Dylan Hunt

Intrinsic dependencies of CT radiomic features on voxel size and number of gray levels

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.

A dosimetric study of polyethylene glycol hydrogel in 200 prostate cancer patients treated with high-dose rate brachytherapy ± intensity modulated radiation therapy

BACKGROUND AND PURPOSE We sought to analyze the effect of polyethylene glycol (PEG) hydrogel on rectal doses in prostate cancer patients undergoing radiotherapy. MATERIALS AND METHODS Between July 2009 and April 2013, we treated 200 clinically localized prostate cancer patients with high-dose rate (HDR) brachytherapy±intensity modulated radiation therapy. Half of the patients received a transrectal ultrasound (TRUS)-guided transperineal injection of 10mL PEG hydrogel (DuraSeal™ Spinal Sealant System; Covidien, Mansfield, MA) in their anterior perirectal fat immediately prior to the first HDR brachytherapy treatment and 5mL PEG hydrogel prior to the second HDR brachytherapy treatment. Prostate, rectal, and bladder doses and prostate-rectal distances were calculated based upon treatment planning CT scans. RESULTS There was a success rate of 100% (100/100) with PEG hydrogel implantation. PEG hydrogel significantly increased the prostate-rectal separation (mean±SD, 12±4mm with gel vs. 4±2mm without gel, p<0.001) and significantly decreased the mean rectal D2 mL (47±9% with gel vs. 60±8% without gel, p<0.001). Gel decreased rectal doses regardless of body mass index (BMI). CONCLUSIONS PEG hydrogel temporarily displaced the rectum away from the prostate by an average of 12mm and led to a significant reduction in rectal radiation doses, regardless of BMI.

Motion as a perturbation: Measurement-guided dose estimates to moving patient voxels during modulated arc deliveries

PURPOSE To present a framework for measurement-guided VMAT dose reconstruction to moving patient voxels from a known motion kernel and the static phantom data, and to validate this perturbation-based approach with the proof-of-principle experiments. METHODS As described previously, the VMAT 3D dose to a static patient can be estimated by applying a phantom measurement-guided perturbation to the treatment planning system (TPS)-calculated dose grid. The fraction dose to any voxel in the presence of motion, assuming the motion kernel is known, can be derived in a similar fashion by applying a measurement-guided motion perturbation. The dose to the diodes in a helical phantom is recorded at 50 ms intervals and is transformed into a series of time-resolved high-density volumetric dose grids. A moving voxel is propagated through this 4D dose space and the fraction dose to that voxel in the phantom is accumulated. The ratio of this motion-perturbed, reconstructed dose to the TPS dose in the phantom serves as a perturbation factor, applied to the TPS fraction dose to the similarly situated voxel in the patient. This approach was validated by the ion chamber and film measurements on four phantoms of different shape and structure: homogeneous and inhomogeneous cylinders, a homogeneous cube, and an anthropomorphic thoracic phantom. A 2D motion stage was used to simulate the motion. The stage position was synchronized with the beam start time with the respiratory gating simulator. The motion patterns were designed such that the motion speed was in the upper range of the expected tumor motion (1-1.4 cm∕s) and the range exceeded the normally observed limits (up to 5.7 cm). The conformal arc plans for X or Y motion (in the IEC 61217 coordinate system) consisted of manually created narrow (3 cm) rectangular strips moving in-phase (tracking) or phase-shifted by 90° (crossing) with respect to the phantom motion. The XY motion was tested with the computer-derived VMAT MLC sequences. For all phantoms and plans, time-resolved (10 Hz) ion chamber dose was collected. In addition, coronal (XY) films were exposed in the cube phantom to a VMAT beam with two different starting phases, and compared to the reconstructed motion-perturbed dose planes. RESULTS For the X or Y motions with the moving strip and geometrical phantoms, the maximum difference between perturbation-reconstructed and ion chamber doses did not exceed 1.9%, and the average for any motion pattern∕starting phase did not exceed 1.3%. For the VMAT plans on the cubic and thoracic phantoms, one point exhibited a 3.5% error, while the remaining five were all within 1.1%. Across all the measurements (N = 22), the average disagreement was 0.5 ± 1.3% (1 SD). The films exhibited γ(3%∕3 mm) passing rates ≥90%. CONCLUSIONS The dose to an arbitrary moving voxel in a patient can be estimated with acceptable accuracy for a VMAT delivery, by performing a single QA measurement with a cylindrical phantom and applying two consecutive perturbations to the TPS-calculated patient dose. The first one accounts for the differences between the planned and delivered static doses, while the second one corrects for the motion.

Quantification and reduction of peripheral dose from leakage radiation on Siemens primus accelerators in electron therapy mode

In this work, leakage radiation from EA200 series electron applicators on Siemens Primus accelerators is quantified, and its penetration ability in water and/or the shielding material Xenolite‐NL established. Initially, measurement of leakage from 10×10−25×25 cm2 applicators was performed as a function of height along applicator and of lateral distance from applicator body. Relative to central‐axis ionization maximum in solid water, the maximum leakage in air observed with a cylindrical ion chamber with 1 cm solid water buildup cap at a lateral distance of 2 cm from the front and right sidewalls of applicators were 17% and 14%, respectively; these maxima were recorded for 18 MeV electron beams and applicator sizes of ≥20×20 cm2. In the patient plane, the applicator leakage gave rise to a broad peripheral dose off‐axis distance peak that shifted closer to the field edge as the electron energy increases. The maximum peripheral dose from normally incident primary electron beams at a depth of 1 cm in a water phantom was observed to be equal to 5% of the central‐axis dose maximum and as high as 9% for obliquely incident beams with angles of obliquity ≤ 40°. Measured depth‐peripheral dose curves showed that the “practical range” of the leakage electrons in water varies from approximately 1.4 to 5.7 cm as the primary electron beam energy is raised from 6 to 18 MeV. Next, transmission measurements of leakage radiation through the shielding material Xenolite‐NL showed a 4 mm thick sheet of this material is required to attenuate the leakage from 9 MeV beams by two‐thirds, and that for every additional 3 MeV increase in the primary electron beam energy, an additional Xenolite‐NL thickness of roughly 2 mm is needed to achieve the aforementioned attenuation level. Finally, attachment of a 1 mm thick sheet of lead to the outer surface of applicator sidewalls resulted in a reduction of the peripheral dose by up to 80% and 74% for 9 and 18 MeV beams, respectively. This sidewall modification had an insignificant effect on the clinical depth dose, cross‐axis beam profiles, and output factors. PACS numbers: 87.53.Bn, 87.56.bd, 87.56.J‐

Health-related quality-of-life changes due to high-dose-rate brachytherapy, low-dose-rate brachytherapy, or intensity-modulated radiation therapy for prostate cancer.

PURPOSE To compare urinary, bowel, and sexual health-related quality-of-life (HRQOL) changes due to high-dose-rate (HDR) brachytherapy, low-dose-rate (LDR) brachytherapy, or intensity-modulated radiation therapy (IMRT) monotherapy for prostate cancer. METHODS AND MATERIALS Between January 2002 and September 2013, 413 low-risk or favorable intermediate-risk prostate cancer patients were treated with HDR brachytherapy monotherapy to 2700-2800 cGy in two fractions (n = 85), iodine-125 LDR brachytherapy monotherapy to 14,500 cGy in one fraction (n = 249), or IMRT monotherapy to 7400-8100 cGy in 37-45 fractions (n = 79) without pelvic lymph node irradiation. No androgen deprivation therapy was given. Patients used an international prostate symptoms score questionnaire, an expanded prostate cancer index composite-26 bowel questionnaire, and a sexual health inventory for men questionnaire to assess their urinary, bowel, and sexual HRQOL, respectively, pretreatment and at 1, 3, 6, 9, 12, and 18 months posttreatment. RESULTS Median follow-up was 32 months. HDR brachytherapy and IMRT patients had significantly less deterioration in their urinary HRQOL than LDR brachytherapy patients at 1 and 3 months after irradiation. The only significant decrease in bowel HRQOL between the groups was seen 18 months after treatment, at which point IMRT patients had a slight, but significant, deterioration in their bowel HRQOL compared with HDR and LDR brachytherapy patients. HDR brachytherapy patients had worse sexual HRQOL than both LDR brachytherapy and IMRT patients after treatment. CONCLUSIONS IMRT and HDR brachytherapy cause less severe acute worsening of urinary HRQOL than LDR brachytherapy. However, IMRT causes a slight, but significant, worsening of bowel HRQOL compared with HDR and LDR brachytherapy.

Evaluating dosimetric accuracy of flattening filter free compensator-based IMRT: Measurements with diode arrays

PURPOSE Compensator-based IMRT coupled with the high dose rate flattening filter free (FFF) beams offers an intriguing possibility of delivering an intensity modulated radiation field in just a few seconds. As a first step, the authors evaluate the dosimetric accuracy of the treatment planning system (TPS) FFF beam model with compensators. METHODS A 6 MV FFF beam from a TrueBeam accelerator (Varian Medical Systems, Palo Alto CA) was modeled in PINNACLE TPS (v. 9.0, Philips Radiation Oncology, Fitchburg WI). Flat brass slabs from 0.3 to 7 cm thick and an 18° brass wedge were used to adjust the beam model. A 2D (MAPCHECK) and 3D (ARCCHECK) diode arrays (Sun Nuclear Corp, Melbourne FL), were investigated for use with the compensator FFF beams. Corrections for diode sensitivity caused by the spectral changes in the beam were introduced. Four compensator plans based on the AAPM TG-119 report were developed. A composite ion chamber measurement, beam by beam MAPCHECK measurements, and a composite ARCCHECK measurement were performed. The array results were analyzed with the same thresholds as in TG-119 report-3%/3 mm with global dose normalization-as well as with the more stringent combinations of the gamma analysis criteria. RESULTS The FFF beam shows a greater variation of the effective attenuation coefficient with brass thickness due to the prevalence of the low energy photons compared to the conventional 6X beam. As a result, a compromise had to be made while trying to achieve dose agreement for a combination of field sizes, brass thicknesses, and measurement depths (≥5 cm in water). An agreement of measured and calculated dose to within 1% was observed for brass thicknesses up to 2 cm. For the 3 cm slab, an error of up to 2.8% was noted for the field sizes above 10 × 10 cm(2), and up to 3.8% for the 5 × 5 cm(2) field. Both diode arrays exhibit a substantial sensitivity drop as the compensator thickness increases, reaching 10% for a 7 cm brass slab. A simple correction based on the brass thickness along the ray was introduced to counteract this effect. Pooled for five profiles, the average ratio of uncorrected and corrected MAPCHECK to ion chamber readings are 0.966 and 1.008, respectively. With the proper correction, all MAPCHECK measurement to calculation comparisons exhibit 100% γ(3%/3 mm) passing rates with global dose-error normalization. For the TG-119-type plans, the average γ(2%/2 mm) passing rate with local normalization is 94% (range 87.8%-98.3%). The lower ARCCHECK γ-analysis passing rates (corrected for diode sensitivity) are predictable based on the observed PDD discrepancies. However, with the 3%/3 mm thresholds and global normalization, the average γ-analysis passing rate is 96.4% (range 89.9%-100%). CONCLUSIONS MAPCHECK analysis demonstrates high passing rates with the stringent γ(2%/2 mm) and local normalization criteria combination. The geometry of the ARCCHECK array creates a stress test for the FFF TPS model because of the shallow depth of the entrance diodes and large air cavity. Hence, the ARCCHECK γ-analysis passing rates are lower than with the MAPCHECK, while still on par with TG-119.

Reproducibility of F18‐FDG PET radiomic features for different cervical tumor segmentation methods, gray‐level discretization, and reconstruction algorithms

Abstract Site‐specific investigations of the role of radiomics in cancer diagnosis and therapy are emerging. We evaluated the reproducibility of radiomic features extracted from 18Flourine–fluorodeoxyglucose (18F‐FDG) PET images for three parameters: manual versus computer‐aided segmentation methods, gray‐level discretization, and PET image reconstruction algorithms. Our cohort consisted of pretreatment PET/CT scans from 88 cervical cancer patients. Two board‐certified radiation oncologists manually segmented the metabolic tumor volume (MTV1 and MTV2) for each patient. For comparison, we used a graphical‐based method to generate semiautomated segmented volumes (GBSV). To address any perturbations in radiomic feature values, we down‐sampled the tumor volumes into three gray‐levels: 32, 64, and 128 from the original gray‐level of 256. Finally, we analyzed the effect on radiomic features on PET images of eight patients due to four PET 3D‐reconstruction algorithms: maximum likelihood‐ordered subset expectation maximization (OSEM) iterative reconstruction (IR) method, fourier rebinning‐ML‐OSEM (FOREIR), FORE‐filtered back projection (FOREFBP), and 3D‐Reprojection (3DRP) analytical method. We extracted 79 features from all segmentation method, gray‐levels of down‐sampled volumes, and PET reconstruction algorithms. The features were extracted using gray‐level co‐occurrence matrices (GLCM), gray‐level size zone matrices (GLSZM), gray‐level run‐length matrices (GLRLM), neighborhood gray‐tone difference matrices (NGTDM), shape‐based features (SF), and intensity histogram features (IHF). We computed the Dice coefficient between each MTV and GBSV to measure segmentation accuracy. Coefficient values close to one indicate high agreement, and values close to zero indicate low agreement. We evaluated the effect on radiomic features by calculating the mean percentage differences (d¯) between feature values measured from each pair of parameter elements (i.e. segmentation methods: MTV1‐MTV2, MTV1‐GBSV, MTV2‐GBSV; gray‐levels: 64‐32, 64‐128, and 64‐256; reconstruction algorithms: OSEM‐FORE‐OSEM, OSEM‐FOREFBP, and OSEM‐3DRP). We used |d¯| as a measure of radiomic feature reproducibility level, where any feature scored |d¯| ±SD ≤ |25|% ± 35% was considered reproducible. We used Bland–Altman analysis to evaluate the mean, standard deviation (SD), and upper/lower reproducibility limits (U/LRL) for radiomic features in response to variation in each testing parameter. Furthermore, we proposed U/LRL as a method to classify the level of reproducibility: High— ±1% ≤ U/LRL ≤ ±30%; Intermediate— ±30% < U/LRL ≤ ±45%; Low— ±45 < U/LRL ≤ ±50%. We considered any feature below the low level as nonreproducible (NR). Finally, we calculated the interclass correlation coefficient (ICC) to evaluate the reliability of radiomic feature measurements for each parameter. The segmented volumes of 65 patients (81.3%) scored Dice coefficient >0.75 for all three volumes. The result outcomes revealed a tendency of higher radiomic feature reproducibility among segmentation pair MTV1‐GBSV than MTV2‐GBSV, gray‐level pairs of 64‐32 and 64‐128 than 64‐256, and reconstruction algorithm pairs of OSEM‐FOREIR and OSEM‐FOREFBP than OSEM‐3DRP. Although the choice of cervical tumor segmentation method, gray‐level value, and reconstruction algorithm may affect radiomic features, some features were characterized by high reproducibility through all testing parameters. The number of radiomic features that showed insensitivity to variations in segmentation methods, gray‐level discretization, and reconstruction algorithms was 10 (13%), 4 (5%), and 1 (1%), respectively. These results suggest that a careful analysis of the effects of these parameters is essential prior to any radiomics clinical application.

High-dose-rate endorectal brachytherapy for locally advanced rectal cancer in previously irradiated patients

PURPOSE Preoperative high-dose-rate (HDR) endorectal brachytherapy is well tolerated among patients with locally advanced rectal cancer. However, these studies excluded patients who previously received pelvic radiation therapy (RT). Because a favorable toxicity profile has been published for HDR endorectal brachytherapy, we evaluated this technique in patients who have previously received pelvic irradiation. METHODS AND MATERIALS We included patients who had received pelvic irradiation for a previous pelvic malignancy and later received preoperative HDR endorectal brachytherapy for rectal cancer. Brachytherapy was delivered to a total dose of 26 Gy in 4 consecutive daily 6.5 Gy fractions. RESULTS We evaluated 10 patients who previously received pelvic external beam radiation therapy (EBRT) alone (n=6), EBRT and brachytherapy (n=2), or brachytherapy alone (n=2). The median interval between the initial course of RT and endorectal brachytherapy was approximately 11 years (range, 1-19 years). Two patients experienced a complete pathologic response while 1 patient had a near complete pathologic response. No acute grade ≥3 toxicity was observed. No intraoperative or postoperative surgical complications were observed. CONCLUSIONS Preoperative HDR endorectal brachytherapy is an alternative to EBRT for patients with locally advanced rectal cancer who have previously received pelvic RT.

Lipiodol as a Fiducial Marker for Image-Guided Radiation Therapy for Bladder Cancer

PURPOSE To evaluate Lipiodol as a liquid, radio-opaque fiducial marker for image-guided radiation therapy (IGRT) for bladder cancer. MATERIALS AND METHODS Between 2011 and 2012, 5 clinical T2a-T3b N0 M0 stage II-III bladder cancer patients were treated with maximal transurethral resection of a bladder tumor (TURBT) and image-guided radiation therapy (IGRT) to 64.8 Gy in 36 fractions ± concurrent weekly cisplatin-based or gemcitabine chemotherapy. Ten to 15mL Lipiodol, using 0.5mL per injection, was injected into bladder submucosa circumferentially around the entire periphery of the tumor bed immediately following maximal TURBT. The authors looked at inter-observer variability regarding the size and location of the tumor bed (CTVboost) on computed tomography scans with versus without Lipiodol. RESULTS Median follow-up was 18 months. Lipiodol was visible on every orthogonal two-dimensional kV portal image throughout the entire, 7-week course of IGRT. There was a trend towards improved inter-observer agreement on the CTVboost with Lipiodol (p = 0.06). In 2 of 5 patients, the tumor bed based upon Lipiodol extended outside a planning target volume that would have been treated with a radiation boost based upon a cystoscopy report and an enhanced computed tomography (CT) scan for staging. There was no toxicity attributable to Lipiodol. CONCLUSIONS Lipiodol constitutes a safe and effective fiducial marker that an urologist can use to demarcate a tumor bed immediately following maximal TURBT. Lipiodol decreases inter-observer variability in the definition of the extent and location of a tumor bed on a treatment planning CT scan for a radiation boost.

Sensitivity of Image Features to Noise in Conventional and Respiratory-Gated PET/CT Images of Lung Cancer: Uncorrelated Noise Effects.

The effect of noise on image features has yet to be studied in depth. Our objective was to explore how significantly image features are affected by the addition of uncorrelated noise to an image. The signal-to-noise ratio and noise power spectrum were calculated for a positron emission tomography/computed tomography scanner using a Ge-68 phantom. The conventional and respiratory-gated positron emission tomography/computed tomography images of 31 patients with lung cancer were retrospectively examined. Multiple sets of noise images were created for each original image by adding Gaussian noise of varying standard deviation equal to 2.5%, 4.0%, and 6.0% of the maximum intensity for positron emission tomography images and 10, 20, 50, 80, and 120 Hounsfield units for computed tomography images. Image features were extracted from all images, and percentage differences between the original image and the noise image feature values were calculated. These features were then categorized according to the noise sensitivity. The contour-dependent shape descriptors averaged below 4% difference in positron emission tomography and below 13% difference in computed tomography between noise and original images. Gray level size zone matrix features were the most sensitive to uncorrelated noise exhibiting average differences >200% for conventional and respiratory-gated images in computed tomography and 90% in positron emission tomography. Image feature differences increased as the noise level increased for shape, intensity, and gray-level co-occurrence matrix features in positron emission tomography and for gray-level co-occurrence matrix and gray-level size zone matrix features in conventional computed tomography. Investigators should be aware of the noise effects on image features.