PMID- 30499597
OWN - NLM
STAT- MEDLINE
DCOM- 20190228
LR  - 20190228
IS  - 2473-4209 (Electronic)
IS  - 0094-2405 (Linking)
VI  - 46
IP  - 2
DP  - 2019 Feb
TI  - Automated data extraction and ensemble methods for predictive modeling of breast 
      cancer outcomes after radiation therapy.
PG  - 1054-1063
LID - 10.1002/mp.13314 [doi]
AB  - PURPOSE: The purpose of this study was to compare the effectiveness of ensemble 
      methods (e.g., random forests) and single-model methods (e.g., logistic 
      regression and decision trees) in predictive modeling of post-RT treatment 
      failure and adverse events (AEs) for breast cancer patients using automatically 
      extracted EMR data. METHODS: Data from 1967 consecutive breast radiotherapy (RT) 
      courses at one institution between 2008 and 2015 were automatically extracted 
      from EMRs and oncology information systems using extraction software. Over 230 
      variables were extracted spanning the following variable segments: patient 
      demographics, medical/surgical history, tumor characteristics, RT treatment 
      history, and AEs tracked using CTCAEv4.0. Treatment failure was extracted 
      algorithmically by searching posttreatment encounters for evidence of local, 
      nodal, or distant failure. Individual models were trained using decision trees, 
      logistic regression, random forests, and boosted decision trees to predict 
      treatment failures and AEs. Models were fit on 75% of the data and evaluated for 
      probability calibration and area under the ROC curve (AUC) on the remaining test 
      set. The impact of each variable segment was assessed by retraining without the 
      segment and measuring change in AUC (DeltaAUC). RESULTS: All AUC values were 
      statistically significant (P < 0.05). Ensemble methods outperformed single-model 
      methods across all outcomes. The best ensemble method outperformed decision trees 
      and logistic regression by an average AUC of 0.053 and 0.034, respectively. Model 
      probabilities were well calibrated as evidenced by calibration curves. Excluding 
      the patient medical history variable segment led to the largest AUC reduction in 
      all models (Average DeltaAUC = -0.025), followed by RT treatment history (-0.021) and 
      tumor information (-0.015). CONCLUSION: In this largest such study in breast 
      cancer performed to date, automatically extracted EMR data provided a basis for 
      reliable outcome predictions across multiple statistical methods. Ensemble 
      methods provided substantial advantages over single-model methods. Patient 
      medical history contributed the most to prediction quality.
CI  - (c) 2018 American Association of Physicists in Medicine.
FAU - Lindsay, William D
AU  - Lindsay WD
AD  - Oncora Medical, Philadelphia, PA, 19103, USA.
FAU - Ahern, Christopher A
AU  - Ahern CA
AD  - Oncora Medical, Philadelphia, PA, 19103, USA.
FAU - Tobias, Jacob S
AU  - Tobias JS
AD  - Oncora Medical, Philadelphia, PA, 19103, USA.
FAU - Berlind, Christopher G
AU  - Berlind CG
AD  - Oncora Medical, Philadelphia, PA, 19103, USA.
FAU - Chinniah, Chidambaram
AU  - Chinniah C
AD  - Department of Radiation Oncology, Hospital of the University of Pennsylvania, 
      Philadelphia, PA, 19104, USA.
FAU - Gabriel, Peter E
AU  - Gabriel PE
AD  - Department of Radiation Oncology, Hospital of the University of Pennsylvania, 
      Philadelphia, PA, 19104, USA.
FAU - Gee, James C
AU  - Gee JC
AD  - Department of Radiology, Perelman School of Medicine, University of Pennsylvania, 
      Philadelphia, PA 19104, USA.
FAU - Simone, Charles B 2nd
AU  - Simone CB 2nd
AD  - Department of Radiation Oncology, University of Maryland School of Medicine, 
      Baltimore, MD, 21201, USA.
LA  - eng
PT  - Journal Article
DEP - 20181228
PL  - United States
TA  - Med Phys
JT  - Medical physics
JID - 0425746
SB  - IM
MH  - Breast Neoplasms/*pathology/*radiotherapy
MH  - Data Mining/*methods
MH  - *Decision Trees
MH  - *Electronic Health Records
MH  - Female
MH  - Humans
MH  - *Machine Learning
MH  - Middle Aged
MH  - Predictive Value of Tests
MH  - Radiotherapy Dosage
MH  - Treatment Outcome
OTO - NOTNLM
OT  - automated data extraction
OT  - ensemble methods
OT  - machine learning
OT  - predictive modeling
OT  - radiotherapy outcomes
EDAT- 2018/12/01 06:00
MHDA- 2019/03/01 06:00
CRDT- 2018/12/01 06:00
PHST- 2018/02/21 00:00 [received]
PHST- 2018/11/11 00:00 [revised]
PHST- 2018/11/12 00:00 [accepted]
PHST- 2018/12/01 06:00 [pubmed]
PHST- 2019/03/01 06:00 [medline]
PHST- 2018/12/01 06:00 [entrez]
AID - 10.1002/mp.13314 [doi]
PST - ppublish
SO  - Med Phys. 2019 Feb;46(2):1054-1063. doi: 10.1002/mp.13314. Epub 2018 Dec 28.