A bstract. Risk estimation of developing breast cancer poses as the first prevention method for early diagnosis. Furthermore, data integration from different departments involved in the process plays a key role. In order to guarantee patient safety, the whole process should be orchestrated and monitored automatically. Support for the solution will be a linked data cloud, composed by all the departments that take part in the process, combined with rule engines.
!
F igure 1. Linked Open Data cloud as of September 2010 2
E sti mating br east cance r r isk
A risk factor is anything that increases your chance of developing breast cancer and a
protective factor is anything that reduces your risk of developing breast cancer. Some
risk factors can be controlled by the patient. For example, if she is overweight, excess
pound losses may reduce patient´s risk of developing breast cancer. But other factors
are beyond their control. For instance, patients cannot change their gender. Women
are much more likely than men to have breast cancer (Giordano et al., 2003). This is
mostly because women have more estrogen and progesterone in their bodies. These
hormones stimulate breast cell growth, both normal and abnormal. Also, aging is the
biggest risk factor for breast cancer, and you cannot stop growing older.
In order to detect the risk factors identified by the current highest level of clinical
evidence, a high quality systematic review of the literature has been carried out.!The
most significant factors identified were! "#$#%&'! &$(#)&%*$'#! (Kauff et al. 2005)+!
#%($&'&%,! (Esther M. John et al., 2007),! menstrual history (Clemons et al. 2006), late
pregnancy or no pregnancy and breastfeeding (Helewa et al. 2002), even risk factors
such as alcohol (Hamajima et al., 2002) or smoking
A common semantic vocabulary, ontology, is needed in order to achieve the goals of this approach, estimating the risk of developing breast cancer and interchanging data between the different departments involved in the process. On the one hand, is necessary to semantically annotate the different risk factors due to the population profiles and risks evolve over the time with different prominent risk factors. On the other hand, a common vocabulary should be shared between different department´s information systems in order to automatically interchange information. Among the different ontologies developed for breast cancer we will focus on the following: cancer ontology at TONES6 repository and the ontology developed by Mathew Hardy Williams7 (Hardy, 2008). This study combines both ontologies to produce the ontology used in this research to support the breast cancer department from Cruces Hospital.
Virtuoso Universal Server8, in its open source edition, storage engine has been used to perform the knowledge base for this study. !"#$% $&'()*+% +,*#,+% -+&$% .$% )//% /#00+(+,&% $+1),)--3% ),,'&)&+/% /)&)% 4#&"#,% +)2"% 567% 8()9":% ;$)*+% '0% 567% 8()9"$% #$% ,+2+$$)(3% 0'(%&"#$%<#,/%'0%)99-#2)&#',% =+2).$+% ".*+%)1'.,&$%'0%/)&)%)(+% 1),)*+/% =3% &"+%$+(>+(: 4
Taking a typical health record stored on any Hospital's database as starting point, the system will gather the information available in each department involved in the process to enrich the record. After this, a set of inference rules to ascertain the risk of developing breast cancer, which is the main aim of this research, will be applied to obtain the suitable monitoring for the patient.
After these and others rules have been applied, patientVUDHN¶OL diagnosed. In case the risk had reached the threshold, patient´s doctor had immediately received an alarm indicating that one of his patients is in high risk of developing breast cancer. 5
F utu r e wor k
Since literature is subject to change, and a systematic review is a hard task, a tool to
automatically distinguish the current medical evidence from the noise in the data will
be developed. In order to obtain these results SPARQL queries must be done over
different endpoints shown in Figure 1. A Linking Transversal
! >6;6]8;6/!Z#/! >8*-&)/! ^#!)'!64!34B-(-4/!'-@6BB-!6.7!@*)6&'! B6.B)*XXB-446@-*6'8F)! *)6.64[&8&! -G! 8.78F87K64! 76'6! G*-;! L\! )=87);8-4-:8B64! &'K78)&/! 8.B4K78.:! LS/L"L! _-;).! _8'(! @*)6&'! B6.B)*! 6.7!RL/PWQ!_-;).!_8'(-K'!'()!78&)6&)#!H*!,!J6.B)*!OPPO2!SQ2!""2!"O\$X$L#! >6*7[! E84486;&/! +#! TOPPSU#! <.'):*6'8.:! `.'-4-:8)&! 6.7! 6*:K;).'6'8-.! G-*! 7)B8&8-.X;6C8.:! 8.! @*)6&'!B6.B)*#!1-B'-*64!a()&8&/!M.8F)*&8'[!J-44):)!D-.7-.#!