(Chemo)Radiotherapy–Immunotherapy Combinations: Time to Get Tailored?

In this issue of Clinical Cancer Research, Chen and colleagues performed an analysis of tumor biopsies and peripheral mononuclear blood cells (PBMC) taken before and during the course of concurrent chemoradiotherapy (CCRT) of cervical cancer patients, representing one of the first clinical studies looking for changes in immune markers and in the interferon response during treatment (1). Not only had they observed large heterogeneity in the tumor immune landscape at baseline, in line with the results of studies on gene expression profiles (2), they also demonstrated that in response to therapy changes are highly variable with increases, decreases or no changes in CD8⁺ T-cell infiltration and activation of the IFN pathway (Fig. 1), two immune parameters that were significantly correlated. The tumor cohort expressed the Granzyme B inhibitory protein SerpinB9 at a high rate, indicating it as a potential target to maximize the effects of CCRT. They also observed an increase in the levels of the “don't eat me” signaling protein CD47, strengthening the rationale of combining CD47 blockade (which is known to act in an interferon-dependent manner, (3)) with radiotherapy to reduce CD47-mediated anti-phagocytosis as previously suggested (4).

Radiotherapy, and a proportion of chemotherapeutic agents, are now widely accepted as immune-stimulating anticancer treatments, with their effects relying (at least partly) on the activation of the immune response. This observation provides a solid rationale to combine radiotherapy with immunotherapies (immunoradiotherapy), and several efforts have been made first in preclinical settings and then moved to the clinics. Despite the large number of investigations, several factors remain to be elucidated: (i) Dose and fractionation regimens which give best results in standard (chemo)radiotherapeutic settings may not be ideal to elicit an important antitumor immune response, and thus not optimal when combined with immunotherapy. For instance, it has been shown in preclinical settings that fractions of 8 Gy induced a strong type I interferon response, which was surprisingly lower when higher doses were used (5). Moreover, most data concerning radiotherapy/immunotherapy combinations in preclinical settings have been generated using hypofractionated or single irradiations, because of the complexity to model a standard fractionation regimen in animal tumor models. This limits the direct translation of these data, making it difficult to predict the efficacy of combination treatments when applied to standard treatment plans used in the clinics. (ii) The efficacy of the treatment may also depend on the timing and sequencing of radiation and immunotherapy treatments. Indeed, the optimal sequencing can be dependent on the mechanisms of induction of the immune response by a combination of treatments, but very few studies so far have directly compared the efficacy of different pre, post, and concurrent radiotherapy/immunotherapy schedules. As an example, a recent report shows that while administration of anti-PD1 after radiotherapy elicited a strong abscopal response in murine tumor models, administration of anti-PD-1 before irradiation generated a suboptimal systemic antitumor response (6). (iii) Radiation and immune-related toxicities may also be increased by the synergy between radiotherapy and immunotherapy, and data concerning the effects of treatment combinations on healthy tissues are still limited, especially given the paucity of dedicated preclinical studies.

The results of Chen and colleagues push these observations further, clearly demonstrating that even in a homogeneous cohort of patients with cervical cancer, the immune activation after CCRT occurs only in a subset of patients, while others even experience a worsening of the immune markers in the course of the treatment. These data raise several questions, as to which are the tumor and host factors that contribute to the triggering of the immunogenicity in response to CCRT, and which are the ones that drives towards a generation of an immunosuppressive environment associated with a therapeutic failure. The data from Chen and colleagues confirm the preclinical observations that the IFN pathway plays a pivotal role in response to radiotherapy. Nevertheless, as for CD8 T-cell infiltration, the responses observed were highly variable among patients, with only approximately one third showing an increase in the tumor IFN signature, and a decrease in the same number of patients. This observation strengthens the rationale for combining radiotherapy with STING agonists, to boost the IFN activation further and enlarge the proportion of responsive patients, especially if the selection of patients will be guided by the use of biomarkers. Indeed, the results reported by Chen and colleagues indicate that the availability of biomarkers to select patients in which CCRT triggered tumor immunogenicity, and thus, will more likely to respond to a combination with immunotherapy, is of outstanding importance. Indeed, the optimal biomarkers to select patients who would benefit most from the (CC)RT/immunotherapy combination treatments remain to be clearly identified, in cervical cancer as well as in other cancer types. Despite the large number of clinical trials evaluating such combinations, only a few involves translational immunological research that could provide mechanistic insights and foster the development of biomarkers. Ideally, the biomarkers should be non or minimally invasive, to allow a longitudinal follow-up in the course of treatment. In this view, Chen and colleagues found that activation of the interferon response in PBMCs reflects the one observed in the tumor and is associated with PD-L1 positivity, opening interesting translational possibilities. Their data, coming from a limited number of patients, needs to be confirmed with larger cohorts, and possibly expanded to other tumor types, however the availability of blood surrogate markers for patient selection is extremely appealing. Moreover, blood-sampling analyses could be corroborated by other non-invasive approaches, such as radiomic analyses to monitor the changes in the tumor immune environment. These approaches have already been successfully demonstrated, although at baseline, for CD8⁺ tumor infiltration (7). Thus, future studies performing concomitant blood-based biomarkers and radiomic analyses in the course of patients' treatments, with the goal to define a clearer and more detailed picture of the tumor immunological parameters, will allow the design of more appropriate and effective combinations with immunomodulators.

Despite improvements in the recent years, treatment of cervical cancer remains a challenging issue, with overall survival of patients with advanced stages limited to less than 18 months, and few therapeutic options for second-line intervention. The results from recent clinical studies are promising, however optimal timing, sequence and choice of immunotherapy delivery are still undetermined and predictive biomarkers are lacking. The clinical relevance of the large heterogeneity observed in squamous cell tumors (2), together with the heterogeneous changes induced by CCRT as demonstrated by Chen and colleagues should be taken into great account when designing studies involving its combination with immunotherapy. For instance, given that only a third of the patients displayed increased PD-L1 levels during CCRT, clinical studies combining anti-PD-1/PD-L1 may give limited improvement in the absence of patient selection and personalization of concurrent immunotherapy beyond PD-1/PD-L1 blockade, with a significant proportion of patients that could benefit from additional inhibitors such as anti-CD47 and/or STING agonists (Fig. 1).

In conclusion, it remains of pivotal interest to further characterize the different aspects that can affect the outcome of (CC)RT/immunotherapy combinations, with the final goal to achieve “tailored” immunoradiotherapies with an optimal therapeutic index.

M. Mondini reports grants from Roche, Boehringer Ingelheim, Eli Lilly, AC Biosciences, Servier, BMS, MSD, Merck-Serono, and AstraZeneca outside the submitted work. E. Deutsch reports grants and personal fees from Roche-Genentech, AstraZeneca, Merck Serono, Boehringer, BMS, and MSD outside the submitted work.

Authors received financial support from INSERM, SIRIC SOCRATE, Fondation ARC pour la recherche sur le cancer, Agence Nationale de la Recherche (ANR), Institut National du Cancer (INCa 2018-1-PL BIO-06-1), and Fondation pour la Recherche Médicale (FRM DIC20161236437).