1. Cascaded Multi-view Canonical Correlation (CaMCCo) for Early Diagnosis of Alzheimer’s Disease via Fusion of Clinical, Imaging and Omic Features
- Author
-
Sterling C. Johnson, Paul Malloy, Joy L. Taylor, Alan J. Lerner, Pradeep Garg, Pierre N. Tariot, David G. Clark, Steven G. Potkin, Franklin Watkins, Howard Bergman, Dana M. Pogorelec, Charles D. Smith, Pradeep Varma, Stephen Pasternack, Betty Lind, Saba Wolday, Douglas W. Scharre, Donna Munic, Marwan N. Sabbagh, Adam S. Fleisher, Joanne S. Allard, Cynthia Hunt, Lidia Glodzik, Charles Bernick, Daniel D'Agostino, Owen T. Carmichael, Geoffrey Tremont, Christopher H. van Dyck, Maria Carroll, Po Lu, Leslie Gordineer, Catherine Mc-Adams-Ortiz, Irina Rachisky, Antero Sarrael, Clifford R. Jack, David Bachman, Dick Trost, Scott Herring, Arthur W. Toga, Evan Fletcher, Christina A. Michel, Lon S. Schneider, Francine Parfitt, Kelly M. Makino, Anahita Adeli, Daniel Varon, Christine M. Belden, Nunzio Pomara, Thomas O. Obisesan, Howard Feldman, Howard Chertkow, Sandra W. Jacobson, Haibo Wang, Greg Jicha, Laura A. Flashman, George Bartzokis, Beau M. Ances, Stacy Schneider, Earl A. Zimmerman, Munir Chowdhury, Bruce L. Miller, Javier Villanueva-Meyer, Kristin Fargher, Michael W. Weiner, Dana Nguyen, Ranjan Duara, T. Y. Lee, Lisa C. Silbert, Benita Mudge, Marilyn S. Albert, James J. Lah, Janet S. Cellar, Gad A. Marshall, Michael Lin, Marc Seltzer, Leslie Shaw, Bojana Stefanovic, Daniel C. Marson, Kyle B. Womack, Liberty Teodoro, Connie Brand, Nadira Trncic, Maria Kataki, Russell H. Swerdlow, Paul S. Aisen, Brigid Reynolds, Mony J. de Leon, Sandra E. Black, Rachelle S. Doody, Paula Ogrocki, Andrew J. Saykin, Raymundo Hernando, Leyla deToledo-Morrell, Anna Burke, Sherye A. Sirrel, Henry W. Querfurth, Jeffrey R. Petrella, Norman R. Relkin, Judith L. Heidebrink, Vernice Bates, Mary L. Creech, David C. Perry, Curtis Caldwell, Sara Dolen, Anton P. Porsteinsson, Patricia Lynn Johnson, Erik D. Roberson, Effie M. Mitsis, Kathleen Johnson, John Q. Trojanowki, Raina Carter, James E. Galvin, Karen Blank, John C. Morris, Bryan M. Spann, Keith A. Johnson, Jared R. Tinklenberg, Stephen Salloway, Ronald J. Killiany, Mimi Dang, Smita Kittur, Mary Quiceno, Kaycee M. Sink, Helen Vanderswag, Erin E. Franklin, Robbartha, Kim Martin, Gaby Thai, Allyson C. Rosen, Karen L. Bell, Tracy Kendall, P. M. Doraiswamy, Kathleen Tingus, Angela Oliver, Adrian Preda, Mary L. Hynes, Laurel A. Beckett, William J. Jagust, Jeffrey M. Burns, Ronald C. Petersen, Allan I. Levey, Balebail Ashok Raj, Lawrence S. Honig, Martin R. Farlow, Richard E. Carson, Dana Mathews, David S. Knopman, Robert C. Green, Jerome A. Yesavage, Elizabeth Finger, Ann Marie Hake, David S. Geldmacher, Yaakov Stern, Raj C. Shah, M.-Marsel Mesulam, Ruth A. Mulnard, Jacobo Mintzer, Howard J. Rosen, Peggy Roberts, Joseph F. Quinn, Raymond Scott Turner, Maria T. Greig, Salvador Borges-Neto, Jeffrey Kaye, Randall Griffith, Diana R. Kerwin, Neill R. Graff-Radford, James B. Brewer, John C. Brockington, Ging-Yuek Robin Hsiung, Anant Madabhushi, Andrew E. Budson, Martha G. MacAvoy, Stephen Correia, Terence Z. Wong, Michelle Rainka, Elizabeth Oates, Alexander Norbash, Chiadi U. Onyike, Gloria Chaing, Kris Johnson, Hillel Grossman, Gary R. Conrad, Nancy Johnson, Lisa D. Ravdin, Mauricio Beccera, Reisa A. Sperling, Heather Johnson, Kristine Lipowski, Charles DeCarli, Barton Lane, Joanne L. Lord, Carl H. Sadowsky, Chris Hosein, Marissa Natelson Love, M. Ismail, Liana G. Apostolova, Dzintra Celmins, Brian R. Ott, Brittany Cerbone, Sanjay Asthana, Alice D. Brown, Neil W. Kowall, Peter A. Hardy, Andrew Kertesz, Sara S. Mason, Horacio Capote, Pauline Maillard, Stephanie Kielb, Henry Rusinek, Ellen Woo, Jeff D. Williamson, Susan De Santi, Amanda Smith, John M Olichney, Michele Assaly, Karen S. Anderson, Parianne Fatica, Brandy R. Matthews, Michael Borrie, Susan Rountree, Chuang Kuo Wu, Curtis Tatsuoka, Teresa Villena, Asha Singanamalli, Borna Bonakdarpour, Colleen S. Albers, Cynthia M. Carlsson, Bonnie S. Goldstein, Sonia Pawluczyk, Edward Coleman, Kenneth M. Spicer, Jared R. Brosch, William Brooks, Partha Sinha, Stephanie Reeder, Daniel Silverman, Robert B. Santulli, Godfrey D. Pearlson, Mark A. Mintun, and Sandra Weintraub
- Subjects
0301 basic medicine ,Male ,Proteomics ,Science ,Neuroimaging ,Disease ,Sensitivity and Specificity ,Article ,03 medical and health sciences ,0302 clinical medicine ,Text mining ,Alzheimer Disease ,medicine ,Humans ,Cognitive Dysfunction ,Set (psychology) ,Aged ,Aged, 80 and over ,Multidisciplinary ,Modalities ,business.industry ,Pattern recognition ,Alzheimer’s Disease Neuroimaging Initiative ,Genomics ,Models, Theoretical ,medicine.disease ,030104 developmental biology ,Categorization ,Case-Control Studies ,Medicine ,Female ,Artificial intelligence ,Alzheimer's disease ,business ,Canonical correlation ,030217 neurology & neurosurgery ,Algorithms ,Biomarkers - Abstract
The introduction of mild cognitive impairment (MCI) as a diagnostic category adds to the challenges of diagnosing Alzheimer’s Disease (AD). No single marker has been proven to accurately categorize patients into their respective diagnostic groups. Thus, previous studies have attempted to develop fused predictors of AD and MCI. These studies have two main limitations. Most do not simultaneously consider all diagnostic categories and provide suboptimal fused representations using the same set of modalities for prediction of all classes. In this work, we present a combined framework, cascaded multiview canonical correlation (CaMCCo), for fusion and cascaded classification that incorporates all diagnostic categories and optimizes classification by selectively combining a subset of modalities at each level of the cascade. CaMCCo is evaluated on a data cohort comprising 149 patients for whom neurophysiological, neuroimaging, proteomic and genomic data were available. Results suggest that fusion of select modalities for each classification task outperforms (mean AUC = 0.92) fusion of all modalities (mean AUC = 0.54) and individual modalities (mean AUC = 0.90, 0.53, 0.71, 0.73, 0.62, 0.68). In addition, CaMCCo outperforms all other multi-class classification methods for MCI prediction (PPV: 0.80 vs. 0.67, 0.63).
- Published
- 2017
- Full Text
- View/download PDF