90Yttrium Selective Internal Radiation Therapy in Unresectable or Otherwise High-Risk Hepatocellular Carcinoma: Single-Centre Experience

Hong Kong J Radiol 2023 Mar;26(1):24-33

 

 

Department of Oncology, Princess Margaret Hospital, Hong Kong

 

  Full Article in PDF

In Hong Kong, hepatocellular carcinoma (HCC) ranks fifth most common cancer and third among the most common causes of cancer death since 2014.[1] Transarterial embolisation or transarterial chemoembolisation (TACE) has been shown to improve the survival of patients with unresectable HCC.[2] [3]

 

Selective internal radiation therapy (SIRT) is a directed liver therapy making use of the tumour vascularity in HCC in which the hepatic artery is usually the sole blood supply. SIRT involves the injection of beta emitters within resin or glass microspheres via the hepatic artery, where the spheres form microemboli, thus giving a very high radiation dose (100 to 1000+ Gy) to the tumour(s) while at the same time minimising the radiation exposure to normal liver tissue by not going through the hepatic veins or the portal system.

 

⁹⁰Yttrium (⁹⁰Y) is a pure beta-emitting isotope (maximum energy 2.28 MeV; mean energy 0.934 MeV), with a mean and maximum penetration range of 2.5 mm and 11 mm, respectively. It is commonly used to treat HCC.[4] ⁹⁰Y SIRT is effective, with one study showing an objective response rate up to 40.0% and a median overall survival (OS) of 16.4 months.[5] It has shown effectiveness in terms of survival, response rates, and safety profile similar to that with TACE in unresectable HCC in several studies and meta-analyses.[6] [7] [8] [9] It was also shown to be an effective treatment to accomplish downstaging as a bridge to transplantation, surgical resection, or radiofrequency ablation.[10] Survival in patients receiving ⁹⁰Y SIRT for intermediate-advanced HCC can vary from 12-24 months (1-year pooled OS = 63%) to 6-12 months (1-year pooled OS = 37%), should portal vein thrombosis be present.[11] [12]

 

Careful selection of suitable candidates for ⁹⁰Y SIRT is necessary. Several prognostic factors, including a low Child–Pugh score, percentage of liver replaced by tumour (≤50%) and alpha-fetoprotein (AFP) level (<400 μg/L) are associated with better survival.[13] Unilobar disease before SIRT and tumour response (complete response/partial response) have also been found to be significant predictors of survival.[14] It is believed that survival can be prolonged in unresectable HCC to a similar extent using TACE[15] with careful selection of candidates.

 

In our hospital (Princess Margaret Hospital), HCC patients are under the care of a multidisciplinary hepatoma team with oncologists, surgeons, and radiologists. Since 2012, ⁹⁰Y SIRT has been offered as a funded treatment by Hospital Authority, a statutory body managing the public healthcare services in Hong Kong, to high-risk HCC patients whose tumours are not resectable or ablatable, with portal vein thrombosis, or with tumour size >8 cm. Patients with infiltrative HCC, Child–Pugh class C disease, ascites, or inadequate liver reserve (with bilirubin level >34 μmol/L) are generally considered ineligible for SIRT. In this study, we report the outcome together with prognostic factors in the use of ⁹⁰Y in the treatment of these advanced cases of HCC in our centre.

 

We retrospectively enrolled HCC patients who received ⁹⁰Y SIRT, either resin microspheres containing ⁹⁰Y (Sirtex, Australia) or ⁹⁰Y-impregnated glass microspheres (TheraSphere; MDS Nordion, Canada), at our hospital between July 2017 and September 2021. Before SIRT, patients underwent hepatic angiography, ^90mtechnetium-macroaggregated albumin scintigraphy, and computed tomography (CT) scans to estimate the potential doses to tumour, liver, and lung. A catheter was guided through the femoral artery and into the hepatic artery by an interventional radiologist. Blood vessels feeding the gastrointestinal tract and extrahepatic sites such as the pancreas were identified and prophylactically embolised if necessary. Patients were deemed ineligible when lung shunting was >20%. The dose activity calculation was based on a partition model.[16] The aim of the treatment was to administer a minimum dose of 120 Gy to the tumour while keeping the dose to normal liver at <40 Gy and to <50 Gy in patients with poor liver reserve. The lung dose was planned to be <20 Gy.

 

Intrahepatic administration of radioactive ⁹⁰Y microspheres using either resin microspheres containing ⁹⁰Y or ⁹⁰Y-impregnated glass microspheres was performed. Understanding of radiation exposure of patients implanted with pure beta emitters is very limited. Patients were kept in a radiation isolation room to wait for assessment by physicists and considered safe if radiation activity was <1.5 GBq according to Radiation Ordinance of Hong Kong and Hospital Authority Code of Practice on Radiation Safety and Protection. They were discharged with medications including pantoprazole, ursodeoxycholic acid, prednisolone, and entecavir if they were viral hepatitis B carriers. A bremsstrahlung scan was performed on day 2 or 3 to document any extrahepatic reflux of ⁹⁰Y microspheres.

 

Patients were followed up by surgeons after ⁹⁰Y treatment with liver function tests and AFP tumour marker test. The first follow-up was within 4 weeks after discharge to assess for any treatment-related toxicities. All patients had reassessment with triphasic CT at approximately 3 months after ⁹⁰Y treatment for an objective evaluation of treatment outcome according to mRECIST (modified Response Evaluation Criteria in Solid Tumors) based on combined assessment of target lesions, non-target lesions, and new lesions. Appearance of one or more new lesions was counted as progression regardless of the response of treated target and non-target lesions classified according to mRECIST. Subsequent follow-up was performed at approximately 1- to 3-month intervals with laboratory tests and/or CT at the discretion of surgeons. Any toxicity or adverse events noted during the first 6 months after completion of ⁹⁰Y treatment were reviewed and graded according to CTCAE (Common Terminology Criteria for Adverse Events) version 5.0 criteria.

 

Treatment responses were assessed radiologically according to mRECIST. Local control rate was defined as the proportion of patients with at least stable disease of an irradiated target lesion. Objective response rate was defined as the proportion of patients with partial or complete response in target lesions and at least stable disease in non-target lesions to ⁹⁰Y SIRT. Progression-free survival (PFS) was defined from the date of SIRT to the date of a radiological sign of progression or death from any cause. OS was calculated from the date of SIRT to the date of death from any cause. Survival curves were determined by the Kaplan-Meier method and comparison between different Barcelona Clinic Liver Cancer (BCLC) stages[17] and AFP levels was done by the log-rank test. Statistical significance was defined at p < 0.05. Univariate and multivariable analyses of different prognostic factors of survival outcomes, including patient and tumour factors, were performed using the Cox proportional hazards analysis. Only factors with p values < 0.05 were considered significant and included in the multivariable analysis. Commercial software SPSS (Windows version 28.0; IBM Corp, Armonk [NY], United States) was used to perform the statistical analyses.

 

From July 2017 to September 2021, 30 HCC patients were treated with ⁹⁰Y, of whom 26 (86.7%) were male. The median age was 66.5 years (range, 40-93). The majority of them (n = 17, 56.7%) were chronic hepatitis B carriers. Liver tumour sizes ranged from 4 cm to 19.7 cm, with a median size of 11.7 cm. Most of them had Child–Pugh class A liver disease (n = 28, 93.3%) and about one-third (n = 11, 36.7%) had AFP level ≥400 μg/L. Eleven (36.7%) patients had portal vein thrombosis with four (13.3%) having thrombosis involving the main portal vein. Half of the patients received TACE with cisplatin and one-third of them received targeted therapy (either sorafenib or lenvatinib) as subsequent treatment. Other characteristics and laboratory investigations are listed in Table 1.

 

Table 1. Baseline characteristics of the patients treated with ⁹⁰Yttrium (n = 29).

 

Median follow-up time was 14.6 months. One patient with ECOG performance status score of 3 and BCLC stage D was excluded from the analysis. Of the 29 patients, 26 had assessable responses on CT (median time = 2.76 months after ⁹⁰Y treatment). Three patients had complete responses (11.5%), four with partial responses (15.4%), 13 with stable disease (50%), and six with progressive disease (23.1%). The objective response rate (defined as the sum of complete and partial responses) was 26.9% while the local control rate (defined as the sum of complete and partial responses and stable disease) was 76.9%. Three patients were lost to follow-up. The median PFS was 6.3 months and 1-year PFS was 40.2% (Figure 1). Median OS was not yet reached and 1-year OS was 57.5% (Figure 2).

 

Figure 1. Kaplan-Meier curves showing progression-free survival (PFS) after ⁹⁰Yttrium selective internal radiation therapy. In all, 20 patients out of 29 had disease progression. Median PFS was 6.3 months. PFSs at 6, 12 and 24 months were 54.8%, 40.2% and 36.5%, respectively.

 

Figure 2. Kaplan-Meier curves showing overall survival (OS) after ⁹⁰Yttrium selective internal radiation therapy. In all, 13 died out of 29 patients. Median OS was not reached. OSs at 6 and 12 months were 75.4% and 57.5%, respectively.

 

Eight out of 29 patients (27.6%) had surgery done after downstaging of disease (Table 2). Median time from ⁹⁰Y treatment to surgery was 6.1 months. One achieved a pathological complete response (Figure 3). Six of them had residual HCC completely resected and one resected with focally involved margin. Length of hospital stay was 5-24 days. One had significant intra-operative blood loss requiring massive blood transfusions. One had postoperative ileus and pulmonary embolism which resolved with anticoagulation. One died early postoperatively due to aspiration pneumonia.

 

Table 2. Demographics of patients amenable to surgery.

 

Figure 3. Computed tomography images of pre-⁹⁰Yttrium selective internal radiation therapy (⁹⁰Y SIRT) [a] and post-⁹⁰Y SIRT (b) in a 67-year-old patient. The known right lobe tumour showed mild reduction in size with no abnormal arterial enhancement. It suggested complete response according to modified Response Evaluation Criteria in Solid Tumors guideline.

 

In univariate analysis (Table 3), BCLC stage C or above (hazard ratio [HR] = 5.733, p = 0.024), and AFP level ≥400 μg/L (HR = 4.270, p = 0.012) were significant prognostic factors for OS whereas BCLC stage C or above (HR = 3.652, p = 0.010), post-SIRT surgery (HR = 0.134, p = 0.008), AFP level ≥400 μg/L (HR = 3.527, p = 0.007), and treatment responder (defined as those with complete response or partial response) [HR = 0.203, p = 0.034; Figure 4] were significant prognostic factors for PFS. In multivariable analysis (Table 4), AFP level ≥400 μg/L remained as a significant prognostic factor for OS (HR = 3.240, p = 0.045; Figure 5) as well as for PFS (HR = 3.930, p = 0.011; Figure 6).

 

Table 3. Univariate analysis.

 

Figure 4. Kaplan-Meier curves showing progression-free survival (PFS) after ⁹⁰Yttrium selective internal radiation therapy in patients with responders and non-responders (median PFS not reached vs. 4.2 months, p = 0.034).

 

Table 4. Multivariable analysis.

 

Figure 5. Kaplan-Meier curves showing overall survival (OS) after ⁹⁰Yttrium selective internal radiation therapy in patients with alpha-fetoprotein (AFP) level ≥400 μg/L and <400 μg/L (median OS = 9.2 months vs. not reached, p = 0.045).

 

Figure 6. Kaplan-Meier curves showing progression-free survival (PFS) after ⁹⁰Yttrium selective internal radiation therapy in patients with alpha-fetoprotein AFP level ≥400 μg/L and <400 μg/L (median PFS = 2.2 months vs. 24.1 months, p = 0.011).

 

In this series, there were four long-term survivors and three complete responders. Those with complete response achieved long survivals ranging from 29.6 to 52.9 months compared to a median of 11.2 months in non-responders. All of them had post-SIRT surgery done with clear resection margins. The median dose of ⁹⁰Y SIRT was higher in responders (200 Gy) than in non-responders (170 Gy). However, OS did not differ significantly with dose (lower dose: p = 0.268, 95% confidence interval = 0.983-1.005; higher dose: p = 0.456, 95% confidence interval = 0.201-2.056). Patient demographics and liver tumour baseline characteristics were investigated in treatment responders and non-responders together with those amenable to post-SIRT surgery and were compared to those that were not. The treatment responder group had better Eastern Cooperative Oncology Group (ECOG) performance status score (p = 0.039) and the group amenable to surgery had significantly more patients with BCLC stage B (p = 0.002) and better ECOG performance status score (p = 0.011).

 

The median postoperative hospital stay was 6.5 days (range, 2-16). Twelve patients (41.4%) had some forms of post-⁹⁰Y treatment complications, in total 29 different kinds of toxicities experienced by them (Table 5). Most side-effects were mild with abdominal pain and fever being the most common. Only one patient had grade 3 abdominal pain requiring hospital admission 6 weeks after ⁹⁰Y SIRT. There was no identifiable cause found in work-up and the patient was discharged the next day after symptom subsided with analgesics. Abnormal liver function with grade 1 hyperbilirubinemia occurred in one patient and was self-limited.

 

Table 5. Radioembolisation-related toxicities (n = 29).

 

There was one case of suspected radiation pneumonitis occurring approximately 7 weeks post-SIRT. The patient was a non-smoker and presented with fever and shortness of breath. The lung dose by the ⁹⁰Y SIRT was 24.4 Gy. Chest CT showed extensive ground-glass opacities and patchy consolidation, which may have represented oedema or infection. Multiple antibiotics and systemic steroids were administered but patient succumbed due to respiratory failure. Since the diagnosis was doubtful, it was not regarded as post-SIRT toxicity.

 

In this study, it was demonstrated that ⁹⁰Y SIRT was a feasible and effective treatment option in our local population who had intermediate- and advanced-stage HCC without serious adverse events. The response rate to ⁹⁰Y SIRT was high and the results were comparable to other Asian series that reported OS and PFS of patients receiving ⁹⁰Y SIRT ranging from 11 to 16.4 months and 2.4 to 11 months, respectively.[5] [18] [19] [20] The wide range of survivals represents heterogeneity of patients’ demographics and disease status, thus making direct comparison of survivals across different studies difficult. In our study, 59% and 34% of patients belong to BCLC stage C and had AFP level ≥400 μg/L, respectively. Both were found to be poor prognostic factors, which is consistent with the findings in a European multicentre analysis.[21] Despite this, the results of our cohort were impressive with encouraging results of local control, PFS, and OS.

 

TACE is commonly used in intermediate-stage HCC but is contraindicated in presence of portal vein thrombosis due to the potential risk of precipitating liver necrosis and failure by its effects on the already compromised hepatic vascular supply. This limitation can be overcome by ⁹⁰Y SIRT due to the small size of the ⁹⁰Y particles which exert microembolic effects on hepatic vascular dynamics.[22] Also, easy application in the left or right hepatic arteries of ⁹⁰Y SIRT makes it attractive for patients compared to superselective TACE with longer intervention times and repeated hospital admissions.[23] Yet, presence of portal vein thrombosis was shown to be associated with worse survival in treatment of ⁹⁰Y SIRT.[24] This was reconfirmed in our study in which the OS after SIRT was shorter in those with portal vein thrombosis compared to those without (median OS = 9.2 months vs. 26.4 months, p = 0.045).

 

In advanced HCC, targeted therapies are the mainstay of treatment with a median survival of approximately 13.6 months with lenvatinib and 12.3 months with sorafenib.[25] Although ⁹⁰Y SIRT failed to demonstrate a statistically significant difference in OS compared with sorafenib in two recent phase III trials, SARAH[26] and SIRveNIB[27], it had significantly fewer severe adverse events and better health-related quality of life. Most of the patients were classified as BCLC stage C in SARAH study, 68% in the ⁹⁰Y group and 67% in the sorafenib group, whereas respective rates were 48.4% and 44.9% in SIRveNIB study. Targeted therapy is associated with numerous side-effects, namely hypertension, diarrhoea, and hand-foot syndrome, and are known to lead to treatment discontinuation permanently in approximately 11% of patients.[28] On the other hand, ⁹⁰Y SIRT has better toxicity profiles with most side-effects being only mild as grades 1 to 2. This is also consistent with the observation in our cohort with the most common side-effects being grades 1 to 2 abdominal pain (23.3%) followed by fever (16.7%). There was only one grade 3 abdominal pain in our study with no identifiable causes. Symptom subsided quickly with analgesics and the patient was discharged the next day with no further complaints noted.

 

⁹⁰Y SIRT is also effective in bridging to liver surgery through tumour shrinkage and inducing future liver remnant hypertrophy in initially unresectable HCC.[29] In our study, 90% of the subjects underwent SIRT as the initial treatment and eight (30%) of them had surgery afterwards. They enjoyed a significantly longer survival (ranging from 24 to 52 months, excluded one died of postoperative complications). One of them demonstrated radiological and pathological complete response in his initial 4-cm tumour in subsequent hepatectomy after SIRT. He remains well without any disease recurrence for over 3 years by now. One of the long survivors received ⁹⁰Y SIRT twice to the right lobe of liver. Radiological complete response was achieved. Right hepatectomy was performed 2 months after the second ⁹⁰Y treatment. Pathology showed residual pT1 grade III HCC with clear resection margins. He has enjoyed >4 years of survival by now without disease recurrence. The above finding illustrated the potential role of downstaging and facilitating curative resection. ⁹⁰Y SIRT outperforms TACE in the role of downstaging from T3 to T2 HCC[30] and patients enjoyed better quality of life with ⁹⁰Y treatment.[31] Our series also showed ⁹⁰Y SIRT is safe with very low rates of grade ≥3 adverse events. Based on our study results, those amenable to surgery mostly were with ECOG performance status score of 0, classified as BCLC stage B with low AFP level, and without portal vein thrombosis, which could further guide our selection of ⁹⁰Y SIRT candidates aiming for surgical resection.

 

The retrospective nature and the small sample size in this cohort might affect survival analysis and determination of the significance of different prognostic factors. Also, the survival is not mature yet where longer follow-up of patients is necessary. Another limitation was the time of the reassessment of CT scans. The mean time to the first response assessment CT was 2.35 months after SIRT and not all patients had regular scans afterwards. Later response might then be underreported. Yet, the first CT was chosen to assess the treatment response to ⁹⁰Y SIRT as most of the patients had subsequent treatment which might confound the response solely due to ⁹⁰Y SIRT. In fact, most of the subjects only had one CT scan done within 6 months of ⁹⁰Y treatment. Furthermore, 23 out of 26 patients had ⁹⁰Y SIRT alone whereas only three cases had planned combination treatment with SIRT and TACE or systemic therapy. Two patients had concurrent TACE and one was taking sorafenib during radioembolisation. Hence, it is believed the CT could reflect the treatment response of ⁹⁰Y SIRT.

 

Our study confirmed the role of ⁹⁰Y SIRT in intermediate- to advance-stage HCC patients. However, careful patient selection is of utmost importance to optimise treatment benefits. There has been evidence suggesting that AFP level ≥400 μg/L predicts a higher rate of dual-tracer positron emission tomography/CT–detected metastasis.[32] Our study further confirms it as a negative prognostic factor, probably due to occult extrahepatic metastases. With discreet use of incorporating staging dual tracer positron emission tomography/CT scan in high-risk cases with AFP level ≥400 μg/L, we might screen out those with extensive distant metastases whom ⁹⁰Y SIRT is not advised as the initial therapy.

 

The result of the series further consolidated the role of ⁹⁰Y SIRT in our local practice. In locally advanced HCC, selection of appropriate treatment modalities has been challenging. The study reflects the importance of careful selection of candidates for ⁹⁰Y SIRT. Good ECOG performance status score and the classification as BCLC stage B HCC are shown to be favourable factors for this expensive radioisotope treatment and should be prioritised when it comes to selection of suitable candidates in multidisciplinary meetings. Other local ablative treatments such as radiofrequency ablation and stereotactic radiotherapy should be reserved for solitary or lower volume disease, whereas systemic therapy is for clearly disseminated disease.

 

⁹⁰Y SIRT is an effective and safe treatment for intermediate-to advanced-stage HCC which provides satisfactory local control with minimal toxicity. Longer survival was observed in patients with AFP level <400 μg/L.