A Novel Framework for the Optimization of Simultaneous ThermoBrachyTherapy

Abstract

In high-dose-rate brachytherapy (HDR-BT) for prostate cancer treatment, interstitial hyperthermia (IHT) is applied to sensitize the tumor to the radiation (RT) dose, aiming at a more efficient treatment. Simultaneous application of HDR-BT and IHT is anticipated to provide maximum radiosensitization of the tumor. With this rationale, the ThermoBrachyTherapy applicators have been designed and developed, enabling simultaneous irradiation and heating. In this research, we present a method to optimize the three-dimensional temperature distribution for simultaneous HDR-BT and IHT based on the resulting equivalent physical dose (EQDphys) of the combined treatment. First, the temperature resulting from each electrode is precomputed. Then, for a given set of electrode settings and a precomputed radiation dose, the EQDphys is calculated based on the temperature-dependent linear-quadratic model. Finally, the optimum set of electrode settings is found through an optimization algorithm. The method is applied on implant geometries and anatomical data of 10 previously irradiated patients, using reported thermoradiobiological parameters and physical doses. We found that an equal equivalent dose coverage of the target can be achieved with a physical RT dose reduction of 20% together with a significantly lower EQDphys to the organs at risk (p-value < 0.001), even in the least favorable scenarios. As a result, simultaneous ThermoBrachyTherapy could lead to a relevant therapeutic benefit for patients with prostate cancer.

Keywords: biological modeling; brachytherapy; hyperthermia; induced; interstitial hyperthermia; linear quadratic model; prostate; prostatic neoplasms; thermoradiotherapy; treatment plan optimization.

Figure 1

Graphical summary of the proposed optimization framework. 1. The planning CT is the initial input of the process. 2. The patient model is generated. 3. The EM field per electrode is precalculated. 4. The temperature distribution per electrode is precalculated. 5. The BT dose distribution is imported from the BT treatment planning software. 6. The TDLQ model is used for the calculation of the combined effect. 7. Both temperature and dose constraints and objectives are used for the optimization process. 8. The TDLQ optimization process optimizes the IHT parameters (8.4) that generate a temperature distribution (8.1) from which an EQD_phys distribution is generated (8.2). This EQD_phys distribution is used to compute the objective function, which needs to be minimized (8.3).

Figure 2

(a) Axial CT slice showing the anatomy of a patient with afterloading catheters (visible as black dots indicated by the arrow) inserted in the prostate. (b) Same axial slice of the resulting patient model after segmentation of all tissues. (c) Lateral 3D view of the prostate, OAR, and simulated TBT applicators in the same patient.

Figure 3

Flowchart illustrating how the TBT EQD_phys distribution is generated from and compared with the original BT-only dose distribution.

Figure 4

Comparison between superpositioned temperature calculation and FEM recalculation on the central axial slice in the prostate. (a) Temperature map using superpositioning of separate FEM calculations for each electrode. (b) Temperature map using a single FEM calculation for the same, combined, electrode settings. (c) γ-index map of the comparison.

Figure 5

Isodose curves for the EQD_phys resulting from different combinations of physical dose and temperature for 1 h: (a) EQD_phys assuming DU-145 data; (b) EQD_phys assuming PC-3 data.

Figure 6

TBT TP results assuming thermoradiobiological parameters equal to the average between DU-145 and PC-3 data. (a) The original clinically applied HDR-BT dose fraction. (b) The applied TBT physical dose (80% of original). (c) Temperature volume histogram showing the temperature coverage in the prostate and OARs. (d) The temperature distribution calculated for the optimal TBT plan. (e) The TBT EQD_phys resulting from the combined treatment. (f) Dose volume histogram of the prostate and OARs for the HDR-BT-only dose (dashed line) and the TBT EQD_phys dose (solid line).

Figure 7

Average values (±standard deviation) over 10 simulated patients of TBT prostate V_100% (a), V_150% (b), and V_200% (c) for different scalings of the original HDR-BT dose. The different colors correspond to the plans generated based on different thermoradiotherapeutic values (red for DU-145, blue for PC-3, and green for the average between DU-145 and PC-3). It is evident that the original prostate coverage is met when the physical dose is scaled over 80% of the original value. The vertical bars correspond to standard deviation. The horizontal dashed lines correspond to the objective (V_100%) and soft constraint (V_150% and V_200%) limits. The green and red areas correspond to targeted and constrained values, respectively.

Figure 8

Average values (±standard deviation) over 10 simulated patients of the optimal TBT prostate T₁₀ (a), T₅₀ (b), and T₉₀ (c) for different scalings of the original HDR-BT dose. The different colors correspond to the plans generated based on different thermoradiotherapeutic values (red for DU-145, blue for PC-3, and green for the average between DU-145 and PC-3). The vertical bars correspond to standard deviation.

Figure 9

Average values (±standard deviation) over 10 simulated patients of the TBT TP parameters for different scaling of the original dose: urethra D_0.1cc (a), rectum D_1cc (b), and bladder D_1cc (c). The different colors correspond to the plans generated based on different thermoradiotherapeutic values (red for DU-145, blue for PC-3, and green for the average between DU-145 and PC-3). The black line shows the lowest possible value, assuming no radiosensitization of the normal tissue. The horizontal dashed lines correspond to the constraint limit for each criterion. The red areas correspond to constrained values for each criterion.