Dynamically prognosticating patients with hepatocellular carcinoma through survival paths mapping based on time-series data

Patients with hepatocellular carcinoma (HCC) always require routine surveillance and repeated treatment, which leads to accumulation of huge amount of clinical data. A predictive model utilizes the time-series data to facilitate dynamic prognosis prediction and treatment planning is warranted. Here we introduced an analytical approach, which converts the time-series data into a cascading survival map, in which each survival path bifurcates at fixed time interval depending on selected prognostic features by the Cox-based feature selection. We apply this approach in an intermediate-scale database of patients with BCLC stage B HCC and get a survival map consisting of 13 different survival paths, which is demonstrated to have superior or equal value than conventional staging systems in dynamic prognosis prediction from 3 to 12 months after initial diagnosis in derivation, internal testing, and multicentric testing cohorts. This methodology/model could facilitate dynamic prognosis prediction and treatment planning for patients with HCC in the future.Patients with hepatocellular carcinoma require regular follow-up. Here, using Cox-based feature selection to identify key prognostic features, the authors convert time-series follow-up data into a cascading survival map, and show that the approach improves dynamic prognosis prediction for patients.

[1]  Yuanxi Li,et al.  Updating Markov models to integrate cross-sectional and longitudinal studies , 2017, Artif. Intell. Medicine.

[2]  Milos Hauskrecht,et al.  A Temporal Abstraction Framework for Classifying Clinical Temporal Data , 2009, AMIA.

[3]  A. Jemal,et al.  Cancer statistics, 2017 , 2017, CA: a cancer journal for clinicians.

[4]  Ma Li,et al.  CURE-SMOTE algorithm and hybrid algorithm for feature selection and parameter optimization based on random forests , 2017, BMC Bioinformatics.

[5]  R. Lencioni,et al.  Treatment of intermediate/advanced hepatocellular carcinoma in the clinic: how can outcomes be improved? , 2010, The oncologist.

[6]  M. Kudo,et al.  Combined sequential use of HAP and ART scores to predict survival outcome and treatment failure following chemoembolization in hepatocellular carcinoma: a multi-center comparative study , 2016, Oncotarget.

[7]  Seungjin Choi,et al.  Inference of dynamic networks using time-course data , 2014, Briefings Bioinform..

[8]  V. Mazzaferro,et al.  EASL-EORTC Clinical Practice Guidelines: Management of hepatocellular carcinoma European Association for the Study of the Liver ⇑ , European Organisation for Research and Treatment of Cancer , 2012 .

[9]  H. Heinzl,et al.  The ART of decision making: Retreatment with transarterial chemoembolization in patients with hepatocellular carcinoma , 2013, Hepatology.

[10]  Riccardo Lencioni,et al.  Modified RECIST (mRECIST) Assessment for Hepatocellular Carcinoma , 2010, Seminars in liver disease.

[11]  J. Bruix,et al.  Treatment of intermediate-stage hepatocellular carcinoma , 2014, Nature Reviews Clinical Oncology.

[12]  J J Deeks,et al.  Systematic reviews of published evidence: miracles or minefields? , 1998, Annals of oncology : official journal of the European Society for Medical Oncology.

[13]  F. Harrell,et al.  Prognostic/Clinical Prediction Models: Multivariable Prognostic Models: Issues in Developing Models, Evaluating Assumptions and Adequacy, and Measuring and Reducing Errors , 2005 .

[14]  M. Dumont,et al.  European Association for the Study of the Liver , 1971 .

[15]  Dimitrios I. Fotiadis,et al.  Machine learning applications in cancer prognosis and prediction , 2014, Computational and structural biotechnology journal.

[16]  G. Dusheiko,et al.  Management of hepatocellular carcinoma. , 1992, Journal of hepatology.

[17]  Riccardo Lencioni,et al.  EASL-EORTC clinical practice guidelines: management of hepatocellular carcinoma. , 2012, Journal of hepatology.

[18]  Sameem Abdul Kareem,et al.  Oral cancer prognosis based on clinicopathologic and genomic markers using a hybrid of feature selection and machine learning methods , 2013, BMC Bioinformatics.

[19]  Cécile Proust-Lima,et al.  Quantifying and comparing dynamic predictive accuracy of joint models for longitudinal marker and time‐to‐event in presence of censoring and competing risks , 2015, Biometrics.

[20]  M. Carvello,et al.  How reliable is current imaging in restaging rectal cancer after neoadjuvant therapy? , 2013, World journal of gastroenterology.

[21]  Daniel B. Mark,et al.  TUTORIAL IN BIOSTATISTICS MULTIVARIABLE PROGNOSTIC MODELS: ISSUES IN DEVELOPING MODELS, EVALUATING ASSUMPTIONS AND ADEQUACY, AND MEASURING AND REDUCING ERRORS , 1996 .

[22]  A review , 2019 .

[23]  S. Holm A Simple Sequentially Rejective Multiple Test Procedure , 1979 .

[24]  Haiyong Wang,et al.  Correlation analysis of preoperative serum alpha-fetoprotein (AFP) level and prognosis of hepatocellular carcinoma (HCC) after hepatectomy , 2013, World Journal of Surgical Oncology.

[25]  Gaëtan MacGrogan,et al.  Variables with time-varying effects and the Cox model: Some statistical concepts illustrated with a prognostic factor study in breast cancer , 2010, BMC medical research methodology.

[26]  P. Goldstraw New TNM classification: achievements and hurdles. , 2013, Translational lung cancer research.

[27]  Cécile Proust-Lima,et al.  Joint latent class models for longitudinal and time-to-event data: A review , 2014, Statistical methods in medical research.

[28]  H. Ye,et al.  Prognostic value of alkaline phosphatase, gamma-glutamyl transpeptidase and lactate dehydrogenase in hepatocellular carcinoma patients treated with liver resection. , 2016, International journal of surgery.

[29]  R. Xu,et al.  Validation and ranking of seven staging systems of hepatocellular carcinoma , 2017, Oncology letters.