We read with interest the recent article by Sangro et al., 1 but we wish to highlight our concern about the inhomogeneity of the radiation dose associated with the angiographic delivery of yttrium 90 (90Y) microspheres. Although the authors claim that 90Y emits “a tumoricidal dose of beta radiation (100-1000+ Gy), far in excess of the doses delivered safely with external beam radiotherapy, over a finite range,” the biological effects of the absorbed dose on tumorous and normal tissues are not simple functions of this dose. The way in which the dose is given to each subvolume (voxel) determines the overall biological effect on the treated tissues. For example, a single hot spot may cause unacceptable damage, and a cold spot may result in a failure to sterilize the tumor. During the angiographic injection of 90Y microspheres, the spatial distribution of the microspheres is very irregular, and the resulting dose distribution is highly heterogeneous. 2 Furthermore, the biologically effective dose is even more heterogeneous because of the effect of changing the dose rate. 3 Therefore, although the overall mean dose distribution may appear satisfactory at a microscopic level, the dose may be highly inhomogeneous, and there may be a considerable risk of small cold spots.
Conventional external beam radiotherapy (EBRT) should also be considered in the context of advanced hepatocellular carcinoma (HCC). EBRT for HCC has significantly advanced in recent years because of improved three-dimensional conformal techniques and improved knowledge of radiation dose–volume relationships. 4 The efficacy and safety of EBRT for advanced HCC have been suggested by a large number of nonrandomized studies. 5 In these studies, EBRT has usually been combined with transarterial chemoembolization for Child A/B patients and has been applied to the tumor thrombus or the primary tumor. Promising nonrandomized data have prompted many calls for randomized studies. Other advantages of EBRT (the dose uniformity, accessibility, cost, noninvasiveness, and outpatient basis) also encourage such studies.
The obvious methodological problems of Sangro et al.'s study 1 (the retrospective analysis and the lack of a control group or randomization) are further concerns, and although the difficulty of randomized controlled trials in this setting is acknowledged, such evidence is essential before the use of 90Y radioembolization can be recommended outside clinical trials. The importance of such trials needs to be highlighted because of the association of 90Y radioembolization with significant costs and harm (including death, as described in this study), the concerns about radiation dose inhomogeneity, and the availability of alternative methods of radiotherapy.