Coil embolization of hepatic arteries to induce redistribution of blood flow and microspheres has a mediocre success rate, concluded a recent study presented by Ahmed Alsultan (University Medical Center, Utrecht, the Netherlands) at the Cardiovascular and Interventional Radiological Society of Europe (CIRSE) annual meeting (22–25 September, Lisbon, Portugal). The data further led Alsultan to advise physicians—when dose planning for a patient—to take into account that the radiation dose in the dependent regions is lower in general.
Alsultan acknowledged that variations in hepatic vasculature can pose a problem for patients undergoing radioembolization. Such variations may require one or more extra injection locations for radioembolization. Yet, according to Alsultan, redistributing the flow by coil-embolizing the problematic artery may be a way of overcoming this issue; eliminating the need for extra injection locations. Therefore, preexisting intrahepatic collaterals would be expected to take over the distribution of blood to a particular area, although Alsultan reported that in practice, himself and his colleagues have observed mixed results.
This led the investigators to carry out a retrospective study in order to screen all radioembolization patients in their affiliated institute, including patients that had a closure of the tumour-feeding artery. By doing so, Alsultan and colleagues aimed to evaluate the effect of coil-embolization of tumour-feeding vessels on the redistribution of blood flow in radioembolization.
To visually assess the effect of redistribution of the microspheres, post-therapy nuclear medicine images were taken from the PET (positron emission tomography) scans—subsequently analysed by two nuclear medicine physicians. The effect of redistribution was quantified by comparing the relative dose to the coil-embolized (dependent) segment relative to the other (non-dependent) segments and to the tumour in that segment on posttreatment PET and SPECT (single photon emission computerised tomography) scans using Simplicit90Y (Mirada Medical) software.
Alsultan pointed to a particular case: a segment four artery that could not be targeted with an injection to the left hepatic artery. The investigators decided to coil this artery and to inject the microspheres proximally to those in the left hepatic artery. According to Alsultan, the nuclear medicine image displayed very good redistribution of the blood flow, while the dosimetry data showed promising ratios: the segment ratio being 1.88 and the tumour ratio 1.21; indicative of good redistribution.
However, he noted that in other instances, less promising outcomes were observed. He discussed another case of a phrenic artery (that supplied three tumours) that was also coiled. Yet in this case, the nuclear medicine image from the PET scan depicted poor distribution and very low tumour rates (segment ratio of 0.93, tumour ratio 0.33).
Of the 37 cases in total, 32 were available for dosimetric analysis and 35 for visual analysis. Regarding the visual analysis, Alsultan reported that around 70% of patients had good redistribution of blood flow (with good interrater agreement, kappa=0.82).
To determine the level of blood redistribution, Alsultan and colleagues came up with three cut-off values: 0.9, 0.8 and 0.7, that correspond to a decrease in the absorbed dose of ten, 20 and 30 per cent, respectively. Therefore, a 30% success rate was observed for the 0.9 cut-off values, and a 55% success rate was observed for the more “liberal” cut-off value of 0.7.
In terms of the segment analysis, the success rate ranged from 46–69%. The median ratios of the dose to the dependent segment and the non-dependent segments were 0.88 (range 0.26 – 2.05) and 0.79 (range 0.19 – 1.62) for dependent tumours compared with non-dependent tumours.
In light of these data, Alsultan concluded that coil embolization of hepatic arteries can be used to redistribute the blood flow and the microspheres. However, he stressed that physicians should keep in mind that the radiation dose in the dependent region is lower in general, which should be taken into account when dose planning.