In conventional treatment, the tolerance of late responding tissues within the field is the main feature that constrains the size of the dose that should be delivered. The a/b ratio is therefore a useful parameter. But it has limitations. It may vary for different cells within a tissue. For example, bone marrow stem cells have a low a/b ratio even though bone marrow failure after whole body irradiation has a high a/b ratio consistent with loss of progenitor cells [18
], and depend upon the steady state kinetics of the tissue compartments [19
]. The LQ model was chosen to fit experimental data well at around 2 Gy dose fractions but predicts a survival curve that continuously bends, and at high doses bD2
in the LQ equation dominates. This means that high ablative doses would be predicted to kill unrealistically large numbers of cells and the LQ model does not necessarily fit data well above around 6–7 Gy fraction sizes [19
]; although some have argued otherwise [21
]. Hybrid isoeffect models have been introduced as a result that are based on in vitro data [22
], although extrapolation from in vitro to in vivo may be inappropriate as vascular effects may contribute to cell kill in vivo and these could differ in dose-dependency from other target cells [23
]. Irrespective of the model used, oligofractionated doses of 3 ´ 20 Gy are very high. Using the LQ model, 3 ´ 20 Gy is equivalent to around 275 Gy in 2 Gy fractions to late responding tissues with an a/b ratio of 3 Gy [11
]; clearly a dose that would not be given clinically. It is hard to escape the conclusion that above about 7 Gy the radiobiology changes.
6.3 Radiobiology and Immunobiology of Radiation Therapy and the Tumor Microenvironment
High single doses or high dose fractions are biologically more cytotoxic than the same dose delivered in smaller fractions. Many factors contribute to this. b-type killing increases with dose as the DNA undergoes more multiple simultaneous hits and suffers more complex unrepairable lesions. Also, if they do not die by rapid apoptosis, cells irradiated with 2 Gy can go through several divisions before life or death is finally decided with the number decreasing with increasing dose [15
]. At the same time, the mode of cell death may change at higher doses, for example enrolling apoptotic machinery [27
]. Reoxygenation of tumor tissue may be compromised by accelerated treatments, although this is a rapid process that is enhanced by tumor cell loss and may be less of a factor.
Throughout the history of RT there has been arguments as to the relative contributions of vascular and parenchymal cell death of cells to damage to normal tissues and tumors. Importantly, high dose fractions have often been thought to preferentially compromise vasculature. Recently, Fuks and Kolesnick noted that “rapid endothelial cell death in tumor displays an apparent threshold at 8–10 Gy and a maximal response at 20–25 Gy” [23
], although others [28
] showed that fractionated protocols could cause vascular loss not too dissimilar to that of high single doses. In the latter study, loss of tumor microvasculature was seen to develop over a period of weeks, with an increase in chronic hypoxia that was associated with an influx of macrophages [28
] and may be a consequence of early endothelial cell loss. Even with such vascular damage, tumors may regrow. Regrowth however appears to depend on vasculogenesis rather than angiogenesis, and is enhanced by radiation-induced infiltration of bone marrow-derived myeloid cells [28
]. Vasculogenesis is a relatively inefficient process and could lead to an increase in hypoxia. The failure of irradiated normal tissue to support angiogenesis has been known for decades as the tumor bed effect [30
]. Microvascular “pruning” by radiation may be a direct cytotoxic effect or indirectly mediated by radiation-induced pro-inflammatory cytokines like tumor necrosis factor (TNF) family members and their associated receptors, as the vasculature is a major target for these agents [31
]. Such radiation-induced pro-inflammatory cytokine production in tissues is generally more marked above 6 Gy [32
]. How common these radiation-induced alterations in the tumor microenvironment are in clinical reality and what their relationship is to tumor cure has yet to be fully established, but for SBRT and SABR the overall radiation delivery time may be critical, as well as the size of dose per fraction. If microvasculature loss occurs slowly over several weeks after the start of tumor irradiation, classically fractionated RT protocols may still be ongoing and be impacted more than accelerated hypofractionated or oligofractionated regimens, which will be completed.
Another consequence of RT that may be impacted by the new delivery techniques is its ability to act as an immunological adjuvant, even in humans [33
]. Clinically, if RT treatments can be optimized to promote anti-tumor immunity, this could increase the odds of achieving local cancer control and combat growth of micrometastases. Multiple mechanisms may operate to mediate radiation-induced immune effects. Radiation can induce production of pro-inflammatory cytokines like TNF and interleukin 1, and cell adhesion molecules by both cells and tissues [34
], especially in the higher dose range. This contributes the “sense of danger” [32
] and immune recognition that is a consequence of maturation of dendritic cells to present antigen and that translates innate into adaptive immunity . Radiation can additionally promote this pathway by increasing MHC class I expression on tumor cells and antigen presenting dendritic cells , along with tumor antigenic peptides and immune co-accessory molecules. The underlying mechanisms are not fully established but may be mediated either directly or indirectly through cytokine induction [36
The immune environment that follows tissue irradiation is therefore generally pro-oxidant, pro-inflammatory, and pro-immune. The microenvironment that is created inevitably involves cell death, caused either directly by radiation or indirectly. In many ways, the events are not dissimilar to those developed in a local infection [41
]. In such pathological circumstances, even cells that die by apoptosis can be immunogenic [42
]. The signals that are sent by cells dying under such circumstances are designed to enhance innate immune recognition of pathogens [43
] and associated damage-associated molecular patterns (DAMPS ) on cells. Chemokines, cytokines , and other wound-associated molecules and a host cell infiltrate are generated to initiate a healing process. This RT-induced scenario would be expected to assist the generation of specific immunity , and this has been reported to occur in many different animal tumor models [44
]. Indeed, irradiation of dendritic cells themselves is sufficient to make them better at tumor antigen cross-presentation, although presentation of endogenous antigen was simultaneously decreased [46
]. The fact that tumor-specific T cells can increase in the peripheral blood of patients during and after RT is an important verification of the potential adjuvanticity of RT in a proportion of patients [33
In spite of the very positive data accumulating to show that RT can act to enhance anti-tumor immunity and aid in tumor cure, there are also examples where there is little evidence of enhanced tumor-specific responses [47
]. The myeloid infiltrate that is generated by many tumors can be highly immunosuppressive [48
]. As already mentioned, following RT-induced vascular damage myeloid cells congregate in hypoxic regions in at least some tumors [49
]. These tumor-associated macrophages secrete arginase, nitric oxide synthetase, and COX2 and can promote tumor growth. Targeting the myeloid compartment has been shown in some cases to enhance the efficacy of RT [2929
The immune system is of course under considerable control by cells other than macrophages and the role for example of regulatory T cells has to be considered, and RT is also able to generate regulatory T cells as a response to damage just as it can generate immunity [50
]. It is likely that the immune balance is critical to the outcome and radiation modulates this. To be effective, RT must overcome these immunosuppressive mechanisms and the long time over which cancer develops is likely to be a detriment to a successful immune outcome, as immune editing of tumor antigen expression and tolerance induction [51
] are likely and will precede the development of immunosuppressive regulatory T cells and macrophages that begin to participate only as tumor becomes palpable [51
The major question is whether RT can translate non-immunogenic cancers into effective immune stimulators and targets for tumor-specific responses. An alternative is that RT simply enhances a preexisting immune response that exists in some patients. Other questions relate to the factors released from the tumor that generate the myeloid cells, such as bone marrow colony stimulating factors. These may be an intrinsic feature and even a marker of tumors that are destined to fail RT. Patients with such tumor may be unable to generate tumor immunity in response to RT that may not be able to counter the growth enhancing effects of the myeloid suppressors and may even encourage their participation. Such patients may require alternative or additional cancer treatments.
Even the extent of “danger” that is elicited by RT may be limited by the tumor microenvironment . Again, here the radiation dose and fraction size may make a critical difference. In general, the evidence presented above show that high doses or dose fractions appear more effective in many regards than standard 2 Gy fraction sizes. Whether SABR very high dose fractions are superior to more moderate sized doses is however a more open question. The preliminary data so far suggest that high single doses of ablative RT may not be optimal for the generation of tumor-specific immunity , and moderately high hypofractionated doses may be superior in this respect [52
], although high single doses can be effective [44
]. This issue requires clinical immune studies to be performed with different radiation protocols for it to be resolved.
Under any circumstances, SABR that is ablative will cause both vascular and parenchymal cell loss. Critical questions relate to the site of this loss and the volume involved and how this will influence the response. Volume may be a variable that has important repercussions on the radiation-induced microenvironmental and immune changes. For example, ablative regimens may allow angiogenesis in surrounding normal tissue to effect better normal tissue recovery, limit the hypoxia that develops, and minimize the immune infiltrate. Since tube-shaped serially organized parenchymal structures like nerves and bronchioles may develop serious complications following ablative therapy, dose is most often moderated for tumors that are centrally located in the lung. In the periphery, loss of tissue function in small areas is less important as there is ample residual lung function.
Kjellberg introduced a volume nomogram for the risk of brain damage from treatment of arterio-venous malformations with SRS [54
] that has been further refined over subsequent years, but the volume constraints that must be applied for ablative therapy in most sites are essentially unknown. IMRT treatment planning does give a read-out for dose-volume histograms that have been used to determine dose limits for conventional treatment in various normal tissues, but these can be misleading as they contain no spatial information and are poor predictors of toxicity. Also, it is unclear to what extent they are relevant for tissue ablative or hypofractionated regimens. The biology of volume effects in RT is likely to be complex. For example, experimental studies on irradiated spinal cord in rats have indicated a “bath-and-shower” effect [55
] for small fields that have a high tolerance for radiation-induced paralysis. In such models if a modest “shower” dose (about 4 Gy) is administered to the area surrounding the targeted “bath” volume, tolerance dramatically decreases. The mechanism underlying this effect is not known but angiogenesis and cell migration are suggestions.
The importance of volume and the nature of the radiation delivery for the generation of tumor immunity and in particular out-of-field abscopal effects [56
] remain to be established. Not only is it likely that more inflammatory responses and “danger” signals for tumor immunity are generated by high ablative doses, but with IMRT these will be more focal and heterogeneous in nature with a steep dose falloff. This contrasts with the large homogeneous fields that tend to be used in more classical treatments. With focused ablative therapy, one could envision a cytokine gradient that will promote more immune cell infiltration and a classical localized inflammatory nidus within the tumor and stimulation of proliferation in the surrounding normal tissue cells that receive a low dose to form a protective barrier [41
]. A negative potential consequence of such proliferation might be carcinogenesis.