skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: SU-E-T-630: Predictive Modeling of Mortality, Tumor Control, and Normal Tissue Complications After Stereotactic Body Radiotherapy for Stage I Non-Small Cell Lung Cancer

Abstract

Purpose: While rates of local control have been well characterized after stereotactic body radiotherapy (SBRT) for stage I non-small cell lung cancer (NSCLC), less data are available characterizing survival and normal tissue toxicities, and no validated models exist assessing these parameters after SBRT. We evaluate the reliability of various machine learning techniques when applied to radiation oncology datasets to create predictive models of mortality, tumor control, and normal tissue complications. Methods: A dataset of 204 consecutive patients with stage I non-small cell lung cancer (NSCLC) treated with stereotactic body radiotherapy (SBRT) at the University of Pennsylvania between 2009 and 2013 was used to create predictive models of tumor control, normal tissue complications, and mortality in this IRB-approved study. Nearly 200 data fields of detailed patient- and tumor-specific information, radiotherapy dosimetric measurements, and clinical outcomes data were collected. Predictive models were created for local tumor control, 1- and 3-year overall survival, and nodal failure using 60% of the data (leaving the remainder as a test set). After applying feature selection and dimensionality reduction, nonlinear support vector classification was applied to the resulting features. Models were evaluated for accuracy and area under ROC curve on the 81-patient test set. Results: Models formore » common events in the dataset (such as mortality at one year) had the highest predictive power (AUC = .67, p < 0.05). For rare occurrences such as radiation pneumonitis and local failure (each occurring in less than 10% of patients), too few events were present to create reliable models. Conclusion: Although this study demonstrates the validity of predictive analytics using information extracted from patient medical records and can most reliably predict for survival after SBRT, larger sample sizes are needed to develop predictive models for normal tissue toxicities and more advanced machine learning methodologies need be consider in the future.« less

Authors:
 [1];  [2];  [3];  [2]; ;  [1]
  1. University of Pennsylvania, Philadelphia, PA (United States)
  2. (United States)
  3. Georgia Institute of Technology, Atlanta, GA (Georgia)
Publication Date:
OSTI Identifier:
22538139
Resource Type:
Journal Article
Resource Relation:
Journal Name: Medical Physics; Journal Volume: 42; Journal Issue: 6; Other Information: (c) 2015 American Association of Physicists in Medicine; Country of input: International Atomic Energy Agency (IAEA)
Country of Publication:
United States
Language:
English
Subject:
62 RADIOLOGY AND NUCLEAR MEDICINE; ANIMAL TISSUES; DATASETS; LUNGS; MEDICAL RECORDS; MORTALITY; NEOPLASMS; PATIENTS; RADIOTHERAPY; SIMULATION

Citation Formats

Lindsay, WD, Oncora Medical, LLC, Philadelphia, PA, Berlind, CG, Oncora Medical, LLC, Philadelphia, PA, Gee, JC, and Simone, CB. SU-E-T-630: Predictive Modeling of Mortality, Tumor Control, and Normal Tissue Complications After Stereotactic Body Radiotherapy for Stage I Non-Small Cell Lung Cancer. United States: N. p., 2015. Web. doi:10.1118/1.4924993.
Lindsay, WD, Oncora Medical, LLC, Philadelphia, PA, Berlind, CG, Oncora Medical, LLC, Philadelphia, PA, Gee, JC, & Simone, CB. SU-E-T-630: Predictive Modeling of Mortality, Tumor Control, and Normal Tissue Complications After Stereotactic Body Radiotherapy for Stage I Non-Small Cell Lung Cancer. United States. doi:10.1118/1.4924993.
Lindsay, WD, Oncora Medical, LLC, Philadelphia, PA, Berlind, CG, Oncora Medical, LLC, Philadelphia, PA, Gee, JC, and Simone, CB. Mon . "SU-E-T-630: Predictive Modeling of Mortality, Tumor Control, and Normal Tissue Complications After Stereotactic Body Radiotherapy for Stage I Non-Small Cell Lung Cancer". United States. doi:10.1118/1.4924993.
@article{osti_22538139,
title = {SU-E-T-630: Predictive Modeling of Mortality, Tumor Control, and Normal Tissue Complications After Stereotactic Body Radiotherapy for Stage I Non-Small Cell Lung Cancer},
author = {Lindsay, WD and Oncora Medical, LLC, Philadelphia, PA and Berlind, CG and Oncora Medical, LLC, Philadelphia, PA and Gee, JC and Simone, CB},
abstractNote = {Purpose: While rates of local control have been well characterized after stereotactic body radiotherapy (SBRT) for stage I non-small cell lung cancer (NSCLC), less data are available characterizing survival and normal tissue toxicities, and no validated models exist assessing these parameters after SBRT. We evaluate the reliability of various machine learning techniques when applied to radiation oncology datasets to create predictive models of mortality, tumor control, and normal tissue complications. Methods: A dataset of 204 consecutive patients with stage I non-small cell lung cancer (NSCLC) treated with stereotactic body radiotherapy (SBRT) at the University of Pennsylvania between 2009 and 2013 was used to create predictive models of tumor control, normal tissue complications, and mortality in this IRB-approved study. Nearly 200 data fields of detailed patient- and tumor-specific information, radiotherapy dosimetric measurements, and clinical outcomes data were collected. Predictive models were created for local tumor control, 1- and 3-year overall survival, and nodal failure using 60% of the data (leaving the remainder as a test set). After applying feature selection and dimensionality reduction, nonlinear support vector classification was applied to the resulting features. Models were evaluated for accuracy and area under ROC curve on the 81-patient test set. Results: Models for common events in the dataset (such as mortality at one year) had the highest predictive power (AUC = .67, p < 0.05). For rare occurrences such as radiation pneumonitis and local failure (each occurring in less than 10% of patients), too few events were present to create reliable models. Conclusion: Although this study demonstrates the validity of predictive analytics using information extracted from patient medical records and can most reliably predict for survival after SBRT, larger sample sizes are needed to develop predictive models for normal tissue toxicities and more advanced machine learning methodologies need be consider in the future.},
doi = {10.1118/1.4924993},
journal = {Medical Physics},
number = 6,
volume = 42,
place = {United States},
year = {Mon Jun 15 00:00:00 EDT 2015},
month = {Mon Jun 15 00:00:00 EDT 2015}
}
  • Purpose: To investigate the dose falloff region for various stereotactic body radiotherapy (SBRT) planning techniques used in the treatment of Stage I non-small-cell lung cancer reporting similar control rates. Methods and Materials: The SBRT plans were constructed on five patient data sets using seven different planning regimens. These regimens varied in the number of beams, number of fractions, prescription target, and prescribed dose used. For each case all regimens were planned using a common gross tumor volume (GTV). To compare dose falloff for the various regimens, resulting physical dose grids were converted into normalized total dose (NTD) grids. Furthermore, tomore » determine the potential coverage of microscopic extension of the various regimens minimal peripheral NTD (NTD-MP{sub 100}) were calculated and plotted as a function of incremental volume expansions of the GTV. Results: Average values for NTD-MP{sub 100} varied over a range of 174 Gy at the GTV periphery, but this range fell to 10 Gy at a distance of 14 mm from the GTV. Of 35 plans, 23 resulted in potential microscopic extension coverage of 78% to 95%. Averages for five of seven regimens fell within the range of 80% to 85%. Results were negligibly affected when intrafraction motion effects were accounted for. Conclusions: Although average NTD-MP{sub 100} varied dramatically at the GTV, periphery values became similar at a distance of 14 mm from the GTV. With the exception of two, potential coverage of microscopic extension was similar for all planning techniques, with averages falling within a 5% range.« less
  • Purpose: Hypofractionated stereotactic body radiation therapy (SBRT) has emerged as an effective treatment option for early-stage non-small cell lung cancer (NSCLC). Using data collected by the Elekta Lung Research Group, we generated a tumor control probability (TCP) model that predicts 2-year local control after SBRT as a function of biologically effective dose (BED) and tumor size. Methods and Materials: We formulated our TCP model as follows: TCP = e{sup [BED10-c Asterisk-Operator L-TCD50]/k} Division-Sign (1 + e{sup [BED10-c Asterisk-Operator L-TCD50]/k}), where BED10 is the biologically effective SBRT dose, c is a constant, L is the maximal tumor diameter, and TCD50 andmore » k are parameters that define the shape of the TCP curve. Least-squares optimization with a bootstrap resampling approach was used to identify the values of c, TCD50, and k that provided the best fit with observed actuarial 2-year local control rates. Results: Data from 504 NSCLC tumors treated with a variety of SBRT schedules were available. The mean follow-up time was 18.4 months, and 26 local recurrences were observed. The optimal values for c, TCD50, and k were 10 Gy/cm, 0 Gy, and 31 Gy, respectively. Thus, size-adjusted BED (sBED) may be defined as BED minus 10 times the tumor diameter (in centimeters). Our TCP model indicates that sBED values of 44 Gy, 69 Gy, and 93 Gy provide 80%, 90%, and 95% chances of tumor control at 2 years, respectively. When patients were grouped by sBED, the model accurately characterized the relationship between sBED and actuarial 2-year local control (r=0.847, P=.008). Conclusion: We have developed a TCP model that predicts 2-year local control rate after hypofractionated SBRT for early-stage NSCLC as a function of biologically effective dose and tumor diameter. Further testing of this model with additional datasets is warranted.« less
  • Purpose: To investigate pulmonary radiologic changes after lung stereotactic body radiotherapy (SBRT), to distinguish between mass-like fibrosis and tumor recurrence. Methods and Materials: Eighty consecutive patients treated with 3- to 5-fraction SBRT for early-stage peripheral non-small cell lung cancer with a minimum follow-up of 12 months were reviewed. The mean biologic equivalent dose received was 150 Gy (range, 78-180 Gy). Patients were followed with serial CT imaging every 3 months. The CT appearance of consolidation was defined as diffuse or mass-like. Progressive disease on CT was defined according to Response Evaluation Criteria in Solid Tumors 1.1. Positron emission tomography (PET)more » CT was used as an adjunct test. Tumor recurrence was defined as a standardized uptake value equal to or greater than the pretreatment value. Biopsy was used to further assess consolidation in select patients. Results: Median follow-up was 24 months (range, 12.0-36.0 months). Abnormal mass-like consolidation was identified in 44 patients (55%), whereas diffuse consolidation was identified in 12 patients (15%), at a median time from end of treatment of 10.3 months and 11.5 months, respectively. Tumor recurrence was found in 35 of 44 patients with mass-like consolidation using CT alone. Combined with PET, 10 of the 44 patients had tumor recurrence. Tumor size (hazard ratio 1.12, P=.05) and time to consolidation (hazard ratio 0.622, P=.03) were predictors for tumor recurrence. Three consecutive increases in volume and increasing volume at 12 months after treatment in mass-like consolidation were highly specific for tumor recurrence (100% and 80%, respectively). Patients with diffuse consolidation were more likely to develop grade {>=}2 pneumonitis (odds ratio 26.5, P=.02) than those with mass-like consolidation (odds ratio 0.42, P=.07). Conclusion: Incorporating the kinetics of mass-like consolidation and PET to the current criteria for evaluating posttreatment response will increase the likelihood of correctly identifying patients with progressive disease after lung SBRT.« less
  • Purpose: To show that a distribution of cell surviving fractions S{sub 2} in a heterogeneous group of patients can be derived from tumor-volume variation curves during radiotherapy for non-small cell lung cancer. Methods: Our analysis was based on two data sets of tumor-volume variation curves for heterogeneous groups of 17 patients treated for nonsmall cell lung cancer with conventional dose fractionation. The data sets were obtained previously at two independent institutions by using megavoltage (MV) computed tomography (CT). Statistical distributions of cell surviving fractions S{sup 2} and cell clearance half-lives of lethally damaged cells T1/2 have been reconstructed in eachmore » patient group by using a version of the two-level cell population tumor response model and a simulated annealing algorithm. The reconstructed statistical distributions of the cell surviving fractions have been compared to the distributions measured using predictive assays in vitro. Results: Non-small cell lung cancer presents certain difficulties for modeling surviving fractions using tumor-volume variation curves because of relatively large fractional hypoxic volume, low gradient of tumor-volume response, and possible uncertainties due to breathing motion. Despite these difficulties, cell surviving fractions S{sub 2} for non-small cell lung cancer derived from tumor-volume variation measured at different institutions have similar probability density functions (PDFs) with mean values of 0.30 and 0.43 and standard deviations of 0.13 and 0.18, respectively. The PDFs for cell surviving fractions S{sup 2} reconstructed from tumor volume variation agree with the PDF measured in vitro. Comparison of the reconstructed cell surviving fractions with patient survival data shows that the patient survival time decreases as the cell surviving fraction increases. Conclusion: The data obtained in this work suggests that the cell surviving fractions S{sub 2} can be reconstructed from the tumor volume variation curves measured during radiotherapy with conventional fractionation. The proposed method can be used for treatment evaluation and adaptation.« less
  • Purpose: Quantification of volume changes on CBCT during SBRT for NSCLC may provide a useful radiological marker for radiation response and adaptive treatment planning, but the reproducibility of CBCT volume delineation is a concern. This study is to quantify inter-scan/inter-observer variability in tumor volume delineation on CBCT. Methods: Twenty earlystage (stage I and II) NSCLC patients were included in this analysis. All patients were treated with SBRT with a median dose of 54 Gy in 3 to 5 fractions. Two physicians independently manually contoured the primary gross tumor volume on CBCTs taken immediately before SBRT treatment (Pre) and after themore » same SBRT treatment (Post). Absolute volume differences (AVD) were calculated between the Pre and Post CBCTs for a given treatment to quantify inter-scan variability, and then between the two observers for a given CBCT to quantify inter-observer variability. AVD was also normalized with respect to average volume to obtain relative volume differences (RVD). Bland-Altman approach was used to evaluate variability. All statistics were calculated with SAS version 9.4. Results: The 95% limit of agreement (mean ± 2SD) on AVD and RVD measurements between Pre and Post scans were −0.32cc to 0.32cc and −0.5% to 0.5% versus −1.9 cc to 1.8 cc and −15.9% to 15.3% for the two observers respectively. The 95% limit of agreement of AVD and RVD between the two observers were −3.3 cc to 2.3 cc and −42.4% to 28.2% respectively. The greatest variability in inter-scan RVD was observed with very small tumors (< 5 cc). Conclusion: Inter-scan variability in RVD is greatest with small tumors. Inter-observer variability was larger than inter-scan variability. The 95% limit of agreement for inter-observer and inter-scan variability (∼15–30%) helps define a threshold for clinically meaningful change in tumor volume to assess SBRT response, with larger thresholds needed for very small tumors. Part of the work was funded by a Kaye award; Disclosure/Conflict of interest: Raymond H. Mak: Stock ownership: Celgene, Inc. Consulting: Boehringer-Ingelheim, Inc.« less