|Year : 2014 | Volume
| Issue : 1 | Page : 21-25
Image guidance in prostate cancer - can offline corrections be an effective substitute for daily online imaging?
Devleena Prasad, Pinaki Das, Niladri S Saha, Sanjoy Chatterjee, Rimpa Achari, Indranil Mallick
Department of Radiation Oncology, Tata Medical Center, Kolkata, West Bengal, India
|Date of Web Publication||23-Apr-2014|
Department of Radiation Oncology, Tata Medical Center, 14 MAR (EW), Newtown, Kolkata - 700 156, West Bengal
Source of Support: None, Conflict of Interest: None
Purpose: This aim of this study was to determine if a less resource-intensive and established offline correction protocol - the No Action Level (NAL) protocol was as effective as daily online corrections of setup deviations in curative high-dose radiotherapy of prostate cancer.
Materials and Methods: A total of 683 daily megavoltage CT (MVCT) or kilovoltage CT (kvCBCT) images of 30 patients with localized prostate cancer treated with intensity modulated radiotherapy were evaluated. Daily image-guidance was performed and setup errors in three translational axes recorded. The NAL protocol was simulated by using the mean shift calculated from the first five fractions and implemented on all subsequent treatments. Using the imaging data from the remaining fractions, the daily residual error (RE) was determined. The proportion of fractions where the RE was greater than 3,5 and 7 mm was calculated, and also the actual PTV margin that would be required if the offline protocol was followed.
Results: Using the NAL protocol reduced the systematic but not the random errors. Corrections made using the NAL protocol resulted in small and acceptable RE in the mediolateral (ML) and superoinferior (SI) directions with 46/533 (8.1%) and 48/533 (5%) residual shifts above 5 mm. However; residual errors greater than 5mm in the anteroposterior (AP) direction remained in 181/533 (34%) of fractions. The PTV margins calculated based on residual errors were 5mm, 5mm and 13 mm in the ML, SI and AP directions respectively.
Conclusion: Offline correction using the NAL protocol resulted in unacceptably high residual errors in the AP direction, due to random uncertainties of rectal and bladder filling. Daily online imaging and corrections remain the standard image guidance policy for highly conformal radiotherapy of prostate cancer.
Keywords: Image guided radiation therapy,prostate cancer,setup errors
|How to cite this article:|
Prasad D, Das P, Saha NS, Chatterjee S, Achari R, Mallick I. Image guidance in prostate cancer - can offline corrections be an effective substitute for daily online imaging?. J Can Res Ther 2014;10:21-5
|How to cite this URL:|
Prasad D, Das P, Saha NS, Chatterjee S, Achari R, Mallick I. Image guidance in prostate cancer - can offline corrections be an effective substitute for daily online imaging?. J Can Res Ther [serial online] 2014 [cited 2019 Sep 17];10:21-5. Available from: http://www.cancerjournal.net/text.asp?2014/10/1/21/131342
| > Introduction|| |
Radiotherapy for prostate cancers requires the delivery of high doses of radiation due to a dose dependence demonstrated in several studies. These high doses increase the likelihood of normal tissue toxicity due to the proximity of structures like the rectum and urinary bladder to the prostate. Modern highly conformal techniques like 3D conformal radiotherapy (3DCRT) and Intensity modulated radiotherapy (IMRT) with limited planning target volume (PTV) margins allow reduction of dose to these structures with associated reduction in toxicity. The clinical benefit of highly conformal techniques has been demonstrated in large studies. 
However, the prostate is a relatively mobile organ and daily variations in its position have been documented.  For highly conformal radiotherapy techniques to be safe, image guidance is extremely important and there is evidence of clinical benefit with the use of image-guided radiation therapy (IGRT) in prostate cancer.  Two commonly employed image-guidance techniques include (a) two dimensional electronic portal imaging (EPI) or (b) cone beam CT, with or without implanted fiducials. Image-guidance can be performed online (decisions for setup corrections based on data acquired during the current treatment session), or offline (overall setup correction decisions made from analyzing a finite number of prior measurements). IGRT for prostate cancer started with several well-studied offline correction protocols. One of the most commonly followed in the No-action-level (NAL) protocol first reported by de Boer and Heijman.  Offline protocols are effective in determining and correcting systematic errors. However; with the availability of new technologies, many institutions have moved on to daily online image-guidance, which allows for the correction of both systematic and random errors. Online corrections are potentially more precise, but substantially more resource-intensive.
The PTV margin that is used for prostate cancer radiotherapy planning is dependent on measurement of positional uncertainties of the target volume with the specific image-guidance protocol being employed. Small PTV margins can only be used if an image-guidance strategy has been put in place and tested.
The image-guidance strategy (modality and frequency) often varies between institutions and is dictated by local constraints-including availability of image-guidance technology, training, manpower and machine time. Resource-constrained hospitals therefore, must find an optimal image-guidance strategy and customize its planning and margins according to this.
In our hospital, we routinely follow online setup corrections for prostate cancer. However; we are constrained in terms of availability of machine time. We did this study in our hospital to determine whether an established offline correction protocol (the NAL protocol) could be an effective substitute for daily online imaging for prostate cancers receiving radical radiotherapy.
| > Materials and Methods|| |
This study was carried out in a newly established tertiary cancer hospital. Image-guidance data of 30 patients with high and intermediate risk prostate cancer treated between June 2011 and July 2012 was analyzed for this study.
Following an agreed institutional policy, patients were simulated in 16 slice CT scanner in supine position with a knee rest. All patients underwent a prescan proctocol for the bladder. This involved passing urine, drinking 500 ml of water and waiting for 30 min prior to imaging. It was also ensured that all patients had cleared their bowels on the day of the planning scan. A Slice thickness of 2.5 mm was used and IV contrast was used as 1ml/kg of body weight. Implanted fiducials were not used.
For all patients, target delineation was performed on a Varian contouring workstation (ARIA v8.6). The clinical target volume (CTV) for intermediate risk patients included the entire prostate and both seminal vesicles. For high risk patients, the pelvic nodes were treated electively. The CTV-PTV margin was 7 mm in all directions. The organs at risk, including the rectum (from the level of ischial tuberosity to recto sigmoid flexure), the bladder, the penile bulb and bilateral femoral heads, were also contoured.
All patients received intensity modulated radiation therapy (IMRT). Treatment planning was performed on either a Tomotherapy (Accuray Inc) platform or the Varian Eclipse platform. Two dose fractionation schedules were used. In schedule A the prescription dose was 65Gy in 25Fr over 5 weeks, with elective volumes treated to 45Gy/25Fr. In schedule B the prescription dose was 60Gy in 20Fr over 4 weeks, with elective volumes treated to 44Gy/20Fr.
All patients received daily online image guidance using MVCT on Tomotherapy or kvCBCT on Novalis Tx (Varian). Automatic soft-tissue matching was done first in the region of interest followed by verification and manual adjustments. Image-matching was performed by radiation therapists and regularly verified by radiation oncologists (Drs. DP and IM). All setup errors were corrected daily with no threshold. Setup errors in the medio-lateral (ML), superoinferior (SI) and antero-posterior (AP) axes were recorded on a daily basis prospectively.
The goal of the study was to determine if the No-action-level (NAL) offline protocol would be as good as daily online imaging. Since the NAL protocol requires imaging for the first 3-5 days and then implementing the average shift on all subsequent treatments, this workflow was simulated as follows [Figure 1]:
- Shifts recorded during treatments from the first five fractions for each patient were used to calculate a mean error in the three axes that would be implemented on all subsequent treatments axes in the NAL protocol
- Using the imaging data from the remaining fractions, the discrepancy between the calculated (or expected) shift and the actual shift, the residual error (RE) was determined
- The proportion of fractions where the RE was greater than 3 mm, 5 mm and 7 mm was counted for each axis
- From the RE the actual PTV margin that would be required if the offline protocol was followed was calculated according to the van Herk formula. 
| > Results|| |
Online images of 30 patients were evaluated. A brief summary of patient and treatment characteristics is presented in [Table 1].
There were a total of 695 total fractions treated. We analyzed 683 daily online volumetric images. Twelve images did not cover the entire area of interest or were unrecoverable.
To simulate the NAL protocol, setup errors from first five fractions of each patient were used to calculate a mean shift that is applied to all remaining fractions. The setup errors from the remaining 533 images were used to calculate the residual errors that would have remained after an offline correction.
Using the data available from daily online imaging, we calculated both the systematic and random errors that would have remained if a) no imaging was done or b) offline correction implemented. This is represented in [Table 2]. Without any imaging, systematic errors were most prominent in the SI axis (6.7 mm), while random errors were the highest in the AP axis (5.2 mm). According to our simulation of the NAL protocol, as compared to 'no imaging', following the offline NAL protocol would have substantially reduced the systematic error in all three axes. However, as expected, offline correction had little or impact on random errors.
|Table 2: Summary of systematic and random errors with no imaging vs. offline correction with the NAL protocol|
Click here to view
Despite the improvement, residual errors remained after the offline correction. While errors in the lateral and longitudinal axes were small, those in the anteroposterior direction remained substantial. This was most evident due to random variations in the AP direction. The number of fractions where the residual error was greater than 3 mm, 5 mm and 7 mm are shown in [Table 3]. The frequency of clinically relevant deviations (>5 mm and >7 mm) was substantially higher in the AP direction compared to the other two axes.
PTV margins were calculated for two scenarios (a) no imaging and (b) after offline correction. These are shown in [Table 4]. PTV margins close to 2cm would be required in the SI and AP axes if no imaging was used. Offline corrections using the NAL protocol helped in reducing PTV margins compared to no imaging. The margin required in the ML and SI directions after offline correction was 5 mm and easily implementable. However, despite offline correction the margin required in the AP direction was quite large. A margin of 13 mm was calculated which is substantially higher than traditionally used margins in conformal radiotherapy.
|Table 4: Planning target volume margins required after shifts according to the NAL protocol|
Click here to view
| > Discussion|| |
We performed this study as a service development project to determine if daily online image guidance was providing additional benefits over a simpler and less resource consuming offline protocol. In a department with long waiting lists where running costs have to be contained and manpower is limited, we questioned the additional benefit of daily imaging vs a well established and easily implementable offline protocol.
Earlier studies have demonstrated that the NAL protocol is an effective offline protocol. In comparison to the alternative Shrinking action level protocol it is simpler to implement and results in more accurate treatment delivery. , Therefore; we chose this as the method to test against daily online imaging. We chose to make one modification. Instead of using the first three fractions to make an estimate of the systematic error, we used the first five fractions. This was done to get a better estimate of the systematic uncertainty and was based on the results of a study by Ludbrook et al., who demonstrated that obtaining offline correction after analyzing the first five images was the most optimal strategy.  Daily online imaging data was already available to us and allowed us to simulate the offline correction protocol easily to derive residual errors after offline correction.
Offline correction did reduce the systematic uncertainty considerably [Table 2]. A study by Johnston et al., also demonstrated a reduction in systemic uncertainty with daily bone matching.  This demonstrates that offline protocols are very effective in scenarios where systematic errors predominate. However, there was no change in the random error which remained substantial in the AP axis. The results of our study shows that for radical intent radiotherapy in prostate cancer, offline protocols, while definitely superior to no image-guidance, is unable to correct the random variations in the AP axis induced by daily variations in rectal and bladder volume. As a result, the PTV margins calculated from the residual errors after the NAL protocol are small in the SI and lateral directions, but large in the AP axis. These findings are not unexpected. Large random variations are well described in prostate radiotherapy. ,, The majority of studies also show that the highest setup variation is in the anteroposterior direction, consistent with our findings.
One approach to reducing random variations in prostate position would be to limit positional variations of the bladder and rectum. Methods to reproduce daily bladder and rectal position have been reported, but with only limited success. ,, We used a bladder filling protocol that was easy to implement. However, it is unclear whether it had any impact on the random uncertainties. For the rectum, we did not use any immobilization device like endorectal balloons. We followed the commonly used protocol of verbally counselling the patient about low residue diets and the importance of passing stools daily. Those who are habitually constipated were advised to take milk of magnesia at night.
There is limited, but equivocal literature on the frequency of imaging required. A study by Beldjoudi et al., performed a study on Helical Tomotherapy testing the number of imaging fractions required for prostate cancer IMRT. They concluded that four imaging fractions were adequate and a margin based on these fractions is not substantially different from daily imaging.  Our observations are different. We found in our study that despite corrections based on five imaging fractions there were significant residual errors requiring substantially large PTV margins. Another larger study by Kupelian et al., has looked at several different image-guidance schedules and compared them to a daily online image-guidance protocol.  They found that any other protocol apart from daily imaging results in significant residual errors that impact PTV margins. Our study validates this latter observation.
Most of the modern treatment planning protocols uses PTV margins that are less than 1cm. Some protocols use a smaller margin in the posterior direction to spare the rectum.  From our observations, we are forced to conclude that without daily image guidance these margins cannot be applied without a significant risk of missing the target and poorer clinical outcomes. In a scenario where daily image guidance is not possible, larger margins will be required. A much larger volume of the rectum and bladder will be exposed to the prescription dose. It is likely that this will result in a higher incidence of clinical toxicity. We have therefore retained our protocol for daily image guidance and online correction as our standard policy.
Our study has some limitations. It is a relatively small study of 30 patients. However, this data is consistent with data from a smaller subgroup of our patients that has previously been presented.  We did not use implanted fiducials for image guidance. We were satisfied with the images obtained from the MVCT and we were able to consistently match images without the need to go through an invasive seed implantation procedure. Image matching with cone beam CT without implanted fiducials has been reported as a reasonable alternative to matching fiducials on KV X-rays.  The image matching data is from the Tomotherapy and Varian platforms and it could be argued that the residual errors and margins required may potentially be different on a different platforms. However, since the systematic errors were accounted for in the offline matching procedure, the residual errors were largely random errors and can be considered independent of the specific treatment or imaging unit.
| > Conclusions|| |
imaging protocols are able to correct for systematic errors in prostate cancer radiotherapy. However, there are substantial residual random errors in the anteroposterior direction and that results in the requirement of large PTV margins in this direction. The NAL offline protocol is therefore not an effective alternative to daily online imaging and correction in prostate cancer radiotherapy.
| > References|| |
|1.||Zelefsky MJ, Levin EJ, Hunt M, Yamada Y, Shippy AM, Jackson A, et al. Incidence of late rectal and urinary toxicities after three-dimensional conformal radiotherapy and intensity-modulated radiotherapy for localized prostate cancer. Int J Radiat Oncol Biol Phys 2008;70:1124-9. |
|2.||Dawson LA, Mah K, Franssen E, Morton G. Target position variability throughout prostate radiotherapy. Int J Radiat Oncol Biol Phys 1998;42:1155-61. |
|3.||Zelefsky MJ, Kollmeier M, Cox B, Fidaleo A, Sperling D, Pei X, et al. Improved clinical outcomes with high-dose image guided radiotherapy compared with non-IGRT for the treatment of clinically localized prostate cancer. Int J Radiat Oncol Biol Phys 2012;84:125-9. |
|4.||de Boer HC, Heijmen BJ. A protocol for the reduction of systematic patient setup errors with minimal portal imaging workload. Int J Radiat Oncol Biol Phys 2001;50:1350-65. |
|5.||van Herk M, Remeijer P, Rasch C, Lebesque JV. The probability of correct target dosage: Dose-population histograms for deriving treatment margins in radiotherapy. Int J Radiat Oncol Biol Phys 2000;47:1121-35. |
|6.||Bel A, van Herk M, Bartelink H, Lebesque JV. A verification procedure to improve patient set-up accuracy using portal images. Radiother Oncol 1993;29:253-60. |
|7.||Ludbrook JJ, Greer PB, Blood P, D›Yachkova Y, Coldman A, Beckham WA, et al. Correction of systematic setup errors in prostate radiation therapy: How many images to perform? Med Dosim 2005;30:76-84. |
|8.||Johnston ML, Vial P, Wiltshire KL, Bell LJ, Blome S, Kerestes Z, et al. Daily online bony correction is required for prostate patients without fiducial markers or soft-tissue imaging. Clin Oncol (R Coll Radiol) 2011;23:454-9. |
|9.||El-Bassiouni M, Davis JB, El-Attar I, Studer GM, Lutolf UM, Ciernik IF. Target motion variability and on-line positioning accuracy during external-beam radiation therapy of prostate cancer with an endorectal balloon device. Strahlenther Onkol 2006;182:531-6. |
|10.||Graf R, Wust P, Budach V, Boehmer D. Potentials of on-line repositioning based on implanted fiducial markers and electronic portal imaging in prostate cancer radiotherapy. Radiat Oncol 2009;4:13. |
|11.||Krengli M, Gaiano S, Mones E, Ballare A, Beldi D, Bolchini C, et al. Reproducibility of patient setup by surface image registration system in conformal radiotherapy of prostate cancer. Radiat Oncol 2009;4:9. |
|12.||O′Doherty UM, McNair HA, Norman AR, Miles E, Hooper S, Davies M, et al. Variability of bladder filling in patients receiving radical radiotherapy to the prostate. Radiother Oncol 2006;79:335-40. |
|13.||Stasi M, Munoz F, Fiorino C, Pasquino M, Baiotto B, Marini P, et al. Emptying the rectum before treatment delivery limits the variations of rectal dose-volume parameters during 3DCRT of prostate cancer. Radiother Oncol 2006;80:363-70. |
|14.||McNair HA, Wedlake L, McVey GP, Thomas K, Andreyev J, Dearnaley DP. Can diet combined with treatment scheduling achieve consistency of rectal filling in patients receiving radiotherapy to the prostate? Radiother Oncol 2011;101:471-8. |
|15.||Beldjoudi G, Yartsev S, Bauman G, Battista J, Van Dyk J. Schedule for CT image guidance in treating prostate cancer with helical tomotherapy. Br J Radiol 2010;83:241-51. |
|16.||Kupelian PA, Lee C, Langen KM, Zeidan OA, Manon RR, Willoughby TR, et al. Evaluation of image-guidance strategies in the treatment of localized prostate cancer. Int J Radiat Oncol Biol Phys 2008;70:1151-7. |
|17.||Prasad D, Das P, Saha N, Nandi M, Guha S, Chatterjee S, et al. Can offline corrections effectively substitute for daily online imaging in radiotherapy of carcinoma of prostate? Radiother Oncol 2012;103 Suppl 1:S142-3. |
|18.||Moseley DJ, White EA, Wiltshire KL, Rosewall T, Sharpe MB, Siewerdsen JH, et al. Comparison of localization performance with implanted fiducial markers and cone-beam computed tomography for on-line image-guided radiotherapy of the prostate. Int J Radiat Oncol Biol Phys 2007;67:942-53. |
[Table 1], [Table 2], [Table 3], [Table 4]