- Email: [email protected]
Weekly Volume and Dosimetric Changes During Chemoradiotherapy With Intensity-Modulated Radiation Therapy for Head and Neck Cancer: A Prospective Observational Study
Int. J. Radiation Oncology Biol. Phys., Vol. 76, No. 5, pp. 1360–1368, 2010 Copyright Ó 2010 Elsevier Inc. Printed in the USA. All rights reserved 0360-3016/10/$–see front matter
doi:10.1016/j.ijrobp.2009.04.005
CLINICAL INVESTIGATION
Head and Neck
WEEKLY VOLUME AND DOSIMETRIC CHANGES DURING CHEMORADIOTHERAPY WITH INTENSITY-MODULATED RADIATION THERAPY FOR HEAD AND NECK CANCER: A PROSPECTIVE OBSERVATIONAL STUDY SHREERANG A. BHIDE, F.R.C.R.,*y MARK DAVIES, B.SC.,z KEVIN BURKE, M.SC.,z HELEN A. MCNAIR, M.SC.,z VIBEKE HANSEN, PH.D.,z Y. BARBACHANO, M.SC.,x I. A. EL-HARIRY, PH.D.,y KATE NEWBOLD, F.R.C.R.,z KEVIN J. HARRINGTON, F.R.C.R.,*y AND CHRISTOPHER M. NUTTING, F.R.C.R.y * Institute of Cancer Research, 237 Fulham Road, London SW6 6JB, United Kingdom; y Head and Neck Unit, Royal Marsden NHS Foundation Trust Hospital, London SW3 6JJ, United Kingdom; z Department of Radiation Oncology, Royal Marsden NHS Foundation Trust Hospital, London and Sutton, United Kingdom; and x Department of Statistics, Royal Marsden NHS Foundation Trust Hospital, London and Sutton, United Kingdom Purpose: The aim of this study was to investigate prospectively the weekly volume changes in the target volumes and organs at risk and the resulting dosimetric changes during induction chemotherapy followed by chemoradiotherapy with intensity-modulated radiation therapy (C-IMRT) for head-and-neck cancer patients. Methods and Materials: Patients receiving C-IMRT for head-and-neck cancer had repeat CT scans at weeks 2, 3, 4, and 5 during radiotherapy. The volume changes of clinical target volume 1 (CTV1) and CTV2 and the resulting dosimetric changes to planning target volume 1 (PTV1) and PTV2 and the organs at risk were measured. Results: The most significant volume differences were seen at week 2 for CTV1 and CTV2. The reductions in the volumes of CTV1 and CTV2 at week 2 were 3.2% and 10%, respectively (p = 0.003 and p < 0.001). The volume changes resulted in a significant reduction in the minimum dose to PTV1 and PTV2 (2 Gy, p = 0.002, and 3.9 Gy, p = 0.03, respectively) and an increased dose range across PTV1 and PTV2 (2.5 Gy, p < 0.001, and 5.1 Gy, p = 0.008, respectively). There was a 15% reduction in the parotid volumes by week 2 (p < 0.001) and 31% by week 4 (p < 0.001). There was a statistically significant increase in the mean dose to the ipsilateral parotid only at week 4 (2.7 Gy, p = 0.006). The parotid glands shifted medially by an average of 2.3 mm (p < 0.001) by week 4. Conclusion: The most significant volumetric changes and dosimetric alterations in the tumor volumes and organs at risk during a course of C-IMRT occur by week 2 of radiotherapy. Further adaptive radiotherapy with replanning, if appropriate, is recommended. Ó 2010 Elsevier Inc. Head and neck cancer, Chemo-IMRT, Volume changes, Dosimetric changes.
INTRODUCTION
dose. Presently, most IMRT plans are generated using data obtained from a single computerized tomography (CT) scan. This presents only a single snapshot of the spatial orientation of the volumes of interest. The spatial relationship is liable to change during the course of treatment for several reasons. The patient might lose weight, the tumor might shrink, and the volume of the OARs (especially parotid glands) might be reduced. When the simultaneousintegrated-boost technique for IMRT is used, sharp dose gradients exist at the margins of target volumes; therefore, the slightest change in anatomy could result in a significant dosimetric change. Previous studies have reported volume changes during head and neck radiotherapy (1-4). However, none of these reports have studied the change in the clinical
The immobilization techniques used for head and neck radiotherapy planning are very effective. Therefore, the principal factors contributing to interfractional errors are the volume changes that occur during the course of the treatment (1, 2). Intensity-modulated radiotherapy (IMRT) permits exquisite radiation dose sculpting around complex target volumes and reduced doses to the organs at risk (OARs). The vital structures in the head and neck region are in close proximity to the tumor volumes. Hence, sharp dose gradients exist at the margins of the target volumes. Therefore, even the slightest change in anatomy can result in target volumes receiving a lower dose or the OARs receiving a higher Reprint requests to: Christopher Nutting, FRCR, Royal Marsden Hospital, Fulham Road, London SW3 6JB, United Kingdom. Tel: (020) 7808 2586; Fax: (020) 7808 2235; E-mail: chris.nutting@ rmh.nhs.uk
This work was presented as an abstract at the 2nd Biennial conference, Innovative Approaches in Head and Neck Oncology (ICHNO). February 26–28, 2009, Barcelona, Spain. Conflict of interest: none. Received Feb 17, 2009, and in revised form March 26, 2009. Accepted for publication April 3, 2009. 1360
Volume changes during head and neck chemo-IMRT d S. A. BHIDE et al.
target volume (CTV) and the resultant dosimetric implications (the dose to the planning target volume [PTV]) in patients with head and neck cancer (HNC) receiving induction chemotherapy, followed by concomitant chemotherapy and IMRT (chemo-IMRT). This study investigates the volumetric changes that the CTV and the parotid glands undergo during a course of radical chemoradiation therapy and the resulting dosimetric alteration. In addition, we quantified the shift in position of the parotid glands during treatment. The study also attempted to correlate the volume changes to patients’ weight and to generate a future strategy for adaptive IMRT for head-and-neck treatment. METHODS AND MATERIALS Patients undergoing induction chemotherapy followed by chemoIMRT (sequential treatment) for HNC were included in this prospective observational study. The study was approved by the Committee for Clinical Research (CCR protocol no. 2909) and the Research and Ethics Committee (REC protocol no. 07/Q0801/29).
Radiotherapy planning Each patient was immobilized on a head support pad by using a customized head-and-shoulder shell attached to the head board, using a positive fixation mechanism with four fixation points, two on either side of the head and two on either side of the shoulders (5). Contrast-enhanced CT scans were performed at 2.5-mm intervals. The scan was performed during the inpatient stay for the second cycle of induction chemotherapy. One anterior and two lateral radiopaque markers were placed on the shell for subsequent isocenter verifications. Gross tumor volume (GTV) was outlined for each patient based on visible tumor and information from the (CT and MRI) diagnostic images and clinical examination. CTV1 was the GTV and involved lymph nodes with at least a 1.5-cm margin to account for microscopic tumor spread. The margins were reduced at the borders of an uninvolved anatomic space (6). CTVs for macroscopic disease (CTV1) and areas at risk of harboring microscopic disease, i.e., lymph node areas II to V not containing gross disease (CTV2) and critical structures (parotid glands, spinal cord, and brain stem) were outlined on each CT slice (7, 8). MR images were used to assist the definition of target volumes, where appropriate. The lymph node areas were defined according to international consensus guidelines (7, 8). PTVs 1 and 2 were generated by adding a margin of 3mm to CTVs 1 and 2 respectively. (5). The PTVs were edited to ensure that they did not come within 5 mm of the skin, and PTV2 was edited out of the PTV1 to avoid conflicting dose constraints. A Helios inverse planning module in Eclipse was used for all IMRT planning. The IMRT planning technique used at our institution has been described in detail in a previously published study (9). The dose objectives for PTV1 and PTV2 were 65 to 70 Gy and 54 to 56 Gy, respectively. Doses were prescribed to the median dosevolume point on the PTV1 dose-volume histogram (DVH) and PTV2 DVH. The following dose constraints were set on the OARs: spinal cord maximum dose, 48 Gy; brain stem maximum dose, 55 Gy; contralateral parotid, 26 Gy; and combined superficial parotid volume, 26 Gy.
Chemotherapy Two cycles (every 21 days) of cisplatin (75 mg/m2) on day 1 and 5-fluorouracil (1,000 mg/m2) on days 1 to 4 was the standard regi-
1361
men used for induction chemotherapy (10). Cisplatin (100 mg/m2) on days 1 and 29 was used for concomitant treatment.
Trial intervention Repeat CT scans (without contrast) were performed at weeks 2, 3, 4, and 5 during radiotherapy (i.e, at 8, 9, 10, and 11 weeks after the start of induction chemotherapy). The effective radiation dose from each scan was 70 mSv. The original position was set using radiopaque markers. The same GTV was contoured on the subsequent scans in relation to fixed anatomical points, with the original planning CT scan as reference (1). CTV1, CTV2, and OARs were reoutlined on each weekly scan and were modified based on the changes in the patient’s anatomy. All the contouring was performed by a single observer (SB) and validated by two other observers (KJH and CMN). This was done to eliminate interobserver variability. The single observer who contoured all the scans was tested for intraobserver variability. The observer delineated the target volumes and the OARs on copies of a single CT scan at weekly intervals for 3 weeks. The variability in the contouring was then quantified as differences between volumes contoured each time. The PTVs were defined as on the original treatment plan. The original treatment plan was transferred onto each of the weekly scans. The alignment was performed using the radiopaque markers, which indicated the marked treatment isocenter. The marked isocenter included any isocenter deviations (from the original treatment plan) that were corrected during the weekly isocenter verification, using portal imaging according to institutional protocol. This method of alignment has been validated for target coverage and setup error variation by O’Daniel et al. (11). The dose distribution and DVHs were generated for each of the weekly scans. Since this was an observational study, new treatment plans were deemed to be experimental, and patients continued to be treated on the original treatment plan. The shift in the position of the parotid glands was calculated as follows. Three fixed bony landmarks were chosen, namely the tip of the mastoid process, the tip of the styloid process, and the inferior end of the lateral pterygoid plate. The coordinates of the center of mass (COM) of each of the parotid glands were documented. The average of the x, y, and z coordinates for each of the bony points generated the coordinates for a hypothetical central bony reference point. The shift of the COM of each parotid gland in each plane was calculated by measuring its distance from the coordinates of the above-mentioned bony point on each of the CT scans. The patient’s weight was documented on a weekly basis.
Statistical analysis In their study with 14 patients, Barker et al. showed that the GTV was reduced by 1.7 to 1.8% per treatment day (1). Assuming that we might have expected to see a similar percent change, 20 patients would have enabled us to estimate the reduction in volume per treatment day to within 0.3% (power 95% = 5%). Hansen et al. (2) reported a reduction in the dose to the PTVGTV D95 of 2.2 Gy (95% confidence interval [CI], 1.4–1.5 Gy). If the variability across patients was similar, we anticipated being able to determine the change in PTV to within approximately 1.2 Gy (power 95% = 5%, 95% CI). Statistical analysis was performed with SPSS 2006 version 15.0 software (SPSS Inc. Chicago, IL). The reduction in volumes of the CTVs and parotid gland; changes in the dose to the target volumes, parotid gland, spinal cord, and brain stem; movement of the parotid glands; and changes in weight were analyzed for each patient and were summarized as means with 95% CI. The volume change and the dosimetric alterations were analyzed together for all the
1362
I. J. Radiation Oncology d Biology d Physics
parotid glands and then separately for the ipsilateral and contralateral glands. The Kolmogorov-Smirnov test was used to test data for normality. Since all the data, except for the maximum dose to the brain stem, were normally distributed, parametric tests were used for testing hypotheses. The Wilcoxon sign ranked test was used as a test of significance for changes in maximum dose to the brain stem. Comparisons were made between the volumes and plans on the scan at a particular time point and the scan performed on the following week, i.e., pretreatment with week 2, week 2 with week 3, and so on. The Bonferroni correction was applied, and an a value was set at 0.012 instead of 0.05. To assess the relationship between weight loss and volume change, the percent weight loss and the percent volume reduction for each patient at each time point were calculated. The correlation between weight loss and volume reduction was estimated using Pearson’s correlation.
RESULTS Twenty patients were entered into the study between August 2007 and May 2009. The concomitant chemo-IMRT commenced 6 weeks after the start of induction chemotherapy. One out of the 20 patients had cisplatin replaced by carboplatin for the final concomitant treatment, due to renal toxicity. The patients completed the treatment protocol without any interruptions. Compliance with the study protocol was excellent, with 19 patients having CT scans according to the protocol. One patient missed the CT scan at week 3, as he was not fit to attend. The patients’ TNM staging distribution is given in Table 1. Ten patients had oropharyngeal cancer, 6 patients had laryngeal/hypopharyngeal cancer, and 4 patients had nasopharyngeal cancer. Intraobserver variability The observed percent difference in volumes of the volumes contoured by the observer at different time points was not significant (data not shown; p > 0.05). Volumetric changes The pretreatment volumes were labeled week 0. The changes in volume were normalized to the percent reduction in volume. The reduction was calculated with respect to the week 0 volumes, which were defined as 100%. CTV1 Comparing the changes in volume of CTV1 at each week with that of the preceding week showed that the greatest reduction in mean absolute volume and mean percent volume was 11 cm3 (95% CI, 3-14) and 3.2% (95% CI, 0.74-5.6), respectively, between week 0 and week 2 (Table 2 and Fig. 1). This difference is statistically significant (p = 0.003). The absolute and percent reductions in the next two-week period, i.e., between week 2 and 4 were, respectively, 7 cm3 (95% CI, 1-13) and 2% (95% CI, 0.1 to 4), which were not statistically significant (p = 0.051). CTV2 Comparing the changes in volume at each week with that of the preceding week showed that the greatest reduction in
Volume 76, Number 5, 2010
Table 1. TNM staging distribution of the trial patients No. of patients with node stage Tumor stage T1 T2 T3 T4 Total no. of patients
N0
N1
N2a
N2b
N2c
N3
Total no. of patients
0 1 2 3 6
1 1 0 0 2
0 0 1 0 1
1 2 2 1 6
0 1 1 2 4
0 0 0 1 1
2 5 6 7 20
mean absolute volume and mean percent volume was 21.6 cm3 (95% CI, 4-18) and 10.5% (95% CI, 7-14), respectively, between week 0 and week 2 (Table 2 and Fig. 2). This difference is statistically significant (p < 0.001). The absolute and percent reduction in the next two-week period, i.e., between week 2 and 4, was, respectively, 12 cm3 (95% CI, 6-17) and 5% (95% CI, 3-8), which was statistically significant (p < 0.001). Parotid glands The greatest absolute and percent reduction in the volume of the parotid glands was 4.2 cm3 (95% CI, 3-5.5) and 14.7% (95% CI, 10-19), respectively, and occurred between week 0 and week 2 (Table 3). This difference was statistically significant (p < 0001). Figure 4 shows the percent reduction if plotted separately for the ipsilateral and the contralateral parotid glands. It shows that the ipsilateral gland shrinks more than the contralateral gland; however, the difference was not statistically significant. The absolute and percent reduction in the next two-week period, i.e., between week 2 and 4, was, respectively, 4 cm3 (95% CI, 3-5) and 16% (95% CI, 16-19), which was statistically significant (p < 0.001). Dosimetric alterations During the critical appraisal of an IMRT treatment plan, two dose parameters were considered most relevant. The first parameter was the minimum dose to the PTV, which Table 2. Comparison of reduction in weekly absolute and relative CTV1 and CTV2*
Target volume
Weekly comparison
Mean absolute difference (cm3)
CTV1
Pre-Rx vs. week 2 Weeks 2 and 3 Weeks 3 and 4 Weeks 4 and 5 Pre-Rx - week 2 Weeks 2 and 3 Weeks 3 and 4 Weeks 4 and 5
11.0 5.6 1.6 8.5 21.6 6.4 5.7 0.58
CTV2
Percent reduction
p valuey
3.2 2.1 0.16 2.5 10.5 2.5 3.4 0.30
0.003 0.020 0.538 0.012 <0.001 0.031 0.034 0.78
* Data are paired t test results comparing the reduction in absolute and relative volumes each week with respect to the volumes on the preceding scan for CTV1 and CTV2. y Result is significant if p is <0.012.
Volume changes during head and neck chemo-IMRT d S. A. BHIDE et al.
1363
Table 3. Comparison of weekly changes in minimum doses for PTV1 and PTV2*
Volume
Mean reduction Mean compared to Treatment minimum preceding plan dose (Gy) week (Gy) p valuey
PTV1 (treatment Week 2 plan, 61.1 Gy) Week 3 Week 4 Week 5 PTV 2 (treatment Week 2 plan, 50.5 Gy) Week 3 Week 4 Week 5
59.1
2.0
0.002
58.3 57.8 58 46.6
1.0 0.64 0.18 3.9
0.180 0.280 0.768 0.03
46.1 47.2 46.5
0.4 1 0.7
0.75 0.39 0.66
* Paired t test comparing the change in minimum dose each week with respect to the preceding week for PTV1 and PTV2. y Result is significant if p is <0.012. Fig. 1. The mean percentage of change in volume of CTV1 with error bars (95% CIs).
quantified the volume receiving a clinically suboptimal dose. The second parameter was the dose range, which indicates the homogeneity of the dose distribution. PTV1 The minimum dose to PTV1 and PTV2 was reduced through the course of the study. When the dose reduction was compared to that of the preceding week for PTV1, the significant difference in dose (2 Gy [95% CI, 0.82-3.2]) was between week 0 and week 2 (p = 0.002). The reduction in the minimum dose to PTV2 was not statistically significant (Table 3). The dose range across PTV1 increased throughout the study period, indicating reducing dose homogeneity
Fig. 2. The mean percentage of change in volume of CTV2 with error bars (95% CIs).
(Table 4). The increase was greatest (2.5 Gy [95% CI, 1.33.8]) between the treatment plan and week 2 (p < 0.001). The corresponding increase in the dose range for PTV2 in the same time interval was 5.1 Gy (95% CI, 1.5-8.8), and this was statistically significant (p = 0.008). There were no significant differences between the minimum dose and the dose range to PTV1 and PTV2 between weeks 2 and 4. Parotid glands The mean dose to the parotid glands increased throughout the study period. However, this increase was not statistically significant (Table 5). The mean dose to the ipsilateral parotid gland was higher than that to the contralateral parotid gland at each time point (Table 5). This is in keeping with the constraints set on each of these glands. The mean dose to the
Fig. 3. The mean percentage of change in volume of the parotid gland with error bars (95%CIs of the means).
I. J. Radiation Oncology d Biology d Physics
1364
Volume 76, Number 5, 2010
Table 5. Weekly increase in mean dose for parotid glands*
Parotid volume tested Combined (treatment plan, 31.7 Gy)
Week 2
32.6
0.9
0.17
Week 3 Week 4 Week 5 Week 2
33 34.3 33.6 38.6
0.3 1.3 -0.7 0.2
0.53 0.01 0.19 0.83
Week 3 Week 4 Week 5 Contralateral (treatment Week 2 plan, 24.3 Gy) Week 3 Week 4 Week 5
39.3 41.1 41.2 26
0.7 2.7 2.8 1.7
0.39 0.006 0.01 0.11
26.2 26.8 25.3
1.9 2.5 1.0
0.07 0.03 0.29
Ipsilateral (treatment plan, 38.4 Gy)
Fig. 4. The change in the percentage of reduction in volume between the week 0 scans and the subsequent scans for ipsilateral (solid squares) and the contralateral (solid circles) glands. Errors bars indicate 95% CI.
ipsilateral parotid gland was significantly higher by 2.7 Gy (95% CI, 0.8-4.4) and 2.8 Gy (95% CI, 0.5-4.9) at weeks 4 and 5, respectively, compared to the original treatment plan (Table 5). There was no significant change in the dose to the contralateral gland (Table 5). There was no significant difference between the mean dose to the parotid glands between weeks 2 and 4. There was a significant medial shift of the parotid glands through the course of treatment, starting at week 2 (p = 0.002). The highest mean movement of the COM was 2.3 mm at week 4 (p < 0.0001). The movement of the COM in the anteroposterior and the inferosuperior directions was not significant.
Table 4. Comparison of weekly dose range increases across PTV1 and PTV2*
Volume PTV1 (treatment plan, 7 Gy)
PTV2 (treatment plan, 11.8 Gy)
Mean reduction Mean compared to Treatment dose preceding plan range (Gy) week (Gy) p valuey Week 2
9.5
2.5
<0.001
Week 3 Week 4 Week 5 Week 2
10 11.3 11 16.9
0.5 1.3 0.3 5.1
0.710 0.069 0.698 0.008
Week 3 Week 4 Week 5
17.1 16.5 16.9
0.2 0.4 0.4
0.929 0.680 0.798
* Paired t test comparing the increase in dose range each week with respect to the preceding week across PTV1 and PTV2. y Result is significant if p is <0.012.
Mean increase Average compared to mean treatment Week dose (Gy) plan (Gy) p valuey
* Paired t test showing the increase in mean dose for the parotid glands at each week with reference to baseline. y Result is significant if p is <0.012.
Spinal cord and brain stem The differences in the maximum doses to the spinal cord and brain stem, when comparing the plan at each week with that of the preceding week, were not significant (Table 6). The increase in cord dose was statistically significant at weeks 4 (p = 0.004) and 5 (p = 0.003) compared to week 0. However the upper limit of the 95% CI on average was lower than the set constraint of 48 Gy (Table 6). Cumulative plan analysis A cumulative plan consisting of the dosimetric change at each week multiplied by the time fraction was calculated and compared to the original plan and also to the dosimetric changes at week 2 (Table 7). This comparison showed that there was a statistically significant difference for the minimum (lower) dose and the dose range (higher) for PTV1 and PTV2 between the original and the cumulative plan. There difference between the plan at week 2 and the cumulative plan was not statistically significant. Similar results were observed for the mean doses to the parotids and the maximum doses to the spinal cord and the brain stem. Weight loss The mean weight loss (95% CI) during treatment was 7.5 kg (3.1 kg). The mean relative weight loss was 9.7% (3.5%) of the original weight. There was no significant correlation between the volume reduction in CTV1 or CTV2 and weight loss. DISCUSSION This is the first study that has investigated the volume changes and the resulting dosimetric implications in patients undergoing sequential treatment for HNC. This study has
Volume changes during head and neck chemo-IMRT d S. A. BHIDE et al.
1365
Table 6. Weekly differences between average maximum doses to OARs Week Organ
Dose*
0
2
3
4
5
Spinal cord
Average maximum dose (Gy) (95% CI) p value Average maximum dose (Gy) (95% CI) p value
44.2 (43-45)
46 (44-47)
45.3 (43-47)
46.3 (44-47)
46.6 (44-48)
46.5 (41-51)
0.05 48.2 (42-53)
0.9 48.5 (43-53)
0.3 48.5 (42-54)
0.5 48.9 (43-54)
0.03
0.7
0.4
0.6
Brain stem
* Result is significant if p is <0.012
demonstrated that the CTVs undergo significant reduction during the course of head and neck IMRT. In a comparison of the change at each time point with respect to the preceding week, the maximum difference was seen between the original treatment scan and the scan at week 2 (i.e., 8 weeks after the first induction chemotherapy) for CTV1 and CTV2. The volume of CTV1 decreased by 3.2% at week 2 and that of CTV2 decreased by 10.5%. The reduction in volume of CTV2 between weeks 2 and 4 was smaller (5%) but statistically significant. The planning scans were performed at the time of administration of the second cycle of induction chemotherapy, i.e., minus day 21 of chemoradiotherapy. The patients had one cycle of induction chemotherapy, the first cycle of concomitant cisplatin, and 10 fractions of radiotherapy between the planning scan and the scan at week 2. This could explain the significant reduction in volume of CTV1 by week 2 of the chemoradiation. However, this does not account for the significant reduction in volume of CTV2 by week 2, as the treatment received by the patient should not have an effect on the volume of CTV2. The reduction in the volume of CTV2 is likely to be due to the anatomical changes taking place in the head and neck region, which is probably due to the loss of adipose tissue in the neck. This is corroborated by further reduction in volume between the next two-week period. The patients could lose weight as a result of receiving induction chemotherapy, which in turn could reduce the amount of adipose tissue in the neck. However, there was no correlation between the weight loss and the change in CTV2, indicating that weight loss probably is not
a sensitive indicator of loss of adipose tissue in the neck. The planning scans could not be performed closer to the start of radiotherapy due to the additional time required for planning and quality assurance of the complex IMRT plans. In retrospect, when this study was designed, provision should have been made for a scan at the start of radiotherapy, which would have helped to elucidate the effects of the second cycle of induction chemotherapy and to differentiate these effects from the effect of chemoradiation. Although we found the intraobserver variability to be nonsignificant, this could still account for the changes in the volumes due to the variability of contouring between different CT sets. The average minimum doses to PTV1 and PTV2 were significantly reduced by 2 Gy and 3.9 Gy, respectively, at week 2 (Table 3). The dose ranges across PTV1 and PTV2 increased by 2.5 Gy and 5.1 Gy, respectively, at week 2 (Table 4). The change for PTV2 was greater, in line with the greater volume loss. The dosimetric changes between weeks 2 and 4 were not significant, probably due to the smaller volume reduction (5% vs. 10.5%, respectively) compared to week 0 and week 2. The dose of the cumulative doses was also significantly different compared to the dose of the original plan. There was no significant difference between the dosimetric changes on the cumulative plan and the plan based on the volume changes at week 2. This indicates that dosimetric alterations resulting from volumetric changes could be minimized by replanning at week 2. It is difficult to comment on the clinical significance of the dosimetric alterations because the data for the alterations were obtained from the
Table 7. Comparison of doses on the cumulative plan, the original plan, and the scan at week 2 Original vs. cumulative plan (time factor correction)
Average dose (Gy)
Week 2 vs. cumulative (time factor correction)
Treatment plan
Original
Week 2
Cumulative
Difference (Gy)
p value
Difference (Gy)
p value
PTV1 minimum PTV1 dose range PTV2 minimum PTV2 dose range Parotid mean Spinal cord maximum Brain stem maximum
61.3 7 50.5 11.8 31.7 44.2 46.5
59.1 9.5 48 16.9 32.6 46 48.2
59.3 9.8 47.9 15.2 32.9 45.4 48.5
1.8 2.8 2.6 3.5 1.2 1.2 2
<0.001 <0.001 0.006 0.001 0.003 0.006 0.004
0.2 0.3 0.1 1.7 0.3 0.6 0.3
0.6 0.6 0.1 0.08 0.6 0.6 0.3
1366
I. J. Radiation Oncology d Biology d Physics
Volume 76, Number 5, 2010
Fig. 5. An example of the volume reduction and the change in dose for the parotid glands (right parotid gland [peach colored] and left parotid gland [yellow]) at different time points. The dark blue line is the 30-Gy isodose line. The dark blue volume is PTV1, and the light blue volume is PTV2. (A) Original plan; (B) week 2; (C) week 3; (D) week 4; (E) week 5.
DVHs, which fail to give an estimate of the spatial aspects of the dose distribution. One possibility is that any areas receiving a lower minimum dose in PTV1 will result in a lower dose to either macroscopic or microscopic disease, resulting in an increased risk of recurrence. This would result if the predominant changes in the patient’s anatomy were asymmetrical. If the changes were symmetrical (due to tumor shrinkage), PTV1 would have been adequately covered by the 95% isodose. However, this would have resulted in an increase in the
minimum dose to PTV1 and not the reduction seen in this study. The dosimetric changes seen in this study are likely to be due to the changes in the patients’ anatomy. The parotid volumes were reduced by 35% through the course of treatment (Table 5). There was a significant increase in mean dose to the ipsilateral gland (2.7 Gy) at week 4 (Table 5). This could be due to CTV1 shrinkage and medial movement of the gland into the high-dose region. The study of the parotid coordinates shows that there is
Volume changes during head and neck chemo-IMRT d S. A. BHIDE et al.
significant medial shift of the parotid glands (0.23 cm by week 4) through the treatment period. Only one study (2), with 13 patients, investigated the volumetric and dosimetric changes in CTVs and OARs in patients treated for HNC with IMRT (no concomitant chemotherapy). The patients were rescanned when there was a clinically observed change in anatomy or significant weight loss. The study demonstrated significant change in volumes of the CTV and parotid glands and significant underdosage to target volumes and overdosage to the OARs. The dosimetric changes were similar to those seen in our study. However, patients were scanned only once and only if that was clinically indicated. In a similar study, Barker et al. scanned 14 patients with tumors or lymph nodes $4 cm in diameter during radiotherapy (1). Only one patient was treated with IMRT. Twelve patients received concurrent chemotherapy. There was a reduction in the GTV at a median rate of 1.8% per treatment day, which was asymmetrical. At the end of treatment, the median parotid volume loss was 28.1% (range, 5.9– 53.6%). The median medial shift of the parotid glands was 3 mm (range, 0.3-9 mm). The median weight loss was 7.1%. The authors did not comment on the dosimetric alterations. That study was limited to patients who had significant lymphadenopathy. It is of clinical relevance to investigate the change in the CTVs based on the patient’s anatomic changes, as in the study by Hansen et al. (2) and our study. This is because there is a lack of evidence for the use of shrinking CTVs generated from reduced GTV for radical treatment of HNC. Ballivy et al. studied the dosimetric implications of weekly volume changes in 8 patients (2 had concomitant chemotherapy) undergoing IMRT for HNC (12). The authors found the differential dosimetric changes in the upper versus the lower neck due to the change in the curvature of the neck. They concluded that a greater PTV margin might be required to account for this, especially when the isocenter is positioned in the lower neck.
1367
The volume changes and the net medial shift reported in the published studies are comparable to those of our study (1-4). However, the mean doses to the parotids when analyzed together did not show a significant increase in mean dose, as would be expected from a net medial shift of the COM, which would move the gland into the high-dose region. Studies by Robar et al. (3) and Vasquez Osorio et al. (4) show that the lateral regions of the glands move during radiotherapy while the medial regions are relatively fixed. This could have occurred in our study, resulting in a net movement of the COM of the parotid glands due to the excessive movement of the lateral region, with the medial region being fixed. Our practice of applying additional constraints on the superficial glands resulted in higher doses to the medial region of the gland. Therefore lateral areas of the gland could have moved medially in the low-dose region, resulting in a nonsignificant increase in the mean dose. CONCLUSIONS In summary, our study is unique in having investigated the serial change in volume and the resulting dosimetric changes in the target volumes and the OARs during head-and-neck chemo-IMRT. The most significant volume change and dose alteration occurs 2 weeks after commencing radiotherapy (i.e., 8 weeks after the start of induction chemotherapy). This could be due to the effect of induction chemotherapy, the first cycle of concomitant chemotherapy, and 10 fractions of radiotherapy, especially in the case of CTV1. In retrospect, this study should have included a CT scan at the start of radiotherapy. This study also shows significant dosimetric implications of the volume change. The sharp dose gradients that exist in the IMRT plans necessitate minimizing dosimetric uncertainties in order to reduce the risk of a geographical miss. An adaptive radiotherapy approach by rescanning the patients, recontouring the target volumes and OARs, and applying the original treatment plan and replanning if necessary is required.
REFERENCES 1. Barker JL Jr., Garden AS, Ang KK, O’Daniel JC, Wang H, Court LE, et al. Quantification of volumetric and geometric changes occurring during fractionated radiotherapy for headand-neck cancer using an integrated CT/linear accelerator system. Int J Radiat Oncol Biol Phys 2004;59(4):960–970. 2. Hansen EK, Bucci MK, Quivey JM, Weinberg V, Xia P. Repeat CT imaging and replanning during the course of IMRT for headand-neck cancer. Int J Radiat Oncol Biol Phys 2006;64(2): 355–362. 3. Robar JL, Day A, Clancey J, Kelly R, Yewondwossen M, Hollenhorst H, et al. Spatial and dosimetric variability of organs at risk in head-and-neck intensity-modulated radiotherapy. Int J Radiat Oncol Biol Phys 2007;68(4):1121–1130. 4. Vasquez Osorio EM, Hoogeman MS, Al-Mamgani A, Teguh DN, Levendag PC, Heijmen BJ. Local anatomic changes in parotid and submandibular glands during radiotherapy for oropharynx cancer and correlation with dose, studied in detail with nonrigid registration. Int J Radiat Oncol Biol Phys 2008; 70(3):875–882.
5. Humphreys M, Guerrero Urbano MT, Mubata C, Miles E, Harrington KJ, Bidmead M, et al. Assessment of a customised immobilisation system for head and neck IMRT using electronic portal imaging. Radiother Oncol 2005;77(1): 39–44. 6. Guerrero Urbano MT, Clark CH, Kong C, Miles E, Dearnaley DP, Harrington KJ, et al. Target volume definition for head and neck intensity modulated radiotherapy: Pre-clinical evaluation of PARSPORT trial guidelines. Clin Oncol 2007; 19(8):604–613. 7. Gregoire V, Eisbruch A, Hamoir M, Levendag P. Proposal for the delineation of the nodal CTV in the node-positive and the post-operative neck. Radiother Oncol 2006;79(1): 15–20. 8. Gregoire V, Levendag P, Ang KK, Bernier J, Braaksma M, Budach V, et al. CT-based delineation of lymph node levels and related CTVs in the node-negative neck: DAHANCA, EORTC, GORTEC, NCIC, RTOG consensus guidelines. Radiother Oncol 2003;69(3):227–236.
1368
I. J. Radiation Oncology d Biology d Physics
9. Bhide S, Clark C, Harrington K, Nutting CM. Intensity modulated radiotherapy improves target coverage and parotid gland sparing when delivering total mucosal irradiation in patients with squamous cell carcinoma of head and neck of unknown primary site. Med Dosim 2007;32(3):188–195. 10. Bhide SA, Ahmed M, Barbachano Y, Newbold K, Harrington KJ, Nutting CM. Sequential induction chemotherapy followed by radical chemo-radiation in the treatment of locoregionally advanced head-and-neck cancer. Br J Cancer 2008;99(1):57–62.
Volume 76, Number 5, 2010
11. O’Daniel JC, Garden AS, Schwartz DL, Wang H, Ang KK, Ahamad A, et al. Parotid gland dose in intensity-modulated radiotherapy for head and neck cancer: Is what you plan what you get? Int J Radiat Oncol Biol Phys 2007;69(4): 1290–1296. 12. Ballivy O, Parker W, Vuong T, Shenouda G, Patrocinio H. Impact of geometric uncertainties on dose distribution during intensity modulated radiotherapy of head-and-neck cancer: the need for a planning target volume and a planning organat-risk volume. Curr Oncol 2006;13(3):108–115.
doi:10.1016/j.ijrobp.2009.04.005
CLINICAL INVESTIGATION
Head and Neck
WEEKLY VOLUME AND DOSIMETRIC CHANGES DURING CHEMORADIOTHERAPY WITH INTENSITY-MODULATED RADIATION THERAPY FOR HEAD AND NECK CANCER: A PROSPECTIVE OBSERVATIONAL STUDY SHREERANG A. BHIDE, F.R.C.R.,*y MARK DAVIES, B.SC.,z KEVIN BURKE, M.SC.,z HELEN A. MCNAIR, M.SC.,z VIBEKE HANSEN, PH.D.,z Y. BARBACHANO, M.SC.,x I. A. EL-HARIRY, PH.D.,y KATE NEWBOLD, F.R.C.R.,z KEVIN J. HARRINGTON, F.R.C.R.,*y AND CHRISTOPHER M. NUTTING, F.R.C.R.y * Institute of Cancer Research, 237 Fulham Road, London SW6 6JB, United Kingdom; y Head and Neck Unit, Royal Marsden NHS Foundation Trust Hospital, London SW3 6JJ, United Kingdom; z Department of Radiation Oncology, Royal Marsden NHS Foundation Trust Hospital, London and Sutton, United Kingdom; and x Department of Statistics, Royal Marsden NHS Foundation Trust Hospital, London and Sutton, United Kingdom Purpose: The aim of this study was to investigate prospectively the weekly volume changes in the target volumes and organs at risk and the resulting dosimetric changes during induction chemotherapy followed by chemoradiotherapy with intensity-modulated radiation therapy (C-IMRT) for head-and-neck cancer patients. Methods and Materials: Patients receiving C-IMRT for head-and-neck cancer had repeat CT scans at weeks 2, 3, 4, and 5 during radiotherapy. The volume changes of clinical target volume 1 (CTV1) and CTV2 and the resulting dosimetric changes to planning target volume 1 (PTV1) and PTV2 and the organs at risk were measured. Results: The most significant volume differences were seen at week 2 for CTV1 and CTV2. The reductions in the volumes of CTV1 and CTV2 at week 2 were 3.2% and 10%, respectively (p = 0.003 and p < 0.001). The volume changes resulted in a significant reduction in the minimum dose to PTV1 and PTV2 (2 Gy, p = 0.002, and 3.9 Gy, p = 0.03, respectively) and an increased dose range across PTV1 and PTV2 (2.5 Gy, p < 0.001, and 5.1 Gy, p = 0.008, respectively). There was a 15% reduction in the parotid volumes by week 2 (p < 0.001) and 31% by week 4 (p < 0.001). There was a statistically significant increase in the mean dose to the ipsilateral parotid only at week 4 (2.7 Gy, p = 0.006). The parotid glands shifted medially by an average of 2.3 mm (p < 0.001) by week 4. Conclusion: The most significant volumetric changes and dosimetric alterations in the tumor volumes and organs at risk during a course of C-IMRT occur by week 2 of radiotherapy. Further adaptive radiotherapy with replanning, if appropriate, is recommended. Ó 2010 Elsevier Inc. Head and neck cancer, Chemo-IMRT, Volume changes, Dosimetric changes.
INTRODUCTION
dose. Presently, most IMRT plans are generated using data obtained from a single computerized tomography (CT) scan. This presents only a single snapshot of the spatial orientation of the volumes of interest. The spatial relationship is liable to change during the course of treatment for several reasons. The patient might lose weight, the tumor might shrink, and the volume of the OARs (especially parotid glands) might be reduced. When the simultaneousintegrated-boost technique for IMRT is used, sharp dose gradients exist at the margins of target volumes; therefore, the slightest change in anatomy could result in a significant dosimetric change. Previous studies have reported volume changes during head and neck radiotherapy (1-4). However, none of these reports have studied the change in the clinical
The immobilization techniques used for head and neck radiotherapy planning are very effective. Therefore, the principal factors contributing to interfractional errors are the volume changes that occur during the course of the treatment (1, 2). Intensity-modulated radiotherapy (IMRT) permits exquisite radiation dose sculpting around complex target volumes and reduced doses to the organs at risk (OARs). The vital structures in the head and neck region are in close proximity to the tumor volumes. Hence, sharp dose gradients exist at the margins of the target volumes. Therefore, even the slightest change in anatomy can result in target volumes receiving a lower dose or the OARs receiving a higher Reprint requests to: Christopher Nutting, FRCR, Royal Marsden Hospital, Fulham Road, London SW3 6JB, United Kingdom. Tel: (020) 7808 2586; Fax: (020) 7808 2235; E-mail: chris.nutting@ rmh.nhs.uk
This work was presented as an abstract at the 2nd Biennial conference, Innovative Approaches in Head and Neck Oncology (ICHNO). February 26–28, 2009, Barcelona, Spain. Conflict of interest: none. Received Feb 17, 2009, and in revised form March 26, 2009. Accepted for publication April 3, 2009. 1360
Volume changes during head and neck chemo-IMRT d S. A. BHIDE et al.
target volume (CTV) and the resultant dosimetric implications (the dose to the planning target volume [PTV]) in patients with head and neck cancer (HNC) receiving induction chemotherapy, followed by concomitant chemotherapy and IMRT (chemo-IMRT). This study investigates the volumetric changes that the CTV and the parotid glands undergo during a course of radical chemoradiation therapy and the resulting dosimetric alteration. In addition, we quantified the shift in position of the parotid glands during treatment. The study also attempted to correlate the volume changes to patients’ weight and to generate a future strategy for adaptive IMRT for head-and-neck treatment. METHODS AND MATERIALS Patients undergoing induction chemotherapy followed by chemoIMRT (sequential treatment) for HNC were included in this prospective observational study. The study was approved by the Committee for Clinical Research (CCR protocol no. 2909) and the Research and Ethics Committee (REC protocol no. 07/Q0801/29).
Radiotherapy planning Each patient was immobilized on a head support pad by using a customized head-and-shoulder shell attached to the head board, using a positive fixation mechanism with four fixation points, two on either side of the head and two on either side of the shoulders (5). Contrast-enhanced CT scans were performed at 2.5-mm intervals. The scan was performed during the inpatient stay for the second cycle of induction chemotherapy. One anterior and two lateral radiopaque markers were placed on the shell for subsequent isocenter verifications. Gross tumor volume (GTV) was outlined for each patient based on visible tumor and information from the (CT and MRI) diagnostic images and clinical examination. CTV1 was the GTV and involved lymph nodes with at least a 1.5-cm margin to account for microscopic tumor spread. The margins were reduced at the borders of an uninvolved anatomic space (6). CTVs for macroscopic disease (CTV1) and areas at risk of harboring microscopic disease, i.e., lymph node areas II to V not containing gross disease (CTV2) and critical structures (parotid glands, spinal cord, and brain stem) were outlined on each CT slice (7, 8). MR images were used to assist the definition of target volumes, where appropriate. The lymph node areas were defined according to international consensus guidelines (7, 8). PTVs 1 and 2 were generated by adding a margin of 3mm to CTVs 1 and 2 respectively. (5). The PTVs were edited to ensure that they did not come within 5 mm of the skin, and PTV2 was edited out of the PTV1 to avoid conflicting dose constraints. A Helios inverse planning module in Eclipse was used for all IMRT planning. The IMRT planning technique used at our institution has been described in detail in a previously published study (9). The dose objectives for PTV1 and PTV2 were 65 to 70 Gy and 54 to 56 Gy, respectively. Doses were prescribed to the median dosevolume point on the PTV1 dose-volume histogram (DVH) and PTV2 DVH. The following dose constraints were set on the OARs: spinal cord maximum dose, 48 Gy; brain stem maximum dose, 55 Gy; contralateral parotid, 26 Gy; and combined superficial parotid volume, 26 Gy.
Chemotherapy Two cycles (every 21 days) of cisplatin (75 mg/m2) on day 1 and 5-fluorouracil (1,000 mg/m2) on days 1 to 4 was the standard regi-
1361
men used for induction chemotherapy (10). Cisplatin (100 mg/m2) on days 1 and 29 was used for concomitant treatment.
Trial intervention Repeat CT scans (without contrast) were performed at weeks 2, 3, 4, and 5 during radiotherapy (i.e, at 8, 9, 10, and 11 weeks after the start of induction chemotherapy). The effective radiation dose from each scan was 70 mSv. The original position was set using radiopaque markers. The same GTV was contoured on the subsequent scans in relation to fixed anatomical points, with the original planning CT scan as reference (1). CTV1, CTV2, and OARs were reoutlined on each weekly scan and were modified based on the changes in the patient’s anatomy. All the contouring was performed by a single observer (SB) and validated by two other observers (KJH and CMN). This was done to eliminate interobserver variability. The single observer who contoured all the scans was tested for intraobserver variability. The observer delineated the target volumes and the OARs on copies of a single CT scan at weekly intervals for 3 weeks. The variability in the contouring was then quantified as differences between volumes contoured each time. The PTVs were defined as on the original treatment plan. The original treatment plan was transferred onto each of the weekly scans. The alignment was performed using the radiopaque markers, which indicated the marked treatment isocenter. The marked isocenter included any isocenter deviations (from the original treatment plan) that were corrected during the weekly isocenter verification, using portal imaging according to institutional protocol. This method of alignment has been validated for target coverage and setup error variation by O’Daniel et al. (11). The dose distribution and DVHs were generated for each of the weekly scans. Since this was an observational study, new treatment plans were deemed to be experimental, and patients continued to be treated on the original treatment plan. The shift in the position of the parotid glands was calculated as follows. Three fixed bony landmarks were chosen, namely the tip of the mastoid process, the tip of the styloid process, and the inferior end of the lateral pterygoid plate. The coordinates of the center of mass (COM) of each of the parotid glands were documented. The average of the x, y, and z coordinates for each of the bony points generated the coordinates for a hypothetical central bony reference point. The shift of the COM of each parotid gland in each plane was calculated by measuring its distance from the coordinates of the above-mentioned bony point on each of the CT scans. The patient’s weight was documented on a weekly basis.
Statistical analysis In their study with 14 patients, Barker et al. showed that the GTV was reduced by 1.7 to 1.8% per treatment day (1). Assuming that we might have expected to see a similar percent change, 20 patients would have enabled us to estimate the reduction in volume per treatment day to within 0.3% (power 95% = 5%). Hansen et al. (2) reported a reduction in the dose to the PTVGTV D95 of 2.2 Gy (95% confidence interval [CI], 1.4–1.5 Gy). If the variability across patients was similar, we anticipated being able to determine the change in PTV to within approximately 1.2 Gy (power 95% = 5%, 95% CI). Statistical analysis was performed with SPSS 2006 version 15.0 software (SPSS Inc. Chicago, IL). The reduction in volumes of the CTVs and parotid gland; changes in the dose to the target volumes, parotid gland, spinal cord, and brain stem; movement of the parotid glands; and changes in weight were analyzed for each patient and were summarized as means with 95% CI. The volume change and the dosimetric alterations were analyzed together for all the
1362
I. J. Radiation Oncology d Biology d Physics
parotid glands and then separately for the ipsilateral and contralateral glands. The Kolmogorov-Smirnov test was used to test data for normality. Since all the data, except for the maximum dose to the brain stem, were normally distributed, parametric tests were used for testing hypotheses. The Wilcoxon sign ranked test was used as a test of significance for changes in maximum dose to the brain stem. Comparisons were made between the volumes and plans on the scan at a particular time point and the scan performed on the following week, i.e., pretreatment with week 2, week 2 with week 3, and so on. The Bonferroni correction was applied, and an a value was set at 0.012 instead of 0.05. To assess the relationship between weight loss and volume change, the percent weight loss and the percent volume reduction for each patient at each time point were calculated. The correlation between weight loss and volume reduction was estimated using Pearson’s correlation.
RESULTS Twenty patients were entered into the study between August 2007 and May 2009. The concomitant chemo-IMRT commenced 6 weeks after the start of induction chemotherapy. One out of the 20 patients had cisplatin replaced by carboplatin for the final concomitant treatment, due to renal toxicity. The patients completed the treatment protocol without any interruptions. Compliance with the study protocol was excellent, with 19 patients having CT scans according to the protocol. One patient missed the CT scan at week 3, as he was not fit to attend. The patients’ TNM staging distribution is given in Table 1. Ten patients had oropharyngeal cancer, 6 patients had laryngeal/hypopharyngeal cancer, and 4 patients had nasopharyngeal cancer. Intraobserver variability The observed percent difference in volumes of the volumes contoured by the observer at different time points was not significant (data not shown; p > 0.05). Volumetric changes The pretreatment volumes were labeled week 0. The changes in volume were normalized to the percent reduction in volume. The reduction was calculated with respect to the week 0 volumes, which were defined as 100%. CTV1 Comparing the changes in volume of CTV1 at each week with that of the preceding week showed that the greatest reduction in mean absolute volume and mean percent volume was 11 cm3 (95% CI, 3-14) and 3.2% (95% CI, 0.74-5.6), respectively, between week 0 and week 2 (Table 2 and Fig. 1). This difference is statistically significant (p = 0.003). The absolute and percent reductions in the next two-week period, i.e., between week 2 and 4 were, respectively, 7 cm3 (95% CI, 1-13) and 2% (95% CI, 0.1 to 4), which were not statistically significant (p = 0.051). CTV2 Comparing the changes in volume at each week with that of the preceding week showed that the greatest reduction in
Volume 76, Number 5, 2010
Table 1. TNM staging distribution of the trial patients No. of patients with node stage Tumor stage T1 T2 T3 T4 Total no. of patients
N0
N1
N2a
N2b
N2c
N3
Total no. of patients
0 1 2 3 6
1 1 0 0 2
0 0 1 0 1
1 2 2 1 6
0 1 1 2 4
0 0 0 1 1
2 5 6 7 20
mean absolute volume and mean percent volume was 21.6 cm3 (95% CI, 4-18) and 10.5% (95% CI, 7-14), respectively, between week 0 and week 2 (Table 2 and Fig. 2). This difference is statistically significant (p < 0.001). The absolute and percent reduction in the next two-week period, i.e., between week 2 and 4, was, respectively, 12 cm3 (95% CI, 6-17) and 5% (95% CI, 3-8), which was statistically significant (p < 0.001). Parotid glands The greatest absolute and percent reduction in the volume of the parotid glands was 4.2 cm3 (95% CI, 3-5.5) and 14.7% (95% CI, 10-19), respectively, and occurred between week 0 and week 2 (Table 3). This difference was statistically significant (p < 0001). Figure 4 shows the percent reduction if plotted separately for the ipsilateral and the contralateral parotid glands. It shows that the ipsilateral gland shrinks more than the contralateral gland; however, the difference was not statistically significant. The absolute and percent reduction in the next two-week period, i.e., between week 2 and 4, was, respectively, 4 cm3 (95% CI, 3-5) and 16% (95% CI, 16-19), which was statistically significant (p < 0.001). Dosimetric alterations During the critical appraisal of an IMRT treatment plan, two dose parameters were considered most relevant. The first parameter was the minimum dose to the PTV, which Table 2. Comparison of reduction in weekly absolute and relative CTV1 and CTV2*
Target volume
Weekly comparison
Mean absolute difference (cm3)
CTV1
Pre-Rx vs. week 2 Weeks 2 and 3 Weeks 3 and 4 Weeks 4 and 5 Pre-Rx - week 2 Weeks 2 and 3 Weeks 3 and 4 Weeks 4 and 5
11.0 5.6 1.6 8.5 21.6 6.4 5.7 0.58
CTV2
Percent reduction
p valuey
3.2 2.1 0.16 2.5 10.5 2.5 3.4 0.30
0.003 0.020 0.538 0.012 <0.001 0.031 0.034 0.78
* Data are paired t test results comparing the reduction in absolute and relative volumes each week with respect to the volumes on the preceding scan for CTV1 and CTV2. y Result is significant if p is <0.012.
Volume changes during head and neck chemo-IMRT d S. A. BHIDE et al.
1363
Table 3. Comparison of weekly changes in minimum doses for PTV1 and PTV2*
Volume
Mean reduction Mean compared to Treatment minimum preceding plan dose (Gy) week (Gy) p valuey
PTV1 (treatment Week 2 plan, 61.1 Gy) Week 3 Week 4 Week 5 PTV 2 (treatment Week 2 plan, 50.5 Gy) Week 3 Week 4 Week 5
59.1
2.0
0.002
58.3 57.8 58 46.6
1.0 0.64 0.18 3.9
0.180 0.280 0.768 0.03
46.1 47.2 46.5
0.4 1 0.7
0.75 0.39 0.66
* Paired t test comparing the change in minimum dose each week with respect to the preceding week for PTV1 and PTV2. y Result is significant if p is <0.012. Fig. 1. The mean percentage of change in volume of CTV1 with error bars (95% CIs).
quantified the volume receiving a clinically suboptimal dose. The second parameter was the dose range, which indicates the homogeneity of the dose distribution. PTV1 The minimum dose to PTV1 and PTV2 was reduced through the course of the study. When the dose reduction was compared to that of the preceding week for PTV1, the significant difference in dose (2 Gy [95% CI, 0.82-3.2]) was between week 0 and week 2 (p = 0.002). The reduction in the minimum dose to PTV2 was not statistically significant (Table 3). The dose range across PTV1 increased throughout the study period, indicating reducing dose homogeneity
Fig. 2. The mean percentage of change in volume of CTV2 with error bars (95% CIs).
(Table 4). The increase was greatest (2.5 Gy [95% CI, 1.33.8]) between the treatment plan and week 2 (p < 0.001). The corresponding increase in the dose range for PTV2 in the same time interval was 5.1 Gy (95% CI, 1.5-8.8), and this was statistically significant (p = 0.008). There were no significant differences between the minimum dose and the dose range to PTV1 and PTV2 between weeks 2 and 4. Parotid glands The mean dose to the parotid glands increased throughout the study period. However, this increase was not statistically significant (Table 5). The mean dose to the ipsilateral parotid gland was higher than that to the contralateral parotid gland at each time point (Table 5). This is in keeping with the constraints set on each of these glands. The mean dose to the
Fig. 3. The mean percentage of change in volume of the parotid gland with error bars (95%CIs of the means).
I. J. Radiation Oncology d Biology d Physics
1364
Volume 76, Number 5, 2010
Table 5. Weekly increase in mean dose for parotid glands*
Parotid volume tested Combined (treatment plan, 31.7 Gy)
Week 2
32.6
0.9
0.17
Week 3 Week 4 Week 5 Week 2
33 34.3 33.6 38.6
0.3 1.3 -0.7 0.2
0.53 0.01 0.19 0.83
Week 3 Week 4 Week 5 Contralateral (treatment Week 2 plan, 24.3 Gy) Week 3 Week 4 Week 5
39.3 41.1 41.2 26
0.7 2.7 2.8 1.7
0.39 0.006 0.01 0.11
26.2 26.8 25.3
1.9 2.5 1.0
0.07 0.03 0.29
Ipsilateral (treatment plan, 38.4 Gy)
Fig. 4. The change in the percentage of reduction in volume between the week 0 scans and the subsequent scans for ipsilateral (solid squares) and the contralateral (solid circles) glands. Errors bars indicate 95% CI.
ipsilateral parotid gland was significantly higher by 2.7 Gy (95% CI, 0.8-4.4) and 2.8 Gy (95% CI, 0.5-4.9) at weeks 4 and 5, respectively, compared to the original treatment plan (Table 5). There was no significant change in the dose to the contralateral gland (Table 5). There was no significant difference between the mean dose to the parotid glands between weeks 2 and 4. There was a significant medial shift of the parotid glands through the course of treatment, starting at week 2 (p = 0.002). The highest mean movement of the COM was 2.3 mm at week 4 (p < 0.0001). The movement of the COM in the anteroposterior and the inferosuperior directions was not significant.
Table 4. Comparison of weekly dose range increases across PTV1 and PTV2*
Volume PTV1 (treatment plan, 7 Gy)
PTV2 (treatment plan, 11.8 Gy)
Mean reduction Mean compared to Treatment dose preceding plan range (Gy) week (Gy) p valuey Week 2
9.5
2.5
<0.001
Week 3 Week 4 Week 5 Week 2
10 11.3 11 16.9
0.5 1.3 0.3 5.1
0.710 0.069 0.698 0.008
Week 3 Week 4 Week 5
17.1 16.5 16.9
0.2 0.4 0.4
0.929 0.680 0.798
* Paired t test comparing the increase in dose range each week with respect to the preceding week across PTV1 and PTV2. y Result is significant if p is <0.012.
Mean increase Average compared to mean treatment Week dose (Gy) plan (Gy) p valuey
* Paired t test showing the increase in mean dose for the parotid glands at each week with reference to baseline. y Result is significant if p is <0.012.
Spinal cord and brain stem The differences in the maximum doses to the spinal cord and brain stem, when comparing the plan at each week with that of the preceding week, were not significant (Table 6). The increase in cord dose was statistically significant at weeks 4 (p = 0.004) and 5 (p = 0.003) compared to week 0. However the upper limit of the 95% CI on average was lower than the set constraint of 48 Gy (Table 6). Cumulative plan analysis A cumulative plan consisting of the dosimetric change at each week multiplied by the time fraction was calculated and compared to the original plan and also to the dosimetric changes at week 2 (Table 7). This comparison showed that there was a statistically significant difference for the minimum (lower) dose and the dose range (higher) for PTV1 and PTV2 between the original and the cumulative plan. There difference between the plan at week 2 and the cumulative plan was not statistically significant. Similar results were observed for the mean doses to the parotids and the maximum doses to the spinal cord and the brain stem. Weight loss The mean weight loss (95% CI) during treatment was 7.5 kg (3.1 kg). The mean relative weight loss was 9.7% (3.5%) of the original weight. There was no significant correlation between the volume reduction in CTV1 or CTV2 and weight loss. DISCUSSION This is the first study that has investigated the volume changes and the resulting dosimetric implications in patients undergoing sequential treatment for HNC. This study has
Volume changes during head and neck chemo-IMRT d S. A. BHIDE et al.
1365
Table 6. Weekly differences between average maximum doses to OARs Week Organ
Dose*
0
2
3
4
5
Spinal cord
Average maximum dose (Gy) (95% CI) p value Average maximum dose (Gy) (95% CI) p value
44.2 (43-45)
46 (44-47)
45.3 (43-47)
46.3 (44-47)
46.6 (44-48)
46.5 (41-51)
0.05 48.2 (42-53)
0.9 48.5 (43-53)
0.3 48.5 (42-54)
0.5 48.9 (43-54)
0.03
0.7
0.4
0.6
Brain stem
* Result is significant if p is <0.012
demonstrated that the CTVs undergo significant reduction during the course of head and neck IMRT. In a comparison of the change at each time point with respect to the preceding week, the maximum difference was seen between the original treatment scan and the scan at week 2 (i.e., 8 weeks after the first induction chemotherapy) for CTV1 and CTV2. The volume of CTV1 decreased by 3.2% at week 2 and that of CTV2 decreased by 10.5%. The reduction in volume of CTV2 between weeks 2 and 4 was smaller (5%) but statistically significant. The planning scans were performed at the time of administration of the second cycle of induction chemotherapy, i.e., minus day 21 of chemoradiotherapy. The patients had one cycle of induction chemotherapy, the first cycle of concomitant cisplatin, and 10 fractions of radiotherapy between the planning scan and the scan at week 2. This could explain the significant reduction in volume of CTV1 by week 2 of the chemoradiation. However, this does not account for the significant reduction in volume of CTV2 by week 2, as the treatment received by the patient should not have an effect on the volume of CTV2. The reduction in the volume of CTV2 is likely to be due to the anatomical changes taking place in the head and neck region, which is probably due to the loss of adipose tissue in the neck. This is corroborated by further reduction in volume between the next two-week period. The patients could lose weight as a result of receiving induction chemotherapy, which in turn could reduce the amount of adipose tissue in the neck. However, there was no correlation between the weight loss and the change in CTV2, indicating that weight loss probably is not
a sensitive indicator of loss of adipose tissue in the neck. The planning scans could not be performed closer to the start of radiotherapy due to the additional time required for planning and quality assurance of the complex IMRT plans. In retrospect, when this study was designed, provision should have been made for a scan at the start of radiotherapy, which would have helped to elucidate the effects of the second cycle of induction chemotherapy and to differentiate these effects from the effect of chemoradiation. Although we found the intraobserver variability to be nonsignificant, this could still account for the changes in the volumes due to the variability of contouring between different CT sets. The average minimum doses to PTV1 and PTV2 were significantly reduced by 2 Gy and 3.9 Gy, respectively, at week 2 (Table 3). The dose ranges across PTV1 and PTV2 increased by 2.5 Gy and 5.1 Gy, respectively, at week 2 (Table 4). The change for PTV2 was greater, in line with the greater volume loss. The dosimetric changes between weeks 2 and 4 were not significant, probably due to the smaller volume reduction (5% vs. 10.5%, respectively) compared to week 0 and week 2. The dose of the cumulative doses was also significantly different compared to the dose of the original plan. There was no significant difference between the dosimetric changes on the cumulative plan and the plan based on the volume changes at week 2. This indicates that dosimetric alterations resulting from volumetric changes could be minimized by replanning at week 2. It is difficult to comment on the clinical significance of the dosimetric alterations because the data for the alterations were obtained from the
Table 7. Comparison of doses on the cumulative plan, the original plan, and the scan at week 2 Original vs. cumulative plan (time factor correction)
Average dose (Gy)
Week 2 vs. cumulative (time factor correction)
Treatment plan
Original
Week 2
Cumulative
Difference (Gy)
p value
Difference (Gy)
p value
PTV1 minimum PTV1 dose range PTV2 minimum PTV2 dose range Parotid mean Spinal cord maximum Brain stem maximum
61.3 7 50.5 11.8 31.7 44.2 46.5
59.1 9.5 48 16.9 32.6 46 48.2
59.3 9.8 47.9 15.2 32.9 45.4 48.5
1.8 2.8 2.6 3.5 1.2 1.2 2
<0.001 <0.001 0.006 0.001 0.003 0.006 0.004
0.2 0.3 0.1 1.7 0.3 0.6 0.3
0.6 0.6 0.1 0.08 0.6 0.6 0.3
1366
I. J. Radiation Oncology d Biology d Physics
Volume 76, Number 5, 2010
Fig. 5. An example of the volume reduction and the change in dose for the parotid glands (right parotid gland [peach colored] and left parotid gland [yellow]) at different time points. The dark blue line is the 30-Gy isodose line. The dark blue volume is PTV1, and the light blue volume is PTV2. (A) Original plan; (B) week 2; (C) week 3; (D) week 4; (E) week 5.
DVHs, which fail to give an estimate of the spatial aspects of the dose distribution. One possibility is that any areas receiving a lower minimum dose in PTV1 will result in a lower dose to either macroscopic or microscopic disease, resulting in an increased risk of recurrence. This would result if the predominant changes in the patient’s anatomy were asymmetrical. If the changes were symmetrical (due to tumor shrinkage), PTV1 would have been adequately covered by the 95% isodose. However, this would have resulted in an increase in the
minimum dose to PTV1 and not the reduction seen in this study. The dosimetric changes seen in this study are likely to be due to the changes in the patients’ anatomy. The parotid volumes were reduced by 35% through the course of treatment (Table 5). There was a significant increase in mean dose to the ipsilateral gland (2.7 Gy) at week 4 (Table 5). This could be due to CTV1 shrinkage and medial movement of the gland into the high-dose region. The study of the parotid coordinates shows that there is
Volume changes during head and neck chemo-IMRT d S. A. BHIDE et al.
significant medial shift of the parotid glands (0.23 cm by week 4) through the treatment period. Only one study (2), with 13 patients, investigated the volumetric and dosimetric changes in CTVs and OARs in patients treated for HNC with IMRT (no concomitant chemotherapy). The patients were rescanned when there was a clinically observed change in anatomy or significant weight loss. The study demonstrated significant change in volumes of the CTV and parotid glands and significant underdosage to target volumes and overdosage to the OARs. The dosimetric changes were similar to those seen in our study. However, patients were scanned only once and only if that was clinically indicated. In a similar study, Barker et al. scanned 14 patients with tumors or lymph nodes $4 cm in diameter during radiotherapy (1). Only one patient was treated with IMRT. Twelve patients received concurrent chemotherapy. There was a reduction in the GTV at a median rate of 1.8% per treatment day, which was asymmetrical. At the end of treatment, the median parotid volume loss was 28.1% (range, 5.9– 53.6%). The median medial shift of the parotid glands was 3 mm (range, 0.3-9 mm). The median weight loss was 7.1%. The authors did not comment on the dosimetric alterations. That study was limited to patients who had significant lymphadenopathy. It is of clinical relevance to investigate the change in the CTVs based on the patient’s anatomic changes, as in the study by Hansen et al. (2) and our study. This is because there is a lack of evidence for the use of shrinking CTVs generated from reduced GTV for radical treatment of HNC. Ballivy et al. studied the dosimetric implications of weekly volume changes in 8 patients (2 had concomitant chemotherapy) undergoing IMRT for HNC (12). The authors found the differential dosimetric changes in the upper versus the lower neck due to the change in the curvature of the neck. They concluded that a greater PTV margin might be required to account for this, especially when the isocenter is positioned in the lower neck.
1367
The volume changes and the net medial shift reported in the published studies are comparable to those of our study (1-4). However, the mean doses to the parotids when analyzed together did not show a significant increase in mean dose, as would be expected from a net medial shift of the COM, which would move the gland into the high-dose region. Studies by Robar et al. (3) and Vasquez Osorio et al. (4) show that the lateral regions of the glands move during radiotherapy while the medial regions are relatively fixed. This could have occurred in our study, resulting in a net movement of the COM of the parotid glands due to the excessive movement of the lateral region, with the medial region being fixed. Our practice of applying additional constraints on the superficial glands resulted in higher doses to the medial region of the gland. Therefore lateral areas of the gland could have moved medially in the low-dose region, resulting in a nonsignificant increase in the mean dose. CONCLUSIONS In summary, our study is unique in having investigated the serial change in volume and the resulting dosimetric changes in the target volumes and the OARs during head-and-neck chemo-IMRT. The most significant volume change and dose alteration occurs 2 weeks after commencing radiotherapy (i.e., 8 weeks after the start of induction chemotherapy). This could be due to the effect of induction chemotherapy, the first cycle of concomitant chemotherapy, and 10 fractions of radiotherapy, especially in the case of CTV1. In retrospect, this study should have included a CT scan at the start of radiotherapy. This study also shows significant dosimetric implications of the volume change. The sharp dose gradients that exist in the IMRT plans necessitate minimizing dosimetric uncertainties in order to reduce the risk of a geographical miss. An adaptive radiotherapy approach by rescanning the patients, recontouring the target volumes and OARs, and applying the original treatment plan and replanning if necessary is required.
REFERENCES 1. Barker JL Jr., Garden AS, Ang KK, O’Daniel JC, Wang H, Court LE, et al. Quantification of volumetric and geometric changes occurring during fractionated radiotherapy for headand-neck cancer using an integrated CT/linear accelerator system. Int J Radiat Oncol Biol Phys 2004;59(4):960–970. 2. Hansen EK, Bucci MK, Quivey JM, Weinberg V, Xia P. Repeat CT imaging and replanning during the course of IMRT for headand-neck cancer. Int J Radiat Oncol Biol Phys 2006;64(2): 355–362. 3. Robar JL, Day A, Clancey J, Kelly R, Yewondwossen M, Hollenhorst H, et al. Spatial and dosimetric variability of organs at risk in head-and-neck intensity-modulated radiotherapy. Int J Radiat Oncol Biol Phys 2007;68(4):1121–1130. 4. Vasquez Osorio EM, Hoogeman MS, Al-Mamgani A, Teguh DN, Levendag PC, Heijmen BJ. Local anatomic changes in parotid and submandibular glands during radiotherapy for oropharynx cancer and correlation with dose, studied in detail with nonrigid registration. Int J Radiat Oncol Biol Phys 2008; 70(3):875–882.
5. Humphreys M, Guerrero Urbano MT, Mubata C, Miles E, Harrington KJ, Bidmead M, et al. Assessment of a customised immobilisation system for head and neck IMRT using electronic portal imaging. Radiother Oncol 2005;77(1): 39–44. 6. Guerrero Urbano MT, Clark CH, Kong C, Miles E, Dearnaley DP, Harrington KJ, et al. Target volume definition for head and neck intensity modulated radiotherapy: Pre-clinical evaluation of PARSPORT trial guidelines. Clin Oncol 2007; 19(8):604–613. 7. Gregoire V, Eisbruch A, Hamoir M, Levendag P. Proposal for the delineation of the nodal CTV in the node-positive and the post-operative neck. Radiother Oncol 2006;79(1): 15–20. 8. Gregoire V, Levendag P, Ang KK, Bernier J, Braaksma M, Budach V, et al. CT-based delineation of lymph node levels and related CTVs in the node-negative neck: DAHANCA, EORTC, GORTEC, NCIC, RTOG consensus guidelines. Radiother Oncol 2003;69(3):227–236.
1368
I. J. Radiation Oncology d Biology d Physics
9. Bhide S, Clark C, Harrington K, Nutting CM. Intensity modulated radiotherapy improves target coverage and parotid gland sparing when delivering total mucosal irradiation in patients with squamous cell carcinoma of head and neck of unknown primary site. Med Dosim 2007;32(3):188–195. 10. Bhide SA, Ahmed M, Barbachano Y, Newbold K, Harrington KJ, Nutting CM. Sequential induction chemotherapy followed by radical chemo-radiation in the treatment of locoregionally advanced head-and-neck cancer. Br J Cancer 2008;99(1):57–62.
Volume 76, Number 5, 2010
11. O’Daniel JC, Garden AS, Schwartz DL, Wang H, Ang KK, Ahamad A, et al. Parotid gland dose in intensity-modulated radiotherapy for head and neck cancer: Is what you plan what you get? Int J Radiat Oncol Biol Phys 2007;69(4): 1290–1296. 12. Ballivy O, Parker W, Vuong T, Shenouda G, Patrocinio H. Impact of geometric uncertainties on dose distribution during intensity modulated radiotherapy of head-and-neck cancer: the need for a planning target volume and a planning organat-risk volume. Curr Oncol 2006;13(3):108–115.