Personalized Radioembolization Improves OS Vs Standard Dosing in Advanced HCC

Beau Toskich, MD

Yttrium-90 (Y-90) glass microspheres (TheraSphere) administered in a personalized, dosimetric approach demonstrated a 16-month improvement in overall survival (OS) compared with a standard dosimetric approach in patients with unresectable hepatocellular carcinoma (HCC), according to findings from the randomized phase 2 DOSISPHERE-01 trial.¹

In the modified intention-to-treat population (n = 56), the median OS was 26.6 months (95% CI, 11.7–not reached) with personalized dosimetry (n = 28) versus 10.7 months (95% CI, 6-16.8) with standard dosimetry (n = 28; HR, 0.421; 95% CI, 0.215-0.826; P = .0096).

The objective response rate (ORR) of index lesions was 71.4% with the personalized approach versus 35.7% with the standard approach (P = .0074), meeting the primary end point of the study. The complete response rates were 21.4% and 10.7%, respectively, and the partial response rates were 50% and 25%, respectively.

Additionally, 36% of patients treated with personalized dosimetry were successfully downstaged to surgery compared with 4% of patients treated with standard dosimetry.

“The DOSISPHERE-01 [trial] has led the field in terms of breaking down what dose [of Y-90] is necessary to achieve a response in HCC,” said Beau Toskich, MD, an interventional radiologist at Mayo Clinic. “Think of [TheraSphere] as a [tolerable] ablative modality for extremely large volumes of tissue.”

The randomized, multicenter, investigator-sponsored DOSISPHERE-01 trial was designed to compare the clinical outcomes of standard and personalized radioembolization with selective internal radiation therapy (SIRT) and TheraSphere, respectively, in patients with advanced HCC. The standard approach consisted of single-compartment dosimetry, defined as a uniform distribution of absorbed Y-90 dose within the perfused tumor and normal liver volume. Comparatively, the personalized approach was defined as a multicompartment distribution of absorbed Y-90 dose within the perfused volume that accounts for preferential blood flow into the tumor versus normal parenchyma.

The dosimetric goal was 120±20 Gy to the perfused lobe with the standard approach versus 205 Gy or more to the index lesion, with 250 Gy to 300 Gy if possible, with the personalized approach. Moreover, personalized dosimetry limited the non-tumor tissue dose to 120 Gy with hepatic reserve of 30% or greater.

Patients were assessed for efficacy and safety 4 to 6 weeks, 3 months, 6 months, and 12 months after treatment.

Additionally, patients were pretreated with arteriogram and 99mTc-MAA using SPECT/CT.

Eligible patients had to have unresectable HCC with 1 or more tumors 7 cm of greater. Patients had to have Barcelona Clinic Liver Cancer (BCLC) A, B, or C disease and 30% or greater hepatic reserve after first radioembolization. Patients with lung shunt fraction greater than 30 Gy, a risk of gastrointestinal (GI) exposure, and/or poor tumor or portal vein thrombosis (PVT) perfusion were excluded.

In both arms, 92.9% of evaluable patients were male, and 78.6% had Child-Pugh A5 disease (remainder, Child-Pugh A6/B7 disease). In the personalized group, 11 patients had BCLC B disease, and 89 patients had BCLC C disease versus 7 and 93 patients in the standard group, respectively. In the personalized group, 42.9% of patients had bilobar disease versus 57.1% in the standard group.

In the personalized dosimetry arm, 26 patients received 1 treatment with TheraSphere, and 2 patients received 2 treatments. In the standard dosimetry arm, 23 patients received 1 treatment, and 5 patients received 2 treatments.

The mean activity administered was 3.6 GBq (range, 2.4-4.8) with personalized TheraSphere versus 2.6 GBq (range, 2.2-3.0) with standard SIRT. The mean absorbed dose (AD) to the perfused liver was 178.4­­±59.9 Gy versus 120.3±15.2 Gy, respectively (P = .0001). Moreover, 67.8% and 3.6% of patients, respectively, had ADs of greater than 150 Gy to the perfused liver (P < .0001).

The mean AD to the index lesion was 331.1±131.5 Gy with personalized dosimetry versus 221.3±139.4 with standard dosimetry (P = .0007). Moreover, 87.5% and 37.5% of patients, respectively, had ADs of greater than 205 Gy to the index lesion. Finally, the mean AD to perfused normal tissue was 92.8±30.1 Gy and 64.5±36.6 Gy, respectively (P = .007).

Notably, the study also evaluated the group of patients with PVT; 64.3% of patients in the personalized group had PVT present versus 75% in the standard group.

In this population, the median OS was 23 months with personalized dosimetry versus 9.5 months with standard dosimetry. Additionally, 44% versus 0% of patients with PVT were successfully downstaged to surgery, respectively.

Regarding grade 3 or greater toxicities related to Y-90, 8.6% (n = 3) of patients who received personalized therapy (n = 35) experienced 1 or more adverse effects (AEs) versus 14.3% (n = 3) of patients who received standard therapy (n = 21). One death was reported in each arm.

Liver AEs were seen in 11.4% (n = 4) of patients in the personalized arm and included ascites (n = 1; 2.9%), bilirubin increase/jaundice (n = 1; 2.9%), and hepatic failure (n = 2; 5.7%). In the standard arm, 23.8% (n = 5) of patients experienced liver AEs, including ascites (n = 2; 9.5%), GI hemorrhage (n = 2; 9.5%), and bilirubin increase/jaundice (n = 2; 9.5%). Notably, no patients experienced encephalopathy while on treatment.

“The DOSISPHERE-01 trial called out the several randomized phase 3 trials that were historically negative,” said Guy E. Johnson, MD, PharmD, a radiologist at the University of Washington (UW) Medical Center, and an assistant professor of radiology at UW Medicine. “[The findings] call into question the dosimetry of those studies and suggest that we had things wrong in terms of the dosimetric approach. [Overall, this] adds to the growing evidence that dose tailoring is really important to outcomes and safety of radioembolization in HCC.”

The well-tolerated nature of the Y-90–based approach suggests that locoregional therapy could play a more significant role in the treatment of patients with HCC.

For example, a single-center, retrospective study published in the Journal of Vascular and Interventional Radiology reported that locoregional therapies given concurrently with nivolumab (Opdivo) had an acceptable safety profile among patients with intermediate/advanced HCC.² Twenty-nine patients underwent treatment with 41 locoregional interventions with transarterial chemoembolization or Y-90 transarterial radioembolization 60 days prior to or concurrently with nivolumab.

At a median follow up of 11.5 months, no grade 3 or 4 AE or deaths attributed to nivolumab were observed. Five cases of grade 3/4 hypoalbuminemia or hyperbilirubinemia were observed within 3 months after locoregional therapy.

“I [recommend] that practices do what is best within their local practice and experience level if it is safe. If you can enroll patients into some of the studies that are coming out, that is encouraged,” concluded Toskich.

References

1. TheraSphere Y-90 glass microspheres: lobar +/- PVT: DOSISPHERE-01 trial summary. Boston Scientific. 2021. Accessed January 22, 2021. http://bit.ly/2Mmc4NU

2. Marinelli B, Cedillo M, Pasik SD, et al. Safety and efficacy of locoregional treatment during immunotherapy with nivolumab for hepatocellular carcinoma: a retrospective study of 41 interventions in 29 patients. J Vasc Interv Radiol. 2020;31(11):1729-1738. doi:10.1016/j.jvir.2020.07.009