Z. Liao

Risk factors for local and regional recurrence in patients with resected N0–N1 non-small-cell lung cancer, with implications for patient selection for adjuvant radiation therapy

BACKGROUND The purpose of this study was to evaluate the actuarial risk of local and regional failure in patients with completely resected non-small-cell lung cancer (NSCLC), and to assess surgical and pathological factors affecting this risk. PATIENTS AND METHODS Between January 1998 and December 2009, 1402 consecutive stage I-III (N0-N1) NSCLC patients underwent complete resection without adjuvant radiation therapy. The median follow-up was 42 months. RESULTS Local-regional recurrence was identified in 9% of patients, with local failure alone in 3% of patients, regional failure alone in 4% of patients, and both local and regional failure simultaneously in 2% of patients. Patients who had local failure were found to be at increased risk of mortality. By multivariate analyses, three variables were shown to be independently significant risk factors for local [surgical procedure (single/multiple wedges+segmentectomy versus lobectomy+bilobectomy+pneumonectomy), tumor size>2.7 cm, and visceral pleural invasion] and regional (pathologic N1 stage, visceral pleural invasion, and lymphovascular space invasion, LVI) recurrence, respectively. CONCLUSION Patients with N0-N1 disease have low rates of locoregional recurrence after surgical resection. However, several prognostic factors can be identified that increase this risk and identify patients who may benefit from adjuvant treatment.

SU‐FF‐J‐22: Impact of Respiratory Motion On Dose Distributions and DVHs of Thoracic Structures — Evaluation Using 4DCT

Purpose: Respiratory motion may have effect on dose distributions and DVHs of thoracic organs and tumors themselves. The purposes of this work are to determine (1) changes of doses and DVHs with respiration; (2) actual doses delivered to these structures upon completion of one treatment fraction. Method and Materials: Ten non‐small‐cell lungcancer cases were selected, all had free‐breathing fast CT and 4DCT scans during normal breathing. These cases had acceptable breathing regularity and 4DCT image quality, sufficient motion of tumor and lung, and inclusion of entire lung. All cases were treated with 3D (8 cases) or IMRT (2 cases) techniques. Lung volumes were outlined on all respiratory phases consistently using a single CT‐number threshold. Dose distributions were re‐calculated for all phases, with DVHs obtained for each structure on all respiratory phases. For cases with significant changes of lung DVHs, deformable image registration was used to calculate total cumulative dose distribution combining all phases. Results: The overall dose distributions were rather insensitive to respiration for both 3D and IMRT beams except extreme cases with diaphragm moved in/out of beams. Thus, change of DVH for static structures (e.g. cord) was often insignificant. Large changes of lung DVHs were observed only for half of the cases, the degree of which was affected by tumor location and volume, and lung motion. Changes of PTV and heart DVHs were mainly caused by their translational movements. Use of free‐breathing CT was inadequate to represent moving anatomies and could cause spurious results, whiles mid‐inspiration CT provided the best estimate of average lung volume, mass, and dose distributions. Conclusion: Though dose distributions did not change significantly with breathing for photon treatments, variations of DVHs for thoracic structures were more complex and had to be assessed for individual patients and structures.

WE-E-J-6C-06: Proton Treatment Planning for Mobile Lung Tumors

Purpose: Traditional treatment planning methods may lead to lung proton treatment plans in which the apparent and actual dose distributions may be significantly different due to respiratory motion. We are developing strategies for designing compensator-based 3D proton treatment plans using 4D CTs (composed of 3D CTs at a sequence of respiratory phases) for mobile lung tumors and assessing the validity of these strategies using 4D dose computation methods. Method and Materials: 4D CTs for a population of lung cancer patients were used to obtain tumor targets and critical structures. The internal target volume (ITV) was the composite of target volumes on the 4D CT. For each patient, we evaluated four compensator design and planning strategies based on (1) the average CT obtained by averaging all phases of the 4D CT; (2) free breathing CT; (3) maximum intensity projection (MIP) CT; and (4) the average CT with the CT numbers inside the tumor volume replaced by the corresponding MIP CT numbers. For each strategy, the resulting apparent dose distribution was compared with the corresponding 4D dose distribution computed by deforming dose distributions of each phase to the reference phase and summing. Results: The composite 4D dose coverage of the target was significantly superior for method (4) while normal tissue doses were slightly higher though well below the limits. A seemingly conservative compensator design using MIP for the entire image, not just the target volume (Method 3), resulted in poor proximal target coverage due to over-estimation of the densities of intervening tissues. Conclusion: 3D proton plans based on the CT obtained by averaging the 3D CTs comprising the 4D CT, and with the CT numbers in the tumor volume replaced by the corresponding MIP CT numbers, is an effective approach to achieve good tumor coverage and acceptable normal tissue sparing.

SU‐FF‐T‐380: Dose Mass Histogram and Its Application for 4D Treatment Planning

Purpose: In evaluating dose distributions of lung treated during respiration, dose volume histogram (DVH) may not be an appropriate term because of variable lung volume with respiration. The purposes of this work are to investigated the use of dose mass histogram (DMH) for lung and assess the differences between DVH and DMH for conventional and 4D treatment planning. Method and Materials: DMH was calculated based on a similar concept of DVH excepted mass of each voxel, which was calculated from CT number to density conversion, was used in tallying dose distributions. For conventional treatment planning, DVHs and DMHs of normal lung (excluding GTV) were calculated and compared for 51 lung cancer cases and 52 esophagus cancer cases. For 4D treatment planning, ten lung cancer cases were analyzed, each of which had 4DCT scans of full lung and optimal CT image quality. Total normal lung was delineated on CT images of all respiratory phases using auto-thresholding with a single CT number. Dose distributions were computed on all respiratory phases with DVH and DMH being calculated for the lung. Results: For conventional plans, difference between DMH and DVH existed for cases with highly inhomogenuous lung tissues. For a majority of cases, such differences were small and may not be clinically significant. For 4D treatment planning, lung volume changed on average by 16.3% between inspiration and expiration. As expected, variation of lung mass was much smaller, only by about 6.6% as assessed from the 4DCT. The change of lung DMH with breathing was often different from that of lung DVH, indicating that deformation of lung mass followed different patterns than that of the lung volume during breathing. Conclusion: DMH may be more relevant than DVH considering varying alveolar-cell density in lung and conservation of lung mass in 4D treatment planning.

MO-DE-207B-07: Assessment of Reproducibility Of FDG-PET-Based Radiomics Features Across Scanners Using Phantom Imaging

PURPOSE Use a NEMA-IEC PET phantom to assess the robustness of FDG-PET-based radiomics features to changes in reconstruction parameters across different scanners. METHODS We scanned a NEMA-IEC PET phantom on 3 different scanners (GE Discovery VCT, GE Discovery 710, and Siemens mCT) using a FDG source-to-background ratio of 10:1. Images were retrospectively reconstructed using different iterations (2-3), subsets (21-24), Gaussian filter widths (2, 4, 6mm), and matrix sizes (128,192,256). The 710 and mCT used time-of-flight and point-spread-functions in reconstruction. The axial-image through the center of the 6 active spheres was used for analysis. A region-of-interest containing all spheres was able to simulate a heterogeneous lesion due to partial volume effects. Maximum voxel deviations from all retrospectively reconstructed images (18 per scanner) was compared to our standard clinical protocol. PET Images from 195 non-small cell lung cancer patients were used to compare feature variation. The ratio of a features standard deviation from the patient cohort versus the phantom images was calculated to assess for feature robustness. RESULTS Across all images, the percentage of voxels differing by <1SUV and <2SUV ranged from 61-92% and 88-99%, respectively. Voxel-voxel similarity decreased when using higher resolution image matrices (192/256 versus 128) and was comparable across scanners. Taking the ratio of patient and phantom feature standard deviation was able to identify features that were not robust to changes in reconstruction parameters (e.g. co-occurrence correlation). Metrics found to be reasonably robust (standard deviation ratios > 3) were observed for routinely used SUV metrics (e.g. SUVmean and SUVmax) as well as some radiomics features (e.g. co-occurrence contrast, co-occurrence energy, standard deviation, and uniformity). Similar standard deviation ratios were observed across scanners. CONCLUSIONS Our method enabled a comparison of feature variability across scanners and was able to identify features that were not robust to changes in reconstruction parameters.

TU-H-CAMPUS-JeP2-05: Can Automatic Delineation of Cardiac Substructures On Noncontrast CT Be Used for Cardiac Toxicity Analysis?

PURPOSE To evaluate the feasibility of using an automatic segmentation tool to delineate cardiac substructures from computed tomography (CT) images for cardiac toxicity analysis for non-small cell lung cancer (NSCLC) patients after radiotherapy. METHODS A multi-atlas segmentation tool developed in-house was used to delineate eleven cardiac substructures including the whole heart, four heart chambers, and six greater vessels automatically from the averaged 4DCT planning images for 49 NSCLC patients. The automatic segmented contours were edited appropriately by two experienced radiation oncologists. The modified contours were compared with the auto-segmented contours using Dice similarity coefficient (DSC) and mean surface distance (MSD) to evaluate how much modification was needed. In addition, the dose volume histogram (DVH) of the modified contours were compared with that of the auto-segmented contours to evaluate the dosimetric difference between modified and auto-segmented contours. RESULTS Of the eleven structures, the averaged DSC values ranged from 0.73 ± 0.08 to 0.95 ± 0.04 and the averaged MSD values ranged from 1.3 ± 0.6 mm to 2.9 ± 5.1mm for the 49 patients. Overall, the modification is small. The pulmonary vein (PV) and the inferior vena cava required the most modifications. The V30 (volume receiving 30 Gy or above) for the whole heart and the mean dose to the whole heart and four heart chambers did not show statistically significant difference between modified and auto-segmented contours. The maximum dose to the greater vessels did not show statistically significant difference except for the PV. CONCLUSION The automatic segmentation of the cardiac substructures did not require substantial modification. The dosimetric evaluation showed no statistically significant difference between auto-segmented and modified contours except for the PV, which suggests that auto-segmented contours for the cardiac dose response study are feasible in the clinical practice with a minor modification to the PV vessel.

SU‐F‐BRD‐01: A Novel 4D Robust Optimization Mitigates Interplay Effect in Intensity‐Modulated Proton Therapy for Lung Cancer

Purpose: To compare the impact of interplay effect on 3D and 4D robustly optimized intensity-modulated proton therapy (IMPT) plans to treat lung cancer. Methods: Two IMPT plans were created for 11 non-small-cell-lung-cancer cases with 6–14 mm spots. 3D robust optimization generated plans on average CTs with the internal gross tumor volume density overridden to deliver 66 CGyE in 33 fractions to the internal target volume (ITV). 4D robust optimization generated plans on 4D CTs with the delivery of prescribed dose to the clinical target volume (CTV). In 4D optimization, the CTV of individual 4D CT phases received non-uniform doses to achieve a uniform cumulative dose. Dose evaluation software was developed to model time-dependent spot delivery to incorporate interplay effect with randomized starting phases of each field per fraction. Patient anatomy voxels were mapped from phase to phase via deformable image registration to score doses. Indices from dose-volume histograms were used to compare target coverage, dose homogeneity, and normal-tissue sparing. DVH indices were compared using Wilcoxon test. Results: Given the presence of interplay effect, 4D robust optimization produced IMPT plans with better target coverage and homogeneity, but slightly worse normal tissue sparing compared to 3D robust optimization (unit: Gy) [D95% ITV: 63.5 vs 62.0 (p=0.014), D5% - D95% ITV: 6.2 vs 7.3 (p=0.37), D1% spinal cord: 29.0 vs 29.5 (p=0.52), Dmean total lung: 14.8 vs 14.5 (p=0.12), D33% esophagus: 33.6 vs 33.1 (p=0.28)]. The improvement of target coverage (D95%,4D – D95%,3D) was related to the ratio RMA3/(TVx10−4), with RMA and TV being respiratory motion amplitude (RMA) and tumor volume (TV), respectively. Peak benefit was observed at ratios between 2 and 10. This corresponds to 125 – 625 cm3 TV with 0.5-cm RMA. Conclusion: 4D optimization produced more interplay-effect-resistant plans compared to 3D optimization. It is most effective when respiratory motion is modest compared to TV. NIH/NCI K25CA168984; Eagles Cancer Research Career Development; The Lawrence W. and Marilyn W. Matteson Fund for Cancer Research; Mayo ASU Seed Grant; The Kemper Marley Foundation

TU‐G‐108‐03: Small Lung Size Increases the Risk of Radiation Pneumonitis in Lung Cancer Patients

PURPOSE There may be certain populations of lung cancer patients who are especially sensitive to radiation pneumonitis. We have examined the relationship of lung size with the risk of pneumonitis. METHODS A total of 509 non-small cell carcinoma patients were included in this study. Patients received normally fractionated radiation (180 - 225 cGy per fraction) with a minimum total dose of 50 Gy. Patients were planned with either 3D conformal (296 pts) or intensity modulated radiation therapy (IMRT, 213 pts). Radiation pneumonitis was graded by the treating physician according to the National Cancer Institute Common Terminology Criteria for Adverse Events (CTCAE) 3.0. Univariate analysis was performed considering mean lung dose, percentage DVH, lung size, and spared lung. The Lyman-Kutcher-Burman (LKB) method was used to fit the normal tissue complication probability (NTCP) curve to the data with respect to effective dose. RESULTS Overall 20% of patients developed Grade 3 or higher pneumonitis, including 24% of the 3D patients and 15% of the IMRT patients. Univariate analysis showed lung size, mean lung dose, percentage treated and spared lung volumes to be significant. An LKB fit to the effective dose showed that it is important to take the volume parameter into account. The risk of pneumonitis predicted by NTCP analysis is underestimated for patients with small total lung sizes (< 2550 cc), where 40% of patients developed Grade 3 or higher pneumonitis. CONCLUSION Patients with small lung sizes may be especially vulnerable to radiation pneumonitis. This new finding could be used during treatment planning to reduce the risk of pneumonitis in patients with small lungs.

SU‐F‐BRCD‐05: Mean Regional Dose to the Esophagus Predicts Acute Toxicity Rate for Lung Cancer Patients

PURPOSE The relationship between spatial aspects of the dose distribution and incidence of acute radiation esophagitis for non-small-cell lung cancer (NSCLC) patients is not well understood. Specifically, the location of dose along the superior-inferior (SI) axis of the esophagus has not been previously considered. We introduce the concept of mean regional dose (MRD) calculated for esophageal subvolumes, and test for significance for prediction of acute esophagitis (AE). METHODS The 3D dose distribution within the esophagus was extracted for 541 NSCLC patients treated with definitive photon therapy. The esophagus contour was divided into equal geometric halves, thirds, and fourths along the SI direction of the structure. MRD in each subvolume was calculated. Univariate logistic regression was performed to determine the correlation between MRD and CTCAE3.0 AE grade = 2 (medical intervention). The MRD was incorporated into an existing NTCP model (based on mean dose for the total esophageal volume) as a separate additive factor. RESULTS Univariate analysis indicated a significant correlation between AE grade = 2 and MRD in each of the esophageal subvolumes except for the inferior third and inferior-most quarter. There was a statistically significant improvement when including the additive MRD factor for the superior/inferior halves, superior/inferior thirds, and superior-most/inferior-most quarters into the NTCP model. CONCLUSIONS This study investigates previously unexplored regional differences in delivered dose to the esophagus of patients treated for NSCLC. There is evidence to suggest that dose to the superior portions of the esophagus is more important as it relates to the potential for acute toxicity. The 541 patient cohort is the largest database used to investigate AE in patients treated for NSCLC, strengthening the power of the statistical results. Additional methods to incorporate dose in individual esophagus voxels (along the SI axis) into the NTCP model are also being explored.

TU‐A‐BRC‐03: A Novel Technique to Use CT Images for in Vivo Detection and Quantification of the Spatial Distribution of Radiation‐Induced Damage to the Esophagus

Purpose: Current dose‐response studies use symptom endpoints (e.g. grade 2 esophagitis) as substitutes for data about actual tissue damage. CTimaging could provide objective spatial data on radiation‐induced damage to the esophagus in lungcancer patients. Methods: Deformable image registration techniques were used to register weekly CTimages taken during radiotherapytreatment with the original planning CTimage. The esophagus contours were automatically mapped. The impact of day‐to‐day variations in the degree of collapse of the esophagus was reduced by using thresholding to remove air from the esophagus. The cross‐sectional area of the esophagus was then calculated for each CT slice. The results from surrounding slices were averaged to reduce apparent abrupt steps in cross‐ sectional area between adjacent CT slices caused by varying undulations/folds in the esophagus. Finally, the relative expansion of the esophagus was calculated as the ratio of the cross‐sectional area of the esophagus in the weekly CT (minus air) to that on the corresponding CT slice in the planning image. This technique was applied to weekly CTimages of 5 lungcancer patients (35 CTs), with acute esophageal toxicity grade 0 to 3. For these patients we examined (1) The correlation between the relative expansion of the esophagus and the clinical toxicity grade, and (2) the correlation between the spatial dose distribution and the spatial variation in esophageal expansion. Results: The average maximum esophageal expansions for toxicity grades 0, 2, and 3 were 1.2, 1.7 and 1.9, respectively. The difference between grade 2 and 0 was statistically significant (p=0.008). The location and degree of variation of the changes in esophagus cross‐section were found to be related to the high dose given at the same location. Conclusion: Radiation‐induced injury to the esophagus can be detected in CTimages. This has potential for use in dose‐response studies.