Radiomic features are potential imaging biomarkers for therapy response assessment in oncology. However, the robustness of features with respect to imaging parameters is not well established. ...Previously identified potential imaging biomarkers were found to be intrinsically dependent on voxel size and number of gray levels (GLs) in a recent texture phantom investigation. Here, we validate the voxel size and GL in-phantom normalizations in lung tumors. Eighteen patients with non-small cell lung cancer of varying tumor volumes were analyzed. To compare with patient data, phantom scans were acquired on eight different scanners. Twenty four previously identified features were extracted from lung tumors. The Spearman rank (r
) and interclass correlation coefficient (ICC) were used as metrics. Eight out of 10 features showed high (r
> 0.9) and low (r
< 0.5) correlations with number of voxels before and after normalizations, respectively. Likewise, texture features were unstable (ICC < 0.6) and highly stable (ICC > 0.8) before and after GL normalizations, respectively. We conclude that voxel size and GL normalizations derived from a texture phantom study also apply to lung tumors. This study highlights the importance and utility of investigating the robustness of radiomic features with respect to CT imaging parameters in radiomic phantoms.
Purpose
Many radiomics features were originally developed for non‐medical imaging applications and therefore original assumptions may need to be reexamined. In this study, we investigated the impact ...of slice thickness and pixel spacing (or pixel size) on radiomics features extracted from Computed Tomography (CT) phantom images acquired with different scanners as well as different acquisition and reconstruction parameters. The dependence of CT texture features on gray‐level discretization was also evaluated.
Methods and materials
A texture phantom composed of 10 different cartridges of different materials was scanned on eight different CT scanners from three different manufacturers. The images were reconstructed for various slice thicknesses. For each slice thickness, the reconstruction Field Of View (FOV) was varied to render pixel sizes ranging from 0.39 to 0.98 mm. A fixed spherical region of interest (ROI) was contoured on the images of the shredded rubber cartridge and the 3D printed, 20% fill, acrylonitrile butadiene styrene plastic cartridge (ABS20) for all phantom imaging sets. Radiomic features were extracted from the ROIs using an in‐house program. Features categories were: shape (10), intensity (16), GLCM (24), GLZSM (11), GLRLM (11), and NGTDM (5), fractal dimensions (8) and first‐order wavelets (128), for a total of 213 features. Voxel‐size resampling was performed to investigate the usefulness of extracting features using a suitably chosen voxel size. Acquired phantom image sets were resampled to a voxel size of 1 × 1 × 2 mm3 using linear interpolation. Image features were therefore extracted from resampled and original datasets and the absolute value of the percent coefficient of variation (%COV) for each feature was calculated. Based on the %COV values, features were classified in 3 groups: (1) features with large variations before and after resampling (%COV >50); (2) features with diminished variation (%COV <30) after resampling; and (3) features that had originally moderate variation (%COV <50%) and were negligibly affected by resampling. Group 2 features were further studied by modifying feature definitions to include voxel size. Original and voxel‐size normalized features were used for interscanner comparisons. A subsequent analysis investigated feature dependency on gray‐level discretization by extracting 51 texture features from ROIs from each of the 10 different phantom cartridges using 16, 32, 64, 128, and 256 gray levels.
Results
Out of the 213 features extracted, 150 were reproducible across voxel sizes, 42 improved significantly (%COV <30, Group 2) after resampling, and 21 had large variations before and after resampling (Group 1). Ten features improved significantly after definition modification effectively removed their voxel‐size dependency. Interscanner comparison indicated that feature variability among scanners nearly vanished for 8 of these 10 features. Furthermore, 17 out of 51 texture features were found to be dependent on the number of gray levels. These features were redefined to include the number of gray levels which greatly reduced this dependency.
Conclusion
Voxel‐size resampling is an appropriate pre‐processing step for image datasets acquired with variable voxel sizes to obtain more reproducible CT features. We found that some of the radiomics features were voxel size and gray‐level discretization‐dependent. The introduction of normalizing factors in their definitions greatly reduced or removed these dependencies.
Even with advanced inverse‐planning techniques, radiation treatment plan optimization remains a very time‐consuming task with great output variability, which prompted the development of more ...automated approaches. One commercially available technique mimics the actions of experienced human operators to progressively guide the traditional optimization process with automatically created regions of interest and associated dose‐volume objectives. We report on the initial evaluation of this algorithm on 10 challenging cases of locoreginally advanced head and neck cancer. All patients were treated with VMAT to 70 Gy to the gross disease and 56 Gy to the elective bilateral nodes. The results of post‐treatment autoplanning (AP) were compared to the original human‐driven plans (HDP). We used an objective scoring system based on defining a collection of specific dosimetric metrics and corresponding numeric score functions for each. Five AP techniques with different input dose goals were applied to all patients. The best of them averaged the composite score 8% lower than the HDP, across the patient population. The difference in median values was statistically significant at the 95% confidence level (Wilcoxon paired signed‐rank test p=0.027). This result reflects the premium the institution places on dose homogeneity, which was consistently higher with the HDPs. The OAR sparing was consistently better with the APs, the differences reaching statistical significance for the mean doses to the parotid glands (p<0.001) and the inferior pharyngeal constrictor (p=0.016), as well as for the maximum doses to the spinal cord (p=0.018) and brainstem (p=0.040). If one is prepared to accept less stringent dose homogeneity criteria from the RTOG 1016 protocol, nine APs would comply with the protocol, while providing lower OAR doses than the HDPs. Overall, AP is a promising clinical tool, but it could benefit from a better process for shifting the balance between the target dose coverage/homogeneity and OAR sparing.
PACS number(s): 87.55.D
Pencil beam (PB) and collapsed cone convolution (CCC) dose calculation algorithms differ significantly when used in the thorax. However, such differences have seldom been previously directly ...correlated with outcomes of lung stereotactic ablative body radiation (SABR).
Data for 201 non-small cell lung cancer patients treated with SABR were analyzed retrospectively. All patients were treated with 50 Gy in 5 fractions of 10 Gy each. The radiation prescription mandated that 95% of the planning target volume (PTV) receive the prescribed dose. One hundred sixteen patients were planned with BrainLab treatment planning software (TPS) with the PB algorithm and treated on a Novalis unit. The other 85 were planned on the Pinnacle TPS with the CCC algorithm and treated on a Varian linac. Treatment planning objectives were numerically identical for both groups. The median follow-up times were 24 and 17 months for the PB and CCC groups, respectively. The primary endpoint was local/marginal control of the irradiated lesion. Gray's competing risk method was used to determine the statistical differences in local/marginal control rates between the PB and CCC groups.
Twenty-five patients planned with PB and 4 patients planned with the CCC algorithms to the same nominal doses experienced local recurrence. There was a statistically significant difference in recurrence rates between the PB and CCC groups (hazard ratio 3.4 95% confidence interval: 1.18-9.83, Gray's test P=.019). The differences (Δ) between the 2 algorithms for target coverage were as follows: ΔD99GITV = 7.4 Gy, ΔD99PTV = 10.4 Gy, ΔV90GITV = 13.7%, ΔV90PTV = 37.6%, ΔD95PTV = 9.8 Gy, and ΔDISO = 3.4 Gy. GITV = gross internal tumor volume.
Local control in patients receiving who were planned to the same nominal dose with PB and CCC algorithms were statistically significantly different. Possible alternative explanations are described in the report, although they are not thought likely to explain the difference. We conclude that the difference is due to relative dosimetric underdosing of tumors with the PB algorithm.
Predicting recurrence after stereotactic body radiotherapy (SBRT) in non-small cell lung cancer (NSCLC) patients is problematic, but critical for the decision of following treatment. This study aims ...to investigate the association of imaging features derived from the first follow-up computed tomography (CT) on lung cancer patient outcomes following SBRT, and identify patients at high risk of recurrence.
Fifty nine biopsy-proven non-small cell lung cancer patients were qualified for this study. The first follow-up CTs were performed about 3 months after SBRT (median time: 91 days). Imaging features included 34 manually scored radiological features (semantics) describing the lesion, lung and thorax and 219 quantitative imaging features (radiomics) extracted automatically after delineation of the lesion. Cox proportional hazard models and Harrel's C-index were used to explore predictors of overall survival (OS), recurrence-free survival (RFS), and loco-regional recurrence-free survival (LR-RFS). Five-fold cross validation was performed on the final prognostic model.
The median follow-up time was 42 months. The model for OS contained Eastern Cooperative Oncology Group (ECOG) performance status (HR = 3.13, 95% CI: 1.17-8.41), vascular involvement (HR = 3.21, 95% CI: 1.29-8.03), lymphadenopathy (HR = 3.59, 95% CI: 1.58-8.16) and the 1st principle component of radiomic features (HR = 1.24, 95% CI: 1.02-1.51). The model for RFS contained vascular involvement (HR = 3.06, 95% CI: 1.40-6.70), vessel attachment (HR = 3.46, 95% CI: 1.65-7.25), pleural retraction (HR = 3.24, 95% CI: 1.41-7.42), lymphadenopathy (HR = 6.41, 95% CI: 2.58-15.90) and relative enhancement (HR = 1.40, 95% CI: 1.00-1.96). The model for LR-RFS contained vascular involvement (HR = 4.96, 95% CI: 2.23-11.03), lymphadenopathy (HR = 2.64, 95% CI: 1.19-5.82), circularity (F13, HR = 1.60, 95% CI: 1.10-2.32) and 3D Laws feature (F92, HR = 1.96, 95% CI: 1.35-2.83). Five-fold cross-validated the areas under the receiver operating characteristic curves (AUC) of these three models were all above 0.8.
Our analysis reveals disease progression could be prognosticated as early as 3 months after SBRT using CT imaging features, and these features would be helpful in clinical decision-making.
To assure accurate treatment delivery on any image-guided radiotherapy system, the relative positions and walkout of the imaging and radiation isocenters must be periodically verified and kept within ...specified tolerances. In this work, we first validated the multiaxis ion chamber array as a tool for finding the radiation isocenter position of a magnetic resonance–guided linear accelerator. The treatment couch with the array on it was shifted in 0.2-mm increments and the reported beam center position was plotted against that shift and fitted to a straight line, in both X and Y directions. From the goodness-of-fit and intercepts of the regression lines, the accuracy and precision were conservatively estimated at 0.2 and 0.1 mm, respectively. This holds true whether the array is irradiated from the front or from the back, which allows efficient collecting the data from the 4 cardinal gantry angles with just 2 array positions. The average isocenter position agreed to within at most 0.4 mm along any cardinal axis with the linac vendor’s film-based procedure, and the maximum walkout radii were 0.32 mm and 0.53 mm, respectively. The magnetic resonance imaging isocenter walkout as a function of gantry angle was studied with 2 different phantoms, one employing a single fiducial at the center and another extracting the rigid displacement values from the distortion map fit of 523 fiducials dispersed over a large volume. The results were close between the 2 phantoms and demonstrated variation in the magnetic resonance imaging isocenter location as high as 1.3 mm along a single axis in the transverse plane. Verification of the magnetic resonance imaging isocenter location versus the gantry angle should be a part of quality assurance for magnetic resonance-guided linear accelerators.
The treatment of central and ultracentral lung tumors with radiotherapy remains an ongoing clinical challenge. The risk of Grade 5 toxicity with ablative radiotherapy doses to these high-risk regions ...is significant as shown in recent prospective studies. Magnetic resonance (MR) image-guided adaptive radiotherapy (MRgART) is a new technology and may allow the delivery of ablative radiotherapy to these high-risk regions safely. MRgART is able to achieve this by utilizing small treatment margins, real-time gating/tracking and on-table plan adaptation to maintain dose to the tumor but limit dose to critical structures. The process of MRgART is complex and has nuances and challenges for the treatment of lung tumors. We outline the critical steps needed for appropriate delivery of MRgART for lung tumors safely and effectively.
Although altered protocols that challenge conventional radiation fractionation have been tested in prospective clinical trials, we still have limited understanding of how to select the most ...appropriate fractionation schedule for individual patients. Currently, the prescription of definitive radiotherapy is based on the primary site and stage, without regard to patient-specific tumor or host factors that may influence outcome. We hypothesize that the proportion of radiosensitive proliferating cells is dependent on the saturation of the tumor carrying capacity. This may serve as a prognostic factor for personalized radiotherapy (RT) fractionation.
We introduce a proliferation saturation index (PSI), which is defined as the ratio of tumor volume to the host-influenced tumor carrying capacity. Carrying capacity is as a conceptual measure of the maximum volume that can be supported by the current tumor environment including oxygen and nutrient availability, immune surveillance and acidity. PSI is estimated from two temporally separated routine pre-radiotherapy computed tomography scans and a deterministic logistic tumor growth model. We introduce the patient-specific pre-treatment PSI into a model of tumor growth and radiotherapy response, and fit the model to retrospective data of four non-small cell lung cancer patients treated exclusively with standard fractionation. We then simulate both a clinical trial hyperfractionation protocol and daily fractionations, with equal biologically effective dose, to compare tumor volume reduction as a function of pretreatment PSI.
With tumor doubling time and radiosensitivity assumed constant across patients, a patient-specific pretreatment PSI is sufficient to fit individual patient response data (R(2) = 0.98). PSI varies greatly between patients (coefficient of variation >128 %) and correlates inversely with radiotherapy response. For this study, our simulations suggest that only patients with intermediate PSI (0.45-0.9) are likely to truly benefit from hyperfractionation. For up to 20 % uncertainties in tumor growth rate, radiosensitivity, and noise in radiological data, the absolute estimation error of pretreatment PSI is <10 % for more than 75 % of patients.
Routine radiological images can be used to calculate individual PSI, which may serve as a prognostic factor for radiation response. This provides a new paradigm and rationale to select personalized RT dose-fractionation.
Abstract Radiomics is being explored for potential applications in radiation therapy. How various imaging protocols affect quantitative image features is currently a highly active area of research. ...To assess the variability of image features derived from conventional three-dimensional (3D) and respiratory-gated (RG) positron emission tomography (PET)/computed tomography (CT) images of lung cancer patients, image features were computed from 23 lung cancer patients. Both protocols for each patient were acquired during the same imaging session. PET tumor volumes were segmented using an adaptive technique which accounted for background. CT tumor volumes were delineated with a commercial segmentation tool. Using RG PET images, the tumor center of mass motion, length, and rotation were calculated. Fifty-six image features were extracted from all images consisting of shape descriptors, first-order features, and second-order texture features. Overall, 26.6% and 26.2% of total features demonstrated less than 5% difference between 3D and RG protocols for CT and PET, respectively. Between 10 RG phases in PET, 53.4% of features demonstrated percent differences less than 5%. The features with least variability for PET were sphericity, spherical disproportion, entropy (first and second order), sum entropy, information measure of correlation 2, Short Run Emphasis (SRE), Long Run Emphasis (LRE), and Run Percentage (RPC); and those for CT were minimum intensity, mean intensity, Root Mean Square (RMS), Short Run Emphasis (SRE), and RPC. Quantitative analysis using a 3D acquisition versus RG acquisition (to reduce the effects of motion) provided notably different image feature values. This study suggests that the variability between 3D and RG features is mainly due to the impact of respiratory motion.
Colorectal cancer is the third most common cancer in men and the second most common in women worldwide, and the incidence is increasing among younger patients. 30% of these malignancies arise in the ...rectum. Patients with rectal cancer have historically been managed with preoperative radiation, followed by radical surgery, and adjuvant chemotherapy, with permanent colostomies in up to 20% of patients. Beginning in the early 2000s, non-operative management (NOM) of rectal cancer emerged as a viable alternative to radical surgery in select patients. Efforts have been ongoing to optimize neoadjuvant therapy for rectal cancer, thereby increasing the number of patients potentially eligible to forgo radical surgery. Magnetic resonance guided radiotherapy (MRgRT) has recently emerged as a treatment modality capable of intensifying preoperative radiation therapy for rectal cancer patients. This technology may also predict which patients will achieve a complete response to preoperative therapy, thereby allowing for more appropriate selection of patients for NOM. The present work seeks to illustrate the potential role MRgRT could play in personalizing rectal cancer treatment thus expanding the role of NOM in rectal cancer.