This review assessed the effect of radioembolism with yttrium-90 microspheres on overall response and survival in patients with colorectal liver metastases or hepatocellular carcinoma and concluded that yttrium-90 resin microspheres delivered high response rates. As many aspects of the review process (including the analysis) were not reported clearly, the reliability of the authors’ conclusion is unclear.

To assess the effect of radioembolism with yttrium-90 microspheres on overall response and survival in patients with liver malignancies.

MEDLINE, EMBASE and The Cochrane Library were searched from 1986 onwards for relevant articles. Only English- and German-language papers were assessed. Search terms were specified. Reference lists of relevant retrieved articles and other sources were searched.

Studies were considered eligible for inclusion if patients had been diagnosed with either colorectal liver metastases (mCRC) or hepatocellular carcinoma (HCC). Eligible studies needed to report tumour response (as measured by computerised tomography scan) or median survival time data in patients who had been treated with either yttrium-90 glass or yttrium-90 resin microspheres. Papers that reported data only on groups with mixed primary disease were excluded.

In the included studies, most patients with hepatic metastatic colorectal cancer received resin microspheres as first-line treatment neoadjuvant to chemotherapy or as salvage therapy. The cytostatic agent received was older generation 5-FU/LV or floxuridine, or newer generation 5-FU/LV + oxaliplatin (FOLFOX) or 5-FU/LV + irinotecan (FOLFIRI). Details of previous treatments received by patients with hepatocellular carcinoma were not reported. Most patients received glass microspheres.

It was unclear how many reviewers performed study selection.

The authors did not state that they assessed validity.

Data on treatment type (glass or resin and other details), tumour type and outcome type (whether there had been any response to treatment, which included complete or partial response, or whether disease had stabilised) were extracted.

The authors did not state how data were extracted for the review.

The proportions of patients who exhibited any response were modelled using an exact binomial likelihood approach and pooled using a random-effects model. P values obtained from this model were used to assess the statistical significance between groups; p values under 0.10 were considered statistically significant.

For the subgroup of patients with mCRC, two subgroup analyses were conducted: comparison of older with newer cytostatic agents; and whether yttrium-90 had/had not been used as either a first-line treatment with adjuvant chemotherapy or salvage therapy. For patients with HCC, differences in any response between patients who received resin microspheres were compared with any response in patients who received glass microspheres.

Results from 30 papers (n=1,217 patients) were included in the meta-analyses.

Overall, the proportion of combined mCRC and HCC patients with any response varied from 29% to 100% (median 82%). Any response with glass microspheres was lower than with resin microspheres, but this difference was not statistically significant at the 5% level.

mCRC: Eighteen studies (681 patients with mCRC) were included in the analysis. Based on this analysis, the specific cytostatic agents used (whether older or newer) were judged not to have affected response (p=0.96). There was a difference in any response when yttrium-90 resin microspheres were used as salvage (any response 79%) compared with first-line therapy (any response 91%); this difference was statistically significant at 10% but not at more conventional 5% level (p=0.07). Median survival duration after resin microsphere treatment ranged from 6.7 to 17.0 months and from diagnosis from 10.8 to 29.4 months.

HCC: Twelve studies (318 patients) were included in the meta-analysis. This indicated that any response proportions were statistically significantly higher (p=0.02) in resin microsphere groups (any response 89%) than glass microsphere groups (any response 78%). Median survival times (seven studies) varied between 7.1 and 21.0 months from treatment and between 9.4 and 24.0 months from diagnosis or recurrence.

The authors concluded that yttrium-90 resin microspheres were associated with high response rates in both salvage and first-line settings. However, the results of relevant ongoing and upcoming phase III trials should be awaited to better understand the true impact on survival of these treatments.

This review addressed a well-defined research question with clear and appropriate inclusion and exclusion criteria. Multiple databases were searched for papers published in more than one language and additional references sought, which reduced the risk of bias. It was unclear whether appropriate methods were used to reduce error and bias in review processes such as study selection and data extraction. No assessment of study quality was reported and few primary study details were reported, which raised the possibility that poor study quality and confounding variables may have biased the results. The results of the statistical analyses were unclearly and incompletely reported, which maked assessment of these results and the conclusions drawn from them challenging. For these reasons, the reliability of the authors' conclusions is unclear.

Practice: Fludeoxyglucose positron emission tomography (¹⁸F-FDG-PET) or functional magnetic resonance imaging (F-MRI) should be added to post-treatment response assessment protocols to improve sensitivity in assessing tumour response.

Research: Not stated.

Not stated.

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.