The application of PET-CT to post-mastectomy regional radiation therapy using a deformable image registration

Abstract

Background: To evaluate the utility of the preoperative PET-CT using deformable image registration (DIR) in the treatment of patients with locally advanced breast cancer and to find appropriate radiotherapy technique for further adequate treatment of axillary nodal area.

Methods: Sixty-five breast cancer patients who had level II, III axillary or supraclavicular lymph node metastasis on ¹⁸F-FDG PET-CT and received postoperative radiotherapy after modified radical mastectomy were enrolled. One radiation oncologist contoured normal organs (axillary vessels, clavicular head, coracoids process and humeral head) and involved lymph nodes on PET-CT and simulation CT slices. After contouring, deformable image registration of PET-CT on simulation CT was carried out. To evaluate the performance of the DIR, Dice similarity coefficient (DSC) and Center of mass (COM) were used. We created two plans, one was the historically designed three field plan and the other was the modified plan based on the location of axillary lymph node, and we compared the doses that irradiated the axillary lymph nodes.

Results: The DSCs for axillary artery, axillary vein, clavicular head, coracoids process and humeral head were 0.43 ± 0.15, 0.39 ± 0.20, 0.85 ± 0.10, 0.72 ± 0.20 and 0.77 ± 0.20, respectively. The distances between the COMs of axillary artery, axillary vein, clavicular head, coracoids process and humeral head in simulation CT and from PET-CT were 13.0 ±7.1, 20.2 ± 11.2, 4.4 ± 6.3, 3.7 ± 6.7, and 9.5 ± 25.0 mm, respectively. In the historically designed plan, only 57.7% of level II lymph nodes received more than 95% of prescribed dose and the coverage was improved to 70.0% with the modified plan (p < 0.01). For level III lymph nodes, the volumes received more than 95% of prescribed dose were similar in both plans (96.8 % vs 97.9%, p = 0.35).

Conclusion: Deformable image registration of PET-CT on simulation CT was helpful in the identification of the location of the preoperatively involved axillary lymph node. Historically designed three-field plan was not adequate to treat the axillary level II lymph node area. Novel treatment technique based on the location of axillary lymph node from PET-CT using DIR can result in more adequate coverage of nodal area.

Figure 1

Normal organs were contoured on simulation CT images. Axillary vessels, clavicular head, coracoids process and humeral head were delineated.

Figure 2

Representative images after deformable image registration. A: Simulation CT images. B: PET-CT images. C: Fusion images of deformed PET and simulation CT.

Figure 3

Digitally reconstructed radiographs (DRR) image of each field. A: Anterior field. It is the historically designed field. B: Tangential fields. They also follow the historically designed fields. C: Modified anterior field. It is 1 cm larger below the junction compared to the historical field. D: Posterior axillary boost beam. Sky blue line represents the area received the dose below 90% of the prescribed dose.

Figure 4

The graphs show Dice similarity coefficient (DSC) and the distance between the Centers of mass (COMs) of structures. The error bars represent 95% confidence interval.

Figure 5

Location of level II axillary lymph node. A. The distribution of the maximal length from the axillary vessels to the level II lymph node. B. The distribution of the maximal depth from surface to level II lymph node.

Figure 6

Representative dose distribution and dose volume histogram (DVH) of axillary lymph nodes and ipsilateral lung in both plans. A. Dose distribution in the historical plan at the level of level II lymph node. Orange colored area is initially involved level II lymph node area. Green line shows the area received 96% of the prescribed dose. B. Dose distribution in the modified plan. C. DVH of axillary LNs and ipsilateral lung in the two plans. The dose irradiated to the level II lymph node increased in the modified plan. The dose to level III lymph node in the modified plan was lower than in the historical plan but the doses were more than prescribed dose in both plans. The maximum dose decreased in the modified plan.