A Deep-Learning Based Prediction of Pancreatic Adenocarcinoma with Electronic Health Records from the State of Maine

Authors: Xiaodong Li, Peng Gao, Chao-Jung Huang, Shiying Hao, Xuefeng B. Ling, Yongxia Han, Yaqi Zhang, Le Zheng, Chengyin Ye, Modi Liu, Minjie Xia, Changlin Fu, Bo Jin, Karl G. Sylvester, Eric Widen

Abstract:

Predicting the risk of Pancreatic Adenocarcinoma (PA) in advance can benefit the quality of care and potentially reduce population mortality and morbidity. The aim of this study was to develop and prospectively validate a risk prediction model to identify patients at risk of new incident PA as early as 3 months before the onset of PA in a statewide, general population in Maine. The PA prediction model was developed using Deep Neural Networks, a deep learning algorithm, with a 2-year electronic-health-record (EHR) cohort. Prospective results showed that our model identified 54.35% of all inpatient episodes of PA, and 91.20% of all PA that required subsequent chemoradiotherapy, with a lead-time of up to 3 months and a true alert of 67.62%. The risk assessment tool has attained an improved discriminative ability. It can be immediately deployed to the health system to provide automatic early warnings to adults at risk of PA. It has potential to identify personalized risk factors to facilitate customized PA interventions.

Keywords: Cancer prediction, deep learning, electronic health records, pancreatic adenocarcinoma.

References:

[1] Bray F, Ferlay J, Soerjomataram I, et al. Global Cancer Statistics 2018: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries: Global Cancer Statistics 2018 (J). CA A Cancer Journal for Clinicians, 2018, 68.
[2] Ferlay J, Colombet M, Soerjomataram I, et al. Estimating the global cancer incidence and mortality in 2018: GLOBOCAN sources and methods (J). International Journal of Cancer, 2019, 144.
[3] Hidalgo M. Pancreatic cancer (J). New England Journal of Medicine, 2010, 362(17): 1605-1617.
[4] Botsis T, Anagnostou VK, Hartvigsen G, Hripcsak G, Weng C. Developing a multivariable prognostic model for pancreatic endocrine tumors using the clinical data warehouse resources of a single institution. Appl Clin Inform 2010;1(1):12.
[5] Verma M. Pancreatic cancer biomarkers and their implication in cancer diagnosis and epidemiology. Cancers 2010;2(4):1830-7.
[6] Chakraborty S, Baine MJ, Sasson AR, Batra SK. Current status of molecular markers for early detection of sporadic pancreatic cancer. BBA-Rev Cancer 2011;1815(1):44-64.
[7] Arslan A A, Helzlsouer K J, Kooperberg C, et al. Anthropometric measures, body mass index, and pancreatic cancer: a pooled analysis from the Pancreatic Cancer Cohort Consortium (PanScan) (J). Archives of internal medicine, 2010, 170(9): 791-802.
[8] Ben Q, Xu M, Ning X, et al. Diabetes mellitus and risk of pancreatic cancer: a meta-analysis of cohort studies (J). European journal of cancer, 2011, 47(13): 1928-1937.
[9] Boursi B, Finkelman B, Giantonio B J, et al. A clinical prediction model to assess risk for pancreatic cancer among patients with new-onset diabetes (J). Gastroenterology, 2017, 152(4): 840-850. e3.
[10] Boursi S B, Finkelman B, Giantonio B J, et al. A clinical prediction model to assess risk for pancreatic cancer among patients with pre-diabetes (J). 2018.
[11] Cai Q C, Chen Y, Xiao Y, et al. A prediction rule for estimating pancreatic cancer risk in chronic pancreatitis patients with focal pancreatic mass lesions with prior negative EUS-FNA cytology (J). Scandinavian journal of gastroenterology, 2011, 46(4): 464-470.
[12] Canto M I, Goggins M, Hruban R H, et al. Screening for early pancreatic neoplasia in high-risk individuals: a prospective controlled study (J). Clinical gastroenterology and hepatology, 2006, 4(6): 766-781.
[13] Chari S T, Leibson C L, Rabe K G, et al. Probability of pancreatic cancer following diabetes: a population-based study (J). Gastroenterology, 2005, 129(2): 504-511.
[14] Gold D V, Goggins M, Modrak D E, et al. Detection of early-stage pancreatic adenocarcinoma (J). Cancer Epidemiology and Prevention Biomarkers, 2010, 19(11): 2786-2794.
[15] Radon T P, Massat N J, Jones R, et al. Identification of a Three-Biomarker Panel in Urine for Early Detection of Pancreatic Adenocarcinoma (J). Clinical Cancer Research An Official Journal of the American Association for Cancer Research, 2015, 21(15):3512.
[16] Grønborg M, Bunkenborg J, Kristiansen T Z, et al. Comprehensive proteomic analysis of human pancreatic juice (J). Journal of proteome research, 2004, 3(5): 1042-1055.
[17] Hart P A, Kamada P, Rabe K G, et al. Weight loss precedes cancer specific symptoms in pancreatic cancer associated diabetes mellitus (J). Pancreas, 2011, 40(5): 768.
[18] Hruban R H, Goggins M, Parsons J, et al. Progression model for pancreatic cancer (J). Clinical cancer research, 2000, 6(8): 2969-2972.
[19] Hsieh M H, Sun L M, Lin C L, et al. Development of a prediction model for pancreatic cancer in patients with type 2 diabetes using logistic regression and artificial neural network models (J). Cancer management and research, 2018, 10: 6317.
[20] Klein A P, Lindström S, Mendelsohn J B, et al. An absolute risk model to identify individuals at elevated risk for pancreatic cancer in the general population (J). PloS one, 2013, 8(9): e72311.
[21] Lowenfels A B, Maisonneuve P, Cavallini G, et al. Pancreatitis and the risk of pancreatic cancer
[J]. New England Journal of Medicine, 1993, 328(20): 1433-1437.
[22] Lucenteforte E, La Vecchia C, Silverman D, et al. Alcohol consumption and pancreatic cancer: a pooled analysis in the International Pancreatic Cancer Case–Control Consortium (PanC4) (J). Annals of oncology, 2012, 23(2): 374-382.
[23] Masahiro N, Yingsong L, Hidemi I, et al. Prediction model for pancreatic cancer risk in the general Japanese population (J). PLoS ONE, 2018, 13(9):e0203386.
[24] Pannala R, Basu A, Petersen G M, et al. New-onset diabetes: a potential clue to the early diagnosis of pancreatic cancer (J). The lancet oncology, 2009, 10(1): 88-95.
[25] Pelaez-Luna M, Takahashi N, Fletcher J G, et al. Resectability of presymptomatic pancreatic cancer and its relationship to onset of diabetes: a retrospective review of CT scans and fasting glucose values prior to diagnosis (J). American Journal of Gastroenterology, 2007, 102(10): 2157-2163.
[26] Sah R P, Nagpal S J S, Mukhopadhyay D, et al. New insights into pancreatic cancer-induced paraneoplastic diabetes (J). Nature reviews Gastroenterology & hepatology, 2013, 10(7): 423.
[27] Permuth-Wey J, Egan K M. Family history is a significant risk factor for pancreatic cancer: results from a systematic review and meta-analysis (J). Familial cancer, 2009, 8(2): 109-117.
[28] Sanoob M U, Madhu A, Ajesh K, et al. Artificial neural network for diagnosis of pancreatic cancer
[J]. International Journal on Cybernetics & Informatics, 2016, 5(2): 40-49.
[29] Verna E C, Hwang C, Stevens P D, et al. Pancreatic cancer screening in a prospective cohort of high-risk patients: a comprehensive strategy of imaging and genetics (J). Clinical cancer research, 2010, 16(20): 5028-5037.
[30] Wang W, Chen S, Brune K A, et al. PancPRO: risk assessment for individuals with a family history of pancreatic cancer (J). Journal of clinical oncology: official journal of the American Society of Clinical Oncology, 2007, 25(11): 1417.
[31] Glorot X, Bengio Y. Understanding the difficulty of training deep feedforward neural networks (J). Journal of Machine Learning Research, 2010, 9:249-256.
[32] Kingma D, Ba J. Adam: A Method for Stochastic Optimization (J). Computer ence, 2014.
[33] Zadrozny B, Elkan C. Transforming classifier scores into accurate multiclass probability estimates (C). The eighth ACM SIGKDD international conference. ACM, 2002.
[34] Mares T, Janouchova E , Kucerova A . Artificial neural networks in calibration of nonlinear mechanical models (J). Advances in Engineering Software, 2016, 95:68-81.