Single-energy metal artefact reduction with CT for carbon-ion radiation therapy treatment planning

Commissioning and Validation of CT Number to SPR Calibration in Carbon Ion Therapy Facility

Purpose

We performed computed tomography (CT)-stopping power ratio (SPR) calibration in a carbon-ion therapy facility and evaluated SPR estimation accuracy.

Materials and Methods

A polybinary tissue model method was used for the calibration of CT numbers and SPR. As a verification by dose calculation, we created a virtual phantom to which the CT-SPR calibration table was applied. Then, SPR was calculated from the change in the range of the treatment planning beam when changing to 19 different CT numbers, and the accuracy of the treatment planning system (TPS) calculation of SPR values from the CT-SPR calibration table was validated. As a verification by measurement, 5 materials (water, milk, olive oil, ethanol, 40% K2HPO4) were placed in a container, and the SPR was obtained by measurement from the change in the range of the beam that passed through the materials.

Results

The results of the dose calculations of the TPS showed that the results agreed within 1% for the lower CT numbers up to 1000 HU, but there was a difference of 3.0% in the higher CT number volume. The difference between the SPR calculated by TPS and the SPR caused by the difference in the energy of the incident particles agreed within 0.51%. The accuracy of SPR estimation was measured, and the error was within 2% for all materials tested.

Conclusion

These results indicate that the SPR estimation errors are within the range of errors that can be expected in particle therapy. From commissioning and verification results, the CT-SPR calibration table obtained during this commissioning process is clinically applicable.

Radiotherapy planning of spine and pelvis using single-energy metal artifact reduction corrected computed tomography sets

Metal artifacts produce incorrect Hounsfield units and impact treatment planning accuracy. This work evaluates the use of single-energy metal artifact reduction (SEMAR) algorithm for treatment planning by comparison to manual artifact overriding. CT datasets of in-house 3D-printed spine and pelvic phantoms with and without metal insert(s) and two treated patients with metal implants were analysed. CT number accuracy improved with the use of SEMAR filter: root mean square deviation (RMSD) from reference (without metal) reduced by 35.4 in spine and 98.8 in hip. The plan dose volume histograms (DVHs) and dosimetric measurements showed comparable results. SEMAR reconstruction improved planning efficiency.

Radiomics analysis of [18F]-fluoro-2-deoxyglucose positron emission tomography for the prediction of cervical lymph node metastasis in tongue squamous cell carcinoma

OBJECTIVES

This study aimed to create a predictive model for cervical lymph node metastasis (CLNM) in patients with tongue squamous cell carcinoma (SCC) based on radiomics features detected by [¹⁸F]-fluoro-2-deoxyglucose (¹⁸F-FDG) positron emission tomography (PET).

METHODS

A total of 40 patients with tongue SCC who underwent ¹⁸F-FDG PET imaging during their first medical examination were enrolled. During the follow-up period (mean 28 months), 20 patients had CLNM, including six with late CLNM, whereas the remaining 20 patients did not have CLNM. Radiomics features were extracted from ¹⁸F-FDG PET images of all patients irrespective of metal artifact, and clinicopathological factors were obtained from the medical records. Late CLNM was defined as the CLNM that occurred after major treatment. The least absolute shrinkage and selection operator (LASSO) model was used for radiomics feature selection and sequential data fitting. The receiver operating characteristic curve analysis was used to assess the predictive performance of the ¹⁸F-FDG PET-based model and clinicopathological factors model (CFM) for CLNM.

RESULTS

Six radiomics features were selected from LASSO analysis. The average values of the area under the curve (AUC), accuracy, sensitivity, and specificity of radiomics analysis for predicting CLNM from ¹⁸F-FDG PET images were 0.79, 0.68, 0.65, and 0.70, respectively. In contrast, those of the CFM were 0.54, 0.60, 0.60, and 0.60, respectively. The ¹⁸F-FDG PET-based model showed significantly higher AUC than that of the CFM.

CONCLUSIONS

The ¹⁸F-FDG PET-based model has better potential for diagnosing CLNM and predicting late CLNM in patients with tongue SCC than the CFM.

Geometric and dosimetric impact of 3D generative adversarial network-based metal artifact reduction algorithm on VMAT and IMPT for the head and neck region

Background

We investigated the geometric and dosimetric impact of three-dimensional (3D) generative adversarial network (GAN)-based metal artifact reduction (MAR) algorithms on volumetric-modulated arc therapy (VMAT) and intensity-modulated proton therapy (IMPT) for the head and neck region, based on artifact-free computed tomography (CT) volumes with dental fillings.

Methods

Thirteen metal-free CT volumes of the head and neck regions were obtained from The Cancer Imaging Archive. To simulate metal artifacts on CT volumes, we defined 3D regions of the teeth for pseudo-dental fillings from the metal-free CT volumes. HU values of 4000 HU were assigned to the selected teeth region of interest. Two different CT volumes, one with four (m4) and the other with eight (m8) pseudo-dental fillings, were generated for each case. These CT volumes were used as the Reference. CT volumes with metal artifacts were then generated from the Reference CT volumes (Artifacts). On the Artifacts CT volumes, metal artifacts were manually corrected for using the water density override method with a value of 1.0 g/cm³ (Water). By contrast, the CT volumes with reduced metal artifacts using 3D GAN model extension of CycleGAN were also generated (GAN-MAR). The structural similarity (SSIM) index within the planning target volume was calculated as quantitative error metric between the Reference CT volumes and the other volumes. After creating VMAT and IMPT plans on the Reference CT volumes, the reference plans were recalculated for the remaining CT volumes.

Results

The time required to generate a single GAN-MAR CT volume was approximately 30 s. The median SSIMs were lower in the m8 group than those in the m4 group, and ANOVA showed a significant difference in the SSIM for the m8 group (p < 0.05). Although the median differences in D_98%, D_50% and D_2% were larger in the m8 group than the m4 group, those from the reference plans were within 3% for VMAT and 1% for IMPT.

Conclusions

The GAN-MAR CT volumes generated in a short time were closer to the Reference CT volumes than the Water and Artifacts CT volumes. The observed dosimetric differences compared to the reference plan were clinically acceptable.

The application of metal artifact reduction methods on computed tomography scans for radiotherapy applications: A literature review

Metal artifact reduction (MAR) methods are used to reduce artifacts from metals or metal components in computed tomography (CT). In radiotherapy (RT), CT is the most used imaging modality for planning, whose quality is often affected by metal artifacts. The aim of this study is to systematically review the impact of MAR methods on CT Hounsfield Unit values, contouring of regions of interest, and dose calculation for RT applications. This systematic review is performed in accordance with the PRISMA guidelines; the PubMed and Web of Science databases were searched using the main keywords "metal artifact reduction", "computed tomography" and "radiotherapy". A total of 382 publications were identified, of which 40 (including one review article) met the inclusion criteria and were included in this review. The selected publications (except for the review article) were grouped into two main categories: commercial MAR methods and research-based MAR methods. Conclusion: The application of MAR methods on CT scans can improve treatment planning quality in RT. However, none of the investigated or proposed MAR methods was completely satisfactory for RT applications because of limitations such as the introduction of other errors (e.g., other artifacts) or image quality degradation (e.g., blurring), and further research is still necessary to overcome these challenges.

Effects of metal implants and a metal artifact reduction tool on calculation accuracy of AAA and Acuros XB algorithms in small fields

PURPOSE

In this study, the dosimetric accuracy of analytical anisotropic algorithm (AAA) and Acuros XB (AXB) dose calculation algorithms (Varian Medical Systems, Palo Alto, CA) was investigated for small radiation fields incident on phantoms of various metals that include stainless steel grade 316L (SS316L) and titanium alloy grade 5 (Ti5) implants. In addition, the effects of using metal artifact reduction for orthopedic implants (O-MAR, Philips Healthcare, Cleveland, OH) were evaluated.

METHODS

The evaluations of AAA and AXB were performed by comparing the crossline profiles calculated by AAA and AXB with GafChromic_TM EBT3 film measurements at the phantom-implant interfaces and in close vicinity of implant materials for small field sizes (1 × 1 cm² , 2 × 2 cm² , 3 × 3 cm² , and 4 × 4 cm² ) of a 6 MV flattening filter free photon beam. O-MAR corrected and uncorrected (UC) computed tomography (CT) images were used for dose calculations. The values of average and standard deviations (SD) of Hounsfield unit (HU) for selected regions of each case were evaluated. The differences in average dose percentages in defined regions were calculated to quantify the relative dosimetric changes between doses calculated on UC and O-MAR corrected CT images.

RESULTS

Compared to UC images, the values of SD were reduced, and the average HU became closer to its reference value in the O-MAR images. There was some discrepancy in average dose percentage differences between calculations using UC and O-MAR images at 1 cm above the SS316L implant (average dose percentage differences were AXB/UC = 5.9% and AXB/O-MAR = -1.2%; AAA/UC = 2.2%, and AAA/O-MAR = -0.8%). Neither AAA nor AXB algorithms predict increase in dose at upper phantom-implant interface (4.9%, 9.9%. 13.5%, and 13.8% for the fields from 1 × 1 cm² to 4 × 4 cm² , respectively). At the side of the SS316L implant (where dark streak artifacts exist), dose difference averages were estimated as - 1.1% and 22.3% when AXB/O-MAR and AXB/UC calculations are compared with EBT3 measurements, respectively. Dose predictions at 1 cm below the SS316L implant were underestimated by AXB/O-MAR (average -0.5%) and AXB/UC (average 2.0%).

CONCLUSIONS

The O-MAR tool was shown to have a favorable dosimetric effect or no effect on the calculations in the upper proximity of the implant materials. The dose differences between EBT3 film measurements and calculations at upper phantom-implant interfaces were smaller when they were calculated using O-MAR images. However, the dose differences increased when O-MAR corrected images were used for AAA calculations at lower phantom-implant interfaces. Use of O-MAR enabled closer agreement for the AXB algorithm, especially in the dark streak artifact regions. The O-MAR algorithm should be used when the dose is calculated with the AXB algorithm in cases of patients with the metal implants. The estimations using AAA and AXB algorithms, in phantom setups, with Ti5 implant material were found to be closer to the EBT3 film measurements, when compared with the same estimations using SS316L implant material.

Evaluation of raw-data-based and calculated electron density for contrast media with a dual-energy CT technique

Objectives

The aim of the current study is to evaluate the accuracy and the precision of raw-data-based relative electron density (RED_raw) and the calibration-based RED (RED_cal) at a range of low-RED to high-RED for tissue-equivalent phantom materials by comparing them with reference RED (RED_ref) and to present the difference of RED_raw and RED_cal for the contrast medium using dual-energy CT (DECT).

Methods

The RED_raw images were reconstructed by raw-data-based decomposition using DECT. For evaluation of the accuracy of the RED_raw, RED_ref was calculated for the tissue-equivalent phantom materials based on their specified density and elemental composition. The RED_cal images were calculated using three models: Lung-Bone model, Lung-Ti model and Lung-Ti (SEMAR) model which used single-energy metal artifact reduction (SEMAR). The difference between RED_raw and RED_cal was calculated.

Results

In the titanium rod core, the deviations of RED_raw and RED_cal (Lung-Bone model, Lung-Ti model and Lung-Ti model with SEMAR) from RED_ref were 0.45%, 50.8%, 15.4% and 15.0%, respectively. The largest differences between RED_raw and RED_cal (Lung-Bone model, Lung-Ti model and Lung-Ti model with SEMAR) in the contrast medium phantom were 8.2%, -23.7%, and 28.7%, respectively. However, the differences between RED_raw and RED_cal values were within 10% at 20 mg/ml. The standard deviation of the RED_raw was significantly smaller than the RED_cal with three models in the titanium and the materials that had low CT numbers.

Conclusion

The RED_cal values could be affected by beam hardening artifacts and the RED_cal was less accurate than RED_raw for high-Z materials as titanium.

Advances in knowledge

The raw-data-based reconstruction method could reduce the beam hardening artifact compared with image-based reconstruction and increase the accuracy for the RED estimation in high-Z materials, such as titanium and iodinated contrast medium.

Metallic dental artifact reduction in computed tomography (Smart MAR): Improvement of image quality and diagnostic confidence in patients with suspected head and neck pathology and oral implants

PURPOSE

We determined whether the Smart MAR metal artifact reduction tool - a three-stage, projection-based, post processing algorithm - improves subjective and objective image quality and diagnostic confidence in patients with dental artifacts and suspected head and neck pathology compared to standard adaptive statistical iterative reconstructions (ASIR V) alone.

METHOD

The study included 100 consecutive patients with nonremovable oral implants or dental fillings and suspected oropharyngeal cancer or abscess. CT raw data of a single-source multislice CT scanner were postprocessed using ASIR V alone and with additional Smart MAR reconstruction. Image quality of baseline ASIR V and Smart MAR-based reconstruction series was compared both quantitatively (5 regions of interest, ROIs) and qualitatively (two independent raters).

RESULTS

Additional Smart MAR reconstruction significantly seems to improve both attenuation and noise adjacent to implants and in more distant areas (all p < 0.001) compared to standard ASIR V reconstructions alone. Signal-to-noise ratio (SNR; p = 0.001) and contrast-to-noise ratio were improved significantly (CNR; p = 0.001). Smart MAR improved visualization of tumor/abscess (detected in 36 of 100 patients, 36%) and representative oropharyngeal tissue (p < 0.001). In 8 of 36 patients (22%), tumor was only detected in Smart MAR series. Mean total DLP was 506.8mGy*cm; average CTDIvol was 5.5 mGy.

CONCLUSIONS

The supplementary use of the Smart MAR post-processing tool seems to significantly improve both subjective and objective image quality as well as diagnostic confidence and lesion detection in CT of the head and neck. In 22% of cases, the tumor was detected only in Smart MAR reconstructed images.

Metal artifact reduction techniques for single energy CT and dual-energy CT with various metal materials

Objective:

The aim of the current study is to evaluate the effectiveness of reduction metal artifacts using kV-CT image with the single-energy based metal artefact reduction (SEMAR) technique by single-energy reconstruction, monochromatic CT and rED reconstructed by dual-energy reconstruction.

Methods:

Seven different metal materials (brass, aluminum, copper, stainless, steel, lead and titanium) were placed inside the water-based PMMA phantom. After DECT-based scan, the artefact index (AI) were evaluated with the kV-CT images with and without SEMAR by single-energy reconstruction, and raw-data based electron density (rED), monochromatic CT images by dual-energy reconstruction. Moreover, the AI with evaluated with rED and the converted ED images from the kV-CT and monochromatic CT images.

Results:

The minimum average value of the AI with all-metal inserts was approximately 80 keV. The AI without SEMAR was larger than that with SEMAR for the 80 kV and 135 kV CT images. In the comparison of the AI for the rED and ED images that were converted from 80 kV and 135 kV CT images with and without SEMAR, the monochromatic CT images of the PMMA phantom with inserted metal materials at 80 keV revealed that the kV-CT with SEMAR reduced the metal artefact substantially.

Conclusion:

The converted ED from the kV-CT and monochromatic CT images could be useful for a comparison of the AI using the same contrast scale. The kV-CT image with SEMAR by single-energy reconstruction was found to substantially reduce metal artefact.

Advances in knowledge:

The effectiveness of reduction of metal artifacts using single-energy based metal artefact reduction (SEMAR) technique and dual-energy CT (DECT) was evaluated the electron density conversion techniques.

Dosimetric assessment of a single-energy metal artifact reduction algorithm for computed tomography images in radiation therapy

This study aimed to evaluate the performance of a single-energy metal artifact reduction (SEMAR) algorithm for radiation therapy treatment using phantom cases with metal inserts, assess improvements in computed tomography (CT) number accuracy, and investigate its effects on treatment planning dosimetry. A standard electron density phantom was scanned with and without metal inserts. The numbers of tissue-equivalent materials on both uncorrected and SEMAR-corrected CT images were compared. Treatment planning accuracy was evaluated by comparing dose distributions computed using true density images (without metal inserts), uncorrected images (with metal inserts), and SEMAR-corrected images (with metal inserts) using three-dimensional gamma analysis. The numbers of the true density and uncorrected and SEMAR-corrected CT images in a muscle plug with unilateral inserts were 25.9 HU, - 281.8 HU, and 26.1 HU, respectively. A similar tendency was obtained for other tissue-equivalent materials, and the numbers on CT images were improved with the SEMAR algorithm. In cases involving 1 portal irradiation, 10-MV X-ray, and the Acuros XB algorithm, the pass ratio between the true density and uncorrected images was 89.89%, while that between the true density and SEMAR-corrected images was 95.03%. Improvements in dose distribution were evident using the SEMAR algorithm. Similar trends were found for different irradiation methods and dose calculation algorithms. The SEMAR algorithm can significantly reduce metal artifacts on CT images used for radiation treatment planning. This aspect influenced dosimetry in the region of the artifact and dose distribution was significantly improved with use of the SEMAR-corrected images.