1. A Crowdsourcing Approach to Developing and Assessing Prediction Algorithms for AML Prognosis
- Author
-
Noren, David P., Long, Byron L., Norel, Raquel, Rrhissorrakrai, Kahn, Hess, Kenneth, Chenyue Wendy, Hu, Bisberg, Alex J., Schultz, Andre, Engquist, Erik, Liu, Li, Lin, Xihui, Chen, Gregory M., Xie, Honglei, Hunter, Geoffrey A. M., Boutros, Paul C., Stepanov, Oleg, Abrams, Zachary, Ambrosini, Giovanna, Anastassiou, Dimitris, Baladandayuthapani, Veerabhadran, Batten, Kimberly, Bucher, Philipp, Buturovic, Ljubomir, Campion, Loic, Creighton, Chad J., Chen, Greg, Cheong, Jae Ho, DI CAMILLO, Barbara, Dreos, René, Estrada, Alan, Fatemi, Seyyed A., Fitzgerald, Andrew, Flynn, Jennifer, Fronczuk, Maciej, Weiyi, Gu, Guha, Subharup, Hosseini, Maryam, Hung, Ling Hong, Hunter, Geoffrey, Hwang, Tae Hyun, Kim, Daniel, Kim, Minsoo, Korra, Jyothi, Krstajic, Damjan, Kumar, Sunil, Kuh, Anthony, Jinpu, Li, Liu, Yashu, Mcmurray, James, Morgan, Daniel, Motiwala, Tasneem, Naegle, Kristen, Niemiec, Rafał, Oehler, Vivian G., Park, Sunho, Pattin, Alejandrina, Peabody, Andrea, Piraino, Scott W., Regan, Kelly, Ronan, Tom, Rościszewski, Antoni, Rudnicki, Witold, Sanavia, Tiziana, Santhanam, Narayana, Shay, Jerry, Tang, Hao, Vilar, Jose M. G., Wang, Tao, Wright, Woodring, Wrzesień, Mariusz, Xiao, Guanghua, Xie, Yang, Yang, Sen, Yang, Tai Hsien Ou, Yang, Tao, Jieping, Ye, Yeung, Ka Yee, Zang, Xiao, Zolfaghar, Kiyana, Żuk, Paweł, Norman, Thea, Friend, Stephen H., Stolovitzky, Gustavo, Kornblau, Steven, Qutub, Amina A., and Tan, Kai
- Subjects
Proteomics ,0301 basic medicine ,Myeloid ,Proteome ,Cancer Treatment ,Bioinformatics ,Patient response ,Biochemistry ,Systems Science ,Mathematical Sciences ,Hematologic Cancers and Related Disorders ,Machine Learning ,Database and Informatics Methods ,0302 clinical medicine ,Medicine and Health Sciences ,lcsh:QH301-705.5 ,Cancer ,Pediatric ,screening and diagnosis ,Ecology ,Proteomic Databases ,Systems Biology ,Myeloid leukemia ,Hematology ,Biological Sciences ,Myeloid Leukemia ,Prognosis ,3. Good health ,Detection ,Prediction algorithms ,Outcome and Process Assessment, Health Care ,Treatment Outcome ,medicine.anatomical_structure ,Oncology ,Computational Theory and Mathematics ,030220 oncology & carcinogenesis ,Modeling and Simulation ,Physical Sciences ,Crowdsourcing ,Risk assessment ,Algorithms ,Research Article ,Acute Myeloid Leukemia ,DREAM 9 AML-OPC Consortium ,Computer and Information Sciences ,Childhood Leukemia ,Pediatric Cancer ,Research and Analysis Methods ,Outcome and Process Assessment ,Risk Assessment ,Sensitivity and Specificity ,03 medical and health sciences ,Cellular and Molecular Neuroscience ,Rare Diseases ,Diagnostic Medicine ,Artificial Intelligence ,Information and Computing Sciences ,Leukemias ,Genetics ,medicine ,Humans ,Acute Myeloid Leukemia, Prediction algorithms, Machine Learning, Bioinformatics ,Molecular Biology ,Ecology, Evolution, Behavior and Systematics ,business.industry ,Amyotrophic Lateral Sclerosis ,Cancers and Neoplasms ,Biology and Life Sciences ,Reproducibility of Results ,Human Genetics ,Outcome and Process Assessment (Health Care) ,medicine.disease ,Human genetics ,4.1 Discovery and preclinical testing of markers and technologies ,Health Care ,Biological Databases ,030104 developmental biology ,lcsh:Biology (General) ,13. Climate action ,Cognitive Science ,business ,Mathematics ,Biomarkers ,Neuroscience - Abstract
Acute Myeloid Leukemia (AML) is a fatal hematological cancer. The genetic abnormalities underlying AML are extremely heterogeneous among patients, making prognosis and treatment selection very difficult. While clinical proteomics data has the potential to improve prognosis accuracy, thus far, the quantitative means to do so have yet to be developed. Here we report the results and insights gained from the DREAM 9 Acute Myeloid Prediction Outcome Prediction Challenge (AML-OPC), a crowdsourcing effort designed to promote the development of quantitative methods for AML prognosis prediction. We identify the most accurate and robust models in predicting patient response to therapy, remission duration, and overall survival. We further investigate patient response to therapy, a clinically actionable prediction, and find that patients that are classified as resistant to therapy are harder to predict than responsive patients across the 31 models submitted to the challenge. The top two performing models, which held a high sensitivity to these patients, substantially utilized the proteomics data to make predictions. Using these models, we also identify which signaling proteins were useful in predicting patient therapeutic response., Author Summary Acute Myeloid Leukemia (AML) is a hematological cancer with a very low 5-year survival rate. It is a very heterogeneous disease, meaning that the molecular underpinnings that cause AML vary greatly among patients, necessitating the use of precision medicine for treatment. While this personalized approach could be greatly improved by the incorporation of high-throughput proteomics data into AML patient prognosis, the quantitative methods to do so are lacking. We held the DREAM 9 AML Outcome Prediction Challenge to foster support, collaboration, and participation from multiple scientific communities in order to solve this problem. The outcome of the challenge yielded several accurate methods (AUROC >0.78, BAC > 0.69) capable of predicting whether a patient would respond to therapy. Moreover, this study also determined aspects of the methods which enabled accurate predictions, as well as key signaling proteins that were informative to the most accurate models.
- Published
- 2016