{"id":4168,"date":"2017-06-29T13:50:49","date_gmt":"2017-06-29T11:50:49","guid":{"rendered":"http:\/\/www.easistent.ro\/?p=4168"},"modified":"2017-07-26T15:38:55","modified_gmt":"2017-07-26T13:38:55","slug":"aplicatii-ale-inteligentei-artificiale-in-practica-clinica","status":"publish","type":"post","link":"https:\/\/www.easistent.ro\/?p=4168","title":{"rendered":"Aplica\u0163ii ale inteligen\u0163ei artificiale \u00een practica clinic\u0103"},"content":{"rendered":"

\"ai1\"<\/a>Un calculator care s\u0103 prognozeze rata de supravie\u0163uire a unui pacient ar fi p\u0103rut de domeniul science-fiction-ului, acum 20 de ani.<\/p>\n

Acum devine realitate, mul\u0163umit\u0103 unei tehnici numit\u0103 \u201c\u00eenv\u0103\u0163are profund\u0103\u201d (\u201cdeep learning\u201d) prin care sistemele de calculatoare pot \u00eenv\u0103\u0163a, \u00een cel mai literal sens al cuv\u00e2ntului, cum s\u0103 \u00een\u0163eleag\u0103, interpreteze \u015fi analizeze imagini.<\/p>\n

Produsele software de acest fel \u00eencearc\u0103 s\u0103 imite activitatea diferitelor straturi de neuroni din neocortex, zona din creier unde au loc procesele de g\u00e2ndire.<\/p>\n

La fel ca un creier uman, software-urile de inteligen\u0163\u0103 artificial\u0103 \u00eenva\u0163\u0103 s\u0103 recunoasc\u0103 tiparele diverselor reprezent\u0103ri digitale ale imaginilor, sunetelor \u015fi ale altor tipuri de date.<\/p>\n

Inteligen\u0163a artificial\u0103 \u00eencepe deja s\u0103 transforme o multitudine de domenii, de la comunica\u0163ii \u015fi transporturi, la medicin\u0103.<\/p>\n

Dup\u0103 cum declar\u0103 speciali\u015ftii \u00een inteligen\u0163\u0103 artificial\u0103, \u00eembun\u0103t\u0103\u0163irea proceselor de \u00eengrijire a pacien\u0163ilor poate reprezenta una dintre cele mai importante aplica\u0163ii practice ale tehnicii de \u201c\u00eenv\u0103\u0163are profund\u0103.\u201d<\/p>\n

De exemplu, profesioni\u015ftii din s\u0103n\u0103tate vor putea folosi inteligen\u0163a artificial\u0103 a calculatoarelor pentru a analiza mai repede \u015fi cu mai mult\u0103 acurate\u0163e imaginile medicale, radiografii, ecografii, RMN-uri, etc. Este o evolu\u0163ie tehnologic\u0103 care accelereaz\u0103 dezvoltarea medicinei personalizate.<\/p>\n

Iar primele rezultate \u00eencurajatoare au \u00eenceput deja s\u0103 apar\u0103. Cercet\u0103torii de la \u015ecoala de S\u0103n\u0103tate Public\u0103 \u015fi Facultatea de Automatic\u0103 a Universit\u0103\u0163ii din Adelaide, Australia au folosit inteligen\u0163a artificial\u0103 pentru a analiza imagistica medical\u0103 a cutiei toracice pentru 48 de pacien\u0163i. Aceast\u0103 analiz\u0103 computerizat\u0103 a reu\u015fit s\u0103 prognozeze care dintre pacien\u0163i vor muri \u00eentr-un interval de 5 ani, cu o acurate\u0163e de 69%, similar\u0103 cu prognoza \u201cmanual\u0103\u201d a clinicienilor. Este primul studiu de acest gen folosind imagistica medical\u0103 \u015fi inteligen\u0163a artificial\u0103 a calculatorului (produse software de tipul \u201cdeep learning\u201d). Cele mai exacte prognoze au fost f\u0103cute \u00een cazul pacien\u0163ilor cu emfizem \u015fi insuficien\u0163\u0103 cardiac\u0103 de tip congestiv.<\/p>\n

Aceast\u0103 cercetare, care a fost publicat\u0103 \u00een revista Nature<\/em><\/strong>, este a\u015fteptat\u0103 s\u0103 deschid\u0103 noi orizonturi \u00een depistarea precoce a bolilor cronice severe, inclusiv a atacurilor de cord.<\/p>\n

\u00cen acela\u015fi timp, \u015fi alte echipe de cercet\u0103tori deruleaz\u0103 proiecte care urm\u0103resc optimizarea procesului de diagnosticare, ajut\u00e2ndu-i pe clinicieni s\u0103 identifice c\u00e2t mai multe elemente relevante din imagistica medical\u0103 disponibil\u0103 (radiografii, ecografii, etc). Dintre aceste ini\u0163iative de cercetare, le not\u0103m pe cele de la Imperial College din Londra (evaluare digital\u0103 a imagisticii traumatismelor craniene) \u015fi ale companiilor Arterys din SUA (dezvoltarea de software pentru vizualizarea \u015fi cuantificarea fluxului sanguin) \u015fi Enlitic din SUA (dezvoltarea de software pentru identificarea de tumori \u015fi fracturi greu detectabile).<\/p>\n

De\u015fi aplica\u0163iile practice ale inteligen\u0163ei artificiale \u00een medicin\u0103 sunt \u00eenc\u0103 la \u00eenceput, cercet\u0103torii sunt \u00eencrez\u0103tori c\u0103 aceasta are poten\u0163ialul s\u0103 revolu\u0163ioneze practica clinic\u0103, acceler\u00e2nd dezvoltarea medicinei personalizate.<\/p>\n

Inteligen\u0163a artificial\u0103 va putea integra \u015fi analiza rapid datele existente \u00een eviden\u0163ele electronice ale pacientului, cele oferite de imagistica medical\u0103 \u015fi cele transmise de diversele dispozitive de tele-medicin\u0103, iar toate vor pune astfel la dispozi\u0163ia clinicienilor elementele necesare pentru crearea unui plan individualizat de \u00eengrijiri.<\/p>\n

Articol realizat de Mirela Musta\u021b\u0103, Redactor Executiv E-Asistent<\/em><\/strong><\/p>\n

 <\/p>\n

Surse de documentare<\/strong>:<\/p>\n

1. https:\/\/www.sciencedaily.com\/releases\/2017\/06\/170601124126.htm<\/a> – Materials<\/a>\u00a0provided by\u00a0University of Adelaide<\/strong><\/a>.<\/p>\n

2. http:\/\/www.nvidia.com\/object\/deep-learning-in-medicine.html<\/a><\/p>\n

3. http:\/\/www.nvidia.com\/object\/deep-learning-in-medicine.html#sthash.JcwVGLeG.dpuf<\/a><\/p>\n

4. https:\/\/www.technologyreview.com\/s\/513696\/deep-learning\/<\/p>\n

5. http:\/\/www.npr.org\/sections\/alltechconsidered\/2014\/02\/20\/280232074\/deep-learning-teaching-computers-to-tell-things-apart<\/a><\/p>\n

6. http:\/\/medicalfuturist.com\/artificial-intelligence-will-redesign-healthcare\/<\/a><\/p>\n

7. http:\/\/www.fiercehealthcare.com\/it\/role-artificial-intelligence-personalized-medicine<\/a><\/p>\n

 <\/p>\n","protected":false},"excerpt":{"rendered":"

Un calculator care s\u0103 prognozeze rata de supravie\u0163uire a unui pacient ar fi p\u0103rut de domeniul science-fiction-ului, acum 20 de ani. Acum devine realitate, mul\u0163umit\u0103 unei tehnici numit\u0103 \u201c\u00eenv\u0103\u0163are profund\u0103\u201d (\u201cdeep learning\u201d) prin care sistemele de calculatoare pot \u00eenv\u0103\u0163a, \u00een cel mai literal sens al cuv\u00e2ntului, cum s\u0103 \u00een\u0163eleag\u0103, interpreteze \u015fi analizeze imagini. Produsele software […]<\/p>\n","protected":false},"author":3,"featured_media":4151,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":[],"categories":[152,3],"tags":[],"_links":{"self":[{"href":"https:\/\/www.easistent.ro\/index.php?rest_route=\/wp\/v2\/posts\/4168"}],"collection":[{"href":"https:\/\/www.easistent.ro\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.easistent.ro\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.easistent.ro\/index.php?rest_route=\/wp\/v2\/users\/3"}],"replies":[{"embeddable":true,"href":"https:\/\/www.easistent.ro\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=4168"}],"version-history":[{"count":2,"href":"https:\/\/www.easistent.ro\/index.php?rest_route=\/wp\/v2\/posts\/4168\/revisions"}],"predecessor-version":[{"id":4170,"href":"https:\/\/www.easistent.ro\/index.php?rest_route=\/wp\/v2\/posts\/4168\/revisions\/4170"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.easistent.ro\/index.php?rest_route=\/wp\/v2\/media\/4151"}],"wp:attachment":[{"href":"https:\/\/www.easistent.ro\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=4168"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.easistent.ro\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=4168"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.easistent.ro\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=4168"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}