High Z-enhanced synchrotron stereotactic radiotherapy relies on the dose-enhancement obtained when tumors, previously filled with a high-Z elements, are irradiated with medium energy x-rays (50–100 keV) in stereotactic conditions. The concept comes initially from the observation in the late 70’s, of additional blood damages in pediatric diagnostic radiology, when using contrast agents 1 . The use of medium energy x-rays to treat cancer could appear surprising nowadays, specially for brain tumors, but as the photoelectric cross section increases proportionally to Z⁴/E³ (where Z is the atomic number of matter and E the energy of photons), there is a subsequent increase of the absorbing properties restricted to the target level, due to the release of secondary particles (photoelectrons, characteristic x-rays and Auger electrons), which deposit most of the initial photon energy in the close vicinity of the primary interaction. Photoelectric effect is the photon interaction that deposits locally the largest part of the photon initial energy (when compared to coherent or incoherent scattering events). This leads to improved dose distributions in comparison with conventional high energy treatments. Numerous studies have been performed for establishing that this method meets dosimetry criteria for patients 2 3 4 5 6 7 8 . From 50 to 80 keV, the brain half value layer increases from 2.93 to 3.64 cm. Although these values are relatively small, the dose is increased by (i) the irradiation geometry and (ii) by the presence of sufficient amount of high Z elements inside the tumor volume (≈ 3–10 mg/mL). LINAC spectra extend from MV to kV energies, however, the contribution of kV radiation in the dose-enhancement is negligible, as shown with Monte carlo simulations or experimentally using gel dosimetry 2 3 4 5 9 . The use of a modified conventional CT scanner, such as proposed by Norman and co-workers 10 , appeared to be the main limitation of the technique due to (i) the beam shape (fan or cone-shaped beam with a high aperture) which broadens the isodoses as the beam goes through the patient; (ii) a low photon flux and a high thermal load which impose long irra-diation times and cooling phases; and (iii) a broad polychromatic spectrum of the beam which increases the dose to the skull and minimizes the dose enhancement effect. For these reasons, nearly parallel intense and tunable monochromatic beams provided by synchrotron source appear to be a must for this radiotherapy technique.

In a series of publications 11 12 13 14 we have reported on the therapeutic efficacy of short-term intracerebral (i.c.) convection enhanced delivery (CED) of either carboplatin or cisplatin or alternatively prolonged intratumoral (i.t.) infusion of carboplatin either alone or in combi-nation with X-irradiation for the treatment of the F98 rat glioma 11 12 13 . Irradiations were carried out at the European Synchrotron Radiation Facility (ESRF) using 78.8 keV synchrotron X-rays or 6 MV photons, by a medical linear accelerator (LINAC), at the University Hospital of Grenoble, France. Carboplatin was selected for these studies because we previously have shown that it was highly effective in treating F98 glioma bearing rats 11 12 13 14 . However, platinum containing drugs have their limitations 15 16 17 for the treatment of brain tumors. These include inadequate dose-limiting toxicity and reduced uptake by brain tumors following systemic administration due to the blood brain barrier (BBB) 18 . Delivery of carboplatin by CED was well tolerated when delivered i.c. to F98 glioma bearing rats 11 12 14 19 20 21 and non-human primates 22 and resulted in prolonged survival and cures of the former.

Cure rates of 20% to 55% were obtained in F98 glioma bearing rats treated with prolonged infusions of either carboplatin or cisplatin using Alzet osmotic pumps alone or in combination with synchrotron X-irradiation. In these studies, the beam energy was tuned at 78.8 keV, which was just above the K-edge of Pt 12 23 . The first study carried out at the ESRF with cisplatin 23 , employed synchrotron X-rays and it was hypothesized that therapeutic efficacy was dependent upon the production of Auger electrons and photoelectrons following irradiation of Pt atoms with monochromatic X-rays. Above the Pt K-edge energy (78.4 keV), extraction of electrons from the K-shell by the photoelectric effect results in the creation of vacancies. The resulting gaps are filled successively by radiative (96%) and non-radiative (4%) transitions from outer shells, thereby resulting in the release of several low energy photons and electrons. If the Pt atoms are located near or within DNA, the emitted low energy electrons can be highly destructive for DNA and lethal to tumor cells 24 , even with small concentrations of Pt. We further hypothesized that drugs such as cisplatin or carboplatin were ideal candidates for this approach, since they had two advantages: first, they were DNA alkylating agents 25 and second, they were efficient carriers of Pt, which has a high Z number. The photoelectric cross-section of Pt at 78.8 keV is 2860 barns/atom, 4.8 times greater than at 78.0 keV. Therefore, if therapeutic efficacy were related to the emission of Auger electrons and photoelectrons from the Pt atoms, a greater therapeutic gain should have been observed with X-irradiation above rather than below the Pt K-edge. Furthermore, and most impor-tantly, almost no enhancement should be observed with 6 MV photons. This is because the Pt photoelectric cross is < 1 barn/atom above 1 MeV (i.e. >2860 × less than the Pt cross section above its K-edge), which is not significantly different from that of water. The Compton interaction process is dominant in this energy range and does not induce local energy deposition, contrary to the photoelectric effect. However, our studies carried out over the past 4 years have called this interpretation into question. The best survival data and cure rate (55%) ever reported with F98 glioma model were obtained by combining i.c. administration of carboplatin by means of Alzet osmotic pumps followed by synchrotron X-irradiation tuned at 78.8 keV. Therefore, it was important to carry out another study under similar conditions with 6 MV X-rays using a LINAC instead of synchrotron radiation to demonstrate that this effect was independent of the X-ray source. In the present study we have shown the equivalency of synchrotron X-rays 11 26 and 6 MV photons in combination with prolonged i.c. administration of carboplatin to produce prolonged survivals and cures of F98 glioma bearing rats.

All operative procedures related to animal care strictly conformed to the Guidelines of the French Government (licenses #380324 and #A3818510002). The protocol was approved by the Grenoble Institute of Neurosciences Ethical Committee (H. Elleaume, PhD, permit #381026). Experiments were performed under anesthesia, and every effort was made to minimize the number of animals used and to alleviate pain and suffering during the experimental procedures.

In the present study we have demonstrated that equivalent survival data were obtained in F98 glioma bearing rats that had been treated with the combination of i.c. infusion of carboplatin in combination with radiation therapy using either 6 MV photons from a LINAC or a monoenergetic beam of 78.8 keV X-rays from a synchrotron. Bernardt et al. have described the influence of relaxations of atoms attached to DNA on radiation-induced cellular DNA damage by low energy photons using Monte Carlo track structure calculations 24 . They found that the number of inner shell relaxations produced by photon irradiation generally was small in comparison to the total number of double strand breaks (DSBs) generated by the radiation itself, when the number of high Z atoms introduced by Pt containing cytotoxic drugs was small and compatible with cell survival studies.

Using atomic absorption spectroscopy, Guarnieri et al. and Kahn et al. have mapped the distribution of platinum after i.c. infusion of carboplatin with ALZET pumps into F98 glioma-bearing rats, with delivery parameters similar to those that we used. Platinum concentrations were maximal in brain sections corresponding to the infusion site, with diminished amounts (5 to 1 μg/g tissue) in sections that were 3 mm from the point of infusion 27 28 . The importance of the DNA damage is dependent on the number of Pt atoms intercalated with DNA molecules. At the molecular level, a larger number of DSBs were detected when cells were pretreated with cisplatin and subsequently irradiated with synchrotron X-rays above the Pt K-edge, compared to those below the K-edge 23 29 . Three times more DSBs were detected when human SQ20B squamous carcinoma cells pretreated with 30 μM cisplatin (3 ×× 10⁸ atoms of Pt atoms per cell) for 6 h 29 , and 1.3 times more DSBs with the same treatment of F98 cells 23 . However, no such an enhancement was observed (even at the molecular level) with the much lower Pt concentrations that would not have been tumoricidal, when the SQ20B cells were pretreated with 3 μM cisplatin (4 × 10⁶ Pt atoms per cell) for 6 h 29 . In our studies, i.t. injection of cisplatin (3 μg in 5 μl), followed 24 h later by 15 Gy of X-irradiation, also produced similar long-term survival of F98 glioma bearing rats, irrespective of whether the synchrotron X-rays had energies below or above the Pt K-edge 23 . Comparable long term cure rates (17% and 18%) also were observed when the animals were irradiated with 78.8 keV synchrotron X-rays or 6 MV photons after cisplatin (6 μg in 20 μl) was administered i.c. by CED 13 . Overall, the present data and those previously reported 11 12 13 23 29 are in good agreement with Bernhardt et al’s. predictions 24 . They strongly suggest that the therapeutic gain obtained by the direct i.c. administration of Pt compounds, followed by X-ray irradiation, was not due to the production of Auger electrons and photoelectrons emitted from the Pt atoms, but rather involved other mechanisms. Only molecular studies performed using extremely high Pt concentrations, which were not attainable in vivo, demonstrated energy dependence. However, this is not an adequate explanation for the in vivo therapeutic efficacy of the combination of Pt based chemotherapy with X-irradiation. In order for synchrotron radiation therapy to be successful, a sufficient, but not lethal, concentration of high Z number atoms must be incorporated into or localized nearby tumor cells, to produce enough photoelectrons or Auger electrons. Other elements, such as iodine, in the form of contrast agents, or iodo-deoxyuridine (IUdR), gold, platinum or gadolinium nanoparticles are under investigations as radiation sensitizers for synchrotron stereotactic radiotherapy 30 31 32 33 34 35 . In a recent review, Kobayashi et al. 36 discussed the enhancement of radiobiological effects by heavy elements, in particular gold and platinum. Auger enhancing phenomena to electron and Hadron therapy is also suggested which broadens furthermore their therapeutic applications.

In another study 37 we have used the same chemotherapy protocol, but a different irradiation scheme: the dose was delivered in three fractions of 5 Gy using 6 MV photons and the whole brain was irradiated, beginning on the day after drug administration, using the same Alzet osmotic pumps. The results are very consis-tent with the data presented here, the chemotherapy groups had the comparable survival rates (MST of 77 d ± 23.0 and 71 d ± 7 and 16%, 14% long term survival rates, respectively). Rats bearing tumors, treated with carboplatin and X-irradiation had MST and (MeST) of 111.8 d (78 d), with 40% surviving more than 180 d (i.e. cured), compared to 77.2 d (59 d) for pump delivery of carboplatin alone and 31.8 d (32 d) for X-irradiated alone. There was no microscopic evidence of residual tumor in the brains of all long-term survivors. The biologically equivalent dose-fraction (BED) can be calculated using the classic linear quadratic equation 38 39 :

BED=n.d[1+d .(α/β ) − 1 ]

where n is the number of fractions, d is the dose per fraction in Gy, and α and β are two variables that indicate the sensitivity of tumor or normal tissue to changes in dose fractionation. The α/β ratio is usually taken to be 10 for tumor and early-reacting tissues and 3 for late-reacting tissues like brain. The biologically effective dose (BED) for 15 Gy, delivered in a single fraction, using the α/β ratios indicated above, is 37.5 Gy in acute and tumor effects and 90 Gy in late effects (37). In comparison, the BEDs for 15 Gy delivered in three fractions of 5 Gy each are largely lower: 22.5 and 40.0 Gy, for tumor and normal brain, respectively. The dose per fraction should be 8 Gy, for obtaining BEDs in a three fractions regimen equivalent to those of 15 Gy delivered in a single fraction 11 . The enhanced survival results obtained using a single fraction of 15 Gy, using either 6 MV X-rays (this study) or synchrotron radiation 12 , in comparison with 15 Gy delivered in 3 fractions 37 is in good agreement with the calculated equivalent BEDS of these irradiation schemes.

The present study firmly establishes the equivalency of i.c. administration of carboplatin either by infusion via osmotic pumps or CED with irradiation with 6 MV X-rays and synchrotron X-rays. Since medical LINACs are widely available worldwide, this could provide the opportunity to clinically evaluate this combination therapy at multiple centers. Intracerebral CED of carboplatin in patients with recurrent glioblastomas (GBM) will be evaluated in a soon to be initiated Phase I at The Ohio State University. If successful, this could lead to a Phase II clinical trial evaluating the combination of i.c. of carboplatin and radiation therapy to treat patients with recurrent GBMs, for whom unfortunately there are presently no good therapeutic options.

The authors declare that they have no competing interests.

LB and HE carried out the studies and drafted the manuscript. ME, PD, JFA and FE participated to the experimental studies. JLR participated in the design of the study and in the drafting. JB participated to the irradiation and help to draft the manuscript. JR and RFB participated in the drafting. All authors read and approved the final manuscript.