Fighting cancer is one of the top public health priorities. Radiotherapy is an effective treatment modality used in patients with different malignancies. However, its efficacy against some radioresistant tumors remains as low as 30–50% [1]. It can be improved by increasing the absorbed radiation dose at the cost of damage to healthy surrounding tissues. One of the methods of increasing the absorbed dose while sparing healthy tissues is contrast-enhanced radiotherapy (CERT), in which dose enhancement is achieved by injecting or otherwise delivering dose-enhancing agents (DEA) to the tumor. DEA are chemical elements with high Z numbers > 52, such as I, Gd, Au, Pt, Bi, etc. These elements readily absorb external X-rays and therefore can be used to enhance the dose absorbed by the tumor at the site of their uptake [2–4]. Unlike conventional radiotherapy in which tumor geometry is important, CERT precision is ensured by the tumor-tropic properties of DEA- based pharmaceuticals. The antitumor effect of CERT has been demonstrated in a number of research studies conducted in animals [5–11]. Unfortunately, the obtained results cannot be translated into clinical practice, because a treatment outcome cannot be predicted without establishing a correlation between the observed therapeutic effect and the absorbed radiation dose/its distribution in the tumor.

The hardest part of both research and clinical studies of CERT efficacy is dosimetry, which is also its least elaborated component. CERT is a binary modality; the absorbed dose and its distribution in the tumor volume are determined not only by the parameters of external radiation, but also by DEA concentration and distribution in the tumor volume.
The intratumoral route of DEA administration is used in the studies of CERT tumor suppressing efficacy both in lab animals [12–17] and in real patients in the clinical setting [18].
The aim of this work was to study the distribution of bismuth during CERT in the volume of a tumor grown from B16F10 murine melanoma cells following the intratumoral administration of its single dose and to analyze the dose-volume histogram data.

METHODS

The study was conducted in С57Bl/6 female mice weighing 20 to 22 g purchased from Stolbovaya breeding and nursery laboratory (Research Center for Biomedical Technologies of FMBA; Russia). The animals were housed in a conventional facility under natural lighting conditions. Murine melanoma B16F10 was used as a tumor model. The 14% cell suspension in 0.2 ml Hanks balanced salts prepared ex tempore was injected subcutaneously in the middle third of the right hind leg. DEA distribution was measured in the tumors of 6 animals once the malignancies reached ~700 mm³ in size. Tumor dimensions were measured with a caliper in three perpendicular planes and the tumor volume was calculated using the ellipsoid volume formula.

Bismuth diethylenetriaminepentaacetic acid (Bi-DTPA) used in our previous research study of CERT efficacy was chosen as DEA [15]. Properties of the Bi-DTPA aqueous solution are shown in the table.
The animals received a single bolus intratumoral injection of 50 μl of the Bi-DTPA solution containing 5 mg of bismuth.
Distribution of DEA in the tumor volume in vivo was studied by micro-CT [19–21] performed on the IVIS Spectrum CT scanner (Perkin Elmer; USA). For the procedure, the mice were anesthetized with 2% isoflurane in air. The whole body scans were obtained before the injection, ~1 min after the injection, and then 3, 5, 10, 15, 20, 25, and 30 min following the injection. Examples of CT tumor images obtained before and after intratumoral DEA administration are shown in fig. 1. Once the scanning was completed, the anesthetized mice were euthanized by cervical dislocation.  

DEA concentrations in the tumors were calculated from the radiopacity per each pixel of the tomographic image. Because the absolute radiopacity values for the homogenous object returned by the IVIS Spectrum CT scanner depend on the tomographic slice number and differ significantly between in the central and peripheral zones of a studied object [22], we calculated DEA concentrations from the difference between the radiopacity of tumor tissues measured before and after the DEA injection. Calibration curves were constructed from the scans of reference tubes containing solutions with known bismuth concentrations (fig. 2). 

To estimate the absorbed radiation dose per pixel on the CT image of the tumor, the dose enhancement factor (DEF) was calculated for the corresponding DEA concentration in this particular pixel measured in every studied time interval. DEF is a ratio of the dose absorbed by an object in the presence of DEA to the dose absorbed in the absence of DEA, irradiation parameters being the same. DEF is calculated by the formula:

DEF = D_DEA (C) / D₀     

where D_DEA(C) is the absorbed dose at a point in an object containing the DEA concertation С irradiated with I₀ intensity at this particular point; D₀ is the absorbed dose at the same point in the same object with zero DEA concentration (С = 0) upon irradiation with the same intensity I₀ [23].

DEF was calculated from measured DEA concentrations based on experimental and theoretical data [2, 24]. DEF values were time-averaged in the 0–30 min interval; the same irradiation time was used in our previous work [15]. Then tumor voxels with the same average DEF values were added up to obtain dose-volume histograms. Because the presence of DEA leads to a local increase of the absorbed dose (the tumor or its part receives a radiation dose exceeding the one expected in the absence of DEA), we used modified cumulative dose-volume histograms different from those constructed for a conventional radiotherapy. The histograms demonstrate the dependency of DEF on the relative tumor volume expressed as percentages. In our study, each relative tumor volume was plotted against the minimal corresponding DEF. This type of dose-volume histograms is more informative for CERT because it visually represents the role of DEA in enhancing the radiation dose absorbed by the organ. The analysis of CT images, DEF computation and construction of dose-volume histograms were done in MATLAB (MathWorks; USA).

RESULTS

The analysis of Bi-DTPA distribution in the tumor injected with a single dose of DEA revealed that the half-life of Bi-DTPA in the tumor was 3 min. By minute 30 the tumor retained only 4% of the injected bismuth (fig. 3).
The volume of Bi-DTPA distribution measured 1 min after the injection was 219 ± 35 mm³ (24 ± 1% relative to the total tumor volume) (fig. 4).
During the first two minutes after the injection there was a competition between Bi-DTPA distribution in tumor tissues and its elimination from the tumor, which kept the DEA-containing tumor volume unchanged. Then it started to decline gradually and by min 25 following the injection was as low as 1–7% of the total tumor volume. The modified dose-volume histograms are shown in fig. 5.
As shown by the histograms, a twofold increase in the absorbed radiation dose was observed for 6% of the total tumor volume. Additional energy release exceeding the nominal radiation dose (DEF > 1) due to the presence of DEA was observed in 29% of the total tumor volume.

DISCUSSION

The obtained dose-volume histograms (fig. 5) reveal a markedly nonuniform distribution of DEA in the tumor volume. The maximal predicted DEF value at the site of the maximum DEA concentration irradiated for 28.5 minutes is 4. This ensures an absorbed radiation dose of 80 Gy at the dose intensity of 0.7 Gy/min. However, such a significant increase in the absorbed dose was observed for only 0.1% of the tumor volume. A 1.5-fold or more dramatic increase in the absorbed dose (> 30 Gy at the same dose rate) caused by the presence of DEA was observed in 10% of the tumor volume.

In order to estimate the maximum growth delay for the irradiated tumor, the following assumptions were made:
1) the irradiated tumor grows exponentially, its doubling time being Td [25];
2) the death of tumor cells is caused only by direct damage induced by radiation;
3) total cell death is observed in that 10% of the tumor volume which was irradiated with a dose over 30 Gy.

The last assumption was made to simplify the estimation of the maximum growth delay time. If some cells in the irradiated volume do survive, the growth delay will be shorter, but the minimum growth delay time will not be affected. If the volume of the cells that survive irradiation in the absence of DEA is taken as V_s, then the volume of the cells surviving irradiation in the presence of DEA injected into the tumor will be 0.9 V_s. If the doubling time is the same in both cases and equals 5 days [17, 26] and the tumor grows exponentially, then the growth delay time will not exceed 0.76 days in the tumor exposed to CERT.

However, our previous experiments [15] demonstrated a longer 10-day tumor growth delay in tumors injected with DEA and exposed to X-ray radiation in comparison with those irradiated in the absence of DEA. Obviously, the antitumor effect observed in the B16F10 melanoma injected with DEA and subsequently exposed to a single fraction of X-ray irradiation cannot be solely linked to the direct damage caused by radiation, but is supported by other mechanisms as well.

CONCLUSIONS

The analysis of DEA and absorbed dose distribution in the tumor volume revealed that a single intratumoral injection of DEA ensures its markedly nonuniform distribution in the tumor and enhances the absorbed radiation dose. To achieve a more uniform distribution of the dose-enhancing agent, multiple US/ CT-guided injections are needed. Our previous findings [15] highlight the importance of studying the mechanisms of CERT tumor suppression efficacy following DEA delivery to the tumor by the same route of administration.