ຊື່ຂອງ : ການເດີນທາງ ຊື່ຫຍໍ້ຂອງ : Holger R. Roth ຊື່ຫຍໍ້ຂອງ : Aoxiao Zhong ເມນູ Harouni Amilcare Gentili ຊື່ຫຍໍ້ຂອງ : Ana Z Abidin ເມນູ ວິທະຍາໄລ Anthony Beardsworth Costa ຊື່ຫຍໍ້ຂອງ : Bradford J Wood ປະເພດ Chien-Sung Tsai Chih-Hung Wang ວິທະຍາໄລ Chun-Nan Hsu ຊື່ຫຍໍ້ຂອງ : C K Lee ພາສາລາວ ພາສາລາວ ປະເພດ Wu ມັງກອນ ພາສາລາວ ລະຫັດ QR ຊື່ຫຍໍ້ຂອງ : Gustavo César de Antônio Corradi ຂໍຂອບໃຈ Hao-Hsin Shin ປະເພດ Hirofumi Obinata ປະເພດ Ren ຊື່ຫຍໍ້ຂອງ : Jason C Crane ຊື່ຫຍໍ້ຂອງ : Jesse Tetreault ຊື່ຫຍໍ້ຂອງ : Jiahui Guan ຊື່ຫຍໍ້ຂອງ : John W. Garrett ຊື່ຫຍໍ້ຂອງ : Joshua D. Kaggie Jung Gil Park ຊື່ຫຍໍ້ຂອງ : Keith Dreyer Krishna ມັງກອນ ລະຫັດ QR ຊື່ຫຍໍ້ຂອງ : Marcio Aloisio Bezerra Cavalcanti ຊື່ຫຍໍ້ຂອງ : Marius George Linguraru ຊື່ຫຍໍ້ຂອງ : A. Haider ຫນ້າທໍາອິດ / Meena AbdelMaseeh ວິທະຍາໄລ ຊື່ຫຍໍ້ຂອງ : Pablo F Damasceno ຊື່ຫຍໍ້ຂອງ : Pedro Mario Cruz e Silva Pochuan Wang Sheng Xu ຊື່ຫຍໍ້ຂອງ : Shuichi Kawano ຊື່ຫຍໍ້ຂອງ : Sira Sriswasdi ດາວໂຫລດ Soo Young Park ຊື່ຫຍໍ້ຂອງ : Thomas M. Grist Varun Buch Watsamon Jantarabenjakul Weichung Wang ຍິນດີຕ້ອນຮັບ Young Tak ວິທະຍາໄລ Lee ຊື່ຫຍໍ້ຂອງ : Xihong Lin Young Joon Kwon ພາສາລາວ ເມນູ Feng ເມນູ N. Priest ຂໍຂອບໃຈ Turkey Benjamin Glicksberg Bernardo Bizzo ຊື່ຫຍໍ້ຂອງ : Seok Kim ພາສາລາວ Chia-Cheng Lee ຊື່ຫຍໍ້ຂອງ : Chia Jung Hsu ຈີນ Lin Chiu-Ling Lai ຊື່ຫຍໍ້ຂອງ : Christopher P. Hess ປະເພດ Colin Compas ຫນ້າທໍາອິດ / Deepeksha Bhatia ຊື່ຫຍໍ້ຂອງ : Eric K Oermann Evan Leibovitz ຊື່ຫຍໍ້ຂອງ : Sasaki Hitoshi Mori ຊື່ຫຍໍ້ຂອງ : Isaac ຊື່ຫຍໍ້ຂອງ : Jae Ho Krishna Nand Keshava ມັງກອນ Li-Chen Fu ຄົ້ນຫາທີ່ດີທີ່ສຸດທີ່ດີທີ່ສຸດທີ່ດີທີ່ສຸດທີ່ດີທີ່ສຸດທີ່ດີທີ່ສຸດ Mike Fralick ມົນ Kyu Kang ພາສາລາວ ຂໍຂອບໃຈ ຊື່ຫຍໍ້ຂອງ : Vateekul ຊື່ຫຍໍ້ຂອງ ຊາຍ Hickman ຊື່ຫຍໍ້ຂອງ : Sharmila Majumdar Shelley L. McLeod ດາວນ໌ໂຫລດ Reed ວິທະຍາໄລ Stefan Gräf ຊື່ຫຍໍ້ຂອງ : Stephanie Harmon ປະເພດ Tatsuya Kodama ຂໍຂອບໃຈ ມັງກອນ 2018 ຊື່ຫຍໍ້ຂອງ : Vitor Lima de Trabor ຊື່ຫຍໍ້ຂອງ : Yothin Rakvongthai ວິທະຍາໄລ Yu Rim Lee ຊື່ຫຍໍ້ຂອງ : Wen Wen ຊື່ຫຍໍ້ຂອງ : Fiona J Gilbert ຊື່ຫຍໍ້ຂອງ : Mona G. Flores ປະເພດ Li ຊື່ຂອງ : ການເດີນທາງ ຊື່ຫຍໍ້ຂອງ : Holger R. Roth ຊື່ຫຍໍ້ຂອງ : Aoxiao Zhong ເມນູ Harouni ປະເພດ Amilcare ຊື່ຫຍໍ້ຂອງ : Ana Z Abidin ເມນູ ວິທະຍາໄລ Anthony Beardsworth Costa ຊື່ຫຍໍ້ຂອງ : Bradford J Wood ປະເພດ Chien-Sung Tsai ວິທະຍາໄລ Chih-Hung Wang ວິທະຍາໄລ Chun-Nan Hsu ຊື່ຫຍໍ້ຂອງ : C K Lee ພາສາລາວ ພາສາລາວ ປະເພດ Wu ມັງກອນ ພາສາລາວ ລະຫັດ QR ຊື່ຫຍໍ້ຂອງ : Gustavo César de Antônio Corradi ຂໍຂອບໃຈ ຊື່ຫຍໍ້ຂອງ : Hao Hsin Shin ປະເພດ Hirofumi Obinata ປະເພດ Ren ຊື່ຫຍໍ້ຂອງ : Jason C Crane ຊື່ຫຍໍ້ຂອງ : Jesse Tetreault ຊື່ຫຍໍ້ຂອງ : Jiahui Guan ຊື່ຫຍໍ້ຂອງ : John W. Garrett ຊື່ຫຍໍ້ຂອງ : Joshua D. Kaggie ວິທະຍາໄລ Jung Gil Park ຊື່ຫຍໍ້ຂອງ : Keith Dreyer Krishna ມັງກອນ ລະຫັດ QR ຊື່ຫຍໍ້ຂອງ : Marcio Aloisio Bezerra Cavalcanti ຊື່ຫຍໍ້ຂອງ : Marius George Linguraru ຊື່ຫຍໍ້ຂອງ : A. Haider ຫນ້າທໍາອິດ / Meena AbdelMaseeh ວິທະຍາໄລ ຊື່ຫຍໍ້ຂອງ : Pablo F Damasceno ຊື່ຫຍໍ້ຂອງ : Pedro Mario Cruz e Silva ພາສາລາວ ປະເພດ ຊື່ຫຍໍ້ຂອງ : Shuichi Kawano ຊື່ຫຍໍ້ຂອງ : Sira Sriswasdi ດາວໂຫລດ Soo Young Park ຊື່ຫຍໍ້ຂອງ : Thomas M. Grist ພາສາລາວ Watsamon Jantarabenjakul ພາສາລາວ ຍິນດີຕ້ອນຮັບ Young Tak ວິທະຍາໄລ Lee ຊື່ຫຍໍ້ຂອງ : Xihong Lin ຊາຍ Joon Kwon ພາສາລາວ ເມນູ Feng ເມນູ N. Priest ຂໍຂອບໃຈ Turkey ຊື່ຫຍໍ້ຂອງ : Benjamin Glicksberg ພາສາລາວ ຊື່ຫຍໍ້ຂອງ : Seok Kim ພາສາລາວ Chia-Cheng Lee ຊື່ຫຍໍ້ຂອງ : Chia Jung Hsu ຈີນ Lin ຊື່ຫຍໍ້ຂອງ : Chiu Ling Lai ຊື່ຫຍໍ້ຂອງ : Christopher P. Hess ປະເພດ Colin Compas ຫນ້າທໍາອິດ / Deepeksha Bhatia ຊື່ຫຍໍ້ຂອງ : Eric K Oermann Evan Leibovitz ຊື່ຫຍໍ້ຂອງ : Sasaki Hitoshi Mori ຊື່ຫຍໍ້ຂອງ : Isaac ຊື່ຫຍໍ້ຂອງ : Jae Ho Krishna Nand Keshava ມັງກອນ ປະເພດ Li-Chen Fu ຄົ້ນຫາທີ່ດີທີ່ສຸດທີ່ດີທີ່ສຸດທີ່ດີທີ່ສຸດທີ່ດີທີ່ສຸດທີ່ດີທີ່ສຸດ ມີນາ Fralick ມົນ Kyu Kang ພາສາລາວ ຂໍຂອບໃຈ ຊື່ຫຍໍ້ຂອງ : Vateekul ຊື່ຫຍໍ້ຂອງ ຊາຍ Hickman ຊື່ຫຍໍ້ຂອງ : Sharmila Majumdar Shelley L. McLeod ດາວນ໌ໂຫລດ Reed ວິທະຍາໄລ Stefan Gräf ຊື່ຫຍໍ້ຂອງ : Stephanie Harmon ປະເພດ Tatsuya Kodama ຂໍຂອບໃຈ ມັງກອນ 2018 ຊື່ຫຍໍ້ຂອງ : Vitor Lima de Trabor ຊື່ຫຍໍ້ຂອງ : Yothin Rakvongthai ວິທະຍາໄລ Yu Rim Lee ຊື່ຫຍໍ້ຂອງ : Wen Wen ຊື່ຫຍໍ້ຂອງ : Fiona J Gilbert ຊື່ຫຍໍ້ຂອງ : Mona G. Flores ປະເພດ Li ອັດຕະໂນມັດ ການຝຶກອົບຮົມການຝຶກອົບຮົມການຝຶກອົບຮົມການຝຶກອົບຮົມການຝຶກອົບຮົມການຝຶກອົບຮົມການຝຶກອົບຮົມການຝຶກອົບຮົມການຝຶກອົບຮົມການຝຶກອົບຮົມການຝຶກອົບຮົມການຝຶກອົບຮົມການຝຶກອົບຮົມການຝຶກອົບຮົມການຝຶກອົບຮົມການຝຶກອົບຮົມການຝຶກອົບຮົມການຝຶກອົບຮົມການຝຶກອົບຮົມການຝຶກອົບຮົມການຝຶກອົບຮົມການຝຶກອົບຮົມການຝຶກອົບຮົມການຝຶກອົບຮົມການຝຶກອົບຮົມການຝຶກອົບຮົມການຝຶກອົບຮົມການຝຶກອົບຮົມການຝຶກອົບຮົມການຝຶກອົບຮົມການຝຶກອົບຮົມການຝຶກອົບຮົມການຝຶກອົບຮົມ ຫນ້າທໍາອິດ ວິທະຍາໄລ, ວິທະຍາໄລ, ວິທະຍາໄລແລະວິທະຍາສາດຂໍ້ມູນໄດ້ຮ່ວມມືໃນປັດຈຸບັນໃນປັດຈຸບັນໃນປັດຈຸບັນໃນປັດຈຸບັນໃນປັດຈຸບັນໃນປັດຈຸບັນໃນປັດຈຸບັນໃນປັດຈຸບັນໃນປັດຈຸບັນໃນປັດຈຸບັນໃນປັດຈຸບັນໃນປັດຈຸບັນໃນປັດຈຸບັນ , ຜູ້ສະຫນອງການຄົ້ນຄວ້າແລະອຸດສາຫະກໍາການຄົ້ນຄວ້າຄົ້ນຄວ້າຄົ້ນຄວ້າຄົ້ນຄວ້າຄົ້ນຄວ້າຄົ້ນຄວ້າຄົ້ນຄວ້າຄົ້ນຄວ້າຄົ້ນຄວ້າຄົ້ນຄວ້າຄົ້ນຄວ້າຄົ້ນຄວ້າຄົ້ນຄວ້າຄົ້ນຄວ້າຄົ້ນຄວ້າຄົ້ນຄວ້າ , , , , , , ການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນ , , ການຄົ້ນຄວ້າຂໍ້ມູນແລະອຸປະກອນການສຸຂະພາບອຸປະກອນການສຸຂະພາບອຸປະກອນການສຸຂະພາບອຸປະກອນການສຸຂະພາບອຸປະກອນການສຸຂະພາບອຸປະກອນການສຸຂະພາບອຸປະກອນການສຸຂະພາບອຸປະກອນການສຸຂະພາບອຸປະກອນການສຸຂະພາບອຸປະກອນການສຸຂະພາບອຸປະກອນການສຸຂະພາບອຸປະກອນການສຸຂະພາບອຸປະກອນການສຸຂະພາບອຸປະກອນການສຸຂະພາບອຸປະກອນການສຸຂະພາບອຸປະກອນການສຸຂະພາບ , ການປິ່ນປົວຂອງການປິ່ນປົວຂອງການປິ່ນປົວຂອງການປິ່ນປົວຂອງການປິ່ນປົວຂອງການປິ່ນປົວຂອງການປິ່ນປົວຂອງການປິ່ນປົວຂອງການປິ່ນປົວຂອງການປິ່ນປົວຂອງການປິ່ນປົວຂອງການປິ່ນປົວຂອງການປິ່ນປົວຂອງການປິ່ນປົວຂອງການປິ່ນປົວຂອງການປິ່ນປົວຂອງການປິ່ນປົວຂອງການປິ່ນປົວຂອງການປິ່ນປົວຂອງການປິ່ນປົວຂອງການປິ່ນປົວຂອງການປິ່ນປົວຂອງການປິ່ນປົວຂອງການປິ່ນປົວຂອງການປິ່ນປົວ , , . 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 ວິທະຍາໄລຂອງວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍາໄລວິທະຍ , , , CXR ໄດ້ຖືກເລືອກເປັນການເຂົ້າລະຫັດການຮູບພາບໂດຍວ່າມັນໄດ້ຖືກເຂົ້າລະຫັດຢ່າງກວ້າງຂວາງແລະໄດ້ຖືກນໍາໃຊ້ຢ່າງກວ້າງຂວາງໂດຍ guidelines ເຊັ່ນ ACR ບໍລິສັດ Fleischner ວິທະຍາໄລ WHO ວິທະຍາໄລ Thoracic ວິທະຍາສາດຄອມພິວເຕີ - National Ministry of Health COVID Handbooks and radiology societies around the world . The output of the CDS model was a score, termed CORISK , ທີ່ສະຫນັບສະຫນູນຄວາມຕ້ອງການການສະຫນັບສະຫນູນ oxygen ແລະທີ່ສາມາດຊ່ວຍເຫຼືອໃນ trialing patients ໂດຍ clinicians frontline , , ສະຫນັບສະຫນູນການຄຸ້ມຄອງຄຸນນະພາບຄຸນນະພາບຄຸນນະພາບຄຸນນະພາບຄຸນນະພາບຄຸນນະພາບຄຸນນະພາບຄຸນນະພາບຄຸນນະພາບຄຸນນະພາບ ໃນປັດຈຸບັນຈໍານວນຫຼາຍຂອງມາດຕະຖານ AI, ລວມທັງມາດຕະຖານ CDS ທີ່ຜ່ານມາ, ໄດ້ຮັບການຝຶກອົບຮົມແລະການຢັ້ງຢືນກ່ຽວກັບຂໍ້ມູນ 'ຍິ່ງໃຫຍ່' ທີ່ບໍ່ມີຄຸນນະສົມບັດ , , potentially resulting in overfitting and lower generalizability. This can be mitigated by training with diverse data from multiple sites without centralization of data using methods such as transfer learning , FL ແມ່ນວິທີທີ່ຖືກນໍາໃຊ້ສໍາລັບການຝຶກອົບຮົມມາດຕະຖານ AI ໃນສະພາບແວດລ້ອມຂໍ້ມູນທີ່ແຕກຕ່າງກັນ, ໂດຍບໍ່ມີການສົ່ງອອກຫຼືຕິດຕໍ່ຂໍ້ມູນພາຍໃຕ້ສະຖານທີ່ເລີ່ມຕົ້ນຂອງເຂົາເຈົ້າ. ນອກເຫນືອໄປຈາກການນໍາໃຊ້ສໍາລັບອຸດສາຫະກໍາຈໍານວນຫຼາຍ, FL ໄດ້ຖືກນໍາສະເຫນີຢ່າງກວ້າງຂວາງສໍາລັບການຄົ້ນຄ້ວາດ້ານວິຊາການດ້ານວິຊາການ . 18 19 20 21 22 23 24 25 26 27 28 29 30 27 31 32 33 34 35 36 ການຝຶກອົບຮົມ federated ສະຫນັບສະຫນູນການເລີ່ມຕົ້ນຢ່າງໄວ້ວາງໃຈຂອງການທົດສອບການທົດສອບທີ່ມີການປັບປຸງ traceability ຂອງຂໍ້ມູນແລະການຄາດຄະເນຂອງການປ່ຽນແປງ algoritmic ແລະຜົນປະໂຫຍດ ວິທີການຫນຶ່ງສໍາລັບ FL, ທີ່ຖືກຊື່ວ່າ client-server, ສະຫນອງຮູບແບບ 'non-trained' ກັບການບໍລິການອື່ນໆ ( 'nodes') ທີ່ປະຕິບັດການຝຶກອົບຮົມ partial, ໃນຂະນະດຽວກັນໃຫ້ຜະລິດຕະພັນຫຼັງຈາກນັ້ນເພື່ອໄດ້ຮັບການເຊື່ອມຕໍ່ໃນການບໍລິການຕົ້ນຕໍ ( 'federated'). ນີ້ໄດ້ຖືກນໍາໃຊ້ເປັນໂຄງການ iterative ໃນເວລາທີ່ການຝຶກອົບຮົມໄດ້ completed . 37 36 ການຄຸ້ມຄອງຂໍ້ມູນສໍາລັບ FL ໄດ້ຖືກປິ່ນປົວໃນສະຖານທີ່, ການປິ່ນປົວບັນຫາຄວາມປອດໄພ, ມີພຽງແຕ່ການເຄື່ອນໄຫວຮູບແບບຫຼື gradients ຂອງການສື່ສານລະຫວ່າງເວັບໄຊທ໌ຂອງລູກຄ້າແລະການບໍລິການ federated , FL ໄດ້ສະແດງໃຫ້ເຫັນຄວາມຍິນດີໃນອຸປະກອນການ imaging ອັດຕະໂນມັດທີ່ຜ່ານມາ , , , ການທົບທວນຄືນ COVID-19 , , ປະເພດທີ່ຍິ່ງໃຫຍ່ແມ່ນຮູບແບບການຄາດຄະເນດິນຟ້າອາກາດໃນປັກສະນະທີ່ຖືກປິ່ນປົວໂດຍ SARS-COV-2 ທີ່ນໍາໃຊ້ຄຸນນະພາບຄຸນນະພາບຄຸນນະພາບຄຸນນະພາບຄຸນນະພາບຄຸນນະພາບຄຸນນະພາບຄຸນນະພາບ . 38 39 40 41 42 43 8 44 45 46 ພວກເຮົາມີຄວາມຕ້ອງການທີ່ຈະພັດທະນາຮູບແບບທີ່ເຂັ້ມແຂງ, ອັດຕະໂນມັດທີ່ສາມາດສະຫນັບສະຫນູນການທົດສອບຜູ້ຊ່ຽວຊານ. ພວກເຮົາມີປະສົບການວ່າຮູບແບບ CDS ສາມາດໄດ້ຮັບການ federated ປະສິດທິພາບໂດຍໃຊ້ການເຂົ້າລະຫັດຂໍ້ມູນທີ່ຖືກນໍາໃຊ້ຢ່າງກວ້າງຂວາງໃນອຸປະກອນການຄົ້ນຄ້ວາແລະທີ່ບໍ່ໄດ້ລົງທະບຽນຢ່າງກວ້າງຂວາງກ່ຽວກັບການຄົ້ນຄວ້າທີ່ກ່ຽວຂ້ອງກັບຜູ້ຊ່ຽວຊານຂອງສະພາບແວດລ້ອມ (ເຊັ່ນດຽວກັນກັບການເຂົ້າລະຫັດສຸຂະພາບຫຼັກສູດຫຼືການເຂົ້າລະຫັດສຸຂະພາບ). ໃນຂະນະທີ່ການນໍາໃຊ້ຮູບແບບ CDS ມີຜົນປະໂຫຍດທີ່ດີເລີດ, ການທົດສອບຮູບພາບທີ່ສໍາຄັນ, ການທົດສອບ ການທົດສອບຂອງພວກເຮົາແມ່ນວ່າ EXAM ຈະເຮັດວຽກດີກວ່າເຊັ່ນດຽວກັນກັບຮູບແບບທ້ອງຖິ່ນແລະຈະທົ່ວໄປດີກວ່າໃນໄລຍະລະບົບການຄຸ້ມຄອງຄຸນນະພາບ. ຄວາມຄິດເຫັນ ລະຫັດ QR ໂມເລກຸນ EXAM ໄດ້ຖືກສ້າງຕັ້ງຂຶ້ນໃນຮູບແບບ CDS ທີ່ຜ່ານມາ ໃນທັງຫມົດ, 20 ຄຸນນະສົມບັດ (19 ຈາກ EMR ແລະຫນຶ່ງ CXR) ໄດ້ຖືກນໍາໃຊ້ເປັນການເຂົ້າລະຫັດກັບມາດຕະຖານ. ຄຸນນະສົມບັດຂອງການເຂົ້າລະຫັດ (ວ່າຈະເປັນ 'ຈິງພື້ນຖານ') ໄດ້ຖືກເຂົ້າລະຫັດໂດຍຜ່ານການປິ່ນປົວຂອງຜູ້ປິ່ນປົວທີ່ມີອຸປະກອນອາກາດຫຼັງຈາກ 24 ແລະ 72 ຊົ່ວໂມງຈາກການເຂົ້າລະຫັດເລີ່ມຕົ້ນໃນບໍລິສັດສະພາບແວດລ້ອມ (ED). ຄຸນນະສົມບັດທີ່ໄດ້ຮັບການເຂົ້າລະຫັດແລະການເຂົ້າລະຫັດທີ່ຖືກຕ້ອງສາມາດໄດ້ຮັບການຊອກຫາໃນ TABLE . 27 1 ປະເພດການປິ່ນປົວທີ່ມີ oxygen ແມ່ນອາຍຸ (RA), oxygen ນ້ໍາຫນັກ (LFO), oxygen ນ້ໍາຫນັກ (HFO) / ventilation noninvasive (NIV) ຫຼື ventilation ເຄື່ອງຈັກ (MV). ຖ້າຫາກວ່າຜູ້ປິ່ນປົວໄດ້ເລີ່ມຕົ້ນການປິ່ນປົວໃນປັດຈຸບັນ, ການປິ່ນປົວທີ່ມີ oxygen ໄດ້ເລີ່ມຕົ້ນການປິ່ນປົວໃນປັດຈຸບັນໄດ້ເລີ່ມຕົ້ນການປິ່ນປົວໃນປັດຈຸບັນ. ສໍາລັບຄຸນນະສົມບັດ EMR, ພຽງແຕ່ຄຸນນະສົມບັດທໍາອິດທີ່ໄດ້ຮັບການຊອກຫາໃນ ED ໄດ້ຖືກນໍາໃຊ້ແລະ preprocessing data ໄດ້ປະກອບດ້ວຍ deidentification, imputation value missing ແລະ normalization ກັບ zero-median ແລະ unit variance. ສໍາລັບຮູບພາບ CXR, ພຽງແຕ່ຄຸນສົມບັດທໍາອິດທີ່ໄດ້ຮັບໃນ ED ໄດ້ຖືກນໍາໃຊ້. ໃນປັດຈຸບັນ, ການນໍາໃຊ້ຮູບແບບນີ້ໄດ້ຖືກນໍາໃຊ້ໂດຍການນໍາໃຊ້ການນໍາໃຊ້ຮູບແບບການນໍາສະເຫນີ EMR ແລະ CXR, ການນໍາໃຊ້ຮູບແບບການນໍາສະເຫນີຂອງການນໍາສະເຫນີຂອງການນໍາສະເຫນີຂອງການນໍາສະເຫນີຂອງການນໍາສະເຫນີຂອງການນໍາສະເຫນີຂອງການນໍາສະເຫນີຂອງການນໍາສະເຫນີຂອງການນໍາສະເຫນີຂອງການນໍາສະເຫນີຂອງການນໍາສະເຫນີຂອງການນໍາສະເຫນີຂອງການນໍາສະເຫນີ. ລະບົບການຜະລິດຂອງມາດຕະຖານແມ່ນຈຸດຄົ້ນບໍ່ແຮ່, ລວມທັງຈຸດຄົ້ນບໍ່ແຮ່ EXAM, ທີ່ເປັນຈຸດຄົ້ນບໍ່ແຮ່ໃນຂະຫນາດ 0-1 ສໍາລັບທຸກປະມານ 24 ແລະ 72 ຊົ່ວໂມງທີ່ກ່ຽວຂ້ອງກັບລາຍລະອຽດທີ່ຜ່ານມາ. ວິທີການ ລະຫັດ QR ໂມເລກຸນ EXAM ໄດ້ຖືກຝຶກອົບຮົມໂດຍໃຊ້ການຄົ້ນຄວ້າຂອງ 16148 cases, ເຮັດໃຫ້ມັນບໍ່ພຽງແຕ່ຫນຶ່ງໃນຕົວແບບ FL ທີ່ຕົ້ນຕໍສໍາລັບ COVID-19 ແຕ່ຍັງເປັນໂຄງການພັດທະນາຂະຫນາດໃຫຍ່ແລະ multi-continent ໃນ AI ທີ່ກ່ຽວຂ້ອງຄົ້ນຄວ້າ (ຮູບ. ) ຂໍ້ມູນລະຫວ່າງເວັບໄຊທ໌ໄດ້ບໍ່ຖືກ harmonized ທີ່ຜ່ານການຊັດເຈນແລະ, ໃນສະພາບແວດລ້ອມຂອງສະພາບແວດລ້ອມການຄອມພິວເຕີຄອມພິວເຕີໃນຊີວິດທີ່ແທ້ຈິງ, ການເຊື່ອມຕໍ່ຢ່າງກວ້າງຂວາງຂອງການເຂົ້າລະຫັດຂໍ້ມູນໄດ້ບໍ່ໄດ້ຮັບການປະຕິບັດໂດຍຜູ້ຊ່ຽວຊານ (ຮູບ. ). ປະເພດ A ລະຫັດ QR ກາຕູນໂລກທີ່ສະແດງໃຫ້ເຫັນ 20 ເວັບໄຊທ໌ທີ່ແຕກຕ່າງກັນຂອງລູກຄ້າທີ່ຊ່ວຍໃຫ້ການສຶກສາ EXAM. , ປະເພດຂອງກໍລະນີທີ່ສະຫນັບສະຫນູນໂດຍບໍລິສັດຫຼືເວັບໄຊທ໌ (client 1 ສະຫນັບສະຫນູນເວັບໄຊທ໌ທີ່ສະຫນັບສະຫນູນຈໍານວນຫຼາຍຂອງກໍລະນີ). , ການຈັດການຄວາມເຂັ້ມແຂງຂອງ X-ray chest ໃນທຸກສະຖານທີ່ຂອງລູກຄ້າ. , ປະເພດຂອງຜູ້ຊ່ຽວຊານໃນທຸກສະຖານທີ່ຜູ້ຊ່ຽວຊານ, ທີ່ສະແດງໃຫ້ເຫັນປະເພດທີ່ຕ່ໍາແລະສູງສຸດ ( asters), ປະເພດປະເພດປະເພດ ( triangles) ແລະປະເພດປະເພດປະເພດ (bars horizontal). . a b c d 1 ການຝຶກອົບຮົມຂອງມາດຕະຖານ FL ໄດ້ຖືກປັບປຸງຢ່າງກວ້າງຂວາງກັບມາດຕະຖານ FL Global ໃນຂໍ້ມູນການທົດສອບຂອງລູກຄ້າ. ການຝຶກອົບຮົມຂອງມາດຕະຖານ FL ໄດ້ຖືກປັບປຸງຢ່າງກວ້າງຂວາງ ( « 1 × 10-3, ການທົດສອບການຢັ້ງຢືນ Wilcoxon) ຂອງ 16% (ຫຼັກສູດໂດຍ AUC ອຸນຫະພູມໃນເວລາທີ່ໃຊ້ຮູບແບບໃນຊຸດທົດສອບພື້ນຖານທີ່ແຕກຕ່າງກັນ: ຈາກ 0.795 ກັບ 0.920, ຫຼື 12.5 ຊົ່ວໂມງ) (ຮູບ. ). It also resulted in 38% generalizability improvement (as defined by average AUC when running the model on all test sets: from 0.667 to 0.920, or 25.3 percentage points) of the best global model for prediction of 24-h oxygen treatment compared with models trained only on a site’s own data (Fig. ສໍາລັບຜົນປະໂຫຍດການຄາດຄະເນຂອງການປິ່ນປົວ oxygen 72 ຊົ່ວໂມງ, ການຝຶກອົບຮົມຮູບແບບທົ່ວໂລກທີ່ດີທີ່ສຸດໄດ້ຮັບການປັບປຸງຜົນປະໂຫຍດປະມານ 18% compared to locally trained models, ໃນຂະນະທີ່ generalizability ຂອງຮູບແບບທົ່ວໂລກໄດ້ປັບປຸງປະມານ 34% (Extended Data Fig. ). ຄວາມປອດໄພຂອງລັກສະນະຂອງພວກເຮົາໄດ້ຖືກຢັ້ງຢືນໂດຍການດາວໂຫລດສາມຊົ່ວໂມງຂອງການຝຶກອົບຮົມທ້ອງຖິ່ນແລະ FL ໃນລາຍລະອຽດຂໍ້ມູນທີ່ແຕກຕ່າງກັນ. P ລະຫັດ QR ພາສາລາວ 1 , ປະສິດທິພາບກ່ຽວກັບການທົດສອບຂອງລູກຄ້າຂອງພວກເຮົາມີການຄາດຄະເນຂອງການປິ່ນປົວ oxygen ຂອງ 24 ຊົ່ວໂມງສໍາລັບຮູບແບບທີ່ໄດ້ຮັບການຝຶກອົບຮົມໂດຍໃຊ້ຂໍ້ມູນພື້ນຖານພຽງແຕ່ (Local) ໃນຂະນະທີ່ຮູບແບບທົ່ວໂລກທີ່ດີທີ່ສຸດທີ່ມີຢູ່ໃນການບໍລິການ (FL (ເບິ່ງທີ່ດີທີ່ສຸດ)). Av., ປະສິດທິພາບການທົດສອບປະມານທັງຫມົດ. , Generalizability (ຄຸນນະພາບປະມານໃນຂໍ້ມູນການທົດສອບຂອງເວັບໄຊທ໌ອື່ນໆ, ເຊັ່ນ: AUC ອຸນຫະພູມ) ເປັນຜົນປະໂຫຍດຂອງຂະຫນາດຊຸດຂໍ້ມູນຂອງລູກຄ້າ (ບໍ່ມີກໍລະນີ). ສາຍ horizontal ສີຂາວສະແດງໃຫ້ເຫັນຄຸນນະພາບການຄຸນນະພາບຂອງຮູບແບບທົ່ວໂລກທີ່ດີທີ່ສຸດ. ການຄຸນນະພາບຂອງ 18 ຂອງ 20ລູກຄ້າແມ່ນສະແດງໃຫ້ເຫັນ, ໃນຂະນະທີ່ລູກຄ້າ 12 ມີຜົນປະໂຫຍດພຽງແຕ່ສໍາລັບ 72 ຊົ່ວໂມງຂອງ oxygen (Extended Data Fig. ) and client 14 had cases only with RA treatment, such that the evaluation metric (av. AUC) was not applicable in either of these cases ( ) Data ສໍາລັບ client 14 ຍັງໄດ້ຖືກກວດສອບຈາກການຄອມພິວເຕີຂອງ generalizability ອັດຕະໂນມັດໃນມາດຕະຖານພື້ນຖານ. a b 1 Methods Local models that were trained using unbalanced cohorts (for example, mostly mild cases of COVID-19) markedly benefited from the FL approach, with a substantial improvement in prediction average AUC performance for categories with only a few cases. This was evident at client site 16 (an unbalanced dataset), with most patients experiencing mild disease severity and with only a few severe cases. The FL model achieved a higher true-positive rate for the two positive (severe) cases and a markedly lower false-positive rate compared to the local model, both shown in the receiver operating characteristic (ROC) plots and confusion matrices (Fig. ແລະ Extended Data Fig. ). More important, the generalizability of the FL model was considerably increased over the locally trained model. ພາສາລາວ 2 , ROC at client site 16, with unbalanced data and mostly mild cases. , ROC of the local model at client site 12 (a small dataset), mean ROC of models trained on larger datasets corresponding to the five client sites in the Boston area (1, 4, 5, 6, 8) and ROC of the best global model in prediction of 72-h oxygen treatment for different thresholds of EXAM score (left, middle, right). The mean ROC is calculated based on five locally trained models while the gray area denotes the ROC standard deviation. ROCs for three different cutoff values ( ການທົດສອບ Risk Score ໄດ້ຖືກສະແດງໃຫ້ເຫັນ. Pos ແລະ neg ສະ ຫນັບ ສະ ຫນັບ ສະ ຫນັບ ສະ ຫນັບ ສະ ຫນັບ ສະ ຫນັບ ສະ ຫນັບ ສະ ຫນັບ ສະ ຫນັບ ສະ ຫນັບ ສະ ຫນັບ ສະ ຫນັບ ສະ ຫນັບ ສະ ຫນູນ ສະ ຫນັບ ສະ ຫນູນ ສະ ຫນັບ ສະ ຫນູນ ສະ ຫນັບ ສະ ຫນູນ ສະ ຫນູນ ສະ ຫນູນ ສະ ຫນູນ ສະ ຫນູນ ສະ ຫນູນ ສະ ຫນູນ ສະ ຫນູນ ສະ ຫນູນ ສະ ຫນູນ ສະ ຫນູນ ສະ ຫນູນ ສະ ຫນູນ ສະ ຫນູນ ສະ ຫນູນ ສະ ຫນູນ ສະ ຫນູນ ສະ ຫນູນ ສະ ຫນູນ ສະ ຫນູນ ສະ ຫນູນ ສະ ຫນູນ ສະ ຫນູນ ສະ ຫນູນ ສະ ຫນູນ ສະ ຫນູນ. a b t In the case of client sites with relatively small datasets, the best FL model markedly outperformed not only the local model but also those trained on larger datasets from five client sites in the Boston area of the USA (Fig. ລະຫັດ QR 3B ໂມເລກຸນທົ່ວໄປໄດ້ເຮັດວຽກດີໃນການຄາດຄະເນຄວາມຕ້ອງການອາກາດໃນ 24 / 72 ຊົ່ວໂມງໃນຜູ້ป่วยທັງ COVID positive ແລະ negative (Extended Data Fig. ລະຫັດ QR 3 Validation at independent sites Following initial training, EXAM was subsequently tested at three independent validation sites: Cooley Dickinson Hospital (CDH), Martha’s Vineyard Hospital (MVH) and Nantucket Cottage Hospital (NCH), all in Massachusetts, USA. The model was not retrained at these sites and it was used only for validation purposes. The cohort size and model inference results are summarized in Table , and the ROC curves and confusion matrices for the largest dataset (from CDH) are shown in Fig. ສະຖານທີ່ການເຮັດວຽກໄດ້ຖືກກໍານົດໃຫ້ເຫັນວ່າການປິ່ນປົວ (ຫຼືຊີວິດ) ໃນລະຫວ່າງການ ventilation nonmechanical ແລະ ventilation mechanical (MV). ຮູບແບບການຝຶກອົບຮົມທັງຫມົດຂອງ FL, EXAM, ໄດ້ຮັບ AUC ອຸນຫະພູມຂອງ 0.944 ແລະ 0.924 ສໍາລັບກິດຈະກໍາການຄາດຄະເນດິນ 24 ແລະ 72 ຊົ່ວໂມງເຊັ່ນດຽວກັນ (ຕາຕະລາງ) ), which exceeded the average performance among sites used in training EXAM. For prediction of MV treatment (or death) at 24 h, EXAM achieved a sensitivity of 0.950 and specificity of 0.882 at CDH, and a sensitivity of 1.000 specificity of 0.934 at MVH. NCH did not have any cases with MV/death at 24 h. In regard to 72-h MV prediction, EXAM achieved a sensitivity of 0.929 and specificity of 0.880 at CDH, sensitivity of 1.000 and specificity of 0.976 at MVH and sensitivity of 1.000 and specificity of 0.929 at NCH. 2 4 2 , , Performance (ROC) (top) and confusion matrices (bottom) of the EXAM FL model on the CDH dataset for prediction of oxygen requirement at 24 h ( ) 72 ຊົ່ວໂມງ ( ). ROCs for three different cutoff values ( ) of the EXAM risk score are shown. a b a b t For MV at CDH at 72 h, EXAM had a low false-negative rate of 7.1%. Representative failure cases are presented in Extended Data Fig. , ສະແດງໃຫ້ເຫັນສອງກໍລະນີ false-negative ຈາກ CDH ໃນຂະນະທີ່ກໍລະນີຫນຶ່ງມີຈໍານວນຫຼາຍຂອງຄຸນນະສົມບັດຂໍ້ມູນ EMR ທີ່ບໍ່ມີແລະອື່ນໆມີ CXR ມີ artefact ການเคลื่อนไหวແລະຄຸນສົມບັດ EMR ທີ່ບໍ່ມີ. 4 ການນໍາໃຊ້ Privacy Differential ການຝຶກອົບຮົມການຝຶກອົບຮົມການຝຶກອົບຮົມການຝຶກອົບຮົມການຝຶກອົບຮົມການຝຶກອົບຮົມ or even the reconstruction of training images from the model gradients themselves ເພື່ອຕອບສະຫນອງຄວາມປອດໄພເຫຼົ່ານີ້, ການປັບປຸງຄວາມປອດໄພໄດ້ຖືກນໍາໃຊ້ເພື່ອປັບປຸງຄວາມປອດໄພໃນກໍລະນີຂອງຂໍ້ມູນ "interception" ໃນໄລຍະການສື່ສານ Site-server ພວກເຮົາມີປະສົບການເຕັກໂນໂລຊີເພື່ອປ້ອງກັນການກວດສອບຂໍ້ມູນຂອງ FL, ແລະເພີ່ມການປະໂຫຍດຄວາມປອດໄພທີ່ພວກເຮົາມີຄວາມຮູ້ສຶກວ່າຈະຊ່ວຍໃຫ້ບໍລິສັດຫຼາຍກວ່າການນໍາໃຊ້ FL. ພວກເຮົາມີຄວາມຮັບຜິດຊອບທີ່ຜ່ານມາທີ່ສະແດງໃຫ້ເຫັນວ່າການກວດສອບຂະຫນາດນ້ອຍ, ແລະເຕັກໂນໂລຊີອື່ນໆທີ່ແຕກຕ່າງກັນຂອງຄວາມປອດໄພ, ສາມາດຖືກນໍາໃຊ້ຢ່າງວ່ອງໄວໃນ FL . Through investigation of a partial weight-sharing scheme , , , we showed that models can reach a comparable performance even when only 25% of weight updates are shared (Extended Data Fig. ). 47 48 49 50 50 51 52 5 ພາສາລາວ This study features a large, real-world healthcare FL study in terms of number of sites and number of data points used. We believe that it provides a powerful proof-of-concept of the feasibility of using FL for fast and collaborative development of needed AI models in healthcare. Our study involved multiple sites across four continents and under the oversight of different regulatory bodies, and thus holds the promise of being provided to different regulated markets in an expedited way. The global FL model, EXAM, proved to be more robust and achieved better results at individual sites than any model trained on only local data. We believe that consistent improvement was achieved owing to a larger, but also a more diverse, dataset, the use of data inputs that can be standardized and avoidance of clinical impressions/reported symptoms. These factors played an important part in increasing the benefits from this FL approach and its impact on performance, generalizability and, ultimately, the model’s usability. For a client site with a relatively small dataset, two typical approaches could be used for fitting a useful model: one is to train locally with its own data, the other is to apply a model trained on a larger dataset. For sites with small datasets, it would have been virtually impossible to build a performant deep learning model using only their local data. The finding, that these two approaches were outperformed on all three prediction tasks by the global FL model, indicates that the benefit for client sites with small datasets arising from participation in FL collaborations is substantial. This is probaby a reflection of FL’s ability to capture more diversity than local training, and to mitigate the bias present in models trained on a homogenous population. An under-represented population or age group in one hospital/region might be highly represented in another region—such as children who might be differentially affected by COVID-19, including disease manifestations in lung imaging . 46 ການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢືນການຢັ້ງຢືນການຢັ້ງຢືນການຢືນການຢັ້ງຢືນການຢ , as well as at different sites that were not a part of the EXAM training. 53 Over 200 prediction models to support decision-making in patients with COVID-19 have been published . Unlike the majority of publications focused on diagnosis of COVID-19 or prediction of mortality, we predicted oxygen requirements that have implications for patient management. We also used cases with unknown SARS-COV-2 status, and so the model could provide input to the physician ahead of receiving a result for PCR with reverse transcription (RT–PCR), making it useful for a real-life clinical setting. The model’s imaging input is used in common practice, in contrast with models that use chest computed tomography, a nonconsensual diagnostic modality. The model’s design was constrained to objective predictors, unlike many published studies that leveraged subjective clinical impressions. The data collected reflect varied incidence rates, and thus the ‘population momentum’ we encountered is more diverse. This implies that the algorithm can be useful in populations with different incidence rates. 19 ການຢັ້ງຢືນ cohort patient ແລະ harmonization data ແມ່ນບໍ່ມີບັນຫາໃຫມ່ໃນຄົ້ນຄວ້າແລະວິທະຍາສາດຂໍ້ມູນ , but are further complicated, when using FL, given the lack of visibility on other sites’ datasets. Improvements to clinical information systems are needed to streamline data preparation, leading to better leverage of a network of sites participating in FL. This, in conjunction with hyperparameter engineering, can allow algorithms to ‘learn’ more effectively from larger data batches and adapt model parameters to a particular site for further personalization—for example, through further fine-tuning on that site . A system that would allow seamless, close-to real-time model inference and results processing would also be of benefit and would ‘close the loop’ from training to model deployment. 54 39 Because data were not centralized they are not readily accessible. Given that, any future analysis of the results, beyond what was derived and collected, is limited. ຄຸນນະພາບຂອງການຝຶກອົບຮົມ EXAM ແມ່ນປະສິດທິພາບທີ່ດີທີ່ສຸດສໍາລັບການຝຶກອົບຮົມ EXAM. ຄຸນນະພາບຂອງການຝຶກອົບຮົມ EXAM ແມ່ນປະສິດທິພາບທີ່ດີທີ່ສຸດສໍາລັບການຝຶກອົບຮົມ EXAM. ຄຸນນະພາບຂອງການຝຶກອົບຮົມ EXAM ແມ່ນປະສິດທິພາບທີ່ດີທີ່ສຸດສໍາລັບການຝຶກອົບຮົມ EXAM. ຄຸນນະພາບຂອງການຝຶກອົບຮົມ EXAM ແມ່ນປະສິດທິພາບທີ່ດີທີ່ສຸດສໍາລັບການຝຶກອົບຮົມ EXAM. ການຝຶກອົບຮົມ EXAM ແມ່ນປະສິດທິພາບທີ່ດີທີ່ສຸດສໍາລັບການຝຶກອົບຮົມ EXAM. Since our data access was limited, we did not have sufficient available information for the generation of detailed statistics regarding failure causes, post hoc, at most sites. However, we did study failure cases from the largest independent test site, CDH, and were able to generate hypotheses that we can test in the future. For high-performing sites, it seems that most failure cases fall into one of two categories: (1) low quality of input data—for example, missing data or motion artifact in CXR; or (2) out-of-distribution data—for example a very young patient. ໃນປັດຈຸບັນ, ພວກເຮົາມີຄວາມຄາດຄະເນດິນດີຕ້ອນຮັບເພື່ອທົດສອບຄວາມສາມາດຂອງການ "ຈຸດທະບຽນ" ໃນໄລຍະເວລາທີ່ແຕກຕ່າງກັນຂອງການປິ່ນປົວ. ພວກເຮົາມີຄວາມຮູ້ສຶກວ່າ, ເນື່ອງຈາກການຄາດຄະເນດິນດີຕ້ອນຮັບໃນໄລຍະ 20 ເວັບໄຊທ໌, ຄວາມປອດໄພນີ້ສາມາດຖືກປິ່ນປົວ. ຄຸນນະສົມບັດທີ່ຈະປັບປຸງປະເພດການຮ່ວມມືຂະຫນາດໃຫຍ່ນີ້ແມ່ນຄວາມສາມາດໃນການຄາດຄະເນຜົນປະໂຫຍດຂອງເວັບໄຊທ໌ລູກຄ້າທັງ ຫມົດ ເພື່ອປັບປຸງຮູບແບບ FL ໃນທົ່ວໂລກ. ນີ້ຈະຊ່ວຍໃນການຄັດເລືອກເວັບໄຊທ໌ລູກຄ້າ, ແລະການຄາດຄະເນການຊື້ຂໍ້ມູນແລະການ annotation ປະຫວັດສາດ. ນີ້ແມ່ນສໍາຄັນຢ່າງກວ້າງຂວາງໃນຂະນະທີ່ຄ່າໃຊ້ຈ່າຍສູງແລະອຸປະກອນເສີມຂະຫນາດໃຫຍ່ຂອງການປະຫວັດສາດຂອງການປະຫວັດສາດຂະຫນາດໃຫຍ່ເຫຼົ່ານີ້, ແລະມັນຈະຊ່ວຍໃຫ້ການປະຫວັດສາດເຫຼົ່ານີ້ໄດ້ຮັບການປະຫວັດສາດຫຼາຍກ່ວາຂະຫນາດໃຫຍ່ຂອງຕົວຢ່າງຂໍ້ມູນ. Future approaches may incorporate automated hyperparameter searching ວິທະຍາໄລ Neural Architecture and other automated machine learning ການປິ່ນປົວເພື່ອຊອກຫາການຝຶກອົບຮົມທີ່ດີທີ່ສຸດສໍາລັບເວັບໄຊທ໌ລູກຄ້າຢ່າງງ່າຍດາຍ. 55 56 57 Known issues of batch normalization (BN) in FL motivated us to fix our basic model for image feature extraction to reduce the divergence between unbalanced client sites. Future work might explore different types of normalization techniques to allow the training of AI models in FL more effectively when client data are nonindependent and identically distributed. 58 49 ການເຮັດວຽກທີ່ຜ່ານມາກ່ຽວກັບການປິ່ນປົວຄວາມປອດໄພໃນອຸປະກອນ FL ໄດ້ສ້າງຄວາມປອດໄພກ່ຽວກັບການປິ່ນປົວຄວາມປອດໄພໃນໄລຍະການຝຶກອົບຮົມຮູບແບບ . Meanwhile, protection algorithms remain underexplored and constrained by multiple factors. While differential privacy algorithms , , ການປົກປັກຮັກສາທີ່ດີ, ພວກເຮົາມີຄວາມສາມາດໃນການປະສິດທິພາບຂອງຮູບແບບ. algorithms ການເຂົ້າລະຫັດ, ເຊັ່ນ: ການເຂົ້າລະຫັດ homomorphic , maintain performance but may substantially increase message size and training time. A quantifiable way to measure privacy would allow better choices for deciding the minimal privacy parameters necessary while maintaining clinically acceptable performance , , . 59 36 48 49 60 36 48 49 Following further validation, we envision deployment of the EXAM model in the ED setting as a way to evaluate risk at both the per-patient and population level, and to provide clinicians with an additional reference point when making the frequently difficult task of triaging patients. We also envision using the model as a more sensitive population-level metric to help balance resources between regions, hospitals and departments. Our hope is that similar FL efforts can break the data silos and allow for faster development of much-needed AI models in the near future. Methods Ethics approval All procedures were conducted in accordance with the principles for human experimentation as defined in the Declaration of Helsinki and International Conference on Harmonization Good Clinical Practice guidelines, and were approved by the relevant institutional review boards at the following validation sites: CDH, MVH, NCH and at the following training sites: MGB, Mass General Hospital (MGH), Brigham and Women’s Hospital, Newton-Wellesley Hospital, North Shore Medical Center and Faulkner Hospital (all eight of these hospitals were covered under MGB’s ethics board reference, no. 2020P002673, and informed consent was waived by the instititional review board (IRB). Similarly, participation of the remaining sites was approved by their respective relevant institutional review processes: Children’s National Hospital in Washington, DC (no. 00014310, IRB certified exempt); NIHR Cambridge Biomedical Research Centre (no. 20/SW/0140, informed consent waived); The Self-Defense Forces Central Hospital in Tokyo (no. 02-014, informed consent waived); National Taiwan University MeDA Lab and MAHC and Taiwan National Health Insurance Administration (no. 202108026 W, informed consent waived); Tri-Service General Hospital in Taiwan (no. B202105136, informed consent waived); Kyungpook National University Hospital in South Korea (no. KNUH 2020-05-022, informed consent waived); Faculty of Medicine, Chulalongkorn University in Thailand (nos. 490/63, 291/63, informed consent waived); Diagnosticos da America SA in Brazil (no. 26118819.3.0000.5505, informed consent waived); University of California, San Francisco (no. 20-30447, informed consent waived); VA San Diego (no. H200086, IRB certified exempt); University of Toronto (no. 20-0162-C, informed consent waived); National Institutes of Health in Bethesda, Maryland (no. 12-CC-0075, informed consent waived); University of Wisconsin-Madison School of Medicine and Public Health (no. 2016-0418, informed consent waived); Memorial Sloan Kettering Cancer Center in New York (no. 20-194, informed consent waived); and Mount Sinai Health System in New York (no. IRB-20-03271, informed consent waived). MI-CLAIM guidelines for reporting of clinical AI models were followed (Supplementary Note ) 2 Study setting ການຄົ້ນຄວ້າໄດ້ນໍາໃຊ້ຂໍ້ມູນຈາກ 20 ວິທະຍາໄລ (fig. ): MGB, MGH, Brigham and Women’s Hospital, Newton-Wellesley Hospital, North Shore Medical Center and Faulkner Hospital; Children’s National Hospital in Washington, DC; NIHR Cambridge Biomedical Research Centre; The Self-Defense Forces Central Hospital in Tokyo; National Taiwan University MeDA Lab and MAHC and Taiwan National Health Insurance Administration; Tri-Service General Hospital in Taiwan; Kyungpook National University Hospital in South Korea; Faculty of Medicine, Chulalongkorn University in Thailand; Diagnosticos da America SA in Brazil; University of California, San Francisco; VA San Diego; University of Toronto; National Institutes of Health in Bethesda, Maryland; University of Wisconsin-Madison School of Medicine and Public Health; Memorial Sloan Kettering Cancer Center in New York; and Mount Sinai Health System in New York. Institutions were recruited between March and May 2020. Dataset curation started in June 2020 and the final data cohort was added in September 2020. Between August and October 2020, 140 independent FL runs were conducted to develop the EXAM model and, by the end of October 2020, EXAM was made public on NVIDIA NGC , , . Data from three independent sites were used for independent validation: CDH, MVH and NCH, all in Massachusetts, USA. These three hospitals had patient population characteristics different from the training sites. The data used for the algorithm validation consisted of patients admitted to the ED at these sites between March 2020 and February 2021, and that satisfied the same inclusion criteria of the data used to train the FL model. 1a 61 62 63 Data collection ເວັບໄຊທ໌ຂອງ 20 client ໄດ້ prepared total of 16,148 cases (both positive and negative) for the purposes of training, validation and testing of the model. ເວັບໄຊທ໌ຂອງຜູ້ຊ່ຽວຊານໄດ້ຊອກຫາຂໍ້ມູນດ້ານວິຊາການທີ່ກ່ຽວຂ້ອງກັບຜູ້ຊ່ຽວຊານທີ່ສະຫນັບສະຫນູນການເຂົ້າລະຫັດການທົດສອບ. ເວັບໄຊທ໌ຂອງຜູ້ຊ່ຽວຊານໄດ້ຊອກຫາຂໍ້ມູນກ່ຽວກັບການເຂົ້າລະຫັດທັງຫມົດຂອງການເຂົ້າລະຫັດ COVID-positive ຈາກເລີ່ມຕົ້ນຂອງການປິ່ນປົວໃນເດືອນມິຖຸນາ 2019 ແລະໃນເວລາທີ່ພວກເຂົາໄດ້ເລີ່ມຕົ້ນການຝຶກອົບຮົມພື້ນຖານສໍາລັບການທົດສອບ EXAM. ການຝຶກອົບຮົມພື້ນຖານທັງຫມົດໄດ້ເລີ່ມຕົ້ນໂດຍ 30 ກັນຍາ 2020. ເວັບໄຊທ໌ຂອງພວກເຂົາຍັງປະກອບມີຜູ້ຊ່ຽວຊານອື່ນໆໃນໄລຍະການທົດສອບ RT-PCR negative. ໃນຂະນະທີ່ຫຼາຍກ່ວາເວັບໄຊທ໌ຂອງພວກເຂົາມີຜູ້ຊ່ຽວຊ 1b A ‘case’ included a CXR and the requisite data inputs taken from the patient’s medical record. A breakdown of the cohort size of the dataset for each client site is shown in Fig. . The distribution and patterns of CXR image intensity (pixel values) varied greatly among sites owing to a multitude of patient- and site-specific factors, such as different device manufacturers and imaging protocols, as shown in Fig. . ປະຫວັດສາດທີ່ແຕກຕ່າງກັນໃນໄລຍະສະຖານທີ່, ປະຫວັດສາດທີ່ແຕກຕ່າງກັນໃນໄລຍະສະຖານທີ່ທີ່ແຕກຕ່າງກັນໃນໄລຍະສະຖານທີ່ແຕກຕ່າງກັນໃນໄລຍະສະຖານທີ່ແຕກຕ່າງກັນ (Extended Data Fig. ). 1b 1c,d 6 Patient inclusion criteria Patient inclusion criteria were: (1) patient presented to the hospital’s ED or equivalent; (2) patient had a RT–PCR test performed at any time between presentation to the ED and discharge from the hospital; (3) patient had a CXR in the ED; and (4) patient’s record had at least five of the EMR values detailed in Table , ທັງຫມົດໄດ້ຮັບໃນ ED, ແລະຜົນປະໂຫຍດທີ່ກ່ຽວຂ້ອງໄດ້ຮັບໃນໄລຍະການຝຶກອົບຮົມ. ເນື່ອງຈາກວ່າ, CXR, ຜົນປະໂຫຍດຫ້ອງທົດລອງແລະ Vital ໄດ້ຖືກນໍາໃຊ້ເປັນຜູ້ທໍາອິດທີ່ສາມາດໄດ້ຮັບການຝຶກອົບຮົມສໍາລັບການຝຶກອົບຮົມໃນໄລຍະການຝຶກອົບຮົມ ED. ຮູບແບບໄດ້ບໍ່ປະກອບມີ CXR, ຜົນປະໂຫຍດຫ້ອງທົດລອງຫຼື Vital ໄດ້ຮັບການຝຶກອົບຮົມຫຼັງຈາກການຝຶກອົບຮົມ ED. 1 ລະຫັດ QR ທັງຫມົດ, 21 ຄຸນນະສົມບັດ EMR ໄດ້ຖືກນໍາໃຊ້ເປັນການເຂົ້າລະຫັດສໍາລັບມາດຕະຖານ. ຄຸນນະສົມບັດຜົນກະທົບ (ທີ່ແມ່ນ, ຄວາມຄິດເຫັນທີ່ພື້ນຖານ) ໄດ້ຖືກເຂົ້າລະຫັດໂດຍຜ່ານຄວາມຕ້ອງການຂອງຜູ້ຊ່ຽວຊານຫຼັງຈາກ 24 ແລະ 72 ຊົ່ວໂມງຈາກການເຂົ້າລະຫັດທໍາອິດກັບ ED. ຄຸນນະສົມບັດ EMR ທີ່ຕ້ອງການແລະຜົນກະທົບທີ່ແຕກຕ່າງກັນສາມາດຊອກຫາໃນທ້ອງຖິ່ນ . 1 ການຈັດການຂອງການປິ່ນປົວ oxygen ໂດຍໃຊ້ອຸປະກອນທີ່ແຕກຕ່າງກັນໃນສະຖານທີ່ຄອມພິວເຕີທີ່ແຕກຕ່າງກັນຖືກສະແດງໃຫ້ເຫັນຢູ່ໃນ Fig Data Extended. ການນໍາໃຊ້ຂອງອຸປະກອນໃນເວລາທີ່ເຂົ້າໄປໃນ ED ແລະຫຼັງຈາກ 24 ແລະ 72 ຊົ່ວໂມງ. ຄວາມແຕກຕ່າງໃນການຈັດການ dataset ໃນລະຫວ່າງທີ່ໃຫຍ່ທີ່ສຸດແລະຂະຫນາດນ້ອຍທີ່ສຸດຂອງເວັບໄຊທ໌ຂອງລູກຄ້າສາມາດເບິ່ງໃນ Fig Extended Data. . 7 8 The number of positive COVID-19 cases, as confirmed by a single RT–PCR test obtained at any time between presentation to the ED and discharge from the hospital, is listed in Supplementary Table . ເວັບໄຊທ໌ລູກຄ້າຂອງພວກເຮົາມີຄວາມຕ້ອງການທີ່ຈະຕັດສິນໃຈເພື່ອຕັດສິນໃຈຂໍ້ມູນຂອງເຂົາເຈົ້າໃນສາມພາກສ່ວນ: 70% ສໍາລັບການຝຶກອົບຮົມ, 10% ສໍາລັບການຢັ້ງຢືນແລະ 20% ສໍາລັບການທົດສອບ. ສໍາລັບຮູບແບບການຄາດຄະເນດິນດີຕ້ອນຮັບ 24 ແລະ 72 ຊົ່ວໂມງ, ການຄາດຄະເນດິນດີຕ້ອນຮັບການຕັດສິນໃຈສໍາລັບທຸກໆສາມການຝຶກອົບຮົມແລະການຄາດຄະເນດິນດີຕ້ອນຮັບການທົດສອບ FL ໄດ້ຖືກສ້າງຂຶ້ນໂດຍອີເມວ. 1 EXAM model development There is wide variation in the clinical course of patients who present to hospital with symptoms of COVID-19, with some experiencing rapid deterioration in respiratory function requiring different interventions to prevent or mitigate hypoxemia , . A critical decision made during the evaluation of a patient at the initial point of care, or in the ED, is whether the patient is likely to require more invasive or resource-limited countermeasures or interventions (such as MV or monoclonal antibodies), and should therefore receive a scarce but effective therapy, a therapy with a narrow risk–benefit ratio due to side effects or a higher level of care, such as admittance to the intensive care unit . In contrast, a patient who is at lower risk of requiring invasive oxygen therapy may be placed in a less intensive care setting such as a regular ward, or even released from the ED for continuing self-monitoring at home . EXAM was developed to help triage such patients. 62 63 64 65 Of note, the model is not approved by any regulatory agency at this time and it should be used only for research purposes. EXAM score EXAM was trained using FL; it outputs a risk score (termed EXAM score) similar to CORISK (Extended Data Fig. ) and can be used in the same way to triage patients. It corresponds to a patient’s oxygen support requirements within two windows—24 and 72 h—after initial presentation to the ED. Extended Data Fig. ສະແດງໃຫ້ເຫັນວິທີທີ່ CORISK ແລະ Score EXAM ສາມາດຖືກນໍາໃຊ້ສໍາລັບການກວດສອບຂອງຜູ້ป่วย. 27 9a 9b Chest X-ray images were preprocessed to select the anterior position image and exclude lateral view images, and then scaled to a resolution of 224 × 224. As shown in Extended Data Fig. , the model fuses information from both EMR and CXR features (based on a modified ResNet34 with spatial attention pretrained on the CheXpert dataset) and the Deep & Cross network . To converge these different data types, a 512-dimensional feature vector was extracted from each CXR image using a pretrained ResNet34, with spatial attention, then concatenated with the EMR features as the input for the Deep & Cross network. The final output was a continuous value in the range 0–1 for both 24- and 72-h predictions, corresponding to the labels described above, as shown in Extended Data Fig. . We used cross-entropy as the loss function and ‘Adam’ as the optimizer. The model was implemented in Tensorflow ການນໍາໃຊ້ NVIDIA Clara Train SDK . AUC ອຸນຫະພູມສໍາລັບການທົດລອງ (≥LFO, ≥HFO / NIV ຫຼື ≥MV) ໄດ້ຖືກຄອມພິວເຕີແລະຖືກນໍາໃຊ້ເປັນມາດຕະຖານການຄາດຄະເນດຽວກັນ, ມີ normalization ກັບ zero-median ແລະ unit variance. ຮູບພາບ CXR ໄດ້ຖືກປິ່ນປົວກ່ອນທີ່ຈະເລືອກຊຸດທີ່ແທ້ຈິງແລະລວມເອົາຮູບພາບການເບິ່ງ lateral, ຫຼັງຈາກນັ້ນ scaled ກັບຄວາມກົດດັນ 224 × 224 (ref. ). 9a 66 67 68 ລະຫັດ QR 69 70 27 Feature imputation and normalization A MissForest algorithm ມັນຖືກນໍາໃຊ້ເພື່ອ impute ຄຸນນະສົມບັດ EMR, ໂດຍທົ່ວໄປ, ໂດຍທົ່ວໄປການຝຶກອົບຮົມ dataset. ຖ້າຫາກວ່າຄຸນນະສົມບັດ EMR ແມ່ນບໍ່ມີຈາກຄຸນນະສົມບັດຂອງຄຸນນະສົມບັດຂອງຄຸນສົມບັດຂອງຄຸນສົມບັດຂອງຄຸນສົມບັດທີ່ຖືກຄຸນສົມບັດໂດຍທົ່ວໄປໂດຍທົ່ວໄປຈາກຂໍ້ມູນຈາກຄຸນສົມບັດຂອງຄຸນສົມບັດຂອງຄຸນສົມບັດຂອງຄຸນສົມບັດຂອງຄຸນສົມບັດຂອງຄຸນສົມບັດຂອງຄຸນສົມບັດຂອງຄຸນສົມບັດຂອງຄຸນສົມບັດຂອງຄຸນສົມບັດຂອງຄຸນສົມບັດຂອງຄຸນສົມບັດຂອງຄຸນສົມບັດຂອງຄຸນສົມບັດຂອງຄຸນສົມບັດຂອງຄຸນສົມບັດຂອງຄຸນສົມບັດຂອງຄຸນສົມບັດຂອງຄຸນສົມບັດຂອງຄຸນສົມບັດຂອງຄຸນສົມບັດຂອງຄຸນສົມບັດຂອງ 71 ຊື່ຫຍໍ້ຂອງ : EMR-CXR data fusion using the Deep & Cross network To model the interactions of features from EMR and CXR data at the case level, a deep-feature scheme was used based on a Deep & Cross network architecture . Binary and categorical features for the EMR inputs, as well as 512-dimensional image features in the CXR, were transformed into fused dense vectors of real values by embedding and stacking layers. The transformed dense vectors served as input to the fusion framework, which specifically employed a crossing network to enforce fusion among input from different sources. The crossing network performed explicit feature crossing within its layers, by conducting inner products between the original input feature and output from the previous layer, thus increasing the degree of interaction across features. At the same time, two individual classic deep neural networks with several stacked, fully connected feed-forward layers were trained. The final output of our framework was then derived from the concatenation of both classic and crossing networks. 68 FL details ປະຫວັດສາດທີ່ດີທີ່ສຸດຂອງ FL ແມ່ນການປະຕິບັດຂອງອຸປະກອນປະມານ federated ເຊັ່ນ McMahan ແລະອື່ນໆ. , or variations thereof. This algorithm can be realized using a client-server setup where each participating site acts as a client. One can think of FL as a method aiming to minimize a global loss function by reducing a set of local loss functions, which are estimated at each site. By minimizing each client site’s local loss while also synchronizing the learned client site weights on a centralized aggregation server, one can minimize global loss without needing to access the entire dataset in a centralized location. Each client site learns locally, and shares model weight updates with a central server that aggregates contributions using secure sockets layer encryption and communication protocols. The server then sends an updated set of weights to each client site after aggregation, and sites resume training locally. The server and client site iterate back and forth until the model converges (Extended Data Fig. ). 72 9c A pseudoalgorithm of FL is shown in Supplementary Note ໃນການທົດສອບຂອງພວກເຮົາ, ພວກເຮົາມີຄວາມຄາດຄະເນຂອງການ federated = 200, ມີ 1 ອາທິດການຝຶກອົບຮົມໃນສະຖານທີ່ at each client. The number of clients, , was up to 20 depending on the network connectivity of clients or available data for a specific targeted outcome period (24 or 72 h). The number of local training iterations, , ປະມານຂະຫນາດຂອງ dataset ໃນທຸກ client and is used to weigh each client’s contributions when aggregating the model weights in federated averaging. During the FL training task, each client site selects its best local model by tracking the model’s performance on its local validation set. At the same time, the server determines the best global model based on the average validation scores sent from each client site to the server after each FL round. After FL training finishes, the best local models and the best global model are automatically shared with all client sites and evaluated on their local test data. 1 T t K ລະຫັດ QR k When training on local data only (the baseline), we set the epoch number to 200. The Adam optimizer was used for both local training and FL with an initial learning rate of 5 × 10–5 and a stepwise learning rate decay with a factor 0.5 after every 40 epochs, which is important for the convergence of federated averaging ການປ່ຽນແປງ affine random, ລວມທັງ rotation, translations, shear, scaling ແລະຄວາມເຂັ້ມແຂງ random ແລະ shifts, ໄດ້ຖືກນໍາໃຊ້ກັບຮູບພາບສໍາລັບການເພີ່ມຂໍ້ມູນໃນໄລຍະການຝຶກອົບຮົມ. 73 ເນື່ອງຈາກຄວາມເຂັ້ມແຂງຂອງ layers BN ໃນເວລາທີ່ເຮັດວຽກກັບລູກຄ້າທີ່ແຕກຕ່າງກັນໃນສະພາບແວດລ້ອມທີ່ບໍ່ແມ່ນອະນຸຍາດແລະຖືກຈັດຂຶ້ນຢ່າງເປັນປະໂຫຍດ, ພວກເຮົາມີຜົນປະໂຫຍດທີ່ດີທີ່ສຸດຂອງຮູບແບບທີ່ເກີດຂຶ້ນໃນເວລາທີ່ຮັກສາ ResNet34 pre-trained ມີຄວາມສົນໃຈສະຖານທີ່ ອັດຕະໂນມັດການຝຶກອົບຮົມຂອງພວກເຮົາແມ່ນການຝຶກອົບຮົມສໍາລັບການຝຶກອົບຮົມຂອງພວກເຮົາ. 58 47 ໃນການຄົ້ນຄວ້ານີ້, ພວກເຮົາມີການຄົ້ນຄວ້າລະບົບການປົກປັກຮັກສາຄວາມເປັນສ່ວນບຸກຄົນທີ່ຂ້າງຂວາງພຽງແຕ່ການປັບປຸງແບບ partial ໃນລະຫວ່າງ server ແລະເວັບໄຊທ໌ຂອງລູກຄ້າ. ການປັບປຸງຂະຫນາດໃຫຍ່ໄດ້ຖືກຈັດຂຶ້ນໃນໄລຍະແຕ່ລະການດໍາເນີນການໂດຍຂະຫນາດໃຫຍ່ຂອງການຊ່ວຍເຫຼືອ, ແລະພຽງແຕ່ຂະຫນາດຫນຶ່ງຂອງການປັບປຸງຂະຫນາດໃຫຍ່ທີ່ສຸດໄດ້ຖືກຈັດຂຶ້ນກັບການບໍລິໂພກ. ສໍາລັບຄວາມແມ່ນຍໍາ, ການປັບປຸງຂະຫນາດໃຫຍ່ (ທີ່ຮູ້ຈັກເປັນ gradients) ໄດ້ຖືກບໍລິໂພກພຽງແຕ່ໃນເວລາທີ່ຄຸນນະພາບ absolut ຂອງພວກເຂົາແມ່ນຫຼາຍກ່ວາຂະຫນາດຫນຶ່ງ percentile, (T) (Data Extended Fig. ), which was computed from all non-zero gradients, Δ , ແລະສາມາດແຕກຕ່າງກັນສໍາລັບທຸກລູກຄ້າ in each FL round ການປ່ຽນແປງຂອງລະບົບນີ້ສາມາດປະກອບດ້ວຍການຕັດເພີ່ມເຕີມຂອງ gradients ຂະຫນາດໃຫຍ່ຫຼືລະບົບຄວາມປອດໄພ differential ການນໍາສະເຫນີຄວາມປອດໄພສໍາລັບການ gradients, ຫຼືເຖິງແມ່ນວ່າການເຂົ້າລະຫັດ raw data, before feeding into the network . k 5 ລະຫັດ QR k t 49 51 Statistical analysis We conducted a Wilcoxon signed-rank test to confirm the significance of the observed improvement in performance between the locally trained model and the FL model for the 24- and 72-h time points (Fig. ແລະ Extended Data Fig. ). The null hypothesis was rejected with one-sided « 1 × 10–3 in both cases. 2 1 P ການຄົ້ນຄວ້າ Pearson ໄດ້ຖືກນໍາໃຊ້ເພື່ອກວດສອບການທົ່ວໄປ (ຄວາມເຂັ້ມແຂງຂອງຄຸນນະພາບ AUC ອຸນຫະພູມກັບຂໍ້ມູນການທົດສອບຂອງເວັບໄຊທ໌ຂອງລູກຄ້າອື່ນໆ) ຂອງມາດຕະຖານທີ່ໄດ້ຮັບການຝຶກອົບຮົມໃນສະຖານະການທີ່ກ່ຽວຂ້ອງກັບຂະຫນາດຂອງ dataset ອິນເຕີເນັດ. ບໍ່ພຽງແຕ່ການຄົ້ນຄວ້າການຄົ້ນຄວ້າທີ່ມີຂະຫນາດນ້ອຍ ( ລະຫັດ QR = 0.035, ອົງປະກອບຂອງຄວາມປອດໄພ (df) = 17 ສໍາລັບຮູບແບບ 24h ແລະ = 0.62, = 0.003, df = 16 ສໍາລັບຮູບແບບ 72-h). ນີ້ສະແດງໃຫ້ເຫັນວ່າຂະຫນາດຂອງ dataset ບໍ່ພຽງແຕ່ເປັນປະໂຫຍດຫນຶ່ງທີ່ຄັດເລືອກຄວາມເຂັ້ມແຂງຂອງຮູບແບບກັບຂໍ້ມູນທີ່ບໍ່ຮູ້ຈັກ. r P r P ການຕັດສິນໃຈຂອງ ROC ຈາກຮູບແບບ FL Global ແລະຮູບແບບທ້ອງຖິ່ນທີ່ໄດ້ຮັບການຝຶກອົບຮົມໃນສະຖານທີ່ທີ່ແຕກຕ່າງກັນ (Extended Data Fig. ), we bootstrapped 1,000 samples from the data and computed the resulting AUCs. We then calculated the difference between the two series and standardized using the formula = (AUC1 – AUC2)/ , where is the standardized difference, is the standard deviation of the bootstrap differences and AUC1 and AUC2 are the corresponding bootstrapped AUC series. By comparing ມີການຈັດການປົກກະຕິ, ພວກເຮົາໄດ້ຮັບການ ຄຸນນະສົມບັດທີ່ແຕກຕ່າງກັນໃນ Table Supplementary ຜົນປະໂຫຍດສະແດງໃຫ້ເຫັນວ່າ hypothesis null ໄດ້ຖືກຕອບສະຫນອງດ້ວຍຄວາມກົດດັນຕ່ໍາ values, indicating the statistical significance of the superiority of FL outcomes. The computation of values was conducted in R with the pROC library . 3 D s D s D P 2 P P 74 Since the model predicts a discrete outcome, a continuous score from 0 to 1, a straightforward calibration evaluation such as a qqplot is not possible. Hence, for a quantified estimate of calibration we quantified discrimination (Extended Data Fig. ). ພວກເຮົາມີການທົດສອບ ANOVA (analysis of variation) ການທົດສອບຫນຶ່ງເສັ້ນທາງເພື່ອທົດສອບ scores model local ແລະ FL ໃນລະຫວ່າງ 4 ປະເພດທີ່ແທ້ຈິງ (RA, LFO, HFO, MV). ການຄາດຄະເນດິນຟ້າອຸປະກອນ, ການຄາດຄະເນດິນຟ້າອຸປະກອນ, ການຄາດຄະເນດິນຟ້າອຸປະກອນ, ການຄາດຄະເນດິນຟ້າອຸປະກອນ, ການຄາດຄະເນດິນຟ້າອຸປະກອນ, ການຄາດຄະເນດິນຟ້າອຸປະກອນ, ການຄາດຄະເນດິນຟ້າອຸປະກອນ, ການຄາດຄະເນດິນຟ້າອຸປະກອນ, ການຄາດຄະເນດິນຟ້າອຸປະກອນ, ການຄາດຄະເນດິນຟ້າອຸປະກອນ, ການຄາດຄະເນດິນຟ້າອຸປະກອນ -values of five different local sites are 245.7, 253.4, 342.3, 389.8 and 634.8, while that of the FL model is 843.5. Given that larger -ຄຸນນະສົມບັດຂໍຂອບໃຈວ່າຄຸນນະສົມບັດທີ່ແຕກຕ່າງກັນຫຼາຍ, ຄຸນນະສົມບັດຈາກຮູບແບບ FL ຂອງພວກເຮົາກໍາລັງສະແດງໃຫ້ເຫັນຜົນກະທົບຂະຫນາດໃຫຍ່ໃນລະຫວ່າງ 4 ປະເພດທີ່ແທ້ຈິງ. ນອກເຫນືອໄປຈາກນີ້, value of the ANOVA test on the FL model is <2 × 10–16, indicating that the FL prediction scores are statistically significantly different among the different prediction classes. 10 F F F P ການທົບທວນຄືນ Further information on research design is available in the linked to this article. ຄວາມຄິດເຫັນທີ່ Nature Research Reporting ການເຂົ້າເຖິງ Data ພວກເຮົາ ກໍາ ລັງເຮັດທຸລະກິດໃນ 2012. ພວກເຮົາແມ່ນບໍລິສັດທີ່ໃຫຍ່ທີ່ສຸດ ສໍາ ລັບຜູ້ຊ່ຽວຊານໃນການຊ່ຽວຊານໃນການຊ່ຽວຊານໃນການຊ່ຽວຊານໃນການຊ່ຽວຊານໃນການຊ່ຽວຊານໃນການຊ່ຽວຊານໃນການຊ່ຽວຊານໃນການຊ່ຽວຊານໃນການຊ່ຽວຊານໃນການຊ່ຽວຊານໃນການຊ່ຽວຊານໃນການຊ່ຽວຊານໃນການຊ່ຽວຊານໃນການຊ່ຽວຊານໃນການຊ່ຽວຊານໃນການຊ່ຽວຊານໃນການຊ່ຽວຊານໃນການຊ່ຽວຊານໃນການຊ່ຽວຊານໃນການຊ່ຽວຊານ. ລະຫັດການເຂົ້າເຖິງ ລະຫັດແລະໂທລະສັບມືຖືທັງຫມົດຖືກນໍາໃຊ້ໃນການຄົ້ນຄວ້ານີ້ແມ່ນສາມາດເຂົ້າເຖິງໃນ NGC. ເພື່ອເຂົ້າເຖິງ, ດາວໂຫລດເປັນຜູ້ຕິດຕໍ່ຫຼືສ້າງໂທລະສັບມືຖື, ຫຼັງຈາກນັ້ນເຂົ້າໄປໃນຫນຶ່ງຂອງ URLs ນີ້. ຮູບແບບການຝຶກອົບຮົມ, ການຝຶກອົບຮົມການຝຶກອົບຮົມ, ລະຫັດການຝຶກອົບຮົມ, ການຢັ້ງຢືນການທົດສອບຂອງຮູບແບບ, ຮູບແບບ readme, ຮູບແບບການຕິດຕັ້ງແລະເອກະສານໃບອະນຸຍາດແມ່ນສາມາດເຂົ້າເຖິງໂດຍທົ່ວໄປໃນ NVIDIA NGC : The federated learning software is available as part of the Clara Train SDK: ດາວນ໌ໂຫລດຕົວແບບ "wget --content-disposition" -O clara_train_covid19_exam_ehr_xray_1.zip”. 61 https://ngc.nvidia.com/catalog/models/nvidia:med:clara_train_covid19_exam_ehr_xray https://ngc.nvidia.com/catalog/containers/nvidia:clara-train-sdk https://api.ngc.nvidia.com/v2/models/nvidia/med/clara_train_covid19_exam_ehr_xray/versions/1/zip References Budd, J. et al. Digital technologies in the public-health response to COVID-19. , 1183–1192 (2020). Nat. Med. 26 Moorthy, V., Henao Restrepo, A. M., Preziosi, M.-P. & Swaminathan, S. Data sharing for novel coronavirus (COVID-19). , 150 (2020). Bull. World Health Organ. 98 Chen, Q., Allot, A. & Lu, Z. Keep up with the latest coronavirus research. , 193 (2020). Nature 579 Fabbri, F., Bhatia, A., Mayer, A., Schlotter, B. & Kaiser, J. BCG IT spend pulse: how COVID-19 is shifting tech priorities. (2020). https://www.bcg.com/publications/2020/how-covid-19-is-shifting-big-it-spend Candelon, F., Reichert, T., Duranton, S., di Carlo, R. C. & De Bondt, M. The rise of the AI-powered company in the postcrisis world. (2020). https://www.bcg.com/en-gb/publications/2020/business-applications-artificial-intelligence-post-covid Chao, H. et al. Integrative analysis for COVID-19 patient outcome prediction. , 101844 (2021). Med. Image Anal. 67 Zhu, X. et al. Joint prediction and time estimation of COVID-19 developing severe symptoms using chest CT scan. , 101824 (2021). Med. Image Anal. 67 Yang, D. et al. Federated semi-supervised learning for Covid region segmentation in chest ct using multi-national data from China, Italy, Japan. , 101992 (2021). Med. Image Anal. 70 Minaee, S., Kafieh, R., Sonka, M., Yazdani, S. & Jamalipour Soufi, G. Deep-COVID: predicting COVID-19 from chest X-ray images using deep transfer learning. , 101794 (2020). Med. Image Anal. 65 COVID-19 Studies from the World Health Organization Database. (2020). https://clinicaltrials.gov/ct2/who_table ACTIV. (2020). https://www.nih.gov/research-training/medical-research-initiatives/activ Coronavirus Treatment Acceleration Program (CTAP). US Food and Drug Administration (2020). https://www.fda.gov/drugs/coronavirus-covid-19-drugs/coronavirus-treatment-acceleration-program-ctap Gleeson, P., Davison, A. P., Silver, R. A. & Ascoli, G. A. A commitment to open source in neuroscience. , 964–965 (2017). Neuron 96 Piwowar, H. et al. The state of OA: a large-scale analysis of the prevalence and impact of open access articles. , e4375 (2018). PeerJ. 6 European Society of Radiology (ESR). What the radiologist should know about artificial intelligence – an ESR white paper. , 44 (2019). Insights Imaging 10 Pesapane, F., Codari, M. & Sardanelli, F. Artificial intelligence in medical imaging: threat or opportunity? Radiologists again at the forefront of innovation in medicine. , 35 (2018). Eur. Radiol. Exp. 2 Price, W. N. 2nd & Cohen, I. G. Privacy in the age of medical big data. , 37–43 (2019). Nat. Med. 25 Liang, W. et al. Development and validation of a clinical risk score to predict the occurrence of critical illness in hospitalized patients with COVID-19. , 1081–1089 (2020). JAMA Intern. Med. 180 Wynants, L. et al. Prediction models for diagnosis and prognosis of covid-19 infection: systematic review and critical appraisal. , m1328 (2020). Brit. Med. J. 369 Zhang, L. et al. D-dimer levels on admission to predict in-hospital mortality in patients with Covid-19. , 1324–1329 (2020). J. Thromb. Haemost. 18 Sands, K. E. et al. Patient characteristics and admitting vital signs associated with coronavirus disease 2019 (COVID-19)-related mortality among patients admitted with noncritical illness. (2020). https://doi.org/10.1017/ice.2020.461 American College of Radiology. CR recommendations for the use of chest radiography and computed tomography (CT) for suspected COVID-19 infection. (2020). https://www.acr.org/Advocacy-and-Economics/ACR-Position-Statements/Recommendations-for-Chest-Radiography-and-CT-for-Suspected-COVID19-Infection Rubin, G. D. et al. The role of chest imaging in patient management during the COVID-19 pandemic: a multinational consensus statement from the Fleischner Society. , 172–180 (2020). Radiology 296 World Health Organization. Use of chest imaging in COVID-19. (2020). https://www.who.int/publications/i/item/use-of-chest-imaging-in-covid-19 Jamil, S. et al. Diagnosis and management of COVID-19 disease. , 10 (2020). Am. J. Respir. Crit. Care Med. 201 Redmond, C. E., Nicolaou, S., Berger, F. H., Sheikh, A. M. & Patlas, M. N. Emergency radiology during the COVID-19 pandemic: The Canadian Association of Radiologists Recommendations for Practice. , 425–430 (2020). Can. Assoc. Radiologists J. 71 Buch, V. et al. Development and validation of a deep learning model for prediction of severe outcomes in suspected COVID-19 Infection. Preprint at (2021). https://arxiv.org/abs/2103.11269 Lyons, C. & Callaghan, M. The use of high-flow nasal oxygen in COVID-19. , 843–847 (2020). Anaesthesia 75 Whittle, J. S., Pavlov, I., Sacchetti, A. D., Atwood, C. & Rosenberg, M. S. Respiratory support for adult patients with COVID-19. , 95–101 (2020). J. Am. Coll. Emerg. Physicians Open 1 Ai, J., Li, Y., Zhou, X. & Zhang, W. COVID-19: treating and managing severe cases. , 370–371 (2020). Cell Res. 30 Esteva, A. et al. A guide to deep learning in healthcare. , 24–29 (2019). Nat. Med. 25 Cahan, E. M., Hernandez-Boussard, T., Thadaney-Israni, S. & Rubin, D. L. Putting the data before the algorithm in big data addressing personalized healthcare. , 78 (2019). NPJ Digit. Med. 2 Thrall, J. H. et al. Artificial intelligence and machine learning in radiology: opportunities, challenges, pitfalls, and criteria for success. , 504–508 (2018). J. Am. Coll. Radiol. 15 Shilo, S., Rossman, H. & Segal, E. Axes of a revolution: challenges and promises of big data in healthcare. , 29–38 (2020). Nat. Med. 26 Gao, Y. & Cui, Y. Deep transfer learning for reducing health care disparities arising from biomedical data inequality. , 5131 (2020). Nat. Commun. 11 Rieke, N. et al. The future of digital health with federated learning. , 119 (2020). NPJ Dig. Med. 3 Yang, Q., Liu, Y., Chen, T. & Tong, Y. Federated machine learning: concept and applications. , 12 (2019). ACM Trans. Intell. Syst. Technol. 10 Ma, C. et al. On safeguarding privacy and security in the framework of federated learning. , 242–248 (2020). IEEE Netw. 34 Brisimi, T. S. et al. Federated learning of predictive models from federated Electronic Health Records. , 59–67 (2018). Int. J. Med. Inform. 112 Roth, H. R. et al. Federated learning for breast density classification: a real-world implementation. In , (eds. Albarqouni, S. et al.) Vol. 12,444, 181–191 (Springer International Publishing, 2020). Proc. Second MICCAI Workshop, DART 2020 and First MICCAI Workshop, DCL 2020 Domain Adaptation and Representation Transfer, and Distributed and Collaborative Learning Sheller, M. J. et al. Federated learning in medicine: facilitating multi-institutional collaborations without sharing patient data. , 12598 (2020). Sci. Rep. 10 Remedios, S. W., Butman, J. A., Landman, B. A. & Pham, D. L. in (eds Remedios, S. W. et al.) (Springer, 2020). Federated Gradient Averaging for Multi-Site Training with Momentum-Based Optimizers Xu, Y. et al. A collaborative online AI engine for CT-based COVID-19 diagnosis. Preprint at (2020). https://www.medrxiv.org/content/10.1101/2020.05.10.20096073v2 Raisaro, J. L. et al. SCOR: A secure international informatics infrastructure to investigate COVID-19. , 1721–1726 (2020). J. Am. Med. Inform. Assoc. 27 Vaid, A. et al. Federated learning of electronic health records to improve mortality prediction in hospitalized patients with COVID-19: machine learning approach. , e24207 (2021). JMIR Med. Inform. 9 Nino, G. et al. Pediatric lung imaging features of COVID-19: a systematic review and meta-analysis. , 252–263 (2021). Pediatr. Pulmonol. 56 Fredrikson, M., Jha, S. & Ristenpart, T. Model inversion attacks that exploit confidence information and basic countermeasures. In 1322–1333, (2015). Proc. 22nd ACM SIGSAC Conference on Computer and Communications Security https://doi.org/10.1145/2810103.2813677 Zhu, L., Liu, Z. & Han, S. in (eds Wallach, H. et al.) 14774–14784 (Curran Associates, Inc., 2019). Advances in Neural Information Processing Systems 32 Kaissis, G. A., Makowski, M. R., Rückert, D. & Braren, R. F. Secure, privacy-preserving and federated machine learning in medical imaging. , 305–311 (2020). Nat. Mach. Intell. 2 Li, W. et al. in 133–141 (Springer, 2019). Privacy-Preserving Federated Brain Tumour Segmentation Shokri, R. & Shmatikov, V. Privacy-preserving deep learning. In (2015). Proc. 53rd Annual Allerton Conference on Communication, Control, and Computing (Allerton) https://doi.org/10.1109/allerton.2015.7447103 Li, X. et al. Multi-site fMRI analysis using privacy-preserving federated learning and domain adaptation: ABIDE results. , 101765 (2020). Med. Image Anal. 65 Estiri, H. et al. Predicting COVID-19 mortality with electronic medical records. , 15 (2021). NPJ Dig. Med. 4 Jiang, G. et al. Harmonization of detailed clinical models with clinical study data standards. , 65–74 (2015). Methods Inf. Med. 54 Yang, D. et al. in . (2019). Searching Learning Strategy with Reinforcement Learning for 3D Medical Image Segmentation https://doi.org/10.1007/978-3-030-32245-8_1 Elsken, T., Metzen, J. H. & Hutter, F. Neural architecture search: a survey. , 1–21 (2019). J. Mach. Learning Res. 20 Yao, Q. et al. Taking human out of learning applications: a survey on automated machine learning. Preprint at (2019). https://arxiv.org/abs/1810.13306 Ioffe, S. & Szegedy, C. Batch normalization: accelerating deep network training by reducing internal covariate shift. In , PMLR , 448–456 (2015). Proc. 32nd International Conf. Machine Learning 37 Kaufman, S., Rosset, S. & Perlich, C. Leakage in data mining: formulation, detection, and avoidance. In , 556–563 (2011). Proc. 17th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining Zhang, C. et al. BatchCrypt: efficient homomorphic encryption for cross-silo federated learning. In , 493–506 (2020). Proc. 2020 USENIX Annual Technical Conference, ATC 2020 . (2020). Nvidia NGC Catalog: COVID-19 Related Models https://ngc.nvidia.com/catalog/models?orderBy=scoreDESC&pageNumber=0&query=covid&quickFilter=models&filters Marini, J. J. & Gattinoni, L. Management of COVID-19 respiratory distress. , 2329–2330 (2020). JAMA 323 Cook, T. M. et al. Consensus guidelines for managing the airway in patients with COVID-19: Guidelines from the Difficult Airway Society, the Association of Anaesthetists the Intensive Care Society, the Faculty of Intensive Care Medicine and the Royal College of Anaesthetist. , 785–799 (2020). Anaesthesia 75 Galloway, J. B. et al. A clinical risk score to identify patients with COVID-19 at high risk of critical care admission or death: an observational cohort study. , 282–288 (2020). J. Infect. 81 Kilaru, A. S. et al. Return hospital admissions among 1419 COVID-19 patients discharged from five U.S. emergency departments. , 1039–1042 (2020). Acad. Emerg. Med. 27 He, K., Zhang, X., Ren, S. & Sun, J. Deep residual learning for image recognition. In (2016). Proc. 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) https://doi.org/10.1109/cvpr.2016.90 Irvin, J. et al. CheXpert: a large chest radiograph dataset with uncertainty labels and expert comparison. , 590–597 (2019). Proc. AAAI Conf. Artif. Intell. 33 Wang, R., Fu, B., Fu, G. & Wang, M. Deep & Cross network for Ad Click predictions. In Article no. 12 (2017). Proc. ADKDD’17 Abadi, M. et al. TensorFlow: asystem for large-scale machine learning. In , USENIX Association 265–283 (2016). 12th USENIX Symposium on Operating Systems Design and Implementation (OSDI 16) . (2020). NVIDIA Clara Imaging https://developer.nvidia.com/clara-medical-imaging Stekhoven, D. J. & Bühlmann, P. MissForest–non-parametric missing value imputation for mixed-type data. , 112–118 (2012). Bioinformatics 28 McMahan, H., Moore, E., Ramage, D., Hampson, S. & y Arcas, B. A. Communication-efficient learning of deep networks from decentralized data. (2017). http://proceedings.mlr.press/v54/mcmahan17a.html Hsieh, K., Phanishayee, A., Mutlu, O. & Gibbons, P. B. The non-IID data quagmire of decentralized machine learning. In PMLR 119 (2020). Proc. 37th International Conf. Machine Learning Robin, X. et al. pROC: an open-source package for R and S+ to analyze and compare ROC curves. , 77 (2011). BMC Bioinformatics 12 ການຮັບຮອງ The views expressed in this study are those of the authors and not necessarily those of the NHS, the NIHR, the Department of Health and Social Care or any of the organizations associated with the authors. MGB thank the following individuals for their support: J. Brink, Department of Radiology, Massachusetts General Hospital, Harvard Medical School, Boston, MA; M. Kalra, Department of Radiology, Massachusetts General Hospital, Harvard Medical School, Boston, MA; N. Neumark, Center for Clinical Data Science, Massachusetts General Brigham, Boston, MA; T. Schultz, Department of Radiology, Massachusetts General Hospital, Boston, MA; N. Guo, Center for Advanced Medical Computing and Analysis, Department of Radiology, Massachusetts General Hospital, Harvard Medical School, Boston, MA; J. K. Cramer, Director, QTIM lab at the Athinoula A. Martinos Center for Biomedical Imaging at MGH; S. Pomerantz, Department of Radiology, Massachusetts General Hospital, Harvard Medical School, Boston, MA; G. Boland, Department of Radiology, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA; W. Mayo-Smith, Department of Radiology, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA. UCSF thank P. B. Storey, J. Chan and J. Block for implementing the UCSF FL client infrastructure, and W. Tellis for providing the source imaging repository for this work. The UCSF EMR and clinical notes for this study were accessed via the COVID-19 Research Data Mart, ວິທະຍາໄລ Chulalongkorn ວິທະຍາໄລ Chulalongkorn ວິທະຍາໄລຂອງ Chulalongkorn ວິທະຍາໄລຂອງ Chulalongkorn ວິທະຍາໄລຂອງ Chulalongkorn ວິທະຍາໄລຂອງ Chulalongkorn ວິທະຍາໄລຂອງ Chulalongkorn ວິທະຍາໄລຂອງ Chulalongkorn ວິທະຍາໄລຂອງ Chulalongkorn ວິທະຍາໄລຂອງ Chulalongkorn ວິທະຍາໄລຂອງ Chulalongkorn ວິທະຍາໄລຂອງ Chulalongkorn ວິທະຍາໄລຂອງ Chulalongkorn ວິທະຍາໄລຂອງ Chulalongkorn ວິທະຍາໄລຂອງ Chulalongkorn ວິທະຍາໄລຂອງ Chulalongkorn ວິທະຍາໄລຂອງ Chulalongkorn ວິທະຍາໄລຂອງ Chulalongkorn ວິທະຍ https://data.ucsf.edu/covid19 ບົດຄວາມນີ້ແມ່ນມີຢູ່ໃນອຸນຫະພູມໂດຍໃບອະນຸຍາດ CC by 4.0 Deed (Attribution 4.0 International). ເອກະສານນີ້ແມ່ນ ໂດຍ CC by 4.0 Deed (Attribution 4.0 International) ອະນຸຍາດ. ສະຫນັບສະຫນູນໂດຍ Nature
