The patient outcomes within this cohort add to the growing literature supporting radiation segmentectomy with glass microspheres for the treatment of early-stage HCC in patients with solitary tumors. The localized objective response rate of 96% aligns with other reported outcomes in literature [1, 12, 13]. Comparably, durable treatment response greater than 2 years was achieved in 98% of treated lesions, and only 11% of patients demonstrated progression at 2 years.
The importance of individual SA is becoming more integral piece of personalized dosimetry [13, 14]. SA ≥ 327Bq and ≥ 446Gy are more likely to generate CPN in patients who have received ablative radiation segmentectomy [13]. In this cohort, SA did not demonstrate a statistically significant difference in TTP, as SA for every case was above the cutoff threshold identified previously for optimal outcomes [13]. On the other hand, early arterial stasis can be seen during the use of resin microspheres, related to its lower SA and need for larger volumes of microspheres to reach the intended dose to target, which may result in early termination of Y90 delivery and failure to completely deliver the prescribed dose [15]. This could imply that limited space exists within the tumor microenvironment for efficient microsphere delivery. Similarly, the results of this study highlight the importance of SA on imaging and pathologic treatment response, as it relates to limited space within the tumor to concentrate microspheres and deliver an ablative dose. In this study, the desired dose was delivered in all cases, which is similar to other publications involving the use of glass microspheres [1, 3, 5, 12, 13]. Based on Fig. 1, as perfused volumes decrease, tumor sphere concentration increases. While this study was not designed to evaluate the maximum sphere concentration allowed in tumor, tumor sphere concentrations rarely increased greater than 20,000 microspheres/mL, even at perfused volumes under 100mL, when spheres are most concentrated in the tumor. This, and studies with resin Y90 microsphere delivery ending in stasis, may suggest that sphere concentration into tumor becomes inefficient above a saturation point.
If 20,000 microspheres/mL were used as the cutoff for efficient microsphere saturation within the tumor microenvironment, the minimum threshold for SA could be calculated to achieve a target dose of 400Gy to the perfused volume. Using MIRD dosimetry, the minimum SA would need to be 412Bq/sphere to achieve a target dose of 400Gy in tumors within the size range of this study, 1.2–6.8 cm. This is equivalent to planning treatment within 8.5 days after initial calibration of glass microspheres.
While a recent study has demonstrated dose thresholds when treating larger liver volumes with resin microspheres, no dose threshold was noted for segmental deliveries despite higher radiation doses to non-tumoral liver, suggesting that AEs are more associated with treatment volume rather than dose [9]. Notably, resin microspheres are much more likely to reach stasis and result in homogeneous saturation of the perfused non-tumoral liver, resulting in a greater amount of hepatic lobules at risk [8, 15]. However, ablative radiation segmentectomy with glass microspheres is well-tolerated with limited AEs in patients with preserved liver function [16]. Similarly, AE rates were very low in this cohort, despite ablative doses to normal tissue. This likely reflects the impact of total liver volume treated and limited non-tumoral liver perfusion as well as smaller spherical concentration and heterogeneous distribution to normal tissue with glass microspheres, resulting in fewer hepatic lobules at risk.
Several limitations in this study exist, including its retrospective design and relatively small sample size. Another potential limitation of this study is the use of SPECT/CT in post- Bremsstrahlung Y90 imaging. SPECT/CT and PET/CT have demonstrated good agreement in regions of treatment on prior studies. However, PET/CT had been shown to overestimate activity in regions of low or no activity [17]. Therefore, any estimates of uptake in areas of low absorbed dose would be at risk for relative underestimation of dose. Additionally, mass balance of total spheres per vial, distributed into to tumor and non-tumor liver, was appropriately accounted for in this cohort. Finally, 4300Bq was used as the mean activity per sphere and some variability may exist between each dose. However, this activity was directly correlated with individual lot numbers from the distributor, and variations between each vial should be relatively small. This number also closely corresponds to a recent study evaluating Bq/sphere with glass microspheres [8].
In conclusion, SA plays a critical role in radiation segmentectomy, as limited space exists within the tumor bed for efficient sphere concentration. Higher individual SA allows reproducible delivery of ablative target radiation doses to tumor with spheres that will undergo more decays per sphere and increase radiation damage to the targeted tumor. Increasing sphere concentration beyond its saturation point results in a higher concentration of microspheres to non-tumoral liver and greater likelihood of adverse events. Therefore, if sphere concentration to tumor begins to become inefficient at 20,000 microspheres/mL, SA should be 412Bq/sphere, or 8.5 days from initial calibration, to achieve a minimum target dose of 400Gy.