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

 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

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

AND

Functional Level

SC1.1.2

Record patient’s new data SC1.1.2

Record patient’s new data

SC1.1.2

Record patient’s old

data

SC1.1.2

Record patient’s old

data

---------------

SC1.2.2.1.1.1

Identify the junior-

radiologist

SC1.2.2.1.1.1

Identify the junior-

radiologist

SC1.2.2.1.1.1

Identify the expert-

radiologist

SC1.2.2.1.1.1

Identify the expert-

radiologist

Refined by

Physical Level

---------------

Refined by

SC1.2.2.

Interpret ROIs located on

mammographic images

SC1.2.2.

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

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

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

Radiological Interpretation

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

Is perform ed on

Reading

Im age

Mammographic Image Echographic Image

ROI

Contains

Is com posed of

Locates

Radiological Interpretation

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