±±¾©»ùÒò×éËù£¨¹ú¼ÒÉúÎïÐÅÏ¢ÖÐÐÄ£©ÏàÖú¿ª·¢ÉäѪ·ÖÊý±£´æÐÍÐÄÁ¦Ë¥½ßÔçÆÚΣº¦Ô¤²âÄ£ÐÍ
½üÈÕ£¬pgµç×Ó¹ÙÍø£¨¹ú¼ÒÉúÎïÐÅÏ¢ÖÐÐÄ£©Óë½â·Å¾ü×ÜÒ½ÔºµÚÁùҽѧÖÐÐÄÏàÖú¿ª·¢µÄÉäѪ·ÖÊý±£´æÐÍÐÄÁ¦Ë¥½ßÔçÆÚΣº¦Ô¤²âÄ£ÐÍHFmeRiskÕýʽ»ñµÃרÀûÊÚȨ£¬×¨ÀûÃû³ÆΪ¡°Ò»ÖÖÓÃÓÚÔ¤²âÉäѪ·ÖÊý±£´æÐÍÐÄ˥Σº¦µÄÄ£ÐÍ¡±£¨CN105586406A£©£¬·¢Ã÷ÈËΪpgµç×Ó¹ÙÍø£¨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»¯ÒòËصÄÅÓ´óÏ໥×÷ÓÃÒýÆðµÄÐÄÔà½á¹¹»ò¹¦Ð§Òì³£±ä¸ï¡£ÂýÐÔÐÄÁ¦Ë¥½ßÒÔÐļ¡ÄÜÁ¿´úлºÍ´úлÖØËÜÕϰΪÌص㣬±£´æ¸ß·¢²¡ÂʺÍËÀÍöÂÊ¡£Ä¿Ç°¹«ÈϵÄÂýÐÔÐÄÁ¦Ë¥½ßÓÐÈýÖÖÑÇÐÍ£¬ÆäÖÐÉäѪ·ÖÊý±£´æÐÍÐÄÁ¦Ë¥½ß£¨Heart failure with preserved ejection fraction£¬HFpEF, ×óÐÄÊÒÉäѪ·ÖÊý>50%£©µÄÔçÆÚΣº¦Ô¤²â¾ßÓÐÌôÕ½ÐÔ£¬½¨Á¢HFpEFÔçÆÚÔ¤²âÄ£ÐͶÔÐÄÁ¦Ë¥½ßµÄΣº¦ÆÀ¹ÀÖÎÀíºÍÁÙ´²¾ö²ßÊ®·ÖÖØÒª¡£
¸ÃÄ£ÐÍÁ¢ÒìÐԵĽáºÏDNA¼×»ù»¯Î»µãºÍÁÙ´²ÌØÕ÷£¬ÀûÓûúеѧϰҪÁìʵÏÖÁËHFpEFÔçÆÚΣº¦Ô¤²â¡£Ä£ÐÍÊÕ¼¯ÁË797Àýδ»¼ÐÄѪ¹Ü¼²²¡¼ÓÈëÕßµÄ97ÏîÁÙ´²ÕïÁÆÊý¾ÝºÍDNA¼×»ù»¯Ð¾Æ¬Êý¾Ý£¬¾¹ý8ÄêËæ·Ã£¬ÆäÖÐ738Ãû¼ÓÈëÕßÎÞÐÄË¥ÌåÏÖ£¬59Ãû¼ÓÈëÕß±»Õï¶ÏΪHFpEF£¬Ñо¿ÈËÔ±½«´ËÊý¾Ý×÷ΪѵÁ·¼¯£¬»ñµÃÁËÒ»×éÓÃÓÚ½¨Á¢Ô¤²âÉäѪ·ÖÊý±£´æÐÍÐÄ˥Σº¦Ä£Ð͵ıê¼ÇÎï×éºÏ¡£Ä£ÐÍ´Ó±í¹ÛÒÅ´«Ñ§µÄ½Ç¶È£¨25¸öCpGsÔÚϸ°û¼äÐźš¢Ï໥×÷ÓúÍÄÜÁ¿´úлÖоßÓÐÒªº¦¹¦Ð§£©ºÍÇé¿ö̻¶µÄ½Ç¶È£¨ÄêÁä¡¢ÀûÄò¼Á¡¢BMI¡¢ÂÑ°×ÄòºÍѪÇ弡ôû£©ÆÀ¹ÀÁËHFpEFµÄÔçÆÚΣº¦£¬¶ÔHFpEFÔçÆÚΣº¦¾ßÓÐÁ¼ºÃµÄ¼ø±ðºÍУÕýÄÜÁ¦£¬AUC½á¹ûΪ0.90£¨95% CI 0.89-0.90£©£¬²âÊÔ¼¯Hosmer-Lemeshowͳ¼ÆÁ¿Îª6.17£¬P=0.632¡£HFmeRiskÀûÓûùÓÚÍƼöϵͳµÄdeepFMËã·¨ºÍ»ùÓÚÌØÕ÷Ñ¡ÔñµÄLASSOºÍXGBoostËã·¨£¬Ñ§Ï°ÕâЩÌØÕ÷±³ºóÒþ²ØµÄÌØÕ÷×éºÏ£¬ÎªHFpEFÔçÆÚΣº¦ÆÀ¹ÀÌṩÁ¢Òì¼û½â¡£
2022Äê1Ô£¬¸ÃÑо¿µÄÀíÂÛ½á¹ûÒÔ¡°A deep learning model for early risk prediction of heart failure with preserved ejection fraction by DNA methylation profiles combined with clinical features¡±ÎªÌ⣬Ðû²¼ÓÚClinical Epigenetics ?ÆÚ¿¯£¬ÕÔѧͮΪµÚÒ»×÷Õߣ¬¶Îµ¡¢ÇþºèÖñºÍÆ«Ïò¶«ÎªÍ¨Ñ¶×÷Õß¡£¸ÃÑо¿»ñµÃÁ˹ú¼Ò×ÔÈ»¿Æѧ»ù½ð¡¢¹ú¼ÒÖصãÑз¢¼Æ»®µÈÏîÄ¿µÄ×ÊÖú¡£
±ðµÄ£¬Æ«Ïò¶«Ñо¿×éÓ뻪ΪÏàÖú¿ª·¢ÁËÂýÐÔ²¡ÒÅ´«Î£º¦ÆÀ¹Àϵͳ£¬ÔÚ¸ÃרÀûÑз¢½×¶Î£¬Ïà¹Ø½á¹ûΪÉÏÊöÆÀ¹ÀϵͳÌṩÁËÀíÂÛÓë¼¼ÊõÖ§³Å¡£
»ùÓÚÉî¶ÈѧϰµÄÉäѪ·ÖÊý±£´æÐÍÐÄË¥Ô¤²âÄ£ÐÍHFmeRisk
ÂÛÎÄÁ´½Ó