Dose escalation with proton or photon radiation treatment for pancreatic cancer
Myriam Bouchard1, Tina M. Briere1, Richard A. Amos1, Sam A. Beddar1 and Christopher H. Crane2
(1) Department of Radiation Physics and (2) Department of Radiation Oncology, University of Texas M. D. Anderson Cancer Center, TX Background
Treatment planning strategy
For each 11 locations, we generated 6 plans, 2 for each technique: 3DCRT (18MV), IMRT (6MV) and proton plans (passive scattering technique). And for each technique, we created a first plan to fully adhere to target coverage objectives while a second aimed to meet OAR sparing constraints with the best coverage possible. We planned to deliver 72 Gy (or CGE) in 36 fractions to the ITV. We created IMRT plans with Pinnacle3 v7.6c with the direct machine parameter optimization (DMPO) calculation and the step-and-shoot delivery method. We created proton plans with Eclipse. Figure 2 illustrates IMRT beam angles, which were the same for all IMRT plans, and a typical choice for proton plan beam angles, which consisted of 2 or 3 oblique posterior beams.
Radiation therapy dose escalation has shown benefits for different cancer sites.[1-2] For pancreatic cancer, it may constitute a solution for better disease control for selected patients. However, proximity of organs at risk (OAR) could hamper dose escalation.
Results
Isodose distribution compared to OAR anatomy
Figure 3 shows results for isodose distribution, expressed by distance in mm (mean ± SD) for all 11 tumor positions. From a practical point of view, we did the following correlation with OAR anatomy.
Purpose
The purpose of this dosimetric study is to determine which pancreatic tumor positions in relation to OAR anatomy would be safe for dose escalation (72 Gy) for the following treatment modalities: 3D conformal radiation therapy (3DCRT), IMRT or protons.
The stomach has a predominant supero-anterior distribution compared to the pancreas. Globally, the distance between the stomach and the GTV needs to be more than 21 ± 4 mm (54 Gy isodose distribution) with IMRT (with the smallest achievable distance in an oblique direction being 14 ± 2 mm) while 20 ± 1 mm anteriorly and 24 ± 2 mm superiorly is needed for proton plans.
Table 2 ITV and CTV V72Gy according to tumor positions
DVH comparison
Figure 4 depicts a DVH for the initial tumor position in the head of the pancreas for the 3 different techniques and the plans that fully respect OAR constraints. Dose escalation was not possible to achieve with 3DCRT at any tumor position. Concerning OAR sparing for all tumor positions, the V15Gy results (mean ± SD), a surrogate for low dose levels, were as follows: the stomach V15Gy, 47.6 ± 3.7% vs 4.8 ± 3.4% (p<0.0001); the duodenum V15Gy 11.3 ± 11.7% vs 4.8 ± 3.4% (p=0.007); and the small bowel V15Gy, 60.7 ± 8.2% vs 9.2 ± 3.6% (p<0.0001) for IMRT and proton plans respectively.
fixed GTV allowing a zero-mm ITV margin, there would be a better chance for safe dose escalation. We justify the choice of proton beam orientation with two arguments: 1- only one of three beams is directly toward the OAR; 2- uncertainty is minimized: gastro-intestinal track gas and breathing variation effects are avoided with proton-beam posterior angles. For proton therapy, the passive scattering technique is a forward planning process, and it represents an important limitation of its potential for dose escalation. Combined to a lateral penumbra (that is similar to the photon beam and which increases with deeply located targets), the potential of sparing besides the downstream Bragg peak is limited by this technique. However, the reduced low dose levels to the gastro -intestinal OAR resulting from proton plans could lead to better treatment tolerance since radiation is usually combined to a radio-sensitizing chemotherapy regimen for pancreatic cancer treatment. Further clinical trial is needed.
Material & Methods
Target and organ at risk contours
We used a free-breathing CT and a 4D-CT data set showing the same 3 cm tumor in the head of the pancreas. A 0.3 cm CTV, a 0.7 cm superior -inferior expansion for the ITV and a 0.5 cm 3D expansion for the PTV were added. No elective nodal region was treated. Then, we studied the same PTV translated every 0.5cm toward the tail of the pancreas for a total of 11 different tumor positions, starting from the initial position up to 5 cm to the patient’s left. The contoured organs at risk (OAR) are illustrated on figure 1.
Dose objectives and constraints
Table 1 lists objectives and constraints that we used for IMRT and also as a forward planning guidance.
Fig 3 Distance in mm (mean±SD) between the GTV margin and isodoses lines (50, 54, 60 Gy) for all plans with ideal target coverage, IMRT (black) vs proton plans (blue)
The duodenum has a “C” shape surrounding the head of pancreas to the patient’s right. Thus, to achieve dose escalation, the distance between the duodenum and the GTV should be at least more than 18 ± 4 mm (60 Gy isodose distribution) with IMRT (with the smallest achievable distance in an oblique direction being 13 ± 2 mm) while 18 ± 1 mm anteriorly and 25 ± 1 mm to the patient’s right is needed for proton plans.
The small bowel has mainly an antero-inferior distribution compared to the pancreas. Thus, for tumors located in the body or tail of pancreas, small bowel volume within 23 ± 5 mm (50 Gy isodose distribution) from the GTV can preclude the possibility of dose escalation with IMRT, and 20 ± 1 mm anteriorly and 23 ± 1 mm inferiorly for proton plans.
Impacts of tumor location along the pancreas
Since the PTV overlapped with one or more OAR at each position, a simultaneous adherence to target objectives and OAR constraints was not possible with any plan. Table 2 shows ITV and CTV V72Gy results according to tumor positions for IMRT, proton and 3DCRT plans that respected OAR constraints. Globally, IMRT plans were more conformal in the higher dose-gradient region circumferentially; this might facilitate dose escalation for cases where OAR are closely surrounding the GTV in almost all directions. On the other hand, one advantage of proton therapy can be appreciated in table 2, where the CTV coverage was better for proton plans than IMRT for tumors in the head of pancreas (#2-3). These positions highlighted the main advantage of proton therapy and its Bragg peak with rapid dose fall-off: the anterior close location of Treitz’s angle (the beginning of the jejunum) compared to the tumor was in the downstream direction of proton beams. Proton therapy would be a good clinical alternative in such anatomical situation. With a
Conclusion
a)
Table 1 Objectives and constraints (* for IMRT plans only)
b)
Discussion
Plan evaluation and comparison
We compared the distance between GTV and specific isodoses linked to OAR tolerance (50 Gy, 54 Gy and 60 Gy) for plans with ideal target coverage. We made measurements at the isocenter level, which was in the middle of each GTV. We also compared DVH for all plans. Finally, we analysed the effect of tumor location on target coverage compromise to respect OAR sparing constraints for that particular case.
The knowledge of isodose distributions according to treatment modality gives an indication of which pancreatic cancer cases allow safe dose escalation according to OAR anatomy. IMRT allows more conformal dose distribution in the high dose regions while proton therapy reduces low dose bath irradiation to the body. The forward planning technique associated with passive scattering proton therapy and proton beams’ large penumbra with depth preclude proton therapy’s full potential for dose escalation, and intensity -modulated proton therapy might be a solution.
References
Fig. 1 Relation between different GTV positions and organs at risk.. a) Target volumes (GTV - in the middle): green, initial tumor (position #1); yellow, position #7; red, position #11. b) Organs at risk = Yellow: the liver; pink: the stomach; violet: the duodenum; blue: small bowel, dark blue: right kidney; light bleu: left kidney; red: the spinal cord.
Fig. 2 Beam orientation for a) all IMRT plans and b) an example of proton plans
Fig. 4 DVH, initial tumor position. Dots:3DCRT; plain: IMRT; dash: proton plans.
1) Pollack et al. Prostate cancer radiation dose response: results of the M.D. Anderson phase III randomized trial. IJROBP 2002; 53:1097-1105. 2) Socinski et al. Dose-escalating conformal thoracic radiation therapy with induction and concurrent carboplatin/paclitaxel in unresectable stage IIIA-B nonsmall cell lung carcinoma. Cancer 2001; 92: 1213-23.