Radioembolization Versus Bland Embolization for Hepatic Metastases from Small Intestinal Neuroendocrine Tumors: Short-Term Results of a Randomized Clinical Trial

Abstract

Background: Radioembolization (RE) with intra-arterial administration of ⁹⁰Y microspheres is a promising technique for the treatment of liver metastases from small intestinal neuroendocrine tumors (SI-NET) not amenable to surgery or local ablation. However, studies comparing RE to other loco-regional therapies are lacking. The aim of this randomized study was to compare the therapeutic response and safety after RE and bland hepatic arterial embolization (HAE), and to investigate early therapy-induced changes with diffusion-weighted MRI (DWI-MRI).

Methods: Eleven patients were included in a prospective randomized controlled pilot study, six assigned to RE and five to HAE. Response according to RECIST 1.1 using MRI or CT at 3 and 6 months post-treatment was recorded as well as changes in DWI-MRI parameters after 1 month. Data on biochemical tumor response, toxicity, and side effects were also collected.

Results: Three months after treatment, all patients in the HAE group showed partial response according to RECIST while none in the RE group did (p = 0.0022). After 6 months, the response rates were 4/5 (80%) and 2/6 (33%) in the HAE and RE groups, respectively (NS). DWI-MRI metrics could not predict RECIST response, but lower pretreatment ADC_(120-800) and larger ADC_(0-800) increase at 1 month were related to larger decrease in tumor diameter when all tumors were counted.

Conclusion: HAE resulted in significantly higher RECIST response after 3 months, but no difference compared to RE remained after 6 months. These preliminary findings indicate that HAE remains a safe option for the treatment of liver metastases from SI-NET, and further studies are needed to establish the role of RE and the predictive value of MR-DWI.

Fig. 1

RECIST response in target lesions in the treated liver at 3 months (dark gray staples) and 6 months (light gray staples) post-treatment. Solid black line indicates threshold for partial response (PR). At 3 months, no responders were seen in the RE group while all patients in the HAE group showed PR (p = 0.0022)

Fig. 2

Relative change in CgA levels compared to baseline values, at 3 months (dark gray staples) and at 6 months (light gray staples) after treatment with HAE or RE. The median decrease in CgA levels at 3 months was 52% in the HAE group and 29% in the RE group, and 47 and 44%, respectively, at 6 months (NS, p = 0.42 and 0.66, respectively)

Fig. 3

Relative change in dU-5HIAA compared to baseline values, at 3 months (dark gray staples) and at 6 months (light gray staples) after treatment with HAE or RE. The median decrease in dU-5HIAA at 3 months was 43% in the HAE group and 25% in the RE group, and 36 and 43%, respectively, at 6 months (NS, p > 0.05). Patient R5 showed no change in dU-5HIAA values at all time points, hence no visible bars in the figure

Fig. 4

Correlation between ADC_(0–800) at 1 month and decrease in LD at 3 months (k = −0.2, p < 0.05) in individual tumors in patients treated with RE (R) and HAE (H), respectively. Values are given as percent of pretreatment value. The blue, red, and black lines show the regression line for HAE patients, RE patients, and all patients, respectively

Fig. 5

Relative individual tumor diameters after treatment. At 3 months after treatment, the relative (compared to baseline) median tumor diameter in measured lesions was significantly smaller after HAE, 45% (range 19–58%), compared to RE, 89% (range 54–106%), (p < 0.001). The difference remained after 6 months, when the relative tumor diameter was 39% (range 7–86%) after HAE and 81% (range 31–103%) after RE (p < 0.001)