.

• Stakeholders’ objectives are met: The team of ex-

perts expressed that they were very satisﬁed with

the application. They indicated that the data man-

agement would help improve the treatment pro-

vided to patients, as they would be able to analyze

and extract all the data.

Medical Record Number.

A Conceptual Model-Based Application for the Treatment and Management of Data in Pediatric Oncology: The Neuroblastoma Use Case

Figure 2: Solution architecture.

Figure 3: Screenshot of ClinGenNBL: Sumamry.

• Possible improvements to the application to facil-

itate its use were identiﬁed: The team of clinical

experts suggested possible visual improvements

that would facilitate the process when entering

data: the grouping of certain toxicities and the cre-

ation of shortcuts in the creation of lookup tables.

In general, stakeholders reported that the use of

the application was beneﬁcial. The tool allows them

to achieve their goals and improve the treatment of

NBL easily and intuitively. They also mentioned that

adopting the tool would take little time and could be

extended to more hospitals.

In conclusion, homogenizing data related to NBL

treatment and its access through a common IS pro-

duces greater satisfaction and beneﬁts than manual

and dispersed treatment (through different applica-

tions). The validation has provided encouraging ﬁnd-

ings, but the results should be studied further.

5.4 Using ClinGenNBL

Thanks to the extensive knowledge of experts in var-

ious areas relevant to this work, such as Information

Systems Engineering (ISE), Bioinformatics, and Web

Engineering, among others, it has been possible to

achieve a high degree of maturity (TRL5 (European

Comission, 2014)) in both the initial design and the

ﬁnal result of the application. It is ready to be trans-

ferred to HUPLF. We present several screenshots of

the platform (Figures 3 and 4) and a demonstration

video, where 831 real patients with relevant data were

introduced. The demonstration can be found here

https://doi.org/10.5281/zenodo.13138339

ENASE 2025 - 20th International Conference on Evaluation of Novel Approaches to Software Engineering

Figure 4: Screenshot of ClinGenNBL: Manage Patient.

All the clinical data was entered in compliance with

GDRP

legislation (GDRP, 2016).

6 CONCLUSIONS

Finally, to conclude, Section 6.1 summarizes the con-

tribution of this work to the deﬁned goals.

6.1 Contribution

This research aimed to generate a conceptual model

for the neuroblastoma domain and to implement an

information system (IS) based on it, using conceptual

modeling techniques to address the identiﬁed prob-

lems. This work has contributed to the development

of the V2 of the CMN and the implementation of ro-

bust software based on it (ClinGenNBL). Clinical ex-

perts continuously validate this solution at a technical

and functional level. Additionally, it has highlighted

the genetic aspects of the domain to advance preci-

sion medicine through the integration of CMN and

CMHG.

Firstly, the CMN improves the understanding of

the domain and establishes a solid foundation on

which to build the solution design. The knowledge

obtained from the domain study can be integrated

and centralized. It also improves stakeholder com-

munication, which is very useful during the require-

ments gathering and validation phase. Secondly, Clin-

GenNBL serves as an information system instance of

General Data Protection Regulation.

the CMN and provides greater efﬁciency to the clin-

icians at HUPLF in managing clinical data and gen-

erating knowledge about neuroblastoma. It comprises

two independent components implemented with mod-

ern and consolidated technologies such as NodeJS,

React, and MySQL. Experts have successfully vali-

dated and tested it with more than 800 real patient

cases.

6.2 Future Work

CMN has been developed to a stable version (V2),

and ClinGenNBL has reached a high degree of matu-

rity (TRL5). However, there is still more work to be

done. Future work includes:

• Maintenance and evolution of the application:

The initial application has been validated satisfac-

torily with the expert opinion of the HUPLF doc-

tors. Based on the needs detected and the possible

improvements identiﬁed, future developments are

proposed to address new functionalities for the ap-

plication.

• Implementation of the integration of the CMN-

CMHG: Given the knowledge gathered through-

out the work, the integration between both models

(CMN and CMHG) is proposed in Section 4 but

not implemented in the application. However, the

integration design and development could be part

of a future engineering cycle.

• Implementation of the tool in the real working en-

vironment: The next step for a successful evolu-

tion is using the application in a realistic environ-

ment. To this end, implementing the HUPLF is

A Conceptual Model-Based Application for the Treatment and Management of Data in Pediatric Oncology: The Neuroblastoma Use Case

only the ﬁrst step, with the possibility of taking

the IS to the rest of the national hospitals and ex-

panding internationally. This must be done in col-

laboration with the technical service of each de-

partment in order to follow the protocols estab-

lished by the healthcare organization.

ACKNOWLEDGEMENTS

The authors would like to thank the PROS Re-

search Center Genome group for their valuable

discussions on the application of CM in medicine,

as well as the staff of the Medical and Pediatric

Oncology Service at Hospital La Fe for their sup-

port and commitment to facilitating our work and

enhancing our understanding of the domain. This

work was supported by the Generalitat Valenciana

through the CoMoDiD project (CIPROM/2021/023),

and the Spanish State Research Agency through

the SREC project (PID2021-123824OB-I00),

MICIN/AEI/10.13039/501 100011033 and coﬁ-

nanced with ERDF and the European Union Next

Generation EU/PRT.

REFERENCES

Arevshatyan, S. et al. (2019). Integration and analysis of

clinical and genomic data of neuroblastoma applying

conceptual modeling. IEEE Journal of Biomedical

and Health Informatics.

Arevshatyan, S. et al. (2020). An application of an ehr based

on conceptual modeling to integrate clinical and ge-

nomic data and guide therapeutic strategy. ANALES.

Bernasconi, A., Canakoglu, A., Masseroli, M., Pinoli, P.,

and Ceri, S. (2020). A review on viral data sources and

search systems for perspective mitigation of covid-19.

Brieﬁngs in Bioinformatics, 22(2):664–675.

Bernasconi, A. et al. (2017). Conceptual modeling for

genomics: Building an integrated repository of open

data. In Conceptual Modeling, pages 325–339, Cham.

Springer International Publishing.

Burriel Coll, V. (2017). Dise

no y Desarrollo de un Sistema

de Informaci

on para la Gesti

on de Informaci

on sobre

ancer de Mama. Universitat Polit

ecnica de Val

encia.

Cahaney, C., Dhir, A., and Ghosh, T. (2022). Role of preci-

sion medicine in pediatric oncology. Pediatric Annals,

51(1):e8–e14.

Castleberry, R. P. (1997). Neuroblastoma. Eur J Cancer,

33(9):1430–7; discussion 1437–8.

nete, A. et al. (2022). Neuroblastoma in spain: Link-

ing the national clinical database and epidemiological

registries – a study by the joint action on rare cancers.

Cancer Epidemiology, 78:102145.

nete Nieto, A. et al. (2023). C

ancer infantil en Espa

na.

Estad

ısticas 1980-2023. Registro Espa

nol de Tumores

Infantiles (RETI-SEHOP). Universitat de Val

encia,

Valencia.

Digital Transition. The digital transition in health-

care: An urgent need. En l

ınea. Available

at: https://digitalforeurope.eu/the-digital-transition-

in-healthcare-an-urgent-need?lang=en. Last access:

28 of October 2024.

European Comission (2014). Technology readdiness levels.

European Commission Decision C (2014)4995 of 22

July 2014). Online. Available at: https://ec.europa.eu.

Evans, R. S. (2016). Electronic health records: Then, now,

and in the future. Yearbook of Medical Informatics,

25(S 01):S48–S61.

Fern

andez Garc

ıa, F., Reyes Rom

an, J. F., Baghiu, C.-

B., P

erez, S., and Pastor L

opez, O. (2022). Clin-

gennbl (v2): Integration and analysis of neurob-

lastoma clinical and genomic data. Available at:

http://hdl.handle.net.

Garc

ıa Sim

on, A., Palacio Le

on, A., Reyes Rom

an, J. F.,

Casamayor R

odenas, J. C., and Pastor, O. (2021). A

conceptual model-based approach to improve the rep-

resentation and management of omics data in preci-

sion medicine. IEEE Access, 9:154071–154085.

GDRP (2016). Regulation (EU) 2016/679 of the European

Parliament and of the Council of 27 April 2016 on the

protection of natural persons with regard to the pro-

cessing of personal data and on the free movement of

such data, and repealing Directive 95/46/EC (General

Data Protection Regulation). Online.

Guerrero Alquicira, R., Rojas Lemus, M., and Fortoul

van der Goes, T. (2017). Histolog

ıa y biolog

ıa celular.

MCGRAW-HILL.

Martin, A. and Celma, M. (2011). Integrating human

genome variation data: An information system ap-

proach. In 22nd International Workshop on Database

and Expert Systems Applications, pages 65–69.

McCabe, M. G. et al. (2024). Precision medicine for

childhood cancer: Current limitations and future per-

spectives. JCO Precision Oncology, 8(8):e2300117.

PMID: 38207228.

Paton, N. et al. (2000). Conceptual modelling of genomic

information. Bioinformatics, 16(6):458–557.

Reyes Rom

an, J. F. and Pastor, O. (2018). Dise

no y desar-

rollo de un sistema de informaci

on gen

omica basado

en un modelo conceptual hol

ıstico del genoma hu-

mano. PhD thesis, Universitat Polit

ecnica de Val

encia.

Sherry, S. T., Ward, M.-H., Kholodov, M., Baker, J., Phan,

L., Smigielski, E. M., and Sirotkin, K. (2001). dbSNP:

the NCBI database of genetic variation. Nucleic Acids

Research, 29(1):308–311.

Siegel, R. L., Giaquinto, A. N., and Jemal, A. (2024). Can-

cer statistics, 2024. CA: A Cancer Journal for Clini-

cians, 74(1):12–49.

Trujillo, J. C. et al., editors (2018). Conceptual Model-

ing: 37th International Conference, ER 2018, Pro-

ceedings. Springer International Publishing.

UNITECO (2019). ¿Qu

e son los protocolos m

edicsos?

Available at: https://www.unitecoprofesional.es.

Wieringa, R. J. (2014). Design Science Methodology

for Information Systems and Software Engineering.

Springer.

ENASE 2025 - 20th International Conference on Evaluation of Novel Approaches to Software Engineering