Abstract
Purpose
This study evaluated pathologic response to ablative transarterial radioembolization (TARE) and compared single-compartment, mean tumor dose, and voxel-based dosimetry thresholds for hepatocellular carcinoma (HCC).
Materials and Methods
This is a retrospective, single-center study which included consecutive patients treated with glass microsphere TARE for a previously untreated HCC followed by liver transplantation between January 2015 – December 2024. A complete response (CR) was assigned if explant pathology showed complete pathologic necrosis (CPN) and if the patient had not undergone further treatments to the tumor. The standard single-compartment, dual-compartment mean tumor, and voxel-based doses were compared for patients with CR and without.
Results
Of the forty-three patients included, 29 had a CR. The only single-compartment dose threshold significantly associated with CR was ≥400 Gy, with a positive predictive value (PPV) of 77%. Mean tumor dose thresholds of ≥300 and ≥400 Gy were associated with CR (PPV 75 and 76%, respectively). The highest PPV for CR observed was for the voxel-based dose threshold of D95 ≥500 Gy, at 88%. Single-compartment, dual-compartment mean tumor, and D95 doses were all significantly associated with the likelihood of CR in separate multivariable logistic regression analyses. The incidence of AEs did not vary significantly based on the tested dose thresholds.
Conclusion
Increasing the single- and dual-compartment dose thresholds beyond ≥400 Gy did not result in significant associations with CR nor improved PPV. In contrast, D95 ≥500 Gy was significantly associated with CR with the highest observed PPV.
Level of Evidence:
Level 3, Local Non-random Sample.
Keywords: Hepatocellular carcinoma, transarterial radioembolization, dosimetry
Introduction
A cornerstone of hepatocellular carcinoma (HCC) treatment is potentially curative hepatic resection or transplantation, though alternative interventions such as percutaneous ablation may also be curative in selected patients. Locoregional therapy (LRT) plays a key role to help bridge or downstage prior to surgery [1,2]. A complete response to LRT on explant pathology has been associated with improved post-transplant outcomes including overall and recurrence-free survival [3,4]. For transarterial radioembolization (TARE) specifically, complete pathologic necrosis (CPN) has been associated with improved recurrence-free survival, disease-specific mortality, and trends in overall survival [5,6].
The LEGACY study (Local radioEmbolization using Glass Microspheres for the Assessment of Tumor Control with Y-90) was a multicenter retrospective study that showed clinically meaningful response rates and overall survival for patients with HCC ≤ 8 cm treated with TARE using a radiation segmentectomy approach (selective delivery up to two Couinaud segments) [7,8]. In a subset dose-pathology analysis, a single-compartment dose >400 Gy was associated with 100% likelihood of CPN on explanted livers [9]. These data built on prior studies to suggest a higher dosimetry standard may increase the likelihood of CPN, though no upper limit of benefit was observed [8,10,11]. With higher doses, there is concern that there may be higher rates of adverse events such as gastrointestinal disorders, laboratory abnormalities (e.g., hyperbilirubinemia, cytopenia), abdominal pain, and nausea [7,12]. More feared complications (i.e. radiation pneumonitis and radioembolization-induced liver disease (REILD)) may be more likely at higher doses and therefore obtaining an upper limit of benefit would be helpful.
The primary aim of this study was to compare the predictive performance of single-compartment, dual-compartment mean tumor, and voxel-based dose thresholds to predict CPN at transplant. As an exploratory aim, post-TARE safety outcomes were also assessed to evaluate for increased toxicity at higher doses.
Materials and Methods
Study Design and Patient Cohort
All HCC patients are evaluated for liver transplantation at this institution. Patients awaiting transplantation are evaluated for LRT as a ‘bridging’ intervention, while those who are initially outside the Milan criteria are evaluated for LRT for ‘downstaging’: specific LRT treatment recommendations are made by multidisciplinary board review including interventional and diagnostic radiology, hepatology, medical oncology, and transplant surgery. This Institutional Review Board-approved HIPAA-compliant retrospective study included all consecutive patients from a single institution who underwent radiation segmentectomy strategy TARE with yttrium-90 [90Y] glass microspheres (Theraspheres®, Boston Scientific; Marlborough, MA) for a previously untreated HCC tumor followed by liver transplantation between January 2015 – December 2024. A similar cohort was used in a study examining voxel-based dosimetry and single-compartment dose as estimated by anatomic imaging.[13] If multiple HCC tumors (by imaging or biopsy-confirmed) were treated during the same TARE treatment session, the largest HCC tumor treated was selected as the index tumor. Patients who did not obtain post-TARE SPECT, were treated with a radiation lobectomy approach, received prior systemic therapy for HCC, or had undergone prior LRT to the index tumor (or same arterial perfusion region) were excluded to avoid confounding effects from prior therapies. Patients who had received prior LRT to other non-index tumors were not excluded. Patients with extrahepatic disease and/or tumor vascular invasion at the time of treatment were excluded from analysis.
Patient and Tumor Baseline Characteristics
Patient demographics, tumor characteristics, and baseline laboratory evaluations were collected. HCC was defined by current practice as tumors meeting LI-RADS criteria for LR-5 or by biopsy confirmation. HCC disease stage by Barcelona Clinic Liver Cancer (BCLC) strategy and the longest pre-TARE tumor diameter was collected [14]. Baseline laboratory evaluations included the most recent pre-TARE Model for End-Stage Liver Disease (MELD-Na) score, and albumin-bilirubin (ALBI) grade [15–17]. Child-Pugh classification was also reported. Tumor markers included total AFP [18].
Dosimetry Calculations
The single-compartment dose was estimated using the post-TARE Bremsstrahlung SPECT/CT with the commercially available Simplicity90 ® software (Mirada Medical, Oxford UK). The multi-compartment dosimetry function in Simplicity90 was then used to estimate the dual-compartment tumor absorbed dose and the voxel-based D95 dose (the minimum dose to 95% of the tumor volume). Tumors were contoured using a freehand tool on axial image slices from at least three separate points: the caudal-most, mid-point, and rostral-most boundaries of the tumor.
The interpolation function in Simplicity90 was utilized to connect these two-dimensional contours into a three-dimensional volume. The resulting tumor contours were assessed visually and adjusted manually as needed. In all cases, additional contours beyond the three boundary points alone were made in order to reduce reliance on the software’s interpolation function.
Given the length of this study period, microspheres were prescribed and ordered according to clinical practice at the time of treatment which, in general, was initially a target single compartment lobar dose of 120 Gy (infused into a single segment), then a single compartment dose of at least 190 Gy and, after the publication of the LEGACY data, at least 400 Gy [7,9,10].
Post-TARE Outcomes
Clinical pathology reports were utilized and samples were not re-examined as part of this study. Explanted livers were examined by pathologists according to institutional practice; the whole explanted liver was serially sectioned in 1 cm slices to identify any radiologically reported observations in addition to any incidental nodules/observations. Pathologists correlated nodules to known tumors and routinely included an estimate of percent tumor necrosis for each nodule. A complete response was assigned if 100% tumor necrosis was reported for the index tumor and the patient had not undergone any additional treatments to the index tumor after the initial TARE. The decision to re-treat the index lesion after the initial TARE was based on a multidisciplinary evaluation using the Liver Imaging Reporting & Data System (LI-RADS) treatment response criteria suggesting residual or recurrent viable disease [19]. Based on these criteria, a non-complete response was assigned if either < 100% necrosis was reported or if the patient had additional treatment(s) to the index tumor following the initial TARE suggesting that complete necrosis was not obtained if residual or recurrent disease was present.
Post-treatment Adverse Events
Imaging, laboratory values, and clinical documentation were evaluated in the 90 days post-treatment session by chart review. AEs were classified retrospectively using the interventional radiology-specific system developed by the Society of Interventional Radiology (modified SIR AE classification system) [20,21].
Statistical Analysis
All statistical analyses were performed using RStudio (version 4.1.0, The R Foundation for Statistical Computing, Vienna, Austria). Patient and tumor characteristics and dose estimates were compared between the complete response (CR) and non-complete response (non-CR) groups by Mann-Whitney test, chi-squared test, or Fisher’s exact test as appropriate. Given the importance of ALBI grade and tumor size as prognostic factors, the relationships between ALBI grade, tumor size, and dose were also evaluated. Additionally, CR rate, accuracy, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were computed for multiple dose cutoffs: 200, 300, 400, 500, 600, 700, and 800 Gy dose thresholds were selected to encompass current clinical practice [7,9,12]. As an exploratory aim, safety outcomes were compared above and below thresholds found to be statistically associated with CR by chi-squared test or Fisher’s exact test as appropriate. The cutoff for statistical significance was p < 0.05 for all analyses.
Results
A total of 45 patients underwent radiation segmentectomy TARE with glass microspheres for a previously untreated HCC tumor followed by liver transplantation within the defined study period, 43 of whom also obtained post-TARE SPECTs as summarized in Figure 1. 29/43 tumors (67%) met the criteria for CR. Patient demographics, laboratory evaluations, and HCC disease characteristics are summarized in Table 1. Child-Pugh classification varied significantly between the CR and non-CR groups (97% class A in CR versus 64% class A in non-CR, p = 0.01). Median maximal tumor size differed significantly between the CR and non-CR groups (2.5 cm versus 3.5 cm, p = 0.046). Of the 14 (33%) tumors that did not meet CR criteria, four were due to repeat treatment for residual/recurrent tumor prior to transplantation and 10 due to <100% necrosis on explant pathology.
Figure 1.
Flow diagram.
Table 1.
Patient demographics, baseline laboratory evaluations, and tumor characteristics by pathologic response.
| Non-complete Response | Complete Response | Significance | |
|---|---|---|---|
|
| |||
| # of Patients | 14 | 29 | |
| Median Time to Transplantation (IQR), days Patient Demographics |
266 (207 – 355) | 228 (184 – 330) | p = 0.61 |
| Median Age at TARE (IQR), years | 62 (57 – 67) | 65 (61 – 68) | p = 0.64 |
| Sex | p = 1 | ||
| Male (%) | 10 (71%) | 20 (69%) | |
| Female (%) | 4 (29%) | 9 (31%) | |
| Race | p = 0.53 | ||
| Asian (%) | 0 | 4 (14%) | |
| Black (%) | 0 | 1 (3%) | |
| Other/Mixed (%)* | 6 (43%) | 12 (41%) | |
| White (%) | 8 (57%) | 12 (41%) | |
| Ethnicity | p = 0.19 | ||
| Non-Hispanic (%) | 4 (29%) | 16 (55%) | |
| Hispanic (%) | 10 (71%) | 13 (45%) | |
| Underlying Liver Disease Etiology ** | |||
| Alcohol | 7 (50%) | 7 (24%) | p = 0.16 |
| NAFLD | 6 (43%) | 9 (31%) | p = 0.51 |
| HCV/HBV | 7 (50%) | 19 (66%) | p = 0.52 |
| Other/Unknown | 1 (7%) | 0 | p = 0.33 |
| HCC Disease Severity | |||
| BCLC Stage | p = 0.60 | ||
| 0 | 0 | 2 (7%) | |
| A | 9 (64%) | 20 (69%) | |
| B | 3 (21%) | 6 (21%) | |
| C | 2 (14%) | 1 (3%) | |
| Lobe Involvement | p = 1 | ||
| Unilobar | 13 (93%) | 26 (90%) | |
| Bilobar | 1 (7%) | 3 (10%) | |
| Solitary/Multifocal | p = 1 | ||
| Solitary | 11 (79%) | 22 (76%) | |
| Multifocal | 3 (21%) | 7 (24%) | |
| Baseline Laboratory Evaluations | |||
| Median MELD-Na (IQR) | 9 (7 – 12) | 9 (8 – 10) | p = 0.75 |
| ALBI Grade | p = 0.13 | ||
| Grade 1 | 4 (29%) | 17 (59%) | |
| Grade 2+ | 10 (71%) | 12 (41%) | |
| Child-Pugh Classification | p = 0.01 | ||
| A | 9 (64%) | 28 (97%) | |
| B | 5 (36%) | 1 (3%) | |
| Median total alpha fetoprotein (IQR), ng/mL | 11.8 (7.7 – 21.5) | 8.0 (4.7 – 20.8) | p = 0.28 |
| Median international normalized ratio (IQR) | 1.2 (1.1 – 1.3) | 1.2 (1.1 – 1.3) | p = 0.94 |
| Tumor Characteristics | |||
| Median longest dimension (IQR), cm | 3.5 (3.0 – 4.4) | 2.5 (2.3 – 3.2) | p = 0.046 |
| Estimated tumor volume (mL)*** | 14.2 (10.3 – 21.4) | 7.8 (3.9 – 12.4) | p = 0.03 |
“Other” or “other/mixed” by self-report.
Patients may have multiple contributing etiologies documented so totals may exceed 100%.
Tumor volume as estimated by the Simplicity90 dosimetry software. IQR: Interquartile range. NAFLD: Non-alcoholic fatty liver disease. HCV: Hepatitis C Virus. HBV: Hepatitis B Virus. BCLC: Barcelona-Clinic Liver Cancer. MELD-Na: Model for End-stage Liver Disease. ALBI Grade: Albumin-Bilirubin Grade.
For each TARE session, the largest treated HCC tumor (LR-5 or biopsy-confirmed HCC) was designated the index tumor. The performance of single-compartment, dual-compartment mean tumor absorbed dose, and voxel-based D95 dose thresholds are summarized in Table 2. The 400 Gy single compartment dose threshold was significantly associated with probability of CR across all three dosimetry models. For the single- and dual-compartment models, higher dose thresholds beyond 400 Gy were not associated with CR nor an increase in PPV. For the voxel-based D95, in addition to the 400 Gy cutoff, D95 ≥500 and ≥600 Gy were also associated with CR and a higher PPV. The highest PPV observed was 88% (21/24) for a D95 ≥500 Gy threshold. As an illustrative case example, Figure 2 demonstrates dose-volume histograms for two patients that received similar (approximately 800 Gy) single-compartment doses but differed along the voxel-based dosing estimates and had different pathologic outcomes.
Table 2.
Association of dose thresholds with pathologic response.
| Single-compartment Dose Thresholds (Gy) | |||||||
|---|---|---|---|---|---|---|---|
|
| |||||||
| ≥200 | ≥300 | ≥400 | ≥500 | ≥600 | ≥700 | ≥800 | |
|
| |||||||
| Association With Response * | |||||||
| Significance* | 0.24 | 0.05 | 0.04 | 0.18 | 0.33 | 0.33 | 0.51 |
| Odds Ratio for CR (95% CI) |
4.48 (0.21–284) |
4.49 (0.83–27.8) |
4.60 (0.93–25.3) |
2.56 (0.57–12.0) |
2.17 (0.50–10.5) |
2.29 (0.51–12.4) |
1.74 (0.38–9.48) |
| Accuracy | 70% (30/43) | 72% (31/43) | 72% (31/43) | 65% (28/43) | 58% (25/43) | 56% (24/43) | 51% (22/43) |
| Sensitivity | 97% (28/29) | 86% (25/29) | 83% (24/29) | 72% (21/29) | 55% (16/29) | 48% (14/29) | 41% (12/29) |
| Specificity | 14% (2/14) | 43% (6/14) | 50% (7/14) | 50% (7/14) | 64% (9/14) | 71% (10/14) | 71% (10/14) |
| PPV | 70% (28/40) | 76% (25/33) | 77% (24/31) | 75% (21/28) | 76% (16/21) | 78% (14/18) | 75% (12/16) |
| NPV | 67% (2/3) | 60% (6/10) | 58% (7/12) | 47% (7/15) | 41% (9/22) | 40% (10/25) | 37% (10/27) |
|
| |||||||
| Dual-compartment Tumor Absorbed Dose Thresholds (Gy) | |||||||
|
| |||||||
| ≥200 | ≥300 | ≥400 | ≥500 | ≥600 | ≥700 | ≥800 | |
|
| |||||||
| Association With Response * | |||||||
| Significance | 0.10 | 0.03 | 0.04 | 0.13 | 0.16 | 0.18 | 0.20 |
| Odds Ratio for CR (95% CI) |
Inf (0.40-Inf) |
7.09
(0.96–87.0) |
6.16
(1.04–47.1) |
3.48 (0.68–19.1) |
2.80 (0.57–14.2) |
2.88 (0.66–13.6) |
2.48 (0.57–11.5) |
| Accuracy | 72% (31/43) | 74% (32/43) | 74% (32/43) | 70% (30/43) | 67% (29/43) | 65% (28/43) | 63% (27/43) |
| Sensitivity | 100% (29/29) | 93% (27/29) | 90% (26/29) | 83% (24/29) | 79% (23/29) | 69% (20/29) | 66% (19/29) |
| Specificity | 14% (2/14) | 36% (5/14) | 43% (6/14) | 43% (6/14) | 43% (6/14) | 57% (8/14) | 57% (8/14) |
| PPV | 71% (29/41) | 75% (27/36) | 76% (26/34) | 75% (24/32) | 74% (23/31) | 77% (20/26) | 76% (19/25) |
| NPV | 100% (2/2) | 71% (5/7) | 67% (6/9) | 55% (6/11) | 50% (6/12) | 47% (8/17) | 44% (8/18) |
|
| |||||||
| Voxel-based D95 Dose Thresholds (Gy) | |||||||
|
| |||||||
| ≥200 | ≥300 | ≥400 | ≥500 | ≥600 | ≥700 | ≥800 | |
|
| |||||||
| Association With Response * | |||||||
| Significance | 0.01 | 0.02 | 0.03 | 0.003 | 0.03 | 0.10 | 0.19 |
| Odds Ratio for CR (95% CI) |
14.4
(1.37–757) |
5.93
(1.13–36.7) |
4.89
(1.04–25.6) |
9.05
(1.78–64.1) |
4.99
(1.02–33.9) |
3.81 (0.78–25.7) |
2.91 (0.59–19.7) |
| Accuracy | 77% (33/43) | 74% (32/43) | 72% (31/43) | 74% (32/43) | 65% (28/43) | 60% (26/43) | 56% (24/43) |
| Sensitivity | 97% (28/29) | 86% (25/29) | 79% (23/29) | 72% (21/29) | 59% (17/29) | 52% (15/29) | 45% (13/29) |
| Specificity | 36% (5/14) | 50% (7/14) | 57% (8/14) | 79% (11/14) | 79% (11/14) | 79% (11/14) | 79% (11/14) |
| PPV | 76% (28/37) | 78% (25/32) | 79% (23/29) | 88% (21/24) | 85% (17/20) | 83% (15/18) | 81% (13/16) |
| NPV | 83% (5/6) | 64% (7/11) | 57% (8/14) | 58% (11/19) | 48% (11/23) | 44% (11/25) | 41% (11/27) |
As measured by Fisher’s Exact Test. CR: Complete Response. PPV: Positive Predictive Value. NPV: Negative Predictive Value.
Figure 2.
Example dose-volume histograms for two cases where single-compartment doses were comparable with disparate voxel-based dose estimates and pathologic outcomes.
Median dose and additional dosimetry metrics are summarized in Table 3 and stratified by ALBI grade and tumor size. Differences in median dose between the CR and non-CR groups were noted when stratified by ALBI and size, though these analyses did not reach statistical significance. To control for ALBI grade and tumor size, multivariable logistic regressions were performed for the best performing single-compartment, dual-compartment tumor dose, and voxel-based D95 dose thresholds. The adjusted ORs are included in Table 4.
Table 3.
Dosimetry by tumor response and stratified by baseline ALBI Grade and by tumor size.
| Non-complete Response | Complete Response | Significance | |
|---|---|---|---|
|
| |||
| All Patients | 14 | 29 | |
| Median single-compartment dose (IQR), Gy | 463 (217 – 764) | 626 (446 – 1234) | p = 0.047 |
| Median dual-compartment tumor absorbed dose (IQR), Gy | 644 (260 – 1177) | 1146 (638 – 1831) | p = 0.035 |
| Median voxel-based D95 (IQR), Gy | 314 (156 – 462) | 727 (452 – 1136) | p = 0.009 |
| Heterogeneity Index D5/D95 (IQR), Gy | 2.62 (1.78 – 3.96) | 2.43 (1.54 – 2.83) | p = 0.17 |
| Lung Shunt Fraction (LSF) | 6.5% (5.3 – 9.0%) | 5.0% (4.0 – 6.0%) | p = 0.026 |
| Median Activity (IQR), GBq | 2.8 (2.2 – 4.0) | 2.2 (1.8 – 3.4) | p = 0.12 |
| Median Perfused Volume (IQR), mL | 330 (217 – 464) | 171 (106 – 267) | p = 0.002 |
| Treatment Week Post-Calibration* | p = 0.65 | ||
| Week 1 | 13 (93%) | 24 (86%) | |
| Week 2 | 1 (7%) | 4 (14%) | |
|
| |||
|
| |||
| ALBI Grade 2+ | 10 | 12 | |
| Median single-compartment dose (IQR), Gy | 463 (229 – 828) | 618 (405 – 1361) | p = 0.34 |
| Median dual-compartment tumor absorbed dose (IQR), Gy | 644 (299 – 1177) | 1189 (587 – 2350) | p = 0.14 |
| Median voxel-based D95 (IQR), Gy | 352 (201 – 462) | 866 (425 – 1323) | p = 0.07 |
| Heterogeneity Index D5/D95 (IQR), Gy | 2.62 (1.82 – 3.55) | 2.00 (1.49 – 2.64) | p = 0.14 |
| Lung Shunt Fraction (LSF) | 6.5% (5.3 – 9.0%) | 5.5% (4.7 – 6.0%) | p = 0.14 |
| Median Activity (IQR), GBq | 2.5 (2.1 – 3.9) | 2.0 (1.7 – 3.0) | p = 0.08 |
| Median Perfused Volume (IQR), mL | 261 (217 – 416) | 148 (104 – 273) | p = 0.04 |
| Treatment Week Post-Calibration* | p = 0.60 | ||
| Week 1 | 9 (90%) | 8 (73%) | |
| Week 2 | 1 (10%) | 3 (27%) | |
|
| |||
| ALBI Grade 1 | 4 | 17 | |
| Median single-compartment dose (IQR), Gy | 400 (213 – 590) | 626 (514 – 1024) | p = 0.12 |
| Median dual-compartment tumor absorbed dose (IQR), Gy | 541 (171 – 994) | 1146 (752 – 1371) | p = 0.12 |
| Median voxel-based D95 (IQR), Gy | 227 (71 – 529) | 695 (500 – 945) | p = 0.14 |
| Heterogeneity Index D5/D95 (IQR), Gy | 2.79 (1.52 – inf) | 2.52 (1.69 – 2.95) | p = 0.62 |
| Lung Shunt Fraction (LSF) | 7.0% (5.3 – 8.3%) | 5.0% (4.0 – 5.0%) | p = 0.20 |
| Median Activity (IQR), GBq | 3.3 (2.8 – 4.3) | 2.4 (2.1 – 3.8) | p = 0.30 |
| Median Perfused Volume (IQR), mL | 510 (381 – 661) | 179 (141 – 244) | p = 0.02 |
| Treatment Week Post-Calibration | p = 1 | ||
| Week 1 | 4 (100%) | 16 (94%) | |
| Week 2 | 0 | 1 (6%) | |
|
| |||
|
| |||
| Tumor Longest Dimension >3.0 cm | 11 | 8 | |
| Median single-compartment dose (IQR), Gy | 575 (209 – 821) | 865 (711 – 1077) | p = 0.09 |
| Median dual-compartment tumor absorbed dose (IQR), Gy | 896 (264 – 1238) | 1231 (1047 – 1837) | p = 0.15 |
| Median voxel-based D95 (IQR), Gy | 437 (177 – 750) | 738 (489 – 895) | p = 0.19 |
| Heterogeneity Index D5/D95 (IQR), Gy | 2.32 (1.77 – 4.08) | 3.01 (2.66 – 3.22) | p = 0.53 |
| Lung Shunt Fraction (LSF) | 6.0% (4.5 – 7.5%) | 5.0% (4.5 – 5.5%) | p = 0.34 |
| Median Activity (IQR), GBq | 2.7 (2.3 – 4.2) | 3.7 (3.4 – 4.0) | p = 0.65 |
| Median Perfused Volume (IQR), mL | 237 (208 – 456) | 194 (169 – 224) | p = 0.06 |
| Treatment Week Post-Calibration* | |||
| Week 1 | 11 (100%) | 8 (100%) | |
| Week 2 | 0 | 0 | |
|
| |||
| Tumor Longest Dimension ≤3.0 cm | 3 | 21 | |
| Median single-compartment dose (IQR), Gy | 270 (247 – 326) | 567 (426 – 1246) | p = 0.04 |
| Median dual-compartment tumor absorbed dose (IQR), Gy | 386 (254 – 517) | 986 (632 – 1670) | p = 0.07 |
| Median voxel-based D95 (IQR), Gy | 197 (146 – 232) | 727 (452 – 1167) | p = 0.01 |
| Heterogeneity Index D5/D95 (IQR), Gy | 3.13 (2.33 – 3.41) | 1.83 (1.46 – 2.57) | p = 0.19 |
| Lung Shunt Fraction (LSF) | 11% (10 – 11.5%) | 5% (4.0 – 6.0%) | p = 0.006 |
| Median Activity (IQR), GBq | 2.9 (2.4 – 3.1) | 2.1 (1.8 – 2.4) | p = 0.48 |
| Median Perfused Volume (IQR), mL | 375 (330 – 478) | 156 (100 – 267) | p = 0.04 |
| Treatment Week Post-Calibration* | p = 0.60 | ||
| Week 1 | 2 (67%) | 16 (80%) | |
| Week 2 | 1 (33%) | 4 (20%) | |
Week of treatment post-calibration was not able to be determined for one case which was excluded from this analysis. LSF: Lung shunt fraction was estimated based on the pre-TARE mapping session 99mTc-MAA (macro-aggregated albumin) distribution on SPECT/CT. IQR: Interquartile range. ALBI Grade: Albumin-Bilirubin Grade.
Table 4.
Multivariable regression analysis of selected dose thresholds.
| Multivariable Logistic Regression | ||
|---|---|---|
| Adjusted OR (95% CI) | Significance | |
|
| ||
| Single-compartment Model | ||
| Perfused Volume Dose, Gy | p = 0.031 | |
| <400 Gy | Ref. | |
| ≥400 Gy | 12.9 (1.70 – 273) | |
| Tumor Size, mm | p = 0.003 | |
| ≤30 mm | Ref. | |
| >30 mm | 0.02 (0.001 – 0.19) | |
| Pre-TARE ALBI Grade | p = 0.047 | |
| Grade 1 | Ref. | |
| Grade 2+ | 0.13 (0.01 – 0.83) | |
| Dual-compartment Model | ||
| Tumor Absorbed Dose, Gy | p = 0.025 | |
| <400 Gy | Ref. | |
| ≥400 Gy | 19.1 (1.98 – 491) | |
| Tumor Size, mm | p = 0.004 | |
| ≤30 mm | Ref. | |
| >30 mm | 0.03 (0.001 – 0.21) | |
| Pre-TARE ALBI Grade | p = 0.046 | |
| Grade 1 | Ref. | |
| Grade 2+ | 0.14 (0.02 – 0.83) | |
| Voxel-based Dosimetry | ||
| D95 Dose, Gy | p = 0.013 | |
| <500 Gy | Ref. | |
| ≥500 Gy | 12.4 (2.02 – 128) | |
| Tumor Size, mm | p = 0.005 | |
| ≤30 mm | Ref. | |
| >30 mm | 0.04 (0.003 – 0.29) | |
| Pre-TARE ALBI Grade | p = 0.069 | |
| Grade 1 | Ref. | |
| Grade 2+ | 0.14 (0.01 – 0.97) | |
ALBI Grade: Albumin-Bilirubin Grade.
One patient was excluded from the safety analysis as no clinical interaction post-TARE was obtained prior to transplantation (20 days after TARE). The most reported AEs (any severity) were fatigue (20/42, 48%), abdominal pain (14/42, 33%), and nausea (14/42, 33%). Overall, 5 (12%) patients had no AEs documented, 7 (17%) had one AE, and 30 (71%) had two or more AEs. Thirteen patients (31%) had elevated alanine aminotransferase (ALT) and/or aspartate aminotransferase (AST) and twelve patients (29%) had elevated total bilirubin within 90 days post-TARE, two of these patients also had new-onset ascites that improved with diuretics. Four patients (10%) had AE’s rising to moderate or severe classification: one case of sepsis requiring hospitalization, two cases of abdominal pain requiring hospitalization and repeated paracenteses, and one case of suspected radiation colitis/peritonitis adjacent to the treatment zone. No cases of REILD, upper gastrointestinal ulceration, or leukopenia were reported.
The incidence of AEs did not vary significantly based on the ≥400 Gy threshold (single-compartment, dual-compartment, or D95): 26/31 (84%) versus 11/11 (100%) (p = 0.30); 29/34 (85%) versus 8/8 (100%) p = 0.56); and 25/29 (86%) versus 12/13 (92%) (p = 1), respectively. The incidence also did not vary based on D95 ≥500 or ≥600 Gy thresholds: 21/24 (88%) versus 16/18 (89%) (p = 1) and 18/20 (90%) versus 19/22 (86%) (p = 1), respectively.
Discussion
This study confirms that ≥400 Gy using the standard single-compartment model was associated with CR, though escalation beyond 400 Gy may not increase rates of pathologic response. The dual-compartment model had similar PPV to the single-compartment model, though again higher dose thresholds beyond ≥400 Gy did not lead to associations with pathologic response. For voxel-based dosimetry, the D95 ≥400, ≥500, and ≥600 Gy were all associated with pathologic response, with the highest PPV observed at 88% for D95 ≥500 Gy. This suggests that, in contrast to the plateauing performance of the single- and dual-compartment models, voxel-based dosimetry may allow for personalized dose escalation strategies that provide greater predictive value than current approaches [13]. D95 represents a balance between mis-registration errors, which would constrain practical usage of D100, while ensuring that nearly the entire tumor volume is evaluated. This is in distinction to both the single- and dual-compartment tumor dose models and may explain the prognostic advantages of D95 reported here.
The observed differences in dose between the CR and non-CR groups remained present when stratified by tumor size and ALBI grade, though these subset analyses are constrained by loss of statistical power. The highest performing dose thresholds for each dosimetry model (single-compartment ≥400 Gy, dual-compartment mean tumor dose of ≥400 Gy, and D95 ≥500 Gy) remained associated with pathologic response in multivariable logistic regression analyses accounting for tumor size and ALBI grade. A median D95 dose of 730 Gy in the CR group was consistent for tumors both larger and smaller than 3 cm. The majority (88%, 21/24) of tumors ≤3 cm in this sample met criteria for complete response. Considering the workflow burdens created by voxel-based dosimetry and difficulty with registration of smaller tumors, it is reasonable to pursue the single-compartment dosimetry strategy in smaller tumors. However, in larger tumors, the D95 thresholds may offer prognostic advantages which would support incorporating voxel-based dosimetry modeling into workflows for these patients.
A potential mechanism by which ALBI impacts response is that patients with advanced liver dysfunction may have increased portal hypertension versus those with relatively preserved liver function. When portal flow is restricted, the liver parenchyma receives compensatory perfusion via the hepatic arterial system. The preferential delivery of microspheres to tumor relies on the assumption that normal liver parenchyma derives most of its blood supply from the portal system whereas tumors tend to rely on arterial flow. It is conceivable that, in patients where the normal liver parenchyma is also dependent on the hepatic arterial system, the distribution of the delivered microspheres may be sub-optimal.
The incidence of AEs did not differ at any of the tested dose thresholds and moderate-severe AEs were rare, with only four occurrences. The median time to transplantation was relatively short in this study at <1 year. The choice of LRT pre-transplantation reflects complex clinical decision making including but not limited to tumor size and location. Given timing of organ availability is not predictable, the results of this study are congruent with current practice to deploy TARE as a bridging modality, however, the choice between LRT modalities was not directly evaluable in a retrospective study.
In current practice, glass microspheres are typically dosed based on a single-compartment dose of ≥400 Gy or a dual-compartment tumor absorbed dose ≥205 Gy. The single-compartment dose threshold was adopted following the LEGACY study (median tumor size: 2.7 cm) and a related dose-pathology analysis of patients that underwent liver transplantation. In this subset, 100% of explanted tumors dosed ≥400 Gy achieved CPN [9]. Toskich et al. reported a 75% CPN rate in patients who received a single-compartment dose >500 Gy, with a median tumor size in that study of 2.3 cm, while Montazeri et al. suggested ≥446 Gy was predictive of CPN [11,22]. In the RASER study (tumors ≤3 cm), for the eight patients with confirmed pathologic necrosis the single-compartment dose was estimated at 776.5 Gy with a mean tumor dose of 1462.6 Gy [23]. In the DOSISPHERE-01 trial, superior outcomes were reported in tumors ≥7 cm treated with >205 Gy tumor absorbed dose [12,24]. The TARGET study for tumors ≥3 cm (~80% were ≥5 cm) also reported a correlation between tumor absorbed dose and imaging response (225.5 Gy versus 188.3 Gy in non-responders) and longer overall survival with increasing tumor dose [25].
Limitations of this study include use of single-institution, retrospective data. Additional dosimetry considerations beyond dose alone such as microsphere concentration, specific activity, and volume fraction are believed to play an important role in tumor response to TARE. Given the institutional preference for week 1 dosing observed in this dataset, these questions were not well-addressed by this analysis. The safety outcomes analysis is limited by the reliance solely on retrospective chart review.
Conclusion
The results of this single-institution study suggest that voxel-based dosimetry standards may offer greater prognostic value for complete pathologic response to radiation segmentectomy compared to single-compartment and dual-compartment dose estimations.
References
- [1].Fujiki M, Aucejo F, Kim R. General overview of neo-adjuvant therapy for hepatocellular carcinoma before liver transplantation: necessity or option? Liver International 2011;31:1081–9. 10.1111/j.1478-3231.2011.02473.x. [DOI] [PubMed] [Google Scholar]
- [2].Galle PR, Forner A, Llovet JM, Mazzaferro V, Piscaglia F, Raoul J-L, et al. EASL Clinical Practice Guidelines: Management of hepatocellular carcinoma. Journal of Hepatology 2018;69:182–236. 10.1016/j.jhep.2018.03.019. [DOI] [PubMed] [Google Scholar]
- [3].DiNorcia J, Florman SS, Haydel B, Tabrizian P, Ruiz RM, Klintmalm GB, et al. Pathologic Response to Pretransplant Locoregional Therapy is Predictive of Patient Outcome After Liver Transplantation for Hepatocellular Carcinoma: Analysis From the US Multicenter HCC Transplant Consortium. Annals of Surgery 2020;271:616. 10.1097/SLA.0000000000003253. [DOI] [PubMed] [Google Scholar]
- [4].Ravaioli M, Grazi GL, Ercolani G, Fiorentino M, Cescon M, Golfieri R, et al. Partial Necrosis on Hepatocellular Carcinoma Nodules Facilitates Tumor Recurrence after Liver Transplantation. Transplantation 2004;78:1780. 10.1097/01.TP.0000145892.97114.EE. [DOI] [PubMed] [Google Scholar]
- [5].Gabr A, Kulik L, Mouli S, Riaz A, Ali R, Desai K, et al. Liver Transplantation Following Yttrium-90 Radioembolization: 15-Year Experience in 207-Patient Cohort. Hepatology 2021;73:998. 10.1002/hep.31318. [DOI] [PubMed] [Google Scholar]
- [6].Gabr A, Abouchaleh N, Ali R, Baker T, Caicedo J, Katariya N, et al. Outcomes of Surgical Resection after Radioembolization for Hepatocellular Carcinoma. Journal of Vascular and Interventional Radiology 2018;29:1502–1510.e1. 10.1016/j.jvir.2018.06.027. [DOI] [PubMed] [Google Scholar]
- [7].Salem R, Johnson GE, Kim E, Riaz A, Bishay V, Boucher E, et al. Yttrium-90 Radioembolization for the Treatment of Solitary, Unresectable HCC: The LEGACY Study. Hepatology 2021;74:2342–52. 10.1002/hep.31819. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [8].Riaz A, Gates VL, Atassi B, Lewandowski RJ, Mulcahy MF, Ryu RK, et al. Radiation Segmentectomy: A Novel Approach to Increase Safety and Efficacy of Radioembolization. International Journal of Radiation Oncology*Biology*Physics 2011;79:163–71. 10.1016/j.ijrobp.2009.10.062. [DOI] [PubMed] [Google Scholar]
- [9].Gabr A, Riaz A, Johnson GE, Kim E, Padia S, Lewandowski RJ, et al. Correlation of Y90-absorbed radiation dose to pathological necrosis in hepatocellular carcinoma: confirmatory multicenter analysis in 45 explants. Eur J Nucl Med Mol Imaging 2021;48:580–3. 10.1007/s00259-020-04976-8. [DOI] [PubMed] [Google Scholar]
- [10].Vouche M, Habib A, Ward TJ, Kim E, Kulik L, Ganger D, et al. Unresectable solitary hepatocellular carcinoma not amenable to radiofrequency ablation: Multicenter radiology-pathology correlation and survival of radiation segmentectomy. Hepatology 2014;60:192–201. 10.1002/hep.27057. [DOI] [PubMed] [Google Scholar]
- [11].Toskich B, Vidal LL, Olson MT, Lewis JT, LeGout JD, Sella DM, et al. Pathologic Response of Hepatocellular Carcinoma Treated with Yttrium-90 Glass Microsphere Radiation Segmentectomy Prior to Liver Transplantation: A Validation Study. Journal of Vascular and Interventional Radiology 2021;32:518–526.e1. 10.1016/j.jvir.2020.12.019. [DOI] [PubMed] [Google Scholar]
- [12].Garin E, Tselikas L, Guiu B, Chalaye J, Edeline J, de Baere T, et al. Personalised versus standard dosimetry approach of selective internal radiation therapy in patients with locally advanced hepatocellular carcinoma (DOSISPHERE-01): a randomised, multicentre, open-label phase 2 trial. The Lancet Gastroenterology & Hepatology 2021;6:17–29. 10.1016/S2468-1253(20)30290-9. [DOI] [PubMed] [Google Scholar]
- [13].Pianka KT, Barahman M, Minocha J, Redmond JW, Schnickel GT, Rose SC, et al. Voxel-based tumor dose correlates to complete pathologic necrosis after transarterial radioembolization for hepatocellular carcinoma. Eur J Nucl Med Mol Imaging 2024. 10.1007/s00259-024-06813-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [14].Reig M, Forner A, Rimola J, Ferrer-Fábrega J, Burrel M, Garcia-Criado Á, et al. BCLC strategy for prognosis prediction and treatment recommendation: The 2022 update. J Hepatol 2022;76:681–93. 10.1016/j.jhep.2021.11.018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [15].Kamath PS, Wiesner RH, Malinchoc M, Kremers W, Therneau TM, Kosberg CL, et al. A model to predict survival in patients with end-stage liver disease. Hepatology 2001;33:464–70. 10.1053/jhep.2001.22172. [DOI] [PubMed] [Google Scholar]
- [16].Biggins SW, Kim WR, Terrault NA, Saab S, Balan V, Schiano T, et al. Evidence-based incorporation of serum sodium concentration into MELD. Gastroenterology 2006;130:1652–60. 10.1053/j.gastro.2006.02.010. [DOI] [PubMed] [Google Scholar]
- [17].Johnson PJ, Berhane S, Kagebayashi C, Satomura S, Teng M, Reeves HL, et al. Assessment of liver function in patients with hepatocellular carcinoma: a new evidence-based approach-the ALBI grade. J Clin Oncol 2015;33:550–8. 10.1200/JCO.2014.57.9151. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [18].Salem R, Lewandowski RJ, Gates VL, Nutting CW, Murthy R, Rose SC, et al. Research Reporting Standards for Radioembolization of Hepatic Malignancies. Journal of Vascular and Interventional Radiology 2011;22:265–78. 10.1016/j.jvir.2010.10.029. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [19].Chernyak V, Fowler KJ, Kamaya A, Kielar AZ, Elsayes KM, Bashir MR, et al. Liver Imaging Reporting and Data System (LI-RADS) Version 2018: Imaging of Hepatocellular Carcinoma in At-Risk Patients. Radiology 2018;289:816–30. 10.1148/radiol.2018181494. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [20].Khalilzadeh O, Baerlocher MO, Shyn PB, Connolly BL, Devane AM, Morris CS, et al. Proposal of a New Adverse Event Classification by the Society of Interventional Radiology Standards of Practice Committee. Journal of Vascular and Interventional Radiology 2017;28:1432–1437.e3. 10.1016/j.jvir.2017.06.019. [DOI] [PubMed] [Google Scholar]
- [21].Baerlocher MO, Nikolic B, Sze DY. Adverse Event Classification: Clarification and Validation of the Society of Interventional Radiology Specialty–Specific System. Journal of Vascular and Interventional Radiology 2023;34:1–3. 10.1016/j.jvir.2022.10.011. [DOI] [PubMed] [Google Scholar]
- [22].Montazeri SA, De la Garza-Ramos C, Lewis AR, Lewis JT, LeGout JD, Sella DM, et al. Hepatocellular carcinoma radiation segmentectomy treatment intensification prior to liver transplantation increases rates of complete pathologic necrosis: an explant analysis of 75 tumors. Eur J Nucl Med Mol Imaging 2022;49:3892–7. 10.1007/s00259-022-05776-y. [DOI] [PubMed] [Google Scholar]
- [23].Kim E, Sher A, Abboud G, Schwartz M, Facciuto M, Tabrizian P, et al. Radiation segmentectomy for curative intent of unresectable very early to early stage hepatocellular carcinoma (RASER): a single-centre, single-arm study. The Lancet Gastroenterology & Hepatology 2022;7:843–50. 10.1016/S2468-1253(22)00091-7. [DOI] [PubMed] [Google Scholar]
- [24].Garin E, Tselikas L, Guiu B, Chalaye J, Rolland Y, de Baere T, et al. Long-Term Overall Survival After Selective Internal Radiation Therapy for Locally Advanced Hepatocellular Carcinomas: Updated Analysis of DOSISPHERE-01 Trial. J Nucl Med 2024;65:264–9. 10.2967/jnumed.123.266211. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [25].Lam M, Garin E, Maccauro M, Kappadath SC, Sze DY, Turkmen C, et al. A global evaluation of advanced dosimetry in transarterial radioembolization of hepatocellular carcinoma with Yttrium-90: the TARGET study. Eur J Nucl Med Mol Imaging 2022;49:3340–52. 10.1007/s00259-022-05774-0. [DOI] [PMC free article] [PubMed] [Google Scholar]