Introduction

Colorectal cancer (CRC) is the third most common cancer in both sexes worldwide, with an estimated 1.4 million new cases each year [1]. It represents one of the leading causes of cancer-related death [1]. The presence of liver metastases substantially affects the prognosis of patients with CRC. Unfortunately, 20–50% of patients already have colorectal liver metastases (CRLM) at diagnosis, and 60–70% of patients develop metachronous disease [2, 3]. Surgical resection is considered the only curative therapeutic strategy for CRLM, but many patients are not suitable surgical candidates due to hepatic involvement, the presence of extrahepatic disease, or co-morbidities [4, 5]. For patients with unresectable disease, systemic treatment represents the standard of care and aims to achieve disease control and improve survival. In some cases, tumor shrinkage may make radical resection possible [6, 7]. Other therapeutic approaches have been investigated in patients unsuitable for surgical resection, including percutaneous thermal ablation [8,9,10], transarterial chemoembolization [11,12,13,14], and transarterial radioembolization [15,16,17].

Particularly, transarterial embolization with drug-eluting beads (DEB) has been proposed as a feasible treatment in patients with CRLM, representing a possible therapeutic option for patients with unresectable disease who progressed to systemic therapy [11, 18]. Compared with standard emulsion of lipiodized oil and chemotherapeutic agents, the use of DEB optimizes the pharmacokinetics of chemotherapeutic drug release [19, 20]. Moreover, the beads also occlude arterial vessels that supply the tumor, resulting in ischemic damage to the tumor cells and subsequent tumor necrosis. In this scenario, smaller beads should provide deeper penetration into the tumor, resulting in increased drug delivery and ischemic effect [21,22,23]. Recently, new small caliber 40 µm drug-eluting beads were tested in animal models and in patients with hepatocellular carcinoma [24, 25]. To our knowledge, use of these beads in the treatment of patients with CRLM has never been reported before. Thus, we designed a prospective single-center pilot study (MIRACLE III) with the aim of testing the safety and efficacy of 40 µm DEB loaded with irinotecan (DEB-IRI) for the treatment of patients with unresectable CRLM pretreated with two or more lines of systemic therapies.

Materials and Methods

Study Design

The MIRACLE III pilot study was an industry-sponsored study designed to evaluate the safety of DEB-IRI embolization with small (40 µm) embolic microspheres. Institutional Review Board approval was obtained, and all patients provided informed consent. The study was conducted in accordance with the European Union regulations, the Declaration of Helsinki, and the ICH Harmonized Tripartite Guideline for Good Clinical Practice (GCP).

Patient Eligibility

Patients aged > 18 years with histologically proven colorectal cancer with mono- or bilobar unresectable liver metastases who showed progression after at least two lines of chemotherapy were considered eligible for the study (i.e., all patients had disease progression after at least two lines of standard therapies before being enrolled in the study). Patients were required to have less than 60% of the liver parenchyma involved by tumor, Eastern Cooperative Oncology Group (ECOG) status 0–1. To be included, the liver had to be the dominant disease site (defined as > 80% tumor body burden), but patients with extrahepatic disease were included as long as this criterion was met. Patients with both synchronous and metachronous metastases were included.

Patients with tumor invasion of the portal vein or the main biliary duct, blood level of bilirubin higher than 3 mg/dl, transaminase values greater than five times upper limit of normal (ULN), the presence of other malignancies or contraindication to general anesthesia, or administration of irinotecan, were excluded.

Treatment Protocol

All patients interrupted chemotherapy before starting the protocol. Monolobar disease treatment was performed at least twice with a 4-week interval. Bilobar disease was treated at least twice per lobe (with a 4-week interval), starting with the lobe with greater disease involvement. Treatments were continued until unacceptable toxicity (i.e., toxicity not acceptable by the patient), death, or computed tomography (CT) demonstrated disease progression.

Pretreatment CT images were analyzed to evaluate the anatomy and to detect the feeding vessels to the target tumor. Treatments were performed with the patients under general anesthesia in a dedicated suite equipped with C-arm and CT. Femoral arterial percutaneous access was obtained under ultrasound guidance. Angiography and CT-angiography acquisitions were performed to identify the feeding vessels of the target lesions. When a small number of feeders were identified, superselective embolization was performed. Otherwise, lobar embolization was performed. Before starting the embolization, 0.06 mg of nitroglycerin (Venitrin®, Methafarma S.p.a., Milan, Italy) and subsequently 1–2 ml of 5% levobupivacaine (Chirocaine®, Abbott S.r.l., Illinois, USA) were administered intraarterially for pain control. Embolization was performed with calibrated 40 µm diameter embolic microspheres (Embozene Tandem Microspheres, Boston Scientific, Marlborough, MA) loaded with irinotecan. Irinotecan was loaded according to the microsphere manufacturer’s instructions at a loading dose of 50 mg irinotecan per milliliter of microspheres, yielding up to 150 mg of irinotecan for a 3-mL syringe. An initial dilution of 2 ml of loaded particles in 9 ml of contrast (Iopromide 370 mg/ml, Ultravist, Bayer, Leverkusen, Germany) and 9 ml of saline solution was performed in a 20-ml syringe. Subsequently, 0.5 ml of the previous solution was hyper-diluted with 2 ml of contrast in a 2.5-ml syringe, and a slow injection was performed. Embolization was continued until blood flow was stopped or the total dose of loaded microspheres was delivered. Femoral access closure was achieved with manual compression. Immediately after embolization, hydration (Isolyte, 2000 ml/24 h), antibiotic prophylaxis (ciprofloxacin 550 mg i.v. for 2 days and then 1 g orally for 3 days), and treatment for nausea and pain (ondansetron 4 mg × 3/day; Ketorolac 10 mg × 3/day) were administered. The day after procedure patients underwent a four-phase contrast-enhanced abdominal CT scan to asses for complications and early result of the treatment.

Study Endpoints

Primary objectives of the study were to investigate treatment safety based on 30-day freedom from serious adverse events (SAE) and local tumor control at three months. Adverse events (AE) were defined as any undesirable experience (e.g., sign, symptom, abnormal laboratory value, or other medical event) occurring in a subject during their participation in the clinical trial, and SAE included death, a life-threatening adverse experience, inpatient hospitalization or prolongation of existing hospitalization, a persistent or significant disability/incapacity, or an event that required medical or surgical intervention to prevent one of these outcomes. AE intensity was graded according to the National Institute of Health (NIH-CTC) criteria V 4.0 (https://www.ctep.cancer.gov). The potential causal relationship between AEs and the study device (drug or microsphere) or procedure was assessed by the site investigators.

Local tumor control was defined as complete response, partial response, or stable disease according to RECIST criteria, version 1.1 [26]. Response to treatment was evaluated on the basis of contrast-enhanced CT follow-up studies. Chest and abdominopelvic CT scans for evaluating tumor response and to assess extrahepatic disease were performed every 2 months (i.e., approximately 1 month after each procedure). After a subject had disease progression, further imaging was not conducted.

Secondary endpoints were local tumor control at 6 and 12 months, time to liver-specific disease progression and overall survival. The official date of death was checked with the national death registry. Follow-up assessments also included routine laboratory blood evaluations every 2–4 weeks during treatment and serum tumor marker (CEA) every 8 weeks.

Statistical Analysis

Descriptive statistics are used to summarize patient baseline characteristics. The AE rate is described according to the number of treatments performed. Local tumor control rates are reported as the percentages of all study patients with response or stable disease on or after the designated time point following the first study treatment. Kaplan–Meier analyses were performed to estimate liver-specific progression-free survival and overall survival. Estimates were made from the first DEB-IRI embolization procedure until disease progression or death from any cause, respectively. Statistical analyses were performed with Statistical Analysis Software (SAS), version 9.2 or later (SAS Institute Inc., Cary, North Carolina).

Results

Patients

From November 2013 to February 2015, 18 patients (11 males, 7 females, mean age 61.2 ± 10.1 years; median 61 years) were prospectively enrolled in the study. Baseline patient characteristics are reported in Table 1. All patients had received at least two lines of systemic therapies.

Table 1 Characteristics of 18 patients treated with DEB-IRI embolization for colorectal liver metastases
Full size table

Two patients with stable disease chose to discontinue the study treatment after three treatments; tumor response information is unavailable for these two patients after about 3 months. Two patients had progression of the target lesion and underwent systemic chemotherapy during the study; one of these patients had two DEB-IRI embolization treatments and progression at 56 days (2 months), the other had three treatments and progression at 217 days (7 months). Three other patients also had systemic chemotherapy but stable target lesion disease. One patient had stable disease by mRECIST criteria after three treatments and underwent surgical resection of the liver metastases 160 days following the first DEB-IRI embolization treatment, after multidisciplinary approval.

Procedures

Overall, 80 embolization procedures were performed (mean of 4.4 treatments per patient, range 2–12). A mean of 1.3 ± 0.6 mL (range 0.1–2.0 mL) of microspheres loaded with 67.3 ± 28.1 mg/mL (range 5.0–100.0 mg) of irinotecan was administered per treatment.

Safety

At 30 days, the SAE-free rate was 100% (18/18). A total of 39 treatment-related AEs occurred across all 80 embolization procedures. Most of these events were G1 (85%; 33/39), with abdominal pain (10 events in seven patients, all G1) or abdominal pain and nausea (6 events in three patients, all G1). No treatment-related significant (2.5% ULN) increases in AST/ALT values were observed and no G4 or G5 treatment-related AEs occurred.

Response

Local tumor control, defined as complete response, partial response, or stable disease, was achieved in 16/18 (88.9%), 7/17 (41.2%), and 3/17 (17.6%) patients at 3, 6, and 12 months, respectively, following the first study treatment. One patient with local tumor control was reported to have a partial response (after two treatments) followed by stable disease. All other patients with local tumor control had stable disease at each measurement. The patient who had liver resection on day 160 was not included in the tumor response analysis at 6 and 12 months. CT images from a patient treated in our protocol are shown in Fig. 1.

Fig. 1
figure 1

Case of a patient with right liver colorectal metastasis treated with small beads DEB-IRI. A Contrast-enhanced computed tomography before treatment demonstrating a right liver lobe metastasis; B Contrast-enhanced computed tomography 24 h after the first treatment. The treated lesion present reduced enhancement and some pooling of contrast material from the day of the treatment (arrows); C Contrast-enhanced computed tomography 24 h after second treatment demonstrate partial reduction of the treated lesion (stable disease according to RECIST 1.1), and some pooling of contrast material from the day of the treatment (arrows). D Contrast-enhanced computed tomography at 4 months from the beginning of the protocol demonstrate further reduction of the enhancement and of the dimensions of the treated lesion (stable disease according to RECIST 1.1)

Full size image

Survival Analyses

As shown in Fig. 2, the median time to hepatic-specific local tumor progression was 5.9 months (range 27–409 days). Ten patients died. Median overall survival was 13.5 months (Fig. 3).

Fig. 2
figure 2

Liver progression-free survival in a series of 18 patients with colorectal liver metastases treated with small beads DEB-IRI

Full size image
Fig. 3
figure 3

Overall survival in a series of 18 patients with colorectal liver metastases treated with small beads DEB-IRI

Full size image

Discussion

Metastatic involvement is the main reason for liver failure, morbidity, and death in patients with CRC. Newer systemic therapy regimens might achieve up to 30 months of overall survival in patients not amenable to radical resection. A multi-institutional phase III randomized trial showed a survival advantage in patients with CRLM refractory to chemotherapy who are treated with regorafenib as third-line chemotherapy, compared with placebo (6.4 vs 5 months) [27]. Alternatively, if surgery is not feasible, liver-directed therapies could be applied in order to improve the outcome of those pretreated patients [11, 18, 28, 29]. Indeed, arterial loco-regional therapies have been shown to be safe and efficacious in the treatment of liver- only or liver-dominant unresectable CRLM, but no evidence level I data were available until recently [30] and their role in the treatment algorithm has not been established.

Arterial directed therapies, which deliver drug directly to the liver, theoretically maximize the local antitumor effect while minimizing systemic side effects (15–18). Moreover, the collateral ischemic effect of transarterial embolization might play a role in achieving tumor cell death. The ischemic effect has been investigated in animal models and clinical studies focused mainly on hepatocellular carcinoma [21,22,23, 31,32,33]. Thus, in the setting of CRLM, the use of small microspheres to perform liver embolization might provide both a deeper penetration of drug into the tumor and improved ischemic effect deriving from a more effective distal embolization. Furthermore, smaller microspheres have been reported to be associated with increased drug delivery and decreased adverse effects and toxicity [23]. In our study, all patients had liver progression after at least two lines of standard systemic treatments. Compared with the patients of 13 different studies collected in a recent review by Akinwande et al. [30], our patient population was in a more advanced stage (extrahepatic disease in 77.8 vs 33%) and pretreated with a larger number of chemotherapy lines (89% with ≥ 3 lines vs weighted average of 2).

Our study reveals that DEB-IRI embolization, performed with the smallest particles available, is a safe treatment that may achieve liver tumor control in a setting of patients heavily pretreated with several systemic lines. Regarding the safety profile in our series, the AEs were mainly represented by post-embolization syndrome (flu-like symptoms such as fever, pain, and fatigue), where 85% of them were G1 and none were high grade (G4 or G5). In the review paper [30], a weighted average of 35.2% AE is reported, but 10.1% were high grade.

This is also in line with previous papers reporting AE after DEB-IRI embolization for the treatment of liver metastases [11, 18]. Aliberti et al., investigating the role of DEB-IRI embolization with 100–300/300–500μm beads, reported PES in the majority of cases, including G3 pain in 25% of patients [18]. Fiorentini et al., comparing DEB-IRI embolization with 100–300μm beads with systemic FOLFIRI, found a similar profile of AE between the two arms, with a significantly lower proportion with G3 neutropenia (4 vs 44%) in the DEB-IRI embolization group [11]. Akinwande et al. [23] who used the smallest particles reported among the studies included in the recent review (75–150μm), did not report any high-grade toxicity. Meanwhile, in the study by Iezzi et al. [34] with the same type of beads, 10% of patients had high-grade AEs. Furthermore, in our series the high number of chemoembolizations reflects the good tolerability of the treatment and the good overall response to the disease. DEB-IRI embolization with small beads seems to present a good profile even if compared with recently FDA-approved systemic treatments for CRLM: regorafenib and TAS 102. In patients receiving regorafenib, 93% had AE of any grade and 54% of patients had treatment-related G3-G4 AE [27]. In patients who received TAS 102, AE of any grade was reported in 98% of cases, with 69% experiencing AE of grade ≥ G3 [27].

In our trial, local tumor control was achieved in 88.2% of cases at 3 months, compared with 56.2% in those seven studies included in the review where RECIST criteria were applied. Moreover, at 6 and 12 months, local liver tumor control was achieved in another 41.2 and 17.6% of cases, obtaining in one patient a tumor shrinkage followed by liver surgery. No objective remissions were observed in the published trials on regorafenib and TAS 102, which describe a patient population similar to ours on third-line systemic chemotherapy [27, 35].

Expected survival of patients with CRLM is particularly poor after the failure of standard treatments. In trials investigating systemic approaches, it has been reported between 5 and 7.1 months [27, 35]. Notably, in our series median overall survival was nearly 14 months, with a mean liver-specific progression-free survival of 6 months. These results are within the range reported in the review by Akinwande et al. (median overall survival: 5.4–25 months; progression-free survival: 4–15.3 months) [30], where better results than in our series were achieved in groups of patients with a lower rate of extrahepatic disease (range 2–50 vs 78%). In the cohorts of studies where patients had ≥ 40% of extrahepatic disease, the median overall survival was 13 months and PFS was 10 months compared with 14 and 6 months, respectively, in our study, but high-grade toxicity, again, was around 10% in these cohorts, compared with 0% in our experience.

This study has some limitations. First, because safety was the primary endpoint of the study, only a small number of patients was included, and our efficacy results should be considered preliminary, and are difficult to be extended to general patients’ population. Further studies on larger patients population are required for better understanding the role of this technique in patients with metastatic CRC. A formal pain assessment (e.g., a visual analog scale rating) was not collected for the study, although this information would be informative for characterizing the DEB-IRI embolization procedure. Comparisons with other studies, especially regarding overall survival, are difficult because our study patients were heavily pretreated with systemic therapies and included a high proportion with distant metastases, whereas in other studies only less-advanced stage patients who received a smaller number of previous lines of systemic therapy were included.

In conclusion, DEB-IRI embolization with small beads is a safe procedure in the treatment of patients with CRLM. This treatment showed promising results in terms of liver-specific progression-free survival and overall survival in patients with advanced heavily pretreated disease. Further larger studies for confirming results in terms of safety and response rate should be planned.