AN ONTOLOGY SUPPORTING THE DAILY PRACTICE
REQUIREMENTS OF RADIOLOGISTS-SENOLOGISTS WITH
THE STANDARD BI-RADS
Souad Demigha
Centre de Recherche en Informatique (C.R.I), Université de Paris 1 – Panthéon-Sorbonne
90 rue de Tolbiac 75634 Paris Cedex 13
Keywords: Ontology, BI-RADS, radiology-senology, conceptual scheme, conceptual model.
Abstract: This paper presents concepts and relationships allowing the development of an ontology that supports the
daily practice requirements of radiologists-senologists with the standard BI-RADS (Breast Imaging
Reporting and Data System). This ontology aims at describing the radiologic-senologic knowledge shared
by the community of technicians, practitioners, gynecologists, radiologists, surgeons and anatomo-
pathologists. It represents a unifying scope for reducing and eliminating ambiguities as well as conceptual
and terminological disarrays. It also ensures the understanding of the concerned community. It allows
communication and dialogue between members of the scientific community even though they are working
in different fields having different requirements and viewpoints. This ontology allowed us to obtain a
conceptual model of the domain. Details concerning the development of the ontology and the generalization
of the conceptual scheme that leads to the design of the conceptual model are described.
1 INTRODUCTION
Current methods are unable to capitalize and to
reuse knowledge acquired from experience. Re-use
is employed with an ad hoc manner. It is a
traditional technique based on experience acquired
during developments of different systems in a
specific domain. The ad hoc manner is inadequate
because it does not allow to share accumulated
experience, contrarily to what medical experts wish
to obtain. In order to achieve investigations on a vast
number of cases, experts only use their own
experience, even though this is a vast amount of
experience. It would be advisable to gather the
experience of the numerous experts for a shared
utilization. Besides the obvious advantages that
could result from this shared knowledge, it would
allow to homogenize the knowledge on the same
topics by standardizing the vocabulary and
definitions. Identical notions should be labelled
using the same terminology so as to compare them.
It is a well-know fact that in medicine, the
development of specialized ontologies is a
mandatory step for elaboration and maintenance of
increasing a thesaurus and not an ambiguous one, for
the sake of communication between terminologists
(Rector, 1999).
As Gruber (Gruber, 1993) wrote, an ontology is
defined as follows: ‘an ontology is a formal, explicit,
specification of a shared conceptualisation’. It
defines concepts used to describe knowledge, their
relationships and their constraints of use.
We have built an ontology based on the standard
BI-RADS (Breast Imaging Reporting and Data
System) (Chabriais and al, 1998), on the scientific
reports of the EBM (Evidence-based on medicine)
and on the reports and experience of radiologists-
senologists of the Necker Hospital in Paris (France)
in view of representation of radiologic-senologic
knowledge and associated clinical reports (Demigha
and al, 2001).
This ontology is fitted to the description of the
senologic knowledge shared by the scientific
community of technicians, practitioners,
gynecologists, radiologists, surgeons and anatomo-
pathologists. It represents a unifying scope for
reducing and eliminating ambiguities as well as
conceptual and terminological confusions. It also
243
Demigha S. (2007).
AN ONTOLOGY SUPPORTING THE DAILY PRACTICE REQUIREMENTS OF RADIOLOGISTS-SENOLOGISTS WITH THE STANDARD BI-RADS.
In Proceedings of the Ninth International Conference on Enterprise Information Systems - ISAS, pages 243-249
DOI: 10.5220/0002354802430249
Copyright
c
SciTePress
ensures that the concerned scientific community
shares a mutual understanding. It allows
communication and dialogue between members of
this scientific community even if though they are
working in different fields that have different
requirements and viewpoints. This ontology allowed
us to obtain the conceptual model of the domain in
radiology-senology, which is structured as cases
using the case-based reasoning approach.
We have analyzed requirements of radiologists-
senologists with the Department of Radiology of the
Necker Hospital in Paris using the Crews-l’Ecritoire
approach (Cooperative Requirements Engineering
With Scenarios) (Ben Achour 1999). Radiologic-
senologic knowledge is made of both text and
images. We have only considered textual
knowledge; images are just associated to patients'
reports for the sake of information. Analysis
performing has allowed to structure the radiologic-
senologic knowledge according to stringent rules. It
is an original approach to solve the issue consisting
in considering the ontology definition as an
engineering issue requirement.
The paper is organized as follows:
- Section 2 positions our work with respect to
existing ontologies.
- Section 3 presents the acquisition of radiologic-
senologic knowledge using the Crews-l’Ecritoire
approach.
- Section 4 presents in details the steps that allowed
us to construct the ontology in the radiology-
senology domain.
- Section 5 provides the conclusion.
2 STATE OF THE ART
The engineering of ontologies (or ontological
engineering) arose from the will to diversify the
applications of the Knowledge-Based-Systems
(KBS). It allows for a representation of knowledge
that does not depend on these various applications,
so as to ensure its portability from an application to
another (Furst, 2004).
At the present time, there is a relative consensus
on the role of ontologies. This consensus is built
around the Gruber formula. "An ontology, is a
formal, explicit, specification of a shared
conceptualisation ". The construction of an ontology
is a conceptualization work. It consists in
identifying, within a corpus, the knowledge specific
to the field of knowledge to be represented, and
consensually acknowledged as pertaining to this
field. Guarino proposes a four level-classification of
ontologies according to the link between the
ontology and the application (Guarino, 1997). High-
level ontologies describe general concepts while
low-level ones describe concepts that depend on a
domain.
The Crews l’Ecritoire approach is based on the
‘‘Requirement Engineering’’ concept and helps
understanding users needs using a semi-automatic
analysis of textual scenarios, i.e. scenarios written in
natural language. Moreover, Crews permits strong
control and verification of the extraction process.
Starting from a high-level problem statement, it
guides the discovery of a complete hierarchy of
goals illustrated by scenarios in a top-down manner.
The approach is based on a set of guidelines to guide
linguistic analysis and verification of scenarios
written in natural language. Use of natural language
allows radiologists to understand scenario meaning
without having expertise in Crews approach and use.
Section 3 presents the acquisition of radiologic-
senologic knowledge using the Crews-l’Ecritoire
approach.
3 ACQUISITION OF SENOLOGIC
KNOWLEDGE
Crews-l’Ecritoire associates concepts of goals and
scenarios to support the requirement elicitation
(Rolland and al, 1999). A goal is defined as
something that some shareholder hopes to achieve in
the future, it is expressed as a verb with eight
optional parameters, each parameter playing a
different role with respect to the verb, and a scenario
is a possible behavior limited to a set of purposeful
interaction taking place among several agents. A
couple <goal, scenario> is a requirement chunk
(RC).
When a goal is discovered, the approach
proposes to author a scenario (coupling in the
forward direction). Then, the approach analyzes
every scenario to yield new goals (coupling in the
backward direction). Starting from a high-level
problem statement, the Crews-l’Ecritoire approach
guides the discovery of a complete hierarchy of
goals illustrated by scenarios to help writing
scenarios in a top-down manner. The approach is
based on a set of guidelines. These guidelines
consist (1) of automated rules to guide goal
discovery and (2) of guidelines to guide linguistic
analysis and verification of scenarios.
Crews introduces three abstraction levels in RCs
specifications: behavior, functional and physical.
ICEIS 2007 - International Conference on Enterprise Information Systems
244
- The behavioral level identifies services to be
provided (Behavioral level::= <Design Goal, Service
Scenario>).
- The functional level focuses between the system
and users to complete the services (Functional
level : := <Service Goal, System Interaction
Scenario>).
- The physical level deals with the actual
performance of the previous level (Physical level ::=
<System Goal, System Internal Scenario>).
Crews allows to organize RCs according to three
strategies in order to construct the RC model,
namely: refinement using the (‘‘Refined by’’
connector), complementary (‘‘AND’’ connector)
and alternative (‘‘OR’’ connector). This
organization is performed thanks to guidelines and
rules which allow to map RC defined at a given
abstraction level into RCs defined at a lower
abstraction level.
Section 4 presents in details the ontology in
radiology-senology domain.
4 CONSTRUCTION OF THE
ONTOLOGY IN SENOLOGY
This section presents detailed steps that allowed us
to construct the ontology in the domain of radiology-
senology.
4.1 Concepts Underlying the Ontology
This section presents the concepts elicited during our
analysis and explains how to translate actions
defined at the behavioral level into actions defined at
the functional level, then into actions defined at the
physical level. This translation mainly uses the
refinement rule ‘‘Refined by’’ and it results in a
structured network of RCs (see Figure 1). An
example about the use of the model is used to
illustrate the three previously mentioned strategy of
RC organization and to explain how to face with
exceptional conditions.
4.1.1 The Behavioral Level
The behavioral level lists services required at the
highest level in order to achieve a goal. A behavioral
RC couples a ‘‘Design Goal’’ to a ‘‘Service
Scenario’’.
Let us consider the ‘‘Design Goal’’ associated
with the RC ‘‘Performing the patient radiological-
senological process from the case base’’ (see Figure
1). This goal means that the case base must provide
radiologists with a mean to support their
mammography-related activity. In the RC, this goal
is coupled with a ‘‘Service Scenario’’ which
contains four basic steps: (1) Performing clinical
examination phase from the case base, (2)
Performing image reading phase with an icon from
the case base, (3) Performing radiological
interpretation with the BI-RADS glossary from the
case base, (4) Performing anatomo-pathological
examination with the BI-RADS glossary from the
case base. These steps are services requested by
radiologists in the course of the daily practice and
are complementary. They are linked through an
“AND'” connector. In the Crews approach, these
four steps refine the initial RC. As they also define
services, the refinement takes place at the contextual
level. This is allowed by Crews when refined
scenarios have the same semantics as the initial
scenario.
4.1.2 The Functional Level
The functional level refines services defined at the
behavioral level when the refinement can be
expressed in terms of user-oriented tasks. They lead
to ‘‘Service Goals’’ and ‘‘Interaction Scenarios’’.
Let us consider the first refined ‘‘Design Goal’’:
“Performing clinical examination phase from the
case base” (see Figure 1). It generates two actions
that correspond to two ‘‘Services Goals’’: (1)
Performing patient’s interrogation from the case
base and (2) Performing patient’s physical
examination from the case base. These actions are
complementing each other and thus are linked
through an ‘‘AND’’ connector. These actions are
complementing each other and thus are linked
through an ‘‘AND’’ connector.
As for the behavioral level, Crews allows to
refine RC belonging to the functional level into new
RCs defined at the same level. For instance, the
second step: ‘‘Performing patient’s physical
examination from the case base’’ can be refined in
two steps: (1) to record old data into the case base,
(2) to record current data into the case base.
AN ONTOLOGY SUPPORTING THE DAILY PRACTICE REQUIREMENTS OF RADIOLOGISTS-SENOLOGISTS
WITH THE STANDARD BI-RADS
245
SC1Perform the patient radiological-senological process from
the knowledge based system
SC1Perform the patient radiological-senological process from
the knowledge based system
SC1.1
Perform clinical examination phase from the
case base
SC1.1
Perform clinical examination phase from the
case base
SC1.2Perform image reading phase with an icon
from the case base
SC1.2Perform image reading phase with an icon
from the case base
SC1.3
Perform radiological interpretation with the BIRADS
glossary from the case base
SC1.3
Perform radiological interpretation with the BIRADS
glossary from the case base
SC1.4Perform anatomo-pathological exam with the BIRADS
glossary from the case base
SC1.4Perform anatomo-pathological exam with the BIRADS
glossary from the case base
Behavioral Level
AND
AND
AND
Refined by
SC1.1.2Perform physical examination
from the case base
SC1.1.2Perform physical examination
from the case base
SC1.1.1Perform interrogation
from the case base
SC1.1.1Perform interrogation
from the case base
SC1.2.1Locate ROIs on
mammographic images
SC1.2.1Locate ROIs on
mammographic images
SC1.2.2Locate ROIs on
echographic images
SC1.2.2Locate ROIs on
echographic images
Refined by
Refined by
Refined by
AND
AND
Functional Level
SC1.1.2
1
Record patient’s new data SC1.1.2
1
Record patient’s new data
SC1.1.2
2
Record patient’s old
data
SC1.1.2
2
Record patient’s old
data
OR
---------------
SC1.2.2.1.1.1
1
Identify the junior-
radiologist
SC1.2.2.1.1.1
1
Identify the junior-
radiologist
SC1.2.2.1.1.1
2
Identify the expert-
radiologist
SC1.2.2.1.1.1
2
Identify the expert-
radiologist
Refined by
OR
Physical Level
---------------
Refined by
SC1.2.2.
1
Interpret ROIs located on
mammographic images
SC1.2.2.
1
Interpret ROIs located on
mammographic images
SC1.2.2.1.1Display BIRADS glossary
menu for the case author
SC1.2.2.1.1Display BIRADS glossary
menu for the case author
SC1.2.2.1.1.1Display identification menu for
the case author
SC1.2.2.1.1.1Display identification menu for
the case author
Refined by
SC1.2.2.2Interpret ROIs located on
echographic images
SC1.2.2.2Interpret ROIs located on
echographic images
Refined by AND
----------
------------------------------
SC1 is a sub-goal of the Information System (IS) building. SC1.X... indicates the scenario number.
Figure 1: Part of the Requirements Chunks Hierarchy.
As these actions are alternative (i.e. new patient vs.
already registered patient), they are connected by an
‘‘OR’’ connector. The scenario SC1.1.2 illustrates
this refinement process through the refinement of the
scenario SC1.1.2. The scenario SC1.1.2 is defined at
the behavioral level as follows:
Initial state:
The case base is online.
The case author is granted access to the case base.
The case author must perform radiological-
senological process.
1 Perform patient’s interrogation from the case base.
2. Perform patient’s physical examination from the
case base.
Final state:
Data of radiological-senological process are
recorded into the case base.
The case base is consistent.
The second step or action is refined at the functional level
by:
1. Record patient’s new data.
2. Record patient’s old data.
4.1.3 The Physical Level
The physical level refines interactions defined at the
functional level. This refinement can be expressed in
system-oriented tasks. They lead to ‘‘System Goals’’
and ‘‘Internal Scenarios’’.
Let us consider the ‘‘Goal Service” ‘‘display
identification menu to the case author” (see Figure
1). It can be refined in two alternatives (i.e.
connected by an ‘‘OR’’ connector) actions: (1) to
identify the junior-radiologist and (2) to identify the
expert-radiologist. Identification means that the case
author is registered as case author by the case base,
and known by the case base. Access authorization is
then granted by the Database Manager System
(DBMS) according access rights as defined by the
administrator. These can be refined into new Internal
Scenarios. Extensive details for level hierarchy are
presented in (Demigha, 2005).
The following scenario illustrates the procedure
by the refinement of the scenario SC1.2.2.1.1.1:
‘‘display identification menu to the case author”.
SC1.2.2.1.1.1is defined at the functional level by:
1. Display identification menu to the case author.
- - - - - - - - - - -
It is refined at the physical level as follows:
1. The system asks the case author to login.
2. The case author introduces his login.
3. If the code is valid then
4. The system continues the login procedure.
When ‘‘the code is not valid’’, system denies the
author case to access database. This cannot be
expressed in a unique scenario as Crews does not
allow the use of the IF/ELSE/THEN structure into a
scenario. In order to manage such a structure, Crews
offers the concept of exceptional scenario. This
scenario is connected to the ‘‘normal scenario’’ with
an alternative connector. In our example, the
associated connection rule is:
ICEIS 2007 - International Conference on Enterprise Information Systems
246
If code is valid then perform SC1.2.2.1.1.1.1
else perform SC1.2.2.1.1.1.1
1
Exceptional scenario SC1.2.2.1.1.1.1managing ‘‘not
valid code’’ is defined at the physical level as
follows: 1. If the code is not valid then
2. The system denies access.
3. The system is consistent.
4.2 Generalization of the Conceptual
Scheme
This section describes the conceptual scheme
obtained from expressions of the three abstraction
levels and illustrated by examples appropriated to
our analysis. We display how we have extracted the
different RCs, concepts and relations gathering
them. These concepts and relationships constitute
the ontology of radiology-senology domain
according to the definition of Gruber (Gruber, 1993).
We display how we have built the ontology of
the radiology-senology domain via the step-to-step
Crews L’Ecritoire approach. It is an approach based
on engineering requirements and on formal
guidelines and rules. It allows to analyze the user’s
requirements, needs and specifics, it defines the
goals at different levels of abstraction, it defines the
scenarios with different levels of abstraction, thus
resulting in a structured network of goals and
scenarios (see Figure 1).
If we take again the behavioral goal at the
highest level illustrated on Figure 1: Performing the
patient radiological-senological process from the
case base.
This goal is refined according to four
complementary actions: (1) Performing clinical
examination phase from the case base, (2)
Performing image reading phase with an icon from
the case base, (3) Performing radiological
interpretation with BI-RADS glossary from the case
base and (4) Performing anatomo-pathological
examination with BI-RADS glossary from the case
base.
The four basic actions will be now detailed.
Starting from these actions, we will show how we
have extracted the concepts and their relationships
and thus, how to construct the ontology.
Let us consider action 1: SC1.1: Performing the
clinical examination phase from the case base.
One key-concept can be extracted at once: clinical
examination.
Key-concept: clinical examination.
This action generates itself two actions:
- SC1.1.1. Performing patient’s interrogation from
the case base.
- SC1.1.2. Performing patient’s physical
examination from the case base.
Let us consider the first sub-action SC1.1.1.
SC1.1.1. Performing patient’s interrogation from the
case base. We can extract two key-concepts: patient
and interrogation.
Key-concepts: patient and interrogation.
For the second sub-action we can extract two
key-concepts: patient and physical examination.
SC1.1.2. Performing patient’s physical
examination from the case base. We can extract two
key-concepts: patient and physical examination.
Graph of Figure 2 represents the first action of
scenario SC1.
From the second sub-action, we also extract two
key-concepts: patient and physical examination.
We have summarized the different extracted key-
concepts on a graph where the concepts are
represented by labels and linked by relationships
symbolized by lines.
Figure 2: The action 1 of the scenario SC1.
Now, the same operation will be carried out on
the second action. We extract two key-concepts:
reading and image.
SC1.2: Performing the image-reading phase with an
icon from the case base.
Key-concepts: reading and image.
This action generates itself two actions:
- SC1.2.1. Locating ROIs on mammograms.
- SC1.2.2. Locating ROIs on echographic images.
The first sub-action contains two key-concepts:
ROI and mammograms.
The second sub-action contains two key-concepts:
ROI and echographic image.
We have summarized the different extracted key-
concepts on a graph where the concepts are
represented by labels and linked by relationships
symbolized by lines.
Let us regroup the concepts extracted from the
second action in a graph: like for the first action, we
have summarized the different extracted key-
concepts on a graph where the concepts are
represented by labels and linked by relationships
symbolized by lines.
Clinical examination
Interrogation
Physical examination
Patient
Contains
Contains
Is performed on
AN ONTOLOGY SUPPORTING THE DAILY PRACTICE REQUIREMENTS OF RADIOLOGISTS-SENOLOGISTS
WITH THE STANDARD BI-RADS
247
Graph of Figure 3 represents the second action of
scenario SC1.
Figure 3: The action 2 of the scenario SC1.
Let us consider action 3 of the scenario:
SC1.3: Performing radiological interpretation with
BI-RADS glossary from the case base. We can
extract from this action one key-concept:
radiological interpretation.
Key-concept: radiological interpretation.
This action generates itself two actions:
- SC1.3.1. Interpreting ROIs with CMs.
- SC1.3.2. Planning radiological report.
The first sub-action contains two key-concepts:
ROI and CM.
The second sub-action contains one key-concept:
radiological report.
Let us regroup the concepts extracted from the
third action in a graph: like for the first and second
actions, we have summarized the different extracted
key-concepts on a graph where the concepts are
represented by labels and linked by relationships
symbolized by lines.
Graph of Figure 4 represents the third action of
scenario SC1.
Figure 4: The action 3 of the scenario SC1.
Now, let us consider the last action of the scenario:
we extract three key-concepts: anatomo-
pathological examination, patient and histological
report. SC1.4. Performing anatomo-pathological
examination BI-RADS glossary from the casebase.
We have extracted three key-concepts: anatomo-
pathological examination, patient and histological
report.
Key-concepts: anatomo-pathological examination,
patient and histological report.
Like for other actions, we have summarized the
different extracted key-concepts on a graph identical
to the previous ones.
Graph of Figure 5 represents the last action of
scenario SC1.
Anatomo-pathological Examination
Patient
Histological Report
Is performed on
Is performed
Figure 5: The action 4 of the scenario SC1.
The graph of Figure 6 consists of fusion of the
four previous graphs. This scheme represents the
overall radiological-senological process. The work
illustrated via this scheme was systematically
achieved using requirements analysis (RCs); it
allowed us to accumulate the knowledge in the
particular field of radiology-senology. The defined
requirements serve as input of the design modelling
of the case base.
The graph of Figure 6 represents the result of this
fusion. It translates the course of the radiological-
senological process in its totality. The work
illustrated here was lead in a systematic manner for
the analysis of requirements chunks (RCs) and it
allows accumulating knowledge in a particular
domain as the radiology-senology. We have
elucidated requirements from the hierarchy of RCs
produced by the application of Crews-l’Ecritoire
approach.
As shown in Figure 6¸ we remark that these
objects can possess two common objects or relations
that it is necessary to factorize or regroup them in
common objects or relations. If we recapture the
definition of an ontology given by Gruber: “an
ontology is a formal, explicit, specification of a
shared conceptualisation’’. This work: consists of
identifying, in a corpus, and specific knowledge-to-
knowledge domain to represent and consensually
recognized as dependent of this domain. This
definition responds
at what we have built.
We have chosen to represent the ontology using
an object-oriented approach (and hence the Unified
Modelling Language (UML) formalism). An
oriented-object approach is acknowledged perfectly
powerful to manage complex data (images, sounds,
temporal data…) such as the nature of radiological-
senological data (text and image).
Reading
Im age
M am m ographic Im age E chographic Im age
ROI
Contains
Is com p osed of
Is com posed of
Locates
Locates
Radiological Interpretation
CM
ROI
R adiological R eport
Interpretes
C haracterizes
In te rp r ete s C M
ICEIS 2007 - International Conference on Enterprise Information Systems
248
Clinical E xam ination
Interrogation
Physical Exam ination
Patient
Contains
Contains
Is perform ed on
Reading
Im age
Mammographic Image Echographic Image
ROI
Contains
Is com posed of
Is com posed of
Locates
Locates
Radiological Interpretation
CM
ROI
Radiological Report
Interp retesCharacterizes
Interpretes C M
Anatom opathological Exam ination
Patient
H istological Report
Is perform ed on
Is perfor m ed
Figure 6: The 4 actions of the scenario SC1.
5 CONCLUSION
The knowledge acquisition process to build our
ontology was guided by requirements with the
Crews L'Ecritoire approach. It couples notions of
goals and scenarios to discover knowledge. With
respect to its contribution, this research has
produced a step of concept extraction and described
their relationships from scenarios. This approach has
efficiently guided the construction of the ontology in
the radiology-senology domain. The orientation
“goal” advocated is relevant since physicians and
their requirements are well within the core of the
process.
This successful experimentation of the “goal,
scenario” approach in order to build the ontology in
radiology-senology allows to conclude that this
approach can apply to other domains and that it
constitutes a systematic approach for ontology
engineering.
Systematic approaches for the construction of
ontologies are scarce. The association of “goal,
scenario” offers, on the one hand, a powerful
approach for the localization of knowledge
underlying the activities of radiologists and, on the
other hand, for generalizing the concepts and their
relationships that make the conceptual model of the
domain.
REFERENCES
Ben Achour, C., Extraction des besoins par analyse de
scénarios textuels, ‘99. Phd thesis, Université de Paris
6, France.
Chabriais, J., Balleyguier, C., Kinkel, K., B Levey, L.,
Stinès, J., Thibault, F., 1998. BI-RADS Illustré.,
Première Edition Française basée sur la troisième
Édition américaine, ‘98, 200 pages.
Demigha, S., Rolland, C., Baum, T.P., Balleyguier, C.,
Vincent, B., Menu, Y., 2001. Crews-l’Ecritoire
Analysis for the Implementation of a Medical Image
Database for Mammography, In Conference
Proceedings – SPIE Medical Imaging, PACS and
Integrated Medical Information: ‘01 vol 4323, pp386-
396, San Diego, California, USA.
Demigha, S., Un système à base de connaissances en
radiologie-sénologie dédié à la formation des
radiologues-sénologues juniors, 2005. Phd thesis in
Informatic, Université de Paris 1, juillet ‘05.
Furst, F, Contribution à l’ingénierie des ontologies : une
méthode et un outil d’opérationnalisation, Thèse de
doctorat de l’Université de Nantes, Ecole des Mines de
Nantes, 2004.
Guarino, N, Semantic matching: Formal ontological
distinctions for information, organization, extraction
and integration. Information extraction: A
multidisciplinary approach to emerging information
technology, International summer school, SCIE-97,
Frascati, Italy, pp. 139-170, 1997.
Gruber, T., 1993, A translation approach to portable
ontology specifications, In Knowledge Acquisition ’93,
vol 5 (2): pages 199-220.
Rector, A.L., 1999. Terminology and concept
representation languages: where are we?. In Artificial
Intelligence in Medicine ’99. vol 15(1): pages 1-4.
Rolland, C., Grosz, C., Kla, R., Experience with goal–
Scenario coupling in Requirements Engineering, 1999.
In the proceedings of the Fourth International
Symposium on Requirements Engineering. ‘99,
Limerick, Ireland.
AN ONTOLOGY SUPPORTING THE DAILY PRACTICE REQUIREMENTS OF RADIOLOGISTS-SENOLOGISTS
WITH THE STANDARD BI-RADS
249