If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Assessing the impact of intrafraction motion correction on PTV margins and target and OAR dosimetry for single-fraction free-breathing lung stereotactic body radiation therapy
Reprint requests to Clara Fallone, PhD, Department of Medical Physics, Nova Scotia Health, Department of Radiation Oncology, Dalhousie University, QEII Health Sciences Centre, Halifax, Nova Scotia, B3H 1V7, Canada.
Department of Medical Physics, Nova Scotia Health (NSH), Halifax, Nova Scotia, B3H2Y9 CanadaDepartment of Radiation Oncology, Dalhousie University, Halifax, Nova Scotia, B3H2Y9 Canada
Department of Radiation Oncology, Nova Scotia Health, Halifax, Nova Scotia, B3H2Y9 CanadaDepartment of Radiation Oncology, Dalhousie University, Halifax, Nova Scotia, B3H2Y9 Canada
Department of Medical Physics, Nova Scotia Health (NSH), Halifax, Nova Scotia, B3H2Y9 CanadaDepartment of Radiation Oncology, Dalhousie University, Halifax, Nova Scotia, B3H2Y9 CanadaDepartment of Physics and Atmospheric Science, Dalhousie University, Halifax, Nova Scotia, B3H2Y9 Canada
The objective of this research is to investigate intrafraction motion correction on planning target volume (PTV) margin requirements and target and organ-at-risk (OAR) dosimetry in single-fraction lung stereotactic body radiation therapy (SBRT). Sixteen patients (15 with upper lobe lesions, 1 with a middle lobe lesion) were treated with single-fraction lung SBRT. Cone-beam computed tomography (CBCT) images were acquired before the treatment, between the arcs, and after the delivery of the treatment fraction. Shifts from the reference images were recorded in anterior-posterior (AP), superior-inferior (SI), and lateral (LAT) dimensions. The deviations from the reference image were calculated for 3 clinical scenarios: not applying intratreatment couch shifts and not correcting for pretreatment deviations < 3 mm ( scenario 1), not applying intratreatment couch shifts and correcting for pretreatment deviations < 3 mm ( scenario 2), and applying all pre- and intratreatment couch shifts (scenario 3). PTV margins were determined using the van Herk formalism for each scenario and maximum and average deviations were assessed. The clinical scenarios were modelled in the treatment planning system based on each patient dataset to assess target and OAR dosimetry. Calculated lower-bound PTV margins in the AP, SI, and LAT dimensions were [4.6, 3.5, 2.3] mm in scenario 1, [4.6, 2.4, 2.2] mm in scenario 2, and [1.7, 1.2, 1.0] mm in scenario 3. The margins are lower bounds because they do not include contributions from nonmotion related errors. Average and maximum intrafraction deviations were larger in the AP dimension compared to the SI and LAT dimensions for all scenarios. A unidimensional movement (several mm) in the negative AP dimension was observed in clinical scenarios 1 and 2 but not scenario 3. Average intrafraction deviation vectors were 1.2, 1.1, and 0.3 mm for scenarios 1, 2, and 3, respectively. Modelled clinical scenarios revealed that using scenario 3 yields significantly fewer treatment plan objective failures compared to scenarios 1 and 2 using a Wilcoxon signed-rank test. Intratreatment motion correction between each arc may enable reductions PTV margin requirements. It may also compensate for unidimensional negative AP movement, and improve target and OAR dosimetry.
It is typically used as a surgery alternative; at least 20 % of patients with early stage lung cancer are deemed inoperable and some operable patients elect not to receive surgery due to possible risks,
such as reactions to anesthesia, excess bleeding, and pneumonia. SBRT delivers high doses of radiation in a small number of sessions and can improve survival
Local consolidative therapy versus maintenance therapy or observation for patients with oligometastatic non-small-cell lung cancer without progression after first-line systemic therapy: a multicentre, randomised, controlled, phase 2 study.
Local consolidative therapy versus maintenance therapy or observation for patients with oligometastatic non-small-cell lung cancer without progression after first-line systemic therapy: a multicentre, randomised, controlled, phase 2 study.
Since single-fraction SBRT introduces a higher delivered dose and potentially longer treatment time, there is greater risk of intrafraction tumor displacement and consequently over-dosing adjacent organs at risk (OARs)
ITV versus mid-ventilation for treatment planning in lung SBRT: a comparison of target coverage and PTV adequacy by using in-treatment 4D cone beam CT.
In this research, several analyses were completed to investigate the effect of intrafraction motion using the CBCTs acquired for positioning (pretreatment), in between treatment arcs (intratreatment), and post-treatment. Three clinical scenarios were assessed: not correcting for intrafraction motion and only correcting for pretreatment deviations larger than 3 mm
(clinical scenario 2) and correcting for intrafraction motion and all pretreatment deviations (clinical scenario 3). The impact of motion corrections was assessed on the following: planning target volume (PTV) margin determination, overall patient deviation from the reference position, and dosimetry. A 3 mm deviation was selected as an evaluation index based on local clinical protocols; some literature suggests a 3 mm action level prior to applying couch shifts in stereotactic lung radiotherapy.
A randomized phase II study comparing 2 stereotactic body radiation therapy (SBRT) schedules for medically inoperable patients with stage I peripheral non-small cell lung cancer: NRG Oncology RTOG 0915 (NCCTG N0927).
Int. J. Radiat. Oncol. Biol. Phys.2015; 93: 757-764
ITV versus mid-ventilation for treatment planning in lung SBRT: a comparison of target coverage and PTV adequacy by using in-treatment 4D cone beam CT.
ITV versus mid-ventilation for treatment planning in lung SBRT: a comparison of target coverage and PTV adequacy by using in-treatment 4D cone beam CT.
used modelling based on patient data to estimate that 3 mm PTV margins would yield 2-year tumor control probability rates of 95.2 % vs 96.1 % with 5 mm margins, while reducing dose to OARs. The Van Herk formalism
Comparison of geometric uncertainties using electronic portal imaging device in focal three-dimensional conformal radiation therapy using different head supports.
ITV versus mid-ventilation for treatment planning in lung SBRT: a comparison of target coverage and PTV adequacy by using in-treatment 4D cone beam CT.
Comparison between target margins derived from 4DCT scans and real-time tumor motion tracking: Insights from lung tumor patients treated with robotic radiosurgery.
This study utilized the van Herk formalism to compute PTV margins from patient data for each clinical scenario.
In addition to assessing the impact of motion correction on calculated PTV margin size, the average and maximum patient deviations from the reference position for the clinical scenarios was also investigated. Dosimetric differences between the treatment plan, and delivering the plan using clinical scenarios 1, 2 and 3, were computed for each patient based on the acquired deviation data and evaluating metrics from single fraction lung SBRT clinical trials.
A randomized phase II study comparing 2 stereotactic body radiation therapy (SBRT) schedules for medically inoperable patients with stage I peripheral non-small cell lung cancer: NRG Oncology RTOG 0915 (NCCTG N0927).
Int. J. Radiat. Oncol. Biol. Phys.2015; 93: 757-764
We hypothesize that correcting for motion during treatment by repositioning the patient between treatment arcs will permit margin reductions in the future and improve dosimetry.
Methods
A total of 16 patients with non-small-cell lung cancer (NSCLC) were treated with free-breathing single-fraction lung SBRT. Patient eligibility criteria included a biopsy-proven peripheral lung metastasis or clinical scenario compatible with metastatic lung cancer, metastasis diameter less than 30 mm, and lesions at least 2 cm away from the chest wall, diaphragm, and lung apex. Only patients with an Eastern Cooperative Oncology Group (ECOG) performance status of 0, 1, or 2 were considered for this technique. The study was approved by the local research ethics board. The patient demographics are summarized in Table A1 in the supplementary appendix. Patient 2 received a 30 Gy prescription dose; all other patients received a 34 Gy prescription dose. All patients underwent a 4DCT simulation on a General Electric Lightspeed CT simulator for treatment planning purposes. Patients were immobilized using an SBRT body device (FreedomX, CDR systems) in combination with a patient-specific vacuum cushion. Treatment plans were created using the Eclipse treatment planning system (Varian Medical Systems, Palo Alto, USA). Internal target volumes (ITVs) were contoured on a maximum intensity projection image of the 4DCT data set. The clinical target volume (CTV) was equivalent to the ITV. Per current local treatment guidelines based on clinical trials,
A randomized phase II study comparing 2 stereotactic body radiation therapy (SBRT) schedules for medically inoperable patients with stage I peripheral non-small cell lung cancer: NRG Oncology RTOG 0915 (NCCTG N0927).
Int. J. Radiat. Oncol. Biol. Phys.2015; 93: 757-764
PTV margins were an isotropic 5 mm expansion from the CTV. Volumetric modulated arc therapy (VMAT) plans were created using the Eclipse photon optimizer (PO) tool (version 13.6.23, Varian Medical Systems, Palo Alto, USA). Target coverage and OAR dose constraints were based on local guidelines and published single-fraction SBRT clinical trials.
A randomized phase II study comparing 2 stereotactic body radiation therapy (SBRT) schedules for medically inoperable patients with stage I peripheral non-small cell lung cancer: NRG Oncology RTOG 0915 (NCCTG N0927).
Int. J. Radiat. Oncol. Biol. Phys.2015; 93: 757-764
Tables A2 and A3 in supplementary material detail the coverage objectives and dose constraints that directed the treatment planning. Dose calculations were performed on the average CT images using the Eclipse Analytical Anisotropic Algorithm (AAA,version 13.6.23, Varian Medical Systems, Palo Alto, USA). The treatment plans consisted of 2-4 axial arcs (couch rotations of 0 degrees).
Single-fraction lung SBRT was delivered using VMAT and a 6 MV flattening filter-free photon beam with a maximum dose rate of 1400 MU/min on a Varian TrueBeam linear accelerator (Varian Medical Systems, Palo Alto, USA). KV-CBCT images were acquired pretreatment, post couch shifts (if applied), in between each treatment arc, and at the end of the session. Patient positioning shifts from the reference CT images were computed using matching of the acquired CBCT images with the reference simulation images at the time of imaging. The radiation therapists completed the matching by prioritizing matching of the target volume. Rotational errors were not accounted for as the institution does not use a 6-degree-of-freedom-couch; however, the therapists repositioned the patient and repeated the process if errors exceeded 3 degrees. The patient was not readjusted if rotational errors were less than 3 degrees. If positional errors in any of the anterior-posterior (AP), superior-inferior (SI), or lateral (LAT) dimensions exceeded 3 mm from the reference CT following any of the CBCTs, the couch position shifts were applied and the patient was reimaged with CBCT. The threshold of 3 mm was selected because it is the institution's local guideline for patient reimaging and repositioning and is also suggested in some literature as an action level prior to applying couch shifts.
In the following analysis, each CBCT data set is assigned a subscript index (i.e. CBCT1…CBCTn). If multiple positioning CBCTs were acquired before beginning the treatment, the final CBCT before treatment was named CBCT1. The treatment duration was recorded as the time between the start of CBCT1 and the end of the post-treatment CBCT using the time stamps from the record and verify system; CBCTs acquired before CBCT1 were excluded from this metric.
Three clinical scenarios were modelled using the data acquired in this study: Clinical scenario 1: no couch position shifts < 3 mm applied after CBCT 1 and no couch position shifts applied during treatment; Clinical scenario 2: couch shifts < 3 mm applied after CBCT 1 and no couch shifts applied during treatment; and Clinical scenario 3: couch shifts following every CBCT. Figures 1 (a), (b), and (c) display the modelled workflows for each clinical scenario. Deviations were calculated using the following in each dimension:
xo = CT reference position
xn = position for nth CBCT
Dn = deviation for nth CBCT
Fig. 1(A). Modelled workflow for clinical scenario 1 (no couch shifts < 3 mm applied after pretreatment CBCT and no intratreatment couch shifts). (B). Modelled workflow for clinical scenario 2 (couch shifts < 3 mm applied after pretreatment CBCT and no intratreatment couch shifts). (C). Modelled workflow for clinical scenario 3 (pre- and intratreatment couch shifts applied).
Although not depicted in the equations above, recorded offset values from the treatment console system were adjusted appropriately to account for couch shifts completed in the clinical workflow that would not have been executed in the modelled scenario, and vice versa.
For each clinical scenario, PTV margins in the AP, SI, and LAT dimensions were assessed for each of the analyses using the van Herk formalism:
where ∑ corresponds to systematic error and represents the quadratic sum of all random error. In this study, the latter includes the standard deviation of intrafraction motion and the standard deviation describing the penumbra :
Comparison between target margins derived from 4DCT scans and real-time tumor motion tracking: Insights from lung tumor patients treated with robotic radiosurgery.
The systematic error (∑) in one dimension and standard deviation of intrafraction motion in the same dimension were calculated from the measured deviations as described by van Herk.
It was shown that for a small number of N fractions (or in this case, a small number of N arcs), the average treatment execution error may not be zero as assumed in Eq. (6)
In this work, N was taken as the average number of arcs over all patients. PTV margins using Eqs. (8) to (11) were computed for all three clinical scenarios.
The average deviations calculated in each scenario were plotted as a function of CBCT number. Maximum and average intrafraction deviations in each dimension were also calculated for the three analyses and vector sums were computed for the averaged values. The dosimetric impacts of delivering the treatment plan using clinical scenarios 1, 2 and 3 were simulated for each patient using the Eclipse treatment planning software. The isocenter of each arc in each plan was adjusted to reflect the calculated patient deviation using the relevant clinical scenario. Isocenter adjustments for each arc were computed by averaging the deviations obtained from the CBCTs acquired immediately before and after the arc. The average of the CBCTs acquired before and after the arc should be most representative of patient positioning during the arc. The arcs were combined into a summed plan for each scenario in Eclipse and the dose distribution was forward-calculated using the MU for each arc from the clinical plan. Multiple target and OAR dose metrics
A randomized phase II study comparing 2 stereotactic body radiation therapy (SBRT) schedules for medically inoperable patients with stage I peripheral non-small cell lung cancer: NRG Oncology RTOG 0915 (NCCTG N0927).
Int. J. Radiat. Oncol. Biol. Phys.2015; 93: 757-764
(Tables A2 and A3, Appendix A) were computed for the original treatment plan and the sum plans with arc-specific isocenters. Target coverage and OAR sparing were compared for the treatment plan and both clinical scenarios based on clinical trial (Radiation Therapy Oncology Group (RTOG) 0915) objectives and constraints
A randomized phase II study comparing 2 stereotactic body radiation therapy (SBRT) schedules for medically inoperable patients with stage I peripheral non-small cell lung cancer: NRG Oncology RTOG 0915 (NCCTG N0927).
Int. J. Radiat. Oncol. Biol. Phys.2015; 93: 757-764
or protocols specific to this institution. The number of failed plan objectives for each scenario was assessed. A Wilcoxon signed-rank test (0.05 confidence level) was completed to determine if differences in number of failures were significant. The CTV coverage in all three scenarios was assessed and compared with the coverage assured using van Herk PTV margin formulation. A Wilcoxon signed-rank test was used to assess significant differences (0.05 confidence level) in CTV coverage between the treatment plan and the clinical scenarios, as well as between the scenarios.
Results
Table A4 in supplementary material indicates the treatment duration, number of CBCTs acquired, number of arcs delivered, and number of couch position shifts given our clinical workflow, as well as any items of note. The average treatment time (excluding any CBCTs acquired before CBCT1 but including the post-treatment CBCT) was 21 ± 5 minutes.
Figure 2 displays the reference CT overlaid with the CBCT for an example patient in the AP, SI, and LAT dimensions. A zoomed axial view of the PTV demonstrates the discrepancy between the reference CT image and the image acquired during the treatment. Table 1 reveals the treatment trajectory and recorded offsets for an example patient. Table 2 lists the van Herk systematic errors (), random errors (), and calculated PTV margins in the AP, SI, and LAT dimensions for the three clinical scenarios. Figure 3(a), (b), and (c) display the average AP, SI , and LAT deviations between the acquired CBCT and the reference CT for the three clinical scenarios, respectively. Table 3 depicts the maximum and average deviations in all dimensions in the three clinical scenarios, with a corresponding average vector.
Fig. 2Reference CT images overlaid with CBCT for a patient. The PTV contour is shown in red. The axial view is zoomed to show the reference image (background), and the CBCT image (window), with the corresponding shift on the edges of the target. Window and level were adjusted for clarity.
Table 1Treatment trajectory and offsets recorded for each CBCT for patient 13
Event
Offset values (from reference CT), mm
Vertical
Longitudinal
Lateral
CBCT1 (pretreatment)
-0.6
2.4
1.1
Arc 1 delivered
CBCT2
-4.5
2.7
1.3
Couch moved due to offset > 3 mm in vertical direction. CBCT acquisition is repeated to confirm couch movement in correct direction. CBCT3 offsets are adjusted to account for couch corrections when calculating deviations in clinical scenario 2.
CBCT3
-1.8
0.2
0.0
Arc 2 delivered
CBCT4
-4.4
-1.7
0.1
Couch moved due to offset > 3 mm in vertical direction. CBCT acquisition is repeated to confirm couch movement in correct direction. CBCT5 offsets are adjusted to account for couch corrections when calculating deviations in clinical scenario 2.
CBCT5
2.2
0.7
0.2
Arc 3 delivered
A post-treatment CBCT was not acquired for this patient.
Fig. 3(A). AP deviations between the acquired CBCT and the reference CT for clinical scenario 1 (no couch shifts < 3 mm applied after pretreatment CBCT and no intratreatment couch shifts). A point corresponding to CBCT6 is not included because only a single patient had a 6th CBCT. Error bars (standard deviations) are not shown on the figure for clarity. Standard deviations for CBCT1 were [1.0, 0.0, 0.0] mm for scenarios 1, 2, and 3 respectively, and increased to [3.3, 2.0, 1.4] mm by CBCT5. (B). SI deviations between the acquired CBCT and the reference CT for clinical scenario 2 (couch shifts < 3 mm applied after pretreatment CBCT and no intratreatment couch shifts). A point corresponding to CBCT6 is not included because only a single patient had a 6th CBCT. Error bars (standard deviations) are not shown on the figure for clarity. Standard deviations for CBCT1 were [1.0, 0.0, 0.0] mm for scenarios 1, 2, and 3 respectively, and increased to [2.2, 1.2, 1.1] mm by CBCT5. (C). LAT deviations between the acquired CBCT and the reference CT for clinical scenario 3 (pre- and intratreatment couch shifts applied). A point corresponding to CBCT6 is not included because only a single patient had a 6th CBCT. Error bars (standard deviations) are not shown on the figure for clarity. Standard deviations for CBCT1 were [0.8, 0.0, 0.0] mm for scenarios 1, 2, and 3 respectively, and increased to [1.3, 0.6, 0.3] mm by CBCT5.
A randomized phase II study comparing 2 stereotactic body radiation therapy (SBRT) schedules for medically inoperable patients with stage I peripheral non-small cell lung cancer.
Int. J. Radiat. Oncol. Biol. Phys.2015; 93: 757-764
Table 4 includes the Wilcoxon signed-rank test statistics (and associated significance) regarding failed objectives for the three scenarios. Figures S1 (a), (b), and (c) illustrate the distribution of failed plan objectives or OAR constraints for clinical scenarios 1, 2, and 3 respectively
Table 4Wilcoxon signed-rank test statistics between the three clinical scenarios for failed dosimetric requirements
Figure 4 demonstrates the % isodose line that covers the full volume of the CTV for the treatment plan and each clinical scenario. Table 5 includes the results of the Wilcoxon signed-rank test regarding CTV isodose line coverage for comparisons between the treatment plan (no motion modelled) and the three clinical scenarios, as well as comparisons between the clinical scenarios.
Fig. 4% Isodose line covering CTV for the treatment plan and modelled delivery using each clinical scenario.
Table 5Wilcoxon signed-rank test statistics between the treatment plan and the 3 clinical scenarios and between the scenarios for CTV isodose line coverage
Lower-bound calculated PTV margins (Table 2) in the AP, SI, and LAT dimensions were [4.6, 3.5, 2.3] mm for clinical scenario 1 (no couch position shifts < 3 mm applied after CBCT1 and no couch shifts applied during treatment), [4.6, 2.4, 2.2] mm for clinical scenario 2 (couch shifts < 3 mm applied after CBCT1 and no couch shifts applied during treatment ) and [1.7, 1.2, 1.0] mm for clinical scenario 3 (couch shifts applied after every CBCT). The calculated PTV margins suggest that the current guidelines for single-fraction lung SBRT of a 5 mm isotropic PTV margin
A randomized phase II study comparing 2 stereotactic body radiation therapy (SBRT) schedules for medically inoperable patients with stage I peripheral non-small cell lung cancer: NRG Oncology RTOG 0915 (NCCTG N0927).
Int. J. Radiat. Oncol. Biol. Phys.2015; 93: 757-764
can potentially be re-evaluated to obtain tighter treatment margins and the corresponding benefits for all clinical scenarios, and even more significantly for clinical scenario 3. However, the margins presented here are a lower bound. The congruency of imaging and radiation isocenters (1 mm for kV- CBCT),
and factors such as inter-observer variability, intraobserver variability, couch motion mechanical limits, potential variations in breathing patterns, CBCT reconstruction and registration error, and variability in target delineation,
ITV versus mid-ventilation for treatment planning in lung SBRT: a comparison of target coverage and PTV adequacy by using in-treatment 4D cone beam CT.
The larger margin requirement in the AP direction suggests a nonisotropic margin should be considered.
To our knowledge, this research is the first to use the van Herk formalism to investigate PTV margins for single-fraction lung SBRT in a non-Cyber-Knife treatment setting.
Comparison between target margins derived from 4DCT scans and real-time tumor motion tracking: Insights from lung tumor patients treated with robotic radiosurgery.
Comparison between target margins derived from 4DCT scans and real-time tumor motion tracking: Insights from lung tumor patients treated with robotic radiosurgery.
reported a 4 mm global PTV margin using van Herk formalisms for the Cyber Knife Xsight. Another Cyber Knife SBRT study used the van Herk formalism to obtain values of 4.7 mm (AP), 6.8 mm (SI), and 4.4 mm (LAT) for treatments of 1-5 fractions.
Comparison between target margins derived from 4DCT scans and real-time tumor motion tracking: Insights from lung tumor patients treated with robotic radiosurgery.
Comparison between target margins derived from 4DCT scans and real-time tumor motion tracking: Insights from lung tumor patients treated with robotic radiosurgery.
Comparison between target margins derived from 4DCT scans and real-time tumor motion tracking: Insights from lung tumor patients treated with robotic radiosurgery.
the anatomy matching was completed using the vertebral column whereas in this study, the target volume was used for matching. Larger PTV margins in the bone-matching case are expected given a study that compared PTV margin requirements for bony anatomy matching vs target matching.
Comparison between target margins derived from 4DCT scans and real-time tumor motion tracking: Insights from lung tumor patients treated with robotic radiosurgery.
This study also employed van Herk parameters for 90 % dose level coverage and 90 % confidence level; selected parameters vary between studies. Moreover, some cases
To our knowledge, no previous study compared van Herk PTV margins for the cases of not correcting vs correcting for intrafraction motion in lung SBRT.
The results shown in Table 2 suggest that correcting for offsets after every acquired CBCT yielded the smallest lower-bound PTV margins (clinical scenario 3). These results imply that obtaining CBCT images between treatment arcs could potentially reduce the impacts of intrafraction motion. Figure 3(a) suggests that there is a tendency for the patient to move in the negative AP dimension (several mm) as the treatment progresses, when couch shifts are not applied between arcs (clinical scenarios 1 and 2). This trend indicates that the AP PTV margin should be larger in the negative AP dimension to account for this movement, which is not reflected in van Herk PTV margin calculations. Increasing the systematic error term in the van Herk formalism would increase the PTV margin in both the positive and negative AP dimensions and thus would not compensate for unidimensional motion. Correcting for motion in between arcs reduced the observed unidimensional offset, as illustrated in Fig. 3(C). Average and maximum intrafraction deviations were overall larger in the AP dimension compared to the SI and LAT dimensions in all scenarios. These findings could reflect that immobilization in the AP dimension was not as stringent as that in the other two dimensions. The vacuum bag immobilizes in the SI and lateral dimensions, but not in the AP dimension. The same trend observed here was also seen in a multi fraction lung SBRT study and attributed to reduced stability in the AP dimension.
In this study, 15 patients had upper lobe nodules and one patient had a middle-lobe nodule (patient 9). Breathing motion is expected to be more significant for lower lobe nodules, especially in the SI direction;
more breathing motion could potentially explain why other studies saw larger required PTV margins in the SI dimension compared to the AP dimension when treating tumors in the lower lobe.
reported 0.17 cm motion in the SI direction for upper lobe tumours vs 0.77 cm for lower lobe tumours. Comparatively, motion in the AP direction was 0.3 cm in both cases.
also describe that lung tumor motion is mainly affected by diaphragm motion and that tumors in the lower lobe exhibit more motion than those in upper lobes, mainly in the SI direction.
In addition, the ITV used in this study is based off a 4DCT maximum intensity projection and thus should account for effects of patient breathing; not all studies employ this method and thus larger deviations in the SI direction due to breathing motion are expected.
In this study, the patient with the middle-lobe nodule was not an outlier. However, given that 15 of 16 patients in this study had upper lobe lesions, the results of this study may not be generalizable to lesions in the lower lobe of the lung. Average intrafraction deviation vectors in this study were 1.2, 1.1, and 0.3 mm for clinical scenarios 1, 2, and 3, respectively. Shah et al.
reported an average intrafraction deviation vector of 1.5 mm when measured as the difference between the post-treatment and positional CBCTs for multi-fraction lung SBRT using a stereotactic frame, consistent with our results for clinical scenarios 1 and 2. For some patients, applying couch position shifts yielded larger average or maximum deviations. Thus, in some cases, attempting to correct couch position more frequently could lead to larger deviations from the reference position, indicating that patient motion is not necessarily systematically unidimensional and some motion is self-corrected. However, intrafraction motion correction yielded less significant deviations for the majority of patients (Table 3).
The PTV margins calculated using the van Herk formalism only provide insight regarding 90% isodose target coverage; additional target and OAR dose metrics were investigated in this study by modelling the clinical scenarios in the treatment planning system for each patient data set. As seen in Table 4, the number of failed objectives or OAR constraints were significantly larger for scenario 1 vs scenario 3, and also for scenario 2 vs 3, indicating that correcting for intrafraction motion improved dose metrics. The parameters chosen in Eq. 6 assert that CTVs will receive 90% of the prescription dose with a 90% confidence interval.
In this study, 81% of patients met the 90% prescription dose coverage criteria in clinical scenarios 1, and 94% met the criteria in scenarios 2 and 3,using the 5 mm isotropic margin. These results suggest that there is potential to reduce PTV margins and still meet the 90% prescription isodose CTV coverage criteria in 90% of patients when using scenarios 2 and or 3. Exploring the coverage obtained and OAR dose using the nonisotropic, smaller margins calculated in this work was outside the scope of this study. However, previous studies reported improvements in OAR dose coupled with adequate target coverage when PTV margins were reduced from 5 mm to 3 mm in multi-fraction lung SBRT.
ITV versus mid-ventilation for treatment planning in lung SBRT: a comparison of target coverage and PTV adequacy by using in-treatment 4D cone beam CT.
Figure 4 indicates that patient 13 was the only patient wherein correcting for intrafraction motion yielded lower CTV isodose coverage. Given that patient 13 yielded the largest motion deviations in the study (Table 3), and required two shifts during treatment due to deviations > 3 mm (Table 1), it appears that some patients are clear outliers with respect to intrafraction motion and the validity of any chosen margin in these cases may be questionable. Patient 12 also showed increased CTV coverage degradation compared to the other patients; however, applying couch shifts reduces the degradation, in line with the other patients.
Several assumptions of this study should be discussed. The van Herk model used in this work assumes an idealized conformation around the 3D surface of the target volume
Moreover, use of the van Herk model for PTV margin determination in lung has previously been validated, demonstrating effective PTV coverage in modelled motion for a range of tumor sizes, motions, treatment techniques, plan conformities, and tissue densities.
Utilizing a mid-ventilation PTV margin was not assessed in this study; however, research has shown that PTV margins can be reduced using a mid-ventilation based PTV margin in combination with image guidance, compared to an ITV approach as used herein.
This technique can be explored further. A limitation is that this study assumed that the couch shifts applied returned the patient to the reference position; however, couch mechanical motions are only accurate within ± 1mm.
Furthermore, all modelled clinical scenarios were derived using the same treatment time; in reality, imaging in between arcs increases total treatment time and could potentially affect intrafraction motion. This inherent limitation creates a bias in favor of clinical scenario 3. On average over all patients, deviations in the AP dimension were 1.14 mm over a mean treatment time of 21 minutes, leading to an estimated average motion of 0.05 mm/min during treatment. If we assume an average of 3 CBCT acquisitions in between arcs, each requiring 2.5 minutes, the motion due to intrafraction imaging would be approximately 0.4 mm. The introduction of the 0.4 mm displacement is quite small compared to the maximum deviations measured using scenario 1 (Table 3), suggesting that applying intrafraction motion corrections is still beneficial, even if it introduces more movement due to a longer treatment time. The argument could also be made that with intrafraction motion correction, any additional introduced deviation would itself be corrected before the next arc. The impact of longer treatment time wasn't investigated in this study. Previously reported multi-fraction SBRT results noted more significant patient drifts in treatments > 20 minutes.
Moreover, a multi-fraction study in breast found that margin requirements for intrafraction motion doubled for a treatment time of 24 minutes compared to 8 minutes.
Other research concluded that mid-treatment CBCT-guided couch repositioning increased treatment duration and did not have a notable impact on target coverage
; however, treatment times in that study were on average greater than 53 minutes (much longer than our study). The effect of treatment time with regards to the benefits of intrafraction motion correction should be investigated further. Rotations were not accounted for in this margin study. Including them in future studies could provide additional insights, though they would not be expected to change the overall assessment that correcting for intrafraction motion yields smaller calculated margin requirements.
Conclusion
Lower-bound PTV margins for single fraction lung SBRT were calculated as [4.6, 3.5, 2.3] mm for no applied intratreatment couch position shifts and no pretreatment shifts for deviations < 3 mm, [4.6, 2.4, 2.2] mm for no applied intratreatment couch shifts and pretreatment couch shifts for deviations < 3 mm, and [1.7, 1.2, 1.0] mm when performing intratreatment couch shifts. These results suggest that PTV margins can be reduced if CBCT imaging is performed between treatment arcs. However, these calculated results represent lower bounds of the PTV margins, as additional uncertainty and variability factors must be considered. For this patient cohort and immobilization system, movement in the anterior-posterior (AP) dimension was most significant, highlighting the potential benefits of a nonisotropic margin. Applying intrafraction motion correction (scenario 3) reduced the unilateral movement in the AP dimension observed when intrafraction corrections were not applied (scenarios 1 and 2). Moreover, correcting for patient motion in between arcs yielded significantly favorable target and OAR dosimetric impacts compared to not correcting for intrafraction motion.
Conflicts of Interest
The authors declare no conflicts of interest to disclose.
Data Availability
The data that support the findings of this study are available from the corresponding author upon reasonable request.
Acknowledgments
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Local consolidative therapy versus maintenance therapy or observation for patients with oligometastatic non-small-cell lung cancer without progression after first-line systemic therapy: a multicentre, randomised, controlled, phase 2 study.
ITV versus mid-ventilation for treatment planning in lung SBRT: a comparison of target coverage and PTV adequacy by using in-treatment 4D cone beam CT.
A randomized phase II study comparing 2 stereotactic body radiation therapy (SBRT) schedules for medically inoperable patients with stage I peripheral non-small cell lung cancer: NRG Oncology RTOG 0915 (NCCTG N0927).
Int. J. Radiat. Oncol. Biol. Phys.2015; 93: 757-764
Comparison of geometric uncertainties using electronic portal imaging device in focal three-dimensional conformal radiation therapy using different head supports.
Comparison between target margins derived from 4DCT scans and real-time tumor motion tracking: Insights from lung tumor patients treated with robotic radiosurgery.
A randomized phase II study comparing 2 stereotactic body radiation therapy (SBRT) schedules for medically inoperable patients with stage I peripheral non-small cell lung cancer.
Int. J. Radiat. Oncol. Biol. Phys.2015; 93: 757-764
00041-9&id=gr1.jpg){kind=link}
00041-9&id=gr2.jpg){kind=link}
00041-9&id=gr3.jpg){kind=link}
00041-9&id=gr4.jpg){kind=link}