±±¾©»ùÒò×éËù£¨¹ú¼ÒÉúÎïÐÅÏ¢ÖÐÐÄ£©ÏàÖú¿ª·¢ÉäѪ·ÖÊý±£´æÐÍÐÄÁ¦Ë¥½ßÔçÆÚΣº¦Õ¹ÍûÄ£×Ó
¿ËÈÕ£¬ÓÀÀû¼¯ÍÅ£¨¹ú¼ÒÉúÎïÐÅÏ¢ÖÐÐÄ£©Óë½â·Å¾ü×ÜÒ½ÔºµÚÁùҽѧÖÐÐÄÏàÖú¿ª·¢µÄÉäѪ·ÖÊý±£´æÐÍÐÄÁ¦Ë¥½ßÔçÆÚΣº¦Õ¹ÍûÄ£×ÓHFmeRiskÕýʽ»ñµÃרÀûÊÚȨ£¬×¨ÀûÃû³ÆΪ¡°Ò»ÖÖÓÃÓÚÕ¹ÍûÉäѪ·ÖÊý±£´æÐÍÐÄ˥Σº¦µÄÄ£×Ó¡±£¨CN105586406A£©£¬·¢Ã÷ÈËΪÓÀÀû¼¯ÍÅ£¨¹ú¼ÒÉúÎïÐÅÏ¢ÖÐÐÄ£©Æ«Ïò¶«Ñо¿Ô±¡¢²©Ê¿Ñо¿ÉúÕÔѧͮ£¨ÒѽáÒµ£¬ÏÖÈιú¼ÒÉúÎïÐÅÏ¢ÖÐÐŤ³Ìʦ£©¡¢ÇþºèÖñ¸±Ñо¿Ô±¡¢½â·Å¾ü×ÜÒ½ÔºµÚÁùҽѧÖÐÐĶεÖ÷ÈÎҽʦ¡£
ÐÄÁ¦Ë¥½ßÊÇÓÉÒÅ´«¡¢Éñ¾¼¤ËØ¡¢´úл¡¢Ñ×Ö¢µÈÉú»¯ÒòËصÄÖØ´óÏ໥×÷ÓÃÒýÆðµÄÐÄÔà½á¹¹»ò¹¦Ð§Ò쳣ת±ä¡£ÂýÐÔÐÄÁ¦Ë¥½ßÒÔÐļ¡ÄÜÁ¿´úлºÍ´úлÖØËÜÕϰΪÌص㣬±£´æ¸ß·¢²¡ÂʺÍéæÃüÂÊ¡£ÏÖÔÚ¹«ÈϵÄÂýÐÔÐÄÁ¦Ë¥½ßÓÐÈýÖÖÑÇÐÍ£¬ÆäÖÐÉäѪ·ÖÊý±£´æÐÍÐÄÁ¦Ë¥½ß£¨Heart failure with preserved ejection fraction£¬HFpEF, ×óÐÄÊÒÉäѪ·ÖÊý>50%£©µÄÔçÆÚΣº¦Õ¹Íû¾ßÓÐÌôÕ½ÐÔ£¬½¨ÉèHFpEFÔçÆÚÕ¹ÍûÄ£×Ó¶ÔÐÄÁ¦Ë¥½ßµÄΣº¦ÆÀ¹ÀÖÎÀíºÍÁÙ´²¾öÒéÊ®·ÖÖ÷Òª¡£
¸ÃÄ£×ÓÁ¢ÒìÐÔµÄÍŽáDNA¼×»ù»¯Î»µãºÍÁÙ´²ÌØÕ÷£¬Ê¹ÓûúеѧϰҪÁìʵÏÖÁËHFpEFÔçÆÚΣº¦Õ¹Íû¡£Ä£×ÓÍøÂçÁË797Àýδ»¼ÐÄѪ¹Ü¼²²¡¼ÓÈëÕßµÄ97ÏîÁÙ´²ÕïÁÆÊý¾ÝºÍDNA¼×»ù»¯Ð¾Æ¬Êý¾Ý£¬¾ÓÉ8ÄêËæ·Ã£¬ÆäÖÐ738Ãû¼ÓÈëÕßÎÞÐÄË¥ÌåÏÖ£¬59Ãû¼ÓÈëÕß±»Õï¶ÏΪHFpEF£¬Ñо¿Ö°Ô±½«´ËÊý¾Ý×÷ΪѵÁ·¼¯£¬»ñµÃÁËÒ»×éÓÃÓÚ½¨ÉèÕ¹ÍûÉäѪ·ÖÊý±£´æÐÍÐÄ˥Σº¦Ä£×ӵıê¼ÇÎï×éºÏ¡£Ä£×Ó´Ó±í¹ÛÒÅ´«Ñ§µÄ½Ç¶È£¨25¸öCpGsÔÚϸ°û¼äÐźš¢Ï໥×÷ÓúÍÄÜÁ¿´úлÖоßÓÐÒªº¦¹¦Ð§£©ºÍÇéÐÎ̻¶µÄ½Ç¶È£¨ÄêËê¡¢ÀûÄò¼Á¡¢BMI¡¢ÂÑ°×ÄòºÍѪÇ弡ôû£©ÆÀ¹ÀÁËHFpEFµÄÔçÆÚΣº¦£¬¶ÔHFpEFÔçÆÚΣº¦¾ßÓÐÓÅÒìµÄÅбðºÍУÕýÄÜÁ¦£¬AUCЧ¹ûΪ0.90£¨95% CI 0.89-0.90£©£¬²âÊÔ¼¯Hosmer-Lemeshowͳ¼ÆÁ¿Îª6.17£¬P=0.632¡£HFmeRiskʹÓûùÓÚÍƼöϵͳµÄdeepFMËã·¨ºÍ»ùÓÚÌØÕ÷Ñ¡ÔñµÄLASSOºÍXGBoostËã·¨£¬Ñ§Ï°ÕâЩÌØÕ÷±³ºóÒþ²ØµÄÌØÕ÷×éºÏ£¬ÎªHFpEFÔçÆÚΣº¦ÆÀ¹ÀÌṩÁ¢Òì¿´·¨¡£
2022Äê1Ô£¬¸ÃÑо¿µÄÀíÂÛЧ¹ûÒÔ¡°A deep learning model for early risk prediction of heart failure with preserved ejection fraction by DNA methylation profiles combined with clinical features¡±ÎªÌ⣬½ÒÏþÓÚClinical Epigenetics ?ÆÚ¿¯£¬ÕÔѧͮΪµÚÒ»×÷Õߣ¬¶Îµ¡¢ÇþºèÖñºÍÆ«Ïò¶«ÎªÍ¨Ñ¶×÷Õß¡£¸ÃÑо¿»ñµÃÁ˹ú¼Ò×ÔÈ»¿Æѧ»ù½ð¡¢¹ú¼ÒÖصãÑз¢ÍýÏëµÈÏîÄ¿µÄ×ÊÖú¡£
±ðµÄ£¬Æ«Ïò¶«Ñо¿×éÓ뻪ΪÏàÖú¿ª·¢ÁËÂýÐÔ²¡ÒÅ´«Î£º¦ÆÀ¹Àϵͳ£¬ÔÚ¸ÃרÀûÑз¢½×¶Î£¬Ïà¹ØЧ¹ûΪÉÏÊöÆÀ¹ÀϵͳÌṩÁËÀíÂÛÓëÊÖÒÕÖ§³Ö¡£
»ùÓÚÉî¶ÈѧϰµÄÉäѪ·ÖÊý±£´æÐÍÐÄË¥Õ¹ÍûÄ£×ÓHFmeRisk
ÂÛÎÄÁ´½Ó