Synergizing liver systemic treatments with interventional oncology: friend or foe?

Interventional radiology techniques provide excellent local tumor control for small tumors in various organs, but several limitations can hamper the oncological outcomes such as the tumor size or the number of lesions. Technical improvements, optimal patient selection and combination with systemic therapies, including immune checkpoint inhibitors, have been successfully developed to overcome these barriers. In this setting, chemotherapy and targeted therapies aim to diminish the tumor burden in addition to local treatments, while immunotherapies may have a synergistic effect in terms of mechanism of action on the tumor cell as well as the immune environment, with multiple treatment combinations being available. Finally, interventional Rrdiology treatments often increase tumor antigen exposure to the immune system, and thus stimulate a specific antitumor immune response that can act beyond the treated site. Notwithstanding their many benefits, combination treatment may also result in complications, the most feared may be auto-immune-related adverse events. In early studies, several combined therapies have shown promising levels of safety and efficacy, particularly in hepatocellular carcinoma. This review provides a comprehensive and up-to-date overview of results of combined therapies for primary and secondary liver malignancies. Recent advances and future perspectives will be discussed.


iNtRoductioN
Interventional radiology (IR), through image-guided locoregional therapies (LRTs) is now the fourth pillar of cancer care with medical oncology, surgery and radiation therapy, especially for liver cancer. 1 IR minimal invasive approaches preserve the liver function without compromising the treatment's efficacy, while image-guidance ensures safe and precise tumor targeting for both percutaneous and intravascular approaches. 2 In liver cancer, LRTs play a major role for both local and advanced stages, although their efficacy can be compromised by tumors' size and number. 3 A growing body of evidence suggests that LRTs might play a critical synergistic role in combination to immunotherapy, targeted therapy (TT) and others systemic treatments (SYST). Indeed, LRTs can trigger both local and systemic immune-responses that are mediated by checkpoint proteins and local release of tumor-associated-antigens (TAA). 4 Recently, the combination of LRTs and SYST has offered new perspectives for the management of liver cancer. 5 Indeed, combined therapies of LRTs and SYST might provide a synergizing antitumoral effect that encompasses the activity of each of these therapies when used alone. Thereby, these combinations might generate a systemic and a local meaningful antitumor response or even prevent tumor recurrences. Additionally, combination treatments may help downstaging tumors before surgical resection Given the importance of LRTs as well as immune and targeted SYST in HCC management, 10 the option of combining both became of particular interest, 17 and is under evaluation as firstor second-line therapies, but also in the neoadjuvant and adjuvant setting.
LRTs and immune status HCC alters the immune balance of patients 18 and can cause local humoral and cellular immune responses in the liver tumor microenvironment (TME). 19 While spontaneous immune responses have been described and may be related to the presence of TAA, 20 HCC has several mechanisms to evade the host's immune response. Indeed, HCC progression triggers many immune regulatory changes that promote tumor tolerance and thus its subsequent progression. 9 For instance, near TME, HCC produces acidity which neutralizes the activity of antitumor immune effectors and activates the regulatory immune cells (Tregs). In addition, TME hypoxia caused by HCC annihilates the T-cells function and their immunologic roles. 19 More precisely, T-cells in the TME are in a hyporesponsive state as they become exhausted after overexpressing inhibitory receptors such as cytolytic T-lymphocyte-associated protein 4 (CTLA-4) or programmed death-1 (PD-1). 21 T-regs cells CD4+CD25+ increase in the HCC TME 22 and in the peripherical blood. 23 All of these immunological changes lead to the reduction of production of proinflammatory cytokines as IFN-γ or TNF-α while increasing the secretion of immunosuppressive cytokines as IL-4 or IL-10 in the TME. 24 Myeloid-derived suppressor cells (MDSCs) also play a major role in tumor immunotolerance as they can negatively influence cytotoxic T-cells population and expand the population of T-regs and consequently boost of the host's immunosuppressive activity. 25 The goal of immunotherapy is to overcome the cancer immunetolerance and create an efficient T-cell-mediated adaptative antitumor immune response. Use of immunotherapy is, however, still limited to advanced HCC stages with a low rate of complete responses. 26 The combination with LRTs both in early and advanced stages could help bypassing the TME obstacle by priming locally an immune response.
Indeed, the immune system may react to LRTs and subsequent cell death. 27 LRTs lead to cell death by hypoxia with Transarterial Chemoembolization (TACE) or necrosis with percutaneous ablation. The intracellular content of cells, various TAA and tumor neoantigens are subsequently released and prime a non-infectious inflammatory immune response. This ability to unmask TAA facilitates the recruitment, proliferation and activation of tumor-specific T-cells and antigen-presenting cells (APCs) such as dendritic cells (DCs) into the TME. 28,29 Although LRTs induce a systemic-specific immune response, as described in many series, 30,31 this has not yet translated into clinically meaningful effects. 32 Indeed, LRTs can also cause a protumorigenic effect. For instance, periablation zone caused by radiofrequency ablation (RFA) or microwave ablation (MWA) increase the production of VEGF, IL-6, hepatocyte growth factor or c-met in liver regeneration and promote the growth of local or distant metastases. [33][34][35] Blocking c-met or vascular endothelial growth factor (VEGF) might suppress these effects and promote a synergizing benefit of LRTs combined with SYST. 36 Therefore, local and systemic inflammation could be decreased or regulated by adding TT or immunotherapy. 37 Very early stage (0) and early stage (A) Three main strategies have been included into the guidelines, with a curative objective: resection with hepatectomy, local ablation, and liver transplantation. 10,13 Other therapies such as superselective TACE, or selective internal radiation therapy (SIRT) can also be proposed as down-taging strategies in order to permit liver transplantation. 10,13,38 The most used local ablation technique is RFA. It represents the first-line curative treatment for single tumors <2 cm, i.e. very early stage (0), or an alternative to surgery for a single 2-5 cm lesion or a multinodular disease (2-3 lesions, each <3 cm). MWA is an option, already promoted by USA guidelines 13 but not yet by European guidelines 10 as data from Phase III RCT are lacking.

of 20 birpublications.org/bjr
Other ablatives techniques such as irreversible electroporation (IRE) or cryoablation (CA), are currently under investigation. 10,13 According to the STORMphase III trial, 39 the combination of surgical resection or RFA with sorafenib as adjuvant therapy did not improve recurrence-free-survival (RFS) (33.3 months with sorafenib and 33.7 months with placebo, p = 0.26) and resulted in more severe AEs noted for the sorafenib group (52% vs 10%). Consequently, no adjuvant therapy is currently recommended following ablation.
The HEAT study 40 investigated RFA plus intravenous lysothermosensitive liposomal doxorubicin (LTLD). No differences were observed in terms of PFS (median PFS : 13.9 months for both) and OS (median OS : 53.7 and 53.4 months for combined therapy and RFA alone respectively, p = 0.67).
In this field, the results of the ongoing OPTIMA study (NCT02112656) may be of significant interest. It explores the LTLD combined with RFA (lasting >45 min) for a single lesion of 3-7 cm in comparison to RFA alone.
Clinical trials have studied the combination of cytokine-induced killer (CIK) with RFA 41 and RFA or percutaneous ethanol injection (PEI) or surgery 42 and yielded interesting oncological results.
Bian et al showed the combination of 131I-metuximab with RFA led to a longer time to recurrence than RFA alone (median time to tumor recurrence: 17 vs 10 months, p = 0.03). 43 Adjuvant immune checkpoint inhibitors (ICI) and vascular targeted agents might be used after a local ablation as in the EMERALD-2 (NCT03847428) trial which is comparing three combinations of durvalumab and/or bevacizumab and/or placebo after resection or local ablation.
Encouraging preliminary results of the NIVOLVE trial (NCT000026648) showed that the combination of RFA or surgical resection with nivolumab as an adjuvant therapy reached 76.7% recurrence-free survival rate at 1 year and a median of 26 months of recurrence-free survival.
Other Phase II or III studies using ICI and/or targeted therapies as neoadjuvant and/or adjuvant treatment after a local ablation or a combination of local ablation with intravascular treatment are underway Tables 1; 2 Indeed, many immunological effects have been suggested in patients undergoing LRTs, not seen in surgical resection patients. 9 For several reasons, 33-35 the risk of intrahepatic or distant recurrence persists even after RFA. Given the potential role of atezolizumab plus bevacizumab for advanced stage disease, 44 AB-LATE02 is exploring the neoadjuvant/adjuvant strategy for RFA treated patients in the hope of reducing the prooncogenic effects of RFA and so the risk of recurrence (RFS). The AB-LATE02 trial (NCT04727307, Table 2.) is investigating in the first arm the combination of atezolizumab (as a neoadjuvant treatment) and RFA followed by atezolizumab plus bevacizumab (as an adjuvant treatment). In the other arm, patients are treated by RFA alone.

Intermediate stage (B)
Conventional transarterial chemoembolization (cTACE) with Lipiodol ® , drug-eluting beads transarterial chemoembolization (DEB-TACE), transarterial radioembolization (TARE) and hepatic arterial infusion of chemotherapy (HAIC) represent the LRTs arsenal for intermediate stage. BCLC B stage is heterogeneous explaining the large spectrum of OS in different studies. Following the results of several RCT, 45,46 TACE has become the main therapeutic approach for this stage. 10,13 TARE and cTACE may share consistent efficacy, as suggested by Phase II randomized trial, 47 with time-to-progression (TTP) of more than 26 months for SIRT vs 6.8 months for cTACE but no statistically significant difference for OS (18.6 months vs 17.7 months, respectively). More recently, the TRACE study 48 showed an improvement of TTP and OS in favor of TARE vs DEB-TACE, with a similar safety profile. Consequently, TARE may be a therapeutical option, as mentioned in the guidelines, but with a lower level of evidence. 10,13 Various Phase II/III randomized trials [49][50][51][52][53] have investigated the combination of TACE with time-to-progression (TKI) (Sorfenib, Orantinib or Brivanib) and have not shown any substantial clinical benefits for the combination compared to TACE alone.
In addition, the embolization in TACE procedure triggers acute tissue hypoxia in part of the targeted liver leading to a proangiogenic growth factor activation such as transient increase of VEGF which could be targeted by TKI. 54 Despite this assumption, Pinter et al did not demonstrate any significant improvement in ORR and OS when combining cTACE with bevacizumab, while reporting a higher risk of severe sepsis and vascular complications. 55 According to Pinter et al, bevacizumab might not be recommended as an adjuvant treatment for cTACE. SPACE trial 51 investigated the combination of DEB-TACE plus sorafenib. The combination was technically feasible, but median TTP was not statistically significantly different between the two arms (169 vs 166 days, respectively, p = 0.072). Of note, modified Response Evaluation Criteria in Solid Tumors (mRECIST) was used to define radiological progression.
Recently, the TACTICS trial showed a significantly improved PFS (25.2 vs 13.5 months, p = 0.006) 56 for the combination of TACE plus sorafenib. Times to vascular invasion, extrahepatic spread and stage progression were longer for the combined therapy. The particularity of the TACTICS study was how authors defined disease progression. Indeed, untreatable (UnTACEable) progression was defined as the patient inability to be treated again by TACE. Intrahepatic tumor progression based on the Response Evaluation Criteria in Cancer of the Liver (RECICL) 57 could be a reason to be "UnTACEeable". For the authors, RECIST 1.1 or 4 of 20 birpublications.org/bjr mRECIST criteria are inappropriate to evaluate tumor progression in patients treated with TACE.
Of note, patients included in the TACTICS trial were treated with sorafenib for a longer time than in the SPACE trial probably due to the use of the RECICL criteria which allow patients to have a longer duration before assessing tumor status progression. Therefore, the combination of TACE plus sorafenib could prevent the progression of intermediate to advanced stage HCC.
In another study, Li et al58 suggested that the combination of DEB-TACE plus Apatinib might enhance significantly OS and PFS with similar adverse effects compared to TACE alone in ≥10 cm HCC lesions.
Other studies have suggested Lenvatinib as a neoadjuvant therapy followed by TACE for initial patients deemed TACE-ineligible. 45,59 Overall, despite the TACTICS results, data supporting sorafenib or other TKI treatments in combination with TACE are still not solid enough for this association to be considered reliable. These results have led to the combination of LRTs with ICI such as PD-1 inhibitors in the hope that such association could improve outcomes in stage (B) patients. Preliminary results of the Phase I/II PETAL trial (NCT03397654) showed high tolerability of pembrolizumab after TACE without cumulative adverse effects (AE). Other trials testing ICI with or without TKI as a neoadjuvant or adjuvant treatment in combination with LRTs are currently ongoing Tables 60 In light of recent trials, stage B2 and B3 of Bolondi's subclassification might be ideal candidates for combination therapy using TACE.
Given this heterogeneity, a personalized approach remains a key factor to improve patients' outcomes Advanced stage (C) The management of advanced stage mostly consists of SYST, either alone or in combinations. 10,13  This was further explored in the Phase III SORAMIC trial, where a non-statistically significant improvement in OS was observed when combining SIRT with Sorafenib (12.1 vs 11.4 months for sorafenib alone, p = 0.9529), with an AE ≥3 rate of 64.8%, compared to 53.8% in the sorafenib-only group. 63 Additionally, a recent meta-analysis confirmed the non-inferiority of SIRT to Sorafenib for stage (C) in terms of OS while offering a better safety profile. 64 Therefore, while waiting for the results of the STOP-HCC trial (NCT01556490), the combination of SIRT with Sorafenib is not currently recommended.
Interestingly, a subgroup analysis of the SORAMIC trial identified a survival benefit in patients under 65 years, as well as non-cirrhotic patients, or patients suffering from non-alcoholic cirrhosis. 63 This in-turn prompted further investigation regarding radiation dose optimization when combining SIRT with TKI, or ICI, in selected patients. Garin et al has recently compared personalized dosimetry, ≥205 Gy targeted at the index lesion, with the standard approach of 120 ± 20 Gy targeted at the perfused lobe, in locally advanced HCC. The comparison showed an ORR of 71% in the personalized dosimetry group, compared to only 36% in the standard irradiation group, with an OS and PFS of 26.6 and 6 months respectively in the experimental arm, vs 10.7 (p = 0.0096) and 3.4 (p = 0.26) months in the latter group. 65 As for TACE, a recent Phase III trial (STAH) demonstrated that the combination of TACE and Sorafenib did in fact prolong TTP, PFS, and improve tumor response rates, despite no statistically significant improvements in OS (12.8 months vs 10.8 months for sorafenib alone, p = 0.290). 66 These results were confirmed in an additional retrospective study, demonstrating a significantly improved TTP of 7 months in the RFA, TACE, and Sorafenib treatment group in stage (C) patients following initial hepatectomy, and an OS of 14 months, compared to a TTP and OS of 4 and 9 months respectively in the Sorafenib As for anti-VEGF therapy, a recent meta-analysis revealed that TACE in combination with Apatinib is both effective and safe in advanced stage, with an OS of 10.0 months in the combined group, compared to 6.2 months in the TACE alone group (p = 0.01). 73 Similarly, combined DEB-TACE and Apatinib have been also shown to improve OS and PFS. 74 As for translational results, it has been recently shown that a CD8 + T cell response against TAA is associated with improved RFS in HCC patients following ablation, meaning that ablation therapy can potentially transform non-infiltrated "cold" tumors into more immunogenic "warm" tumors. This indicates a potential synergy between the immunological impact of local ablation and immunotherapy. 75  Terminal stage (D) The impact of LRTs combined with systemic therapies is yet to be explored in this population.
iNtRAhepAtic cholANGiocARciNoMA CCA represents a heterogeneous group of malignancies arising from the biliary tract, the second most common primary liver cancer and are commonly classified as either perihilar, distal, or intrahepatic (iCCA). 3,79 The carcinogenesis of CCA is yet to be fully understood. It often arises in biliary tissue characterized by chronic inflammation and cholestasis, with a resulting mixture of proinflammatory cytokines, growth factors, and bile acids.
The presence of such complex components in the TME ultimately leads to increased and/or aberrant activation of various cell surface receptors, and to the down-regulation of intracellular signaling, all of which contribute to the carcinogenesis process. 80 Similar to HCC, 70% of CCA cases are diagnosed at a metastatic or locally advanced stage, mainly due to the lack of specific symptoms during early stages. 81 Current guidelines suggest Cisplatin-Gemcitabine (CisGem) as a first-line treatment for unresectable CCA cases, with a median OS of 11.4 months. 82 The identification of isocitrate dehydrogenase-1 (IDH-1) mutations and fibroblast growth factor receptor-2 (FGFR-2) fusions as the most common targetable alternations in CCA have made them the subject of recent therapeutic interest. 83 Indeed, selective FGFR inhibitors have been shown to be effective in patients with advanced refractory iCCA harboring the above-mentioned FGFR-2 alteration. [84][85][86][87] Unsurprisingly, both Pemigatinib and Infigratinib have been approved by the FDA as second-line treatment options for refractory iCCA. 85 Furthermore, the emergence of immunotherapy has prompted the initiation of several trials assessing the safety and efficacy of adaptive immune cell transfer or ICI in the context of iCCA treatment. Initial vaccine-based immunotherapies did not show improvements in clinical outcomes in iCCA patients. 88 Nowadays, Pembrolizumab is indicated in iCCA patients with mismatch repair defects and microsatellite instability, as well as iCCA cases with a high mutational burden. 89 LRTs can also be used as neoadjuvant treatment options in iCCA patients, mainly for tumor burden reduction and downstaging purposes in patients likely to undergo curative resection. This was demonstrated in the MISPHEC trial, which combined SIRT with CisGem as a first-line treatment in patients with unresectable iCCA, during which a significant proportion of patients were down-staged, allowing for surgical intervention. 90 In addition, a recent Phase II trial by Cercek et al combining Gemcitabine plus Oxaliplatin and HAIC with floxuridine in unresectable iCCA patients has demonstrated an achieved OS of 25 months, rendering the combination as a suitable first-line treatment option. 91 It is evident that the management of iCCA will keep evolving, thanks to the multiple ongoing trials (KEYNOTE-966, KEYNOTE-158…) involving SYST and/or LRTs Tables 7; 8  Despite all this, relapse rates in iCCA remain far high and prognosis far too low. Hence, the need for further exploration in neoadjuvant and adjuvant combinatorial strategies, as well as a better method of patient selection, in the hope of improving long-term oncological outcomes.

Secondary liver cancer
Metastatic disease is the most frequent form of liver malignancies and cancer recurrence, 93 and is often associated 9 of 20 birpublications.org/bjr  (Continued) with a poor prognosis despite all therapeutic advances. 94 Interestingly, even for cancers responding to immunotherapies, liver involvement is considered as an independent factor of poor prognosis and poor response. In this setting, LRTs, in combination with other systemic approaches, have shown promises in terms of improving clinical outcomes. [95][96][97] Unlike chemotherapy, which has witnessed a decline due to a high rate of morbidity and local recurrence, LRTs are incrementally gaining practice as first-line, second-line, or adjuvant treatment options for liver metastases to improve the OS or even allow a down-staging for surgery. 98 Colorectal carcinoma liver metastases Colorectal cancer (CRC) is the third most common malignancy, the leading cause of liver metastasis (LM) and the second most deadly cancer. 99 Approximately, half of CRC patients experience synchronous or metachronous metastatic liver disease, with only one quarter of patients being eligible for surgical intervention. 100 The standard of care is currently a combination of chemotherapy, most commonly FOLFOX or FOLFIRI, in combination with targeted therapies, mainly Bevacizumab or Cetuximab, with mFOLFOX and Bevacizumab combination being approved for clinical use. 101 In a study by Pilati et al, in order to combine first-line therapeutics with LRTs, patients with unresectable LM received either Floxuridine-Leucovorin HAI alone or in combination with systemic chemotherapy. There were no differences between the two arms, with ORRs and OS at 52.7% and 18.0 months in HAI alone arm compared to 50.6% and 19.1 months in the combination group (p > 0.05). 102 Recently, the OPTILIV trial demonstrated a 15% increase of conversion rate to surgery in patients with previously unresectable LM when combining systemic Cetuximab with Irinotecan-Oxaliplatin-5-FU HAI. 103 Two other trials using modern systemic chemotherapy in combination with HAI demonstrated similar results. 104,105 Further RCT are required to confirm such improvements. Luckily, the Phase III PUMP (NTR7493), and PACHA-01 (NCT02494973) trials are assessing the role of HAI as an adjuvant therapy to surgical resection in CRC-LM patients.
As for ablation therapies, the recent Phase II CLOCC trial demonstrated an improvement of OS when combining RFA with systemic chemotherapy, with a median OS of 45.6 months in patients receiving combined treatment vs 40.5 months in patients only treated with chemotherapy (p = 0.01). 106 As of today, there are no current trials evaluating the combination of MWA and systemic therapy in CRC-LM patients.
Regarding immunotherapy, it has been well established that ICI can be highly effective in mismatch repair deficient (dMMR) and microsatellite instability-high (MSI) CRC cancers. This is in part due to the associated high tumor mutational burden and consequential increased immunogenicity, abundant expression of IC regulators, and Immune cell infiltration. [107][108][109] Indeed, Pembrolizumab has been recently shown to extend PFS by 8.3 months in dMMR/MSI-CRC patients, compared to patients receiving chemotherapy alone. 110 Unfortunately, only 2-4% of CRC cases are diagnosed as dMMR or MSI. 111 However, similar to the rationale in HCC, the immunological impacts of LRTs in CRC-LM are anticipated, and synergy between IO approaches and immunotherapy is being considered. Following this logic, Lemdami et al had demonstrated   (Continued) an increase in specific T cell response, and significant tumor regression in CRC-LM patients receiving RFA in combination with in-situ immunotherapy or PD1/PDL1 blockade as adjuvant therapy. 112 In addition, an increase in tumor antigen specific T cell infiltration and PDL1 expression in primary CRC tumors was observed in CRC-LM patients receiving RFA. 113 Further explorations of the potential synergy of RFA and PD1/PDL1 blockade are required.
As for alternatives to current systemic therapies, DEB-TACE combined with Irinotecan (DEBIRI) resulted in a median OS and PFS for 22 and 7 months, a significant improvement from a median of OS and PFS of 15 and 4 months in FOLFIRI treated CRC-LM patients in addition to quality-of-life amelioration. Moreover, further analysis demonstrated a greater benefit for KRAS-wild type patients within the DEBIRI treatment arm, with an OS of 26 months, compared to 14 months in KRAS-mutated patients. 114 Furthermore, significant improvements in ORRs, hepatic PFS, and a more durable overall PFS were observed when combining DEBIRI with FOLFOX, compared to FOLFOX alone, in patients undergoing down-staging chemotherapy. 115 Similarly, the PARAGON-II trial asserted that DEBIRI was comparable to systemic neoadjuvant chemotherapy in terms of OS and anti tumor response. 116 Despite these encouraging results, more data is required from RCT prior to enlisting such combinations per guideline Tables 9; 10. TARE has also been proven efficacious in CRC patients with LM ineligible for resection, ablation, or further systemic therapies. For instance, the retrospective MORE study had demonstrated a survival benefit, with an OS of 10 months. 117 Combining TARE with fluorouracil as a first-line treatment option resulted in an overall PFS of 4.5 months and a liver PFS of 5.5 months, compared to 2.1 months as PFS and liver PFS in fluorouracil only treated patients, without differences in OS. 118 SIRLFOX trial noted a comparable PFS benefit when combining TARE with FOLFOX, with or without Bevacizumab, as well as drastically delaying local progression. 119 Furthermore, the following post hoc analysis revealed a statistically significant increase in the eligibility for surgical resection in patients receiving SIRT, compared to the control group, 119 suggesting a more effective downstaging strategy when combining SIRT with systemic chemotherapy in CRC-LM patients. Besides, Mulcahy et al has demonstrated that the addition of TARE to second-line chemotherapy had extended both PFS and hepatic PFS (8 and 9.1 months, p = 0.0013 and p < 0.0001 respectively) compared to the chemotherapy alone treatment group. 120 In this context, there is further need for better patient selection prior to incorporating TARE into second-line treatment options in CRC patients with unresectable LM. With the help of LRTs and advancements in down-staging approaches, patients with LM disease are becoming more readily eligible for surgical intervention, and are consequently experiencing improvements in oncological end points. The inclusion of DEBIRI and SIRT into guidelines, either alone or in combination with well-established SYST, has resulted in better management and survival for LM patients.

Conclusion
In conclusion, locoregional image-guided therapies can stimulate the immune response and/or allow a pause in systemic treatments. Recent studies show promising results regarding their combination with systemic treatments. Although these combination treatments are still being evaluated, with mostly early phase trials published to date, this innovative approach is expected to play an increasing role in patient management, particularly in the era of "personalized care". In this context, multidisciplinarity is key to define the best therapeutic strategy to help achieve a durable oncological response, including the types of combinations, the right therapeutic sequence, and the appropriate choice of targets.
Even if it is too early to give a clear answer to the title of this review, recent trials suggest that image-guided locoregional therapies and new systemic treatments tend to be new allies. In order to develop a sustainable friendship, carefully designed largescale randomized trials are needed, with a particular attention regarding patient selection.