MACHINE LEARNING-BASED METHOD FOR PREDICTION OF BREAST CANCER PROGNOSIS USING NEXT-GENERATION SEQUENCING, AND PREDICTION SYSTEM THEREFOR

• Main Authors: JO, Jeong Hee, HAN, Won Shik, PARK, In Ae, LEE, Jong Won, LEE, Hee Jin, LEE, Han Byoel, LEE, Sae Byul, AHN, Sei Hyun, KIM, Ae Ree, KIM, Min Su, RYU, Han Suk, KIM, Chung Yeul, YOON, Sung Roh, KWON, Sun Young, KIM, Sun
• Format: Patent
• Language: English; French; Korean
• Published: 11.07.2019
• Subjects: HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATIONTECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING ORPROCESSING OF MEDICAL OR HEALTHCARE DATA; INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTEDFOR SPECIFIC APPLICATION FIELDS; PHYSICS

A machine learning-based method for prediction of breast cancer prognosis using next-generation sequencing comprises the steps of: allowing a computer device to use RNA sequencing data of a subject tissue to measure an expression level of a target gene; allowing the computer device to input the expression level of the target gene into an artificial neural network adapted beforehand; and allowing the computer device to estimate the prognosis of breast cancer of the subject on the basis of the output value of the artificial neural network. The artificial neural network is adapted beforehand to have target gene expression levels of multiple samples as input values and to output results depending on oncotype DX recurrence scores for the multiple samples. L'invention concerne un procédé basé sur l'apprentissage automatique pour la prédiction du pronostic du cancer du sein utilisant un séquençage de nouvelle génération et comprenant les étapes consistant : à ce qu'un dispositif informatique utilise des données de séquençage d'ARN d'un tissu d'un sujet pour mesurer le niveau d'expression d'un gène cible ; à ce que le dispositif informatique entre le niveau d'expression du gène cible dans un réseau neuronal artificiel adapté au préalable ; et à ce que le dispositif informatique estime le pronostic du cancer du sein du sujet sur la base de la valeur de sortie du réseau neuronal artificiel. Le réseau neuronal artificiel est adapté au préalable pour présenter des niveaux d'expression de gènes cibles de multiples échantillons en tant que valeurs d'entrée et pour délivrer des résultats en fonction des scores de récidive du test Oncotype DX pour les multiples échantillons. 차세대 염기서열분석을 이용한 기계학습 기반 유방암 예후 예측 방법은 컴퓨터 장치가 피험자 조직(tissue)의 RNA 시퀀싱 데이터를 이용하여 타겟 유전자의 발현량을 측정하는 단계, 상기 컴퓨터 장치가 상기 타겟 유전자의 발현량을 사전에 마련한 인공신경망(Artificial Neural Network)에 입력하는 단계 및 상기 컴퓨터 장치가 상기 인공신경망의 출력값을 기준으로 상기 피험자에 대한 유방암 예후를 추정하는 단계를 포함한다. 상기 인공신경망은 복수의 샘플의 타겟 유전자 발현량을 입력값으로 갖고, 상기 복수의 샘플에 대한 온코타입(Oncotype DX)의 재발 점수(recurrence score)에 따른 결과를 출력하도록 사전에 마련된다.