scieee Science in your language
[sp] (orig)

Towards the intelligent diagnosis of hematological diseases

Abstract

In traditional medicine, patient diagnosis usually implies an in depth study of its state and symptoms that a specialist has to carry out. The adaptation and customization of the medical treatment to those individual characteristics of each patient is what we know as Precision Medicine. Furthermore, in the case of multidisciplinary fields such as haematology, the identification of several diseases usually implies complex analyses in order to have a high degree of certainty in the diagnosis. A better understanding of the clinical tests and their relationship and the finding of new patterns between them will enable us to avoid a significant amount of such tests by supporting the specialist with new information. In this line, Artificial Intelligence has proven to be a useful methodology for data analytics in general whose main drawback is the need of huge amounts of data to achieve high accuracy. In the particular case of clinical data, it is widely generated in hospitals but the lack of standardization and the difficulties of availability require complex preprocessing. Therefore, we have collected 100,000 complete blood counts and developed a method to 1) automatically label textual diagnosis using deep neural networks with Long short-term memory cells. In this approach, a group of specialists has manually labelled 1,000 CBCs through a mobile application, which have then been used to feed the network in order to learn to interpret the diagnosis, and 2) to make an intelligent diagnosis of new samples in which a subset of 10,000 CBCs has been used as an input to a Support Vector Machine. In summary, in this work we present two different prototypes of architectures in order to define methods for the collection, preprocessing and intelligent classification of clinical data, focusing in haematological disease. Our proposal presents encouraging results with accuracies greater than 90% in both cases.

Read accessible full text

Towards the intelligent diagnosis of hematological diseases

Author: Díaz-del-Pino, Sergio,Trelles-Martínez, Roberto,Perez-Wohlfeil, Esteban,Trelles-Salazar, Oswaldo Rogelio
Year: 2019
Source: https://riuma.uma.es/xmlui/bitstream/10630/18838/1/ECTB2019-Abstract.pdf
Towa ds he in elligen diagnosis o hema ological diseases
Se gio Díaz-del-Pino, Robe o T elles-Ma inez, Es eban Pe ez-Wohl eil, Oswaldo T elles
In adi ional medicine, pa ien diagnosis usually implies an in dep h s udy o i s s a e and
symp oms ha a specialis has o ca y ou . The adap a ion and cus omiza ion o he medical
ea men o hose indi idual cha ac e is ics o each pa ien is wha we know as P ecision
Medicine. Howe e , he whole p ocess equi es se e al es s, analysis and adminis a ion
p ocedu es which in he end consumes ime and makes i expensi e o he public heal h
sys em. The e o e, he esea ch o new me hods ha assis s he specialis s could lead us o
educe no only he economic e ec o he p ocess bu he wo kload while p oducing mo e
accu a e diagnosis and be e quali y o li e o he pa ien s.
Fu he mo e, in he case o mul idisciplina y ields such as haema ology, he iden i ica ion o
se e al diseases usually implies complex analyses such as pe iphe al blood smea , capilla y
elec opho esis and ch oma og aphy, in o de o ha e a high deg ee o ce ain y in he
diagnosis. A be e unde s anding o he clinical es s and hei ela ionship and he inding o
new pa e ns be ween hem will enable us o a oid a signi ican amoun o such es s by
suppo ing he specialis wi h new in o ma ion.
In his line, A i icial In elligence has p o en o be a use ul me hodology o da a analy ics in
gene al whose main d awback is he need o huge amoun s o da a o achie e high
accu acy. In he pa icula case o clinical da a, i is widely gene a ed in hospi als bu he lack
o s anda diza ion and he di icul ies o a ailabili y equi e complex p ep ocessing.
The e o e, we ha e collec ed 100,000 comple e blood coun s (CBC) om he Hospi al
Clínico San Ca los (Mad id) and de eloped a me hod o 1) au oma ically label ex ual
diagnosis using deep neu al ne wo ks wi h Long sho - e m memo y cells. In his app oach,
a g oup o specialis s has manually labelled 1,000 CBCs h ough a mobile applica ion, which
ha e hen been used o eed he ne wo k in o de o lea n o in e p e he diagnosis, and 2)
o make an in elligen diagnosis o new samples in which a subse o 10,000 CBCs has been
used as an inpu o a Suppo Vec o Machine.
In summa y, in his wo k we p esen wo di e en p o o ypes o a chi ec u es in o de o
de ine me hods o he collec ion, p ep ocessing and in elligen classi ica ion o clinical da a,
ocusing in haema ological disease. Ou p oposal p esen s encou aging esul s wi h
accu acies g ea e han 90% in bo h cases.