Abstract
Hepatic metastases of colorectal carcinoma are a leading cause of cancer-related mortality. Most colorectal liver metastases become refractory to chemotherapy and biologic agents, at which point the median overall survival declines to 4–5 mo. Radioembolization with 90Y has been used in the salvage setting with favorable outcomes. This study reports the survival and safety outcomes of 531 patients treated with glass-based 90Y microspheres at 8 institutions, making it the largest 90Y study for patients with colorectal liver metastases. Methods: Data were retrospectively compiled from 8 institutions for all 90Y glass microsphere treatments for colorectal liver metastases. Exposure to chemotherapeutic or biologic agents, prior liver therapies, biochemical parameters before and after treatment, radiation dosimetry, and complications were recorded. Uni- and multivariate analyses for predictors of survival were performed. Survival outcomes and clinical or biochemical adverse events were recorded. Results: In total, 531 patients received 90Y radioembolization for colorectal liver metastases. The most common clinical adverse events were fatigue (55%), abdominal pain (34%), and nausea (19%). Grade 3 or 4 hyperbilirubinemia occurred in 13% of patients at any time. The median overall survival from the first 90Y treatment was 10.6 mo (95% confidence interval, 8.8–12.4). Performance status, no more than 25% tumor burden, no extrahepatic metastases, albumin greater than 3 g/dL, and receipt of no more than 2 chemotherapeutic agents independently predicted better survival outcomes. Conclusion: This multiinstitutional review of a large cohort of patients with colorectal liver metastases treated with 90Y radioembolization using glass microspheres has demonstrated promising survival outcomes with low toxicity and low side effects. The outcomes were reproducible and consistent with prior reports of radioembolization.
Metastases of colorectal carcinoma are the third most common cause of cancer-related mortality worldwide. Advances in chemotherapy regimens, biologic agents, and liver resection in select cases have prolonged overall survival after the diagnosis of hepatic metastases. Nonetheless, most colorectal liver metastases become refractory or resistant to these regimens, at which point survival estimates range from 4 to 5 mo. 90Y radioembolization of hepatic metastases has been increasingly used in this salvage setting and has shown favorable survival outcomes—exceeding 10 mo—after 90Y radioembolization (1–3). Most of the published outcomes have been after 90Y radioembolization using resin microspheres. However, a substantial number of patients with colorectal liver metastases undergo 90Y radioembolization with glass microspheres. This paper reports the survival and safety outcomes of 531 patients with hepatic metastases of colorectal cancer treated with glass 90Y microspheres at several institutions.
MATERIALS AND METHODS
Patient Cohort
Between 2001 and 2014, 531 consecutive patients underwent radioembolization of hepatic metastases of colorectal carcinoma at 8 institutions. The study was approved by Institutional Review Boards and was compliant with the Health Insurance Portability and Accountability Act. Data were retrospectively compiled into a common database and analyzed at a single institution. The clinical trials registration number was NCT00532740.
This report complies with the research reporting standards for radioembolization (4). Clinical side effects and biochemical toxicity according to version 4.0 of National Cancer Institute common terminology criteria were recorded at follow-up. Clinical adverse events and biochemical toxicity occurring at any time after treatment were recorded (not limited to 30 d).
The study represents retrospectively collected data at cancer centers with expertise in locoregional therapies. The indications for 90Y radioembolization included unresectable metastases from colorectal cancer; imaging-confirmed progressive disease refractory to previous systemic or locoregional therapy; an Eastern Cooperative Oncology Group (ECOG) status of no more than 2; the ability to undergo angiography and selective visceral catheterization; and adequate hematology counts (granulocytes ≥ 1.5 × 109/L and platelets ≥ 50 × 109/L), renal function (creatinine ≤ 2.0 mg/dL), and liver function (bilirubin ≤ 2.0 mg/dL). Exclusion criteria included significant extrahepatic disease (life expectancy < 3 mo); angiographic evidence or 99mTc-macroaggregated albumin scan evidence of uncorrectable gastrointestinal flow; or an estimated lung dose of more than 30 Gy in a single session.
Patient Evaluation and Workup
All patients underwent baseline laboratory tests and radiologic imaging within 1 mo of treatment. Imaging information was used to determine baseline tumor burden, uni- or multifocality, and presence or absence of extrahepatic metastases. Pretreatment angiography was performed to determine proper catheter positioning and to identify any collateral flow to the gastrointestinal tract (5–9). 99mTc-macroaggregated albumin scanning was performed to detect gastrointestinal flow and lung shunt fraction (6). Prophylactic embolization of aberrant vessels was performed when appropriate.
Treatment Plan
All radioembolization procedures were performed with a glass-based 90Y device (TheraSphere; BTG International Ltd.). This device is currently approved for hepatocellular carcinoma in the United States and for liver neoplasia worldwide (10). The method for determining the injected activity required to deliver 120 Gy has been published previously (4,6,11,12).
Overall Survival
Median overall survival was calculated from the dates of diagnosis of the primary cancer, hepatic metastases, and first 90Y treatment, censored to the date of last follow-up. Survival analyses were substratified on the basis of exposure to cytotoxic chemotherapeutics (5-fluorouracil/capecitabine, oxaliplatin, or irinotecan) and biologic agents (bevacizumab, cetuximab/panitumumab, or regorafenib).
Uni- and Multivariate Analyses and Statistical Plan
Uni- and multivariate analyses were performed using the Cox proportional hazards model. The variables entered into the univariate analyses included patient demographics (sex and age), performance status (ECOG), prior therapy (cytotoxic or biologic agents), presence of metastatic disease at diagnosis, tumor burden, presence of extrahepatic metastases, and liver function. Variables with a P value of no more than 0.25 by the univariate model were included in the multivariate model. Analyses were performed using SPSS Statistics (version 22.0; IBM), with a P value of less than 0.05 considered significant.
RESULTS
Baseline Characteristics
Most patients (63%) were less than 65 y old at the time of treatment (59% were male), most (96%) had an ECOG status of 0 or 1, most (70%) had tumor in no more than 25% of the liver volume, and most (62%) had liver-only disease (38% had limited extrahepatic disease). Eighteen percent of patients had prior hepatic resection, 14% had prior liver ablation, and 4% had prior transarterial chemoembolization (Table 1).
Baseline Characteristics
Before 90Y radioembolization, 56% of patients received 3 cytotoxic chemotherapeutics (5-fluorouracil, oxaliplatin, and irinotecan), whereas 41% received only 1 or 2 of these agents and 3% received none of these agents. Twenty-two percent of patients received no biologic agents, 56% received 1 biologic agent, and 22% received 3 biologic agents (Table 2).
Cytotoxic Chemotherapies and Biologic Agents
Nearly all patients underwent lobar or selective radioembolization at the first treatment. Only 2% of patients received whole-liver treatment in a single setting.
90Y Dosimetry and Delivery
The median radiation dose delivered to the liver was 120.2 Gy (range, 35–391 Gy). In all cases, 90% or more of the dose was delivered. Extrahepatic arterial coil embolization was performed in 25% of patients.
Side Effects and Biochemical Toxicity
The most frequent clinical side effects included fatigue (55%), abdominal pain or discomfort (34%), and nausea (19%). Vomiting, anorexia, and fever occurred in less than 10% of patients (Table 3). No gastrointestinal ulcers were reported.
Clinical Side Effects
Grade 3 or 4 biochemical toxicity, recorded at any time after treatment, included effects on the levels of bilirubin (13%), alkaline phosphatase (9%), albumin (8%), aspartate transaminase (3%), and alanine transaminase (<1%) (Table 4).
Grade 3–4 Biochemical Toxicity*
Overall Survival
At the time of data compilation, 284 patients had died. Survival analysis is provided in Tables 5–8. Median overall survival was 48.7 mo (95% confidence interval [95% CI], 44.2–53.2) from the date of diagnosis of the primary tumor, censored to the last follow-up; 37.7 mo (95% CI, 33.7–41.7) from diagnosis of hepatic metastases; and 10.6 mo (95% CI, 8.8–12.4) from the first 90Y treatment. The median time from diagnosis of hepatic metastases to the first 90Y treatment was 17.5 mo (95% CI, 15.2–19.7). Median overall survival from the first 90Y treatment was longer in patients without extrahepatic disease than in those with extrahepatic disease (14.4 vs. 6.6 mo, P < 0.001).
Overall Survival
Survival by Exposure to Cytotoxic Agents (5-Fluorouracil, Oxaliplatin, or Irinotecan)
Survival by Exposure to Biologic Agents (Bevacizumab, Cetuximab, Panitumumab, or Regorafenib)
Survival by Era
The time from diagnosis of hepatic metastases to the first 90Y treatment was longer for patients who received 3 cytotoxic chemotherapeutics than for those who received 2 or fewer (22.6 vs. 10.9 mo, P < 0.001). Median overall survival from the first 90Y treatment was shorter for patients who received 3 cytotoxic chemotherapeutics than for those who received 2 or fewer (9.2 vs. 14.7 mo, P < 0.001). Regarding patients without extrahepatic disease at the time of the first 90Y treatment, median overall survival was longer for those receiving 2 or fewer cytotoxic chemotherapeutics than for those receiving 3 (16.5 vs. 13.1 mo, P < 0.05). Regarding patients with extrahepatic disease at the first 90Y treatment, median overall survival was longer for those receiving 2 or fewer cytotoxic chemotherapeutics than for those receiving 3 (9.6 vs. 5.4 mo, P = 0.003).
Median overall survival after the first 90Y treatment did not significantly differ between patients who received no biologic agents and those who received one (11.5 vs. 12.9 mo); was longer for patients who received 0 or 1 biologic agent than for those who received 2 or more (12.9 vs. 7.0 mo, P < 0.001); and was longer for patients treated after 2004 than for those treated before 2004 (10.9 vs. 7.0 mo, P < 0.05).
Uni- and Multivariate Analyses
Better survival outcomes were predicted by ECOG performance status 0 (P < 0.001), American Joint Committee on Cancer stage 3 or less at the time of diagnosis (P = 0.113), lack of extrahepatic metastases (P < 0.001), hepatic tumor burden of no more than 25% (P < 0.001), 2 or fewer cytotoxic chemotherapeutics (P < 0.001), no biologic therapy (P < 0.001), bilirubin less than 1.3 mg/dL (P < 0.001), and albumin greater than 3 g/dL (P < 0.001).
On multivariate analysis, ECOG performance status 0 (hazard ratio, 0.61; 95% CI, 0.46–0.79), hepatic tumor burden of no more than 25% (hazard ratio, 0.37; 95% CI, 0.28–0.49), no extrahepatic metastatic disease (hazard ratio, 0.50; 95% CI, 0.38–0.64), albumin greater than 3 g/dL (hazard ratio, 0.47; 95% CI, 0.35–0.63), and 2 or fewer cytotoxic chemotherapeutics (hazard ratio, 0.61; 95% CI, 0.46–0.79) independently predicted better survival outcomes (Table 9).
Multivariate Analysis for Survival
DISCUSSION
Colorectal carcinoma remains one of the most common cancers worldwide, with approximately half the affected individuals eventually developing liver metastases. Only a minority of these patients are candidates for potentially curative therapies such as surgical resection or ablation. Modern chemotherapy regimens and the advent of biologic agents have significantly prolonged the median overall survival of patients with hepatic metastases to approximately 29–32 mo (13,14). Nonetheless, once hepatic metastases become chemorefractory, survival estimates are poor, typically between 4 and 5 mo (15–17).
Transarterial 90Y radioembolization has increasingly been used as a locoregional therapy for chemorefractory hepatic metastases and was recently included in the European Society for Medical Oncology clinical practice guidelines for the treatment of liver-limited colorectal metastases failing chemotherapeutic options (18). Radioembolization relies on the fact that the blood supply of hepatic tumors is different from that of normal liver parenchyma. Primary and metastatic hepatic tumors receive most of their blood supply from the hepatic arteries, as opposed to the predominantly portal venous blood supply of the liver parenchyma. Injection of radioactive microspheres into the hepatic artery leads to preferential deposition within tumor capillaries, providing antitumoral radiation effects that are not available in the standard treatment paradigms for colorectal liver metastases.
Several studies have shown 90Y radioembolization to be beneficial in the treatment of metastatic colorectal cancer. A randomized phase II analysis of the addition of a single treatment of 90Y resin microspheres to the standard regimen of 5-fluorouracil/leucovorin for patients with colorectal liver metastases found significantly prolonged median overall survival compared with the standard regimen alone (29.4 vs. 12.8 mo, P = 0.02) (19). The addition of 90Y radioembolization to systemic chemotherapy has been shown to prolong the time to liver tumor progression in patients with chemorefractory colorectal liver metastases (5.5 vs. 2.1 mo, P = 0.003) (20), and matched-pair analysis of patients with chemorefractory colorectal liver metastases demonstrated a survival benefit for patients receiving 90Y radioembolization compared with those receiving best supportive care (8.3 vs. 3.5 mo, P < 0.001) (21). A prospective multicenter phase II analysis of 90Y radioembolization for chemorefractory liver metastases reported a hepatic progression–free survival of 3.0 mo for patients with colorectal metastases. Median overall survival was 8.8 mo for all patients with metastatic colorectal cancer and increased to 10.5 mo for patients with liver-only colorectal metastases (22).
However, most of the published literature on 90Y radioembolization for colorectal liver metastases has reported on resin microspheres, in part because of differences in the approved indications for glass and resin microspheres (20,21,23,24). Nonetheless, a significant number of patients undergo 90Y radioembolization of colorectal liver metastases with glass microspheres. The results of this multiinstitutional review of 531 patients with colorectal liver metastases treated with 90Y radioembolization using glass microspheres underscore the consistent, reproducible, and favorable survival outcomes of the therapy in the salvage setting. The data show that radioembolization using glass microspheres provides reliable dose delivery and is safe and well tolerated.
Once colorectal liver metastases become chemorefractory, patients’ median overall survival approaches 4–5 mo (15–17). In addition to the studies cited previously, several large cohort studies have reported survival data for patients undergoing salvage 90Y radioembolization. These data are remarkably consistent and nearly identical to our own findings. Median overall survival from the first 90Y treatment in our cohort was 10.6 mo. In the next largest published study, of 302 patients with chemorefractory colorectal liver metastases treated using resin microspheres, Saxena et al. (1) found a median overall survival of 10.5 mo after the first 90Y treatment, which mirrors the 10.6-mo median overall survival of 214 patients treated with glass microspheres reported by Lewandowski et al. (2). Kennedy et al. (3) also reported a median overall survival of 10.5 mo among responders in a multiinstitutional review of 208 patients treated with resin 90Y microspheres. Although the 214 patients reported by Lewandowski et al. were included in our multiinstitutional cohort, removal of these 214 patients from the survival analysis still indicates a median overall survival of 10.5 mo for the remaining 317 patients.
The significance of these survival outcomes in the salvage setting should not be overlooked. Although data from retrospective review certainly cannot be equated with data from prospective, randomized trials, these results resonate favorably considering the median overall survival outcomes achieved in the salvage setting with other agents. Cetuximab demonstrated a median overall survival of 9.5 mo in the salvage setting for patients with the wild-type KRAS gene, compared with 4.8 mo achieved with best supportive care (16). Subsequent trials investigating its use earlier in the treatment of metastatic colorectal cancer have led to its acceptance as a cornerstone of colorectal cancer treatment (18,25). More recently, regorafenib has become standard-of-care treatment for chemorefractory colorectal liver metastases by demonstrating a prolongation in median overall survival from 5 mo with best supportive care and placebo to 6.4 mo with best supportive care and regorafenib (15). Several prospective, randomized trials are currently evaluating the role of 90Y radioembolization at different points in the treatment of colorectal liver metastases, including the EPOCH trial (26) and the combined 1,100-patient SIRFLOX (27), FOXFIRE (28), and FOXFIRE global (29) studies.
The precise and reliable dose delivery achieved with glass 90Y microspheres is a critical observation. Glass microspheres have a low embolic load, which does not limit delivery of the prescribed activity of 90Y (30). In our cohort, delivery of at least 90% of the dose was achieved in all cases. The heavy embolic load of resin microspheres, on the other hand, can result in arterial stasis, limiting the actual 90Y dose delivered. In a recent phase I trial of resin microspheres for treatment of colorectal metastases, arterial stasis that limited the total administered activity occurred in 37.5% of treatment sessions (31). Such reliability of dose delivery with glass microspheres not only is a vital component of consistent and reproducible brachytherapeutic treatments but also is central to the principles of oncologic clinical trials, as the safety and effectiveness of a therapy cannot be evaluated if the actual dose delivered varies. This principle was confirmed in a recent phase I study of the radiosensitizing chemotherapy capecitabine used with escalating whole-liver doses of 90Y glass microspheres and controlled, predictable escalation of the 90Y dose (32). During the same period, a phase I study of resin 90Y radioembolization with capecitabine therapy was published, but the 90Y dose could not be escalated because dose delivery with resin microspheres is particularly unpredictable in the salvage setting for colorectal liver metastases, in which vessels can be adversely affected by chemotherapy exposure. In addition, only 41.7% of patients received whole-liver radioembolization, limiting assessment of the true liver tolerance to radioembolization during concomitant radiosensitizing chemotherapy.
Given the changes in hepatic vasculature and flow dynamics seen in patients who have been heavily pretreated with chemotherapy and biologic agents, the high activity and low embolic load of glass microspheres offer tremendous value. A frequent criticism of 90Y radioembolization is the need for complex coil embolization of extrahepatic vessels in order to prevent nontarget 90Y radioembolization, or the need for coil embolization of hepatic vasculature for intrahepatic flow redistribution in order to achieve dose delivery, both of which not only prolong the procedures but also carry a risk of promoting tumor-perfusing collateral artery networks—potentially complicating future treatments. In our cohort, however, only 25% of patients underwent arterial coil embolization, which is in line with the increasingly low rates of extrahepatic arterial embolization during glass microsphere radioembolization previously reported by our group (33). In our group’s more contemporary patient population, prophylactic coil embolization approaches 10%. The safety of this practice is underscored by the fact that no gastrointestinal ulcers were reported.
The low rate of clinical side effects and biochemical toxicity in this large patient cohort confirms the safety of 90Y when performed across institutions. The most common clinical side effects—fatigue, abdominal pain or discomfort, and nausea—were comparable to the known and expected side effects of radioembolization. Grade 3 and 4 biochemical toxicity after radioembolization affected only a minority of patients and may in fact have been overestimated in this series, which, by recording hepatic toxicity at any time after radioembolization, inevitably also captured the effects of hepatic decompensation related to tumor progression.
This study had limitations. Retrospective data collection limits the accuracy of patient-subgroup comparisons. For this reason, our study focused on the outcomes that are least prone to misinterpretation and bias, namely survival and the safety of a treatment when performed at diverse institutions. In addition, this study was confounded by variability in the number of treatments that patients received. Also, many patients did not receive all available systemic options, for reasons including poor tolerance and KRAS mutation status. Nonetheless, despite differences in local oncology practice and in the nuances of the procedure, 90Y radioembolization with glass microspheres is safe and offers consistent survival outcomes for patients with chemorefractory colorectal liver metastases.
These results add to growing evidence, now comprising more than 1,000 patients, in support of 90Y for treatment of colorectal liver metastases. Collaboration among oncologists and interventional radiologists is needed to conduct the large-scale, prospective trials required to more precisely define the role of 90Y radioembolization in the treatment of colorectal liver metastases.
CONCLUSION
The multiinstitutional cohort of colorectal liver metastasis patients treated with 90Y glass microspheres reviewed in our study is the largest yet analyzed. Our results demonstrate promising survival outcomes that are reproducible and consistent with prior reports. Glass microspheres provide reliable and precise radiation dose delivery that is safe and well tolerated. The results of the large-scale randomized studies that are under way are awaited.
DISCLOSURE
The costs of publication of this article were defrayed in part by the payment of page charges. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 USC section 1734. Ryan Hickey, Robert Lewandowski, Jean-Francois Geschwind, Siddarth Padia, Mary Mulcahy, and Riad Salem are advisors to BTG. No other potential conflict of interest relevant to this article was reported.
Footnotes
Published online Dec. 3, 2015.
- © 2016 by the Society of Nuclear Medicine and Molecular Imaging, Inc.
REFERENCES
- Received for publication August 26, 2015.
- Accepted for publication November 30, 2015.
