Prognostic Value and Reproducibility of Pretreatment CT Texture Features in Stage III Non-Small Cell Lung Cancer

doi:10.1016/j.ijrobp.2014.07.020

International Journal of Radiation OncologyBiologyPhysics

Volume 90, Issue 4, 15 November 2014, Pages 834-842

https://doi.org/10.1016/j.ijrobp.2014.07.020 Get rights and content

Purpose

To determine whether pretreatment CT texture features can improve patient risk stratification beyond conventional prognostic factors (CPFs) in stage III non-small cell lung cancer (NSCLC).

Methods and Materials

We retrospectively reviewed 91 cases with stage III NSCLC treated with definitive chemoradiation therapy. All patients underwent pretreatment diagnostic contrast enhanced computed tomography (CE-CT) followed by 4-dimensional CT (4D-CT) for treatment simulation. We used the average-CT and expiratory (T50-CT) images from the 4D-CT along with the CE-CT for texture extraction. Histogram, gradient, co-occurrence, gray tone difference, and filtration-based techniques were used for texture feature extraction. Penalized Cox regression implementing cross-validation was used for covariate selection and modeling. Models incorporating texture features from the 33 image types and CPFs were compared to those with models incorporating CPFs alone for overall survival (OS), local-regional control (LRC), and freedom from distant metastases (FFDM). Predictive Kaplan-Meier curves were generated using leave-one-out cross-validation. Patients were stratified based on whether their predicted outcome was above or below the median. Reproducibility of texture features was evaluated using test-retest scans from independent patients and quantified using concordance correlation coefficients (CCC). We compared models incorporating the reproducibility seen on test-retest scans to our original models and determined the classification reproducibility.

Results

Models incorporating both texture features and CPFs demonstrated a significant improvement in risk stratification compared to models using CPFs alone for OS (P=.046), LRC (P=.01), and FFDM (P=.005). The average CCCs were 0.89, 0.91, and 0.67 for texture features extracted from the average-CT, T50-CT, and CE-CT, respectively. Incorporating reproducibility within our models yielded 80.4% (±3.7% SD), 78.3% (±4.0% SD), and 78.8% (±3.9% SD) classification reproducibility in terms of OS, LRC, and FFDM, respectively.

Conclusions

Pretreatment tumor texture may provide prognostic information beyond that obtained from CPFs. Models incorporating feature reproducibility achieved classification rates of ∼80%. External validation would be required to establish texture as a prognostic factor.

Introduction

Lung cancer is currently the most common cause of death from cancer in the United States (1). Frequently, patients present with stage III disease and are not candidates for surgical resection. For these patients, standard of care consists of definitive chemoradiation therapy. Even when patients are treated aggressively, patient 3-year survival rate is approximately 27% (2). Inoperable non-small cell lung cancer (NSCLC) patients are a very heterogeneous population with varying degrees of tumor extent, comorbidity, and other characteristics. This presents a significant challenge to clinicians attempting to provide optimal treatment. Traditional TNM staging is not ideal for stratifying patients, and there is a tremendous need to develop better tools for assessing prognosis.

Efforts have been made to address this issue by identifying prognostic genetic expression signatures and using functional imaging techniques such as fluorodeoxyglucose-labeled positron emission tomography (PET) 3, 4, 5. Recently, tumor heterogeneity as assessed by computed tomography (CT) has yielded promising preliminary results in a variety of cancers 6, 7, 8. These techniques assess the spatial variation of tumor density within a patient's tumor. Because CT scans are routinely obtained for all patients undergoing radiation therapy, prognostic markers generated in this manner would be less costly and less time consuming than genetic or functional imaging-based techniques.

In this study we examined the impact of CT texture features to enhance patient risk stratification, beyond conventional prognostic factors (CPFs) for patients with stage III NSCLC.

Section snippets

Patients

We retrospectively reviewed the medical records of patients with stage III NSCLC treated with definitive radiation therapy between July 2004 and January 2012. These dates were chosen in order to include patients receiving 4-dimensional (4D)CT, which our institution implemented in early 2004, and provide adequate follow-up time. We excluded all patients receiving induction chemotherapy and proton-based radiation therapy and those with <5 years' posttreatment for solid tumor, multiple primary

Model development and analysis

Using L1 penalized Cox proportional hazards regression, models for OS, LRC, and FFDM were generated. Covariates with non-zero coefficients are shown in Table 3. To illustrate patient stratification using these models, cross-validated Kaplan-Meier curves for OS, LRC, and FFDM were developed along with their associated 95% confidence intervals (CI). Models with and without texture features are shown in Figure 1. Models using both texture features and CPFs demonstrated a statistically significant

Discussion

We showed that models incorporating both texture features and CPFs demonstrated a statistically significant improvement in stratification compared to models using CPFs alone in cross-validated Kaplan-Meier curves in terms of OS (P=.046), LRC (P=.01), and FFDM (P=.005). Furthermore, we showed that even when incorporating the reproducibility of our metrics within our models, classification rates of approximately 80% are still achieved.

We used cross-validation for both model selection and model

Conclusions

Patient outcome modeling has significant implications in many fields of medicine and particularly in oncology. Our study found that the combination of CT texture and CPFs can be used to generate superior outcome models when compared to CPFs alone in terms of OS, LRC, and FFDM. This additional information could be of use to physicians, patients, and caregivers. Further work needs to be done in order to generate widely applicable, accurate risk prediction tools capable of being implemented

References (25)

D.M. Hari et al.
AJCC Cancer Staging Manual 7th edition criteria for colon cancer: Do the complex modifications improve prognostic assessment?
J Am Coll Surg
(2013)
R. Siegel et al.
Cancer statistics, 2013
CA Cancer J Clin
(2013)
N. Hanna et al.
Phase III study of cisplatin, etoposide, and concurrent chest radiation with or without consolidation docetaxel in patients with inoperable stage III non-small-cell lung cancer: The Hoosier Oncology Group and U.S. Oncology
J Clin Oncol
(2008)
H.Y. Chen et al.
A five-gene signature and clinical outcome in non-small-cell lung cancer
N Engl J Med
(2007)
M. Machtay et al.
Prediction of survival by [18F]fluorodeoxyglucose positron emission tomography in patients with locally advanced non-small-cell lung cancer undergoing definitive chemoradiation therapy: Results of the ACRIN 6668/RTOG 0235 trial
J Clin Oncol
(2013)
D.J. Raz et al.
A multigene assay is prognostic of survival in patients with early-stage lung adenocarcinoma
Clin Cancer Res
(2008)
O.S. Al-Kadi et al.
Texture analysis of aggressive and nonaggressive lung tumor CE CT images
IEEE Trans Biomed Eng
(2008)
B. Ganeshan et al.
Texture analysis of non-small cell lung cancer on unenhanced computed tomography: Initial evidence for a relationship with tumour glucose metabolism and stage
Cancer Imaging
(2010)
T. Win et al.
Tumor heterogeneity and permeability as measured on the ct component of PET/CT predict survival in patients with non-small cell lung cancer
Clin Cancer Res
(2013)
T. Berghmans et al.
Prognostic factors in stage III non-small cell lung cancer: A review of conventional, metabolic and new biological variables
Ther Adv Med Oncol
(2011)

M. Amadasun et al.

Textural features corresponding to textural properties

IEEE Trans Syst Man Cybern

(1989)

R.M. Haralick et al.

Textural features for image classification

IEEE T Syst Man Cyb

(1973)

Cited by (162)

Predicting the BRAF mutation with pretreatment MRI radiomics features for melanoma brain metastases receiving Gamma Knife radiosurgery
2023, Clinical Radiology
To develop a model using radiomics features extracted from magnetic resonance imaging (MRI) images of Gamma Knife radiosurgery (GKRS) to predict the BRAF mutation in patients with melanoma brain metastases (MBM).
Data of 220 tumours were classified into two groups. One was a group whose BRAF mutation was identified, and the other group whose BRAF mutation was not identified. We extracted 1,962 radiomics features from gadolinium contrast-enhanced T1-weighted MRI treatment-planning images. Synthetic Minority Over-sampling TEchnique (SMOTE) was performed to address the unbalanced data-related issues. A single-layer neural network (NN) was used to build predictive models with radiomics features. The sensitivity, specificity, accuracy, and the area under the curve (AUC) were evaluated to assess the model performance.
The prediction performance for the final evaluation without the SMOTE had an accuracy of 77.14%, a specificity of 82.44%, a sensitivity of 81.85%, and an AUC of 0.79. The application of SMOTE improved the prediction model to an accuracy of 83.1%, a specificity of 87.07%, a sensitivity of 78.82%, and an AUC of 0.82.
The current study showed the feasibility of generating a highly accurate NN model for the BRAF mutation prediction. The prediction performance improved with SMOTE. The model assists physicians to obtain more accurate expectations of the treatment outcome without a genetic test.
Development and optimisation of a preclinical cone beam computed tomography-based radiomics workflow for radiation oncology research
2023, Physics and Imaging in Radiation Oncology
Radiomics features derived from medical images have the potential to act as imaging biomarkers to improve diagnosis and predict treatment response in oncology. However, the complex relationships between radiomics features and the biological characteristics of tumours are yet to be fully determined. In this study, we developed a preclinical cone beam computed tomography (CBCT) radiomics workflow with the aim to use in vivo models to further develop radiomics signatures.
CBCT scans of a mouse phantom were acquired using onboard imaging from a small animal radiotherapy research platform (SARRP, Xstrahl). The repeatability and reproducibility of radiomics outputs were compared across different imaging protocols, segmentation sizes, pre-processing parameters and materials. Robust features were identified and used to compare scans of two xenograft mouse tumour models (A549 and H460).
Changes to the radiomics workflow significantly impact feature robustness. Preclinical CBCT radiomics analysis is feasible with 119 stable features identified from scans imaged at 60 kV, 25 bin width and 0.26 mm slice thickness. Large variation in segmentation volumes reduced the number of reliable radiomics features for analysis. Standardization in imaging and analysis parameters is essential in preclinical radiomics analysis to improve accuracy of outputs, leading to more consistent and reproducible findings.
We present the first optimised workflow for preclinical CBCT radiomics to identify imaging biomarkers. Preclinical radiomics has the potential to maximise the quantity of data captured in in vivo experiments and could provide key information supporting the wider application of radiomics.
Radiomics for Predicting Lung Cancer Outcomes Following Radiotherapy: A Systematic Review
2022, Clinical Oncology
Citation Excerpt :
Only 11% of features in the final model also predicted for overall survival, however, perhaps suggesting that distant disease and patient survival may be influenced by separate elements of the underlying biological phenotype. Incorporation of texture features stable across both three- and four-dimensional CT scans ± contrast in another study improved clinical models of distant control with average classification rates of about 0.89 [49]. Although no studies explored features predictive for nodal relapse specifically following conventional radiotherapy as a primary focus, a few studies did examine features that corresponded with likelihood of ‘regional failure’, as recorded in Table 1 [46,47,49].
Lung cancer's radiomic phenotype may potentially inform clinical decision-making with respect to radical radiotherapy. At present there are no validated biomarkers available for the individualisation of radical radiotherapy in lung cancer and the mortality rate of this disease remains the highest of all other solid tumours. MEDLINE was searched using the terms ‘radiomics’ and ‘lung cancer’ according to the Preferred Reporting Items for Systematic Reviews and Met-Analyses (PRISMA) guidance. Radiomics studies were defined as those manuscripts describing the extraction and analysis of at least 10 quantifiable imaging features. Only those studies assessing disease control, survival or toxicity outcomes for patients with lung cancer following radical radiotherapy ± chemotherapy were included. Study titles and abstracts were reviewed by two independent reviewers. The Radiomics Quality Score was applied to the full text of included papers. Of 244 returned results, 44 studies met the eligibility criteria for inclusion. End points frequently reported were local (17%), regional (17%) and distant control (31%), overall survival (79%) and pulmonary toxicity (4%). Imaging features strongly associated with clinical outcomes include texture features belonging to the subclasses Gray level run length matrix, Gray level co-occurrence matrix and kurtosis. The median cohort size for model development was 100 (15–645); in the 11 studies with external validation in a separate independent population, the median cohort size was 84 (21–295). The median number of imaging features extracted was 184 (10–6538). The median Radiomics Quality Score was 11% (0–47). Patient-reported outcomes were not incorporated within any studies identified. No studies externally validated a radiomics signature in a registered prospective study. Imaging-derived indices attained through radiomic analyses could equip thoracic oncologists with biomarkers for treatment response, patterns of failure, normal tissue toxicity and survival in lung cancer. Based on routine scans, their non-invasive nature and cost-effectiveness are major advantages over conventional pathological assessment. Improved tools are required for the appraisal of radiomics studies, as significant barriers to clinical implementation remain, such as standardisation of input scan data, quality of reporting and external validation of signatures in randomised, interventional clinical trials.
Development and Validation of a CT-Based Signature for the Prediction of Distant Metastasis Before Treatment of Non-Small Cell Lung Cancer
2022, Academic Radiology
Citation Excerpt :
In their research, feature selection via concave minimization (FSV) is used to select effective radiomics features, finally the selected features are mainly composed of first-order wavelet characteristics and auxiliary classification. While in our study, all radiomics features of the training cohort are standardized at first, and then the LASSO and RFE algorithm are used to select features, all selected features are high-order features, suggesting that the high-order features of the tumor can effectively assist in the diagnosis of DM (30). Although the effect of Junior's (17) radiomics model was better than that reported in our research, the sample size was inconsistent; absence of feature repeatability analyses; and independent validation may lead to unreliable results.
To develop and validate a radiomics model, a clinical-semantic model and a combined model by using standard methods for the pretreatment prediction of distant metastasis (DM) in patients with non-small-cell lung cancer (NSCLC) and to explore whether the combined model provides added value compared to the individual models.
This retrospective study involved 356 patients with NSCLC. According to the image biomarker standardization initiative reference manual, we standardized the image processing and feature extraction using in-house software. Finally, 6692 radiomics features were extracted from each lesion based on contrast-enhanced chest CT images. The least absolute shrinkage selection operator and the recursive feature elimination algorithm were used to select features. The logistic regression classifier was used to build the model. Three models (radiomics model, clinical-semantic model and combined model) were constructed to predict DM in NSCLC. Area under the receiver operating characteristic curves were used to validate the ability of the three models to predict DM. A visual nomogram based on the combined model was developed for DM risk assessment in each patient.
The receiver operating characteristic curve showed predictive performance for DM of the radiomics model (area under the curve [AUC] values for training and validation were 0.76 [95% CI, 0.704 - 0.820] and 0.76 [95% CI, 0.653 - 0.858], respectively). The combined model had AUCs of 0.78 (95% CI, 0.723 - 0.835) and 0.77 (95% CI, 0.673 - 0.870) in the training and validation cohorts, respectively. Both the radiomics model and combined model performed better than the clinical-semantic model (0.70 [95% CI, 0.634 - 0.760] and 0.67 [95% CI, 0.554 - 0.787] in the training and validation cohorts, respectively).
The radiomics model and combined model may be useful for the prediction of DM in patients with NSCLC.
Radiomics Nomogram for Predicting Locoregional Failure in Locally Advanced Non–small Cell Lung Cancer Treated with Definitive Chemoradiotherapy
2022, Academic Radiology
To develop and validate a computed tomography (CT)-based radiomics nomogram for predicting locoregional failure (LRF) in patients with locally advanced non–small cell lung cancer (NSCLC) treated with definitive chemoradiotherapy (CRT).
A total of 141 patients with locally advanced NSCLC treated with definitive CRT from January 2014 to December 2017 were included and divided into testing cohort (n = 100) and validation (n = 41) cohort. Radiomics features were extracted from pretreatment contrast enhanced CT. The least absolute shrinkage and selection operator logistic regression was processed to select predictive features from the testing cohort and constructed a radiomics signature. Clinical characteristics and the radiomics signature were analyzed using univariable and multivariate Cox regression. The radiomics nomogram was established with the radiomics signature and independent clinical factors. Harrell's C-index, calibration curves and decision curves were used to assess the performance of the radiomics nomogram.
The radiomics signature, which consisted of eight selected features, was an independent factor of LRF. The clinical predictors of LRF were the histologic type and clinical stage. The radiomics nomogram combined with the radiomics signature and clinical prognostic factors showed good performance with C-indexes of 0.796 (95% confidence interval [CI]: 0.709–0.883) and 0.756 (95% CI: 0.674–0.838) in the testing and validation cohorts respectively. Additionally, the combined nomogram resulted in better performance (p < 0.001) for the estimation of LRF than the nomograms with the radiomics signature (C-index: 0.776; 95% CI: 0.686–0.866) or clinical predictors (C-index: 0.641; 95% CI: 0.542–0.740) alone.
The radiomics nomogram provided the best performance for LRF prediction in patients with locally advanced NSCLC, which may help optimize individual treatments.
Predicting the radiotherapeutic treatment response of non-small cell lung cancer
2022, Computer Methods and Programs in Biomedicine Update
Predicting the treatment response of a patient undergoing treatment for NSCLC (non-small cell lung cancer) is useful in the estimation of effectiveness of the treatment protocol. To assess the prognostic value of radiomic features extracted from CBCT (cone-beam computed tomography) images of NSCLC patients, an efficient classification model is developed. The study aims to predict the early response (predicting the response at the third week of the radiation therapy) of the patients using the first-week and third-week data, of the acquired CBCT images.
This retrospective study includes 99 NSCLC patients having undergone conventionally fractionated radiation therapy (RT) over six weeks. The radiomic features are computed using the Python pyradiomics package. Initially 107 radiomic features were chosen. A triage-based learning paradigm is implemented, wherein the machine decides upon the difficulty of a sample. If the machine recognizes an instance to be hard to classify, then it is left for human expert assistance, otherwise, a feed-forward neural network is trained which classifies a patient’s response to treatment.
The proposed model achieves accuracy $0.75 \pm 0.03$ , precision $0.79 \pm 0.03$ , sensitivity $0.82 \pm 0.05$ , f-score $0.80 \pm 0.03$ and specificity $0.61 \pm 0.08$ to predict the patient’s early response. Using, ensemble learning technique an accuracy of $0.79 \pm 0.03$ , precision $0.91 \pm 0.06$ , sensitivity $0.80 \pm 0.08$ , f-score $0.84 \pm 0.03$ and specificity $0.82 \pm 0.13$ is achieved.
The experimental results revealed that the accuracy of proposed model is better than the state-of-the-art technique. The results show that CBCT radiomic features can prognosticate the tumor response for NSCLC patients.

View all citing articles on Scopus

: This work was supported in part by the American Legion Auxiliary, the American Association of Physicists in Medicine Graduate Fellowship, the University of Texas Graduate School of Biomedical Sciences at Houston, and National Cancer Institute grant R03CA178495-01. The University of Texas MD Anderson Cancer Center is supported by National Institutes of Health core grant CA 16672.

: Conflict of interest: none.

View full text

Clinical InvestigationPrognostic Value and Reproducibility of Pretreatment CT Texture Features in Stage III Non-Small Cell Lung Cancer

Purpose

Methods and Materials

Results

Conclusions

Introduction

Section snippets

Patients

Model development and analysis

Discussion

Conclusions

J Am Coll Surg

Cancer statistics, 2013

CA Cancer J Clin

Phase III study of cisplatin, etoposide, and concurrent chest radiation with or without consolidation docetaxel in patients with inoperable stage III non-small-cell lung cancer: The Hoosier Oncology Group and U.S. Oncology

J Clin Oncol

A five-gene signature and clinical outcome in non-small-cell lung cancer

N Engl J Med

Prediction of survival by [18F]fluorodeoxyglucose positron emission tomography in patients with locally advanced non-small-cell lung cancer undergoing definitive chemoradiation therapy: Results of the ACRIN 6668/RTOG 0235 trial

J Clin Oncol

A multigene assay is prognostic of survival in patients with early-stage lung adenocarcinoma

Clin Cancer Res

Texture analysis of aggressive and nonaggressive lung tumor CE CT images

IEEE Trans Biomed Eng

Texture analysis of non-small cell lung cancer on unenhanced computed tomography: Initial evidence for a relationship with tumour glucose metabolism and stage

Cancer Imaging

Tumor heterogeneity and permeability as measured on the ct component of PET/CT predict survival in patients with non-small cell lung cancer

Clin Cancer Res

Prognostic factors in stage III non-small cell lung cancer: A review of conventional, metabolic and new biological variables

Ther Adv Med Oncol

Textural features corresponding to textural properties

IEEE Trans Syst Man Cybern

Textural features for image classification

IEEE T Syst Man Cyb

Clinical Investigation
Prognostic Value and Reproducibility of Pretreatment CT Texture Features in Stage III Non-Small Cell Lung Cancer