Pulmonary arteriovenous malformations (PAVMs) can persist after embolization 25-49% of the time. The reason for treatment failure is not well understood. This study examines a cohort of recanalized (flow through previous embolic treatment) or reperfused (flow through a recruited feeding vessel) PAVMs, including primary treatment characteristics, retreatment techniques and outcomes of repeat embolization. This study retrospectively reviewed 19 patients who underwent 24 consecutive PAVM embolization procedures for persistent PAVMs with follow-up cross-sectional imaging at a single center between August 2014 and January 2022. Characteristics of the PAVMs were collected including simple vs complex, saccular vs non-saccular, primary embolic material, embolization of the nidus, type of persistence, retreatment data and reperfusion/recanalization on follow-up imaging. The success of repeat treatment was assessed on follow-up imaging. Chi-square test was used to compare groups. Median patient age was 59 (range 9-71) and 42% were male. Seven patients with 10/24 (42%) PAVMs were primarily treated at an outside institution. Most PAVMs were simple (71%), saccular (71%) and primarily treated with coils (88%), rather than plug (4%) or coils and plug (8%). The nidus was primarily treated in only 4/24 (17%). Primary treatment recanalization was seen in 15/24 (62%), reperfusion in 7/24 (29%) and both in 2/24 (8%). Retreatment was performed with non-fibered coils (42%), fibered coils (29%), AVP and coils (21%) or Onyx (8%). Retreatment technical failure was 4/24 (7%) and an additional 3/24 (12%) reperfused on follow-up. Thus, 7/24 (29%) PAVMs had persistent flow on follow-up imaging with median follow-up of 31 months. Retreatment embolization into the nidus through the existing embolic or via a recruited vessel was achieved in 11/24 (46%). None of these demonstrated persistence on follow-up; however, there was persistence when the nidus was not treated (7/13, 54%), (p=0.005). No correlation was seen between PAVM persistence and prior recanalization (p=0.68), prior reperfusion (p=0.54), simple (p=0.68) or saccular PAVM (p=0.68), retreatment with non-fibered coils (p=0.26), fibered coils (p=0.36) or AVP (p=0.54). No major procedural complications occurred. Retreatment of persistent PAVMs can achieve a high rate (83%) of technical success. Embolization of the nidus either through the existing embolization material or via a recruited pulmonary vessel with coils, plugs or liquid embolic can achieve high long-term occlusion.

Hypovascular tumors are particularly difficult to treat effectively with trans-arterial therapies due to low tumor-to-normal liver (T/N) tissue perfusion. This study aims to investigate T/N ratio of patients receiving yttrium-90 (Y90) radioembolization glass microspheres for tumors that are hypovascular on pre-treatment CT or MR. Institutional review board approved this retrospective study of patients who underwent radioembolization with 90Y glass microspheres at a single institution for hypovascular liver tumors between 2022 and 2023. Patients were included if tumors were hypoattenuating relative to normal liver on properly timed arterial phase CT or MRI. Patients were excluded if tumors were necrotic. Immediately following planning angiography with 99mTc-MAA administration, all patients received SPECT/CT scans. Multi-compartment voxel-based dosimetry was performed using Y90 dosimetry software (Simplicit90Y, Boston Scientific Corporation) and T/N ratios were calculated from dosimetry data. The final cohort included 23 patients (10 female, 13 male, median age 69, range 52-85 years) with hypovascular liver tumors that underwent Y90 glass microsphere treatments. Ten patients had cirrhosis of varying etiologies, including HCV (n=4), alcohol (n=5), and non-alcoholic fatty liver disease (n=1). Three of the patients had previously received Y90. Patients had either hepatocellular carcinoma (n=9), intrahepatic cholangiocarcinoma (n=6), or liver metastasis (n=8). All patients were Child-Pugh A, with ECOG status 0 (n=11), 1 (n=11), or 2 (n=1). Treatment strategies included lobar (5/23, 22%), lobar plus segment (3/23, 13%), multiple radiation segmentectomy (7/23, 30%), and single radiation segmentectomy (8/23, 35%). The mean index tumor size was 5.2 ± 3.2 cm. The mean lung shunt fraction was 5.5 ± 3.3% for the patient cohort. Prescribed dose to the perfused volume ranged from 120 to 625 Gy with a mean of 286.4 ± 154.0 Gy. The mean tumor absorbed dose was 425.2 ± 282.5 Gy, and the mean normal liver absorbed dose was 85.4 ± 43.0 Gy. T/N ratios ranged from 0.28 to 4.02, with a mean of 1.61 ± 0.83. T/N ratios calculated using multi-compartment dosimetry were on average greater than 1.00 for the patient cohort analyzed, indicating that hypovascular tumor uptake of Y90 dose may be greater than previously assumed and patients with such tumor morphologies could benefit from proceeding with treatment.

Few studies have established the safety and toxicity of repeated radioembolization in the same arterial territory. Cumulative radiation exposure can lead to liver toxicity, particularly in patients with limited functional hepatic reserve. The risk of complications such as vascular injury and liver fibrosis has been suggested to be higher with retreatment. We aim to describe efficacy and safety in a cohort of patients that received multiple radioembolization procedures to the same vascular territory.

Pulmonary arteriovenous malformations (PAVMs) are abnormal direct connections between pulmonary arteries and veins. Embolization is the preferred treatment for PAVMs; however, there is ongoing debate regarding the best approach to ensure long-term occlusion. This study investigates which techniques or characteristics correlate with treatment failure. Review of PAVM embolization procedures between September 2014 and December 2022 at a hereditary hemorrhagic telangiectasia (HHT) center of excellence was performed. PAVMs undergoing initial treatment with follow up cross-sectional imaging were included. Data was collected for PAVM angioarchitecture, embolization of the nidus, embolic material, technical success, and recanalization/reperfusion on follow-up cross sectional imaging. Categorical variables were compared using chi-square test and continuous variables using independent samples t-test. The final cohort included 93 PAVMs in 57 patients (median age 47, 65% female). The majority of patients had HHT (49/57, 86%). Most PAVMs were simple (77/93, 82%) and saccular (68/93, 73%). The nidus was treated in 63/93 (68%), and more often with simple PAVMs (56/77, 73%) than complex PAVMs (7/16, 44%), which was statistically significant (P=0.03). Treatment was performed with coils (64/93, 69%), coils and plugs (23/93, 25%), plugs alone (5/93, 5%), or Onyx (1/93, 1%). Complex PAVMs were more likely to be treated with a plug (10/16, 62%) than simple PAVMS (18/77, 23%), which was statistically significant (P< 0.005). Treatment technical success was achieved in 85/93 (91%) of PAVMs after the index procedure, and 90/93 (97%) after all subsequent procedures, including 3 that developed delayed thrombosis on follow-up. Recanalization occurred in 2/93 (2%) after 3 and 35 months, and reperfusion occurred in 1/93 (1%) after 2.5 months. Mean follow-up was 30 ± 26 months. Of the 11 that had primary technical failure or recanalization/reperfusion, statistically significant associations were seen with complex angioarchitecture (5/11, 45%, P=0.02) and fibered coils (7/11, 63%, P=0.05). Also in this subgroup, mean feeding artery diameter and sac diameter were bigger (4.5 vs. 3.9 mm and 16 vs. 13.3 mm, respectively), but not statistically significant. No major procedural or post-procedural complications occurred. Embolization of simple and complex PAVMs can result in high rates of technical success. Complex PAVMs may be more difficult to treat the entire nidus, and may have higher rates of technical failure or reperfusion/recanalization.

The safety and efficacy of liver resection after Y-90 radioembolization remains controversial. This study aims to investigate the clinical, surgical and pathological outcomes of patients who underwent liver resection after Y-90 radioembolization for the treatment of hepatocellular carcinoma (HCC), intrahepatic cholangiocarcinoma (ICC) or liver metastases in a contemporary clinical practice. Single-institution review was performed of all patients who underwent hepatic resection or liver transplant after Y-90 radioembolization between January 2019 and July 2023. Patient charts and imaging were reviewed to collect baseline demographics, disease and treatment data, and pathological explant data. The primary endpoints were surgical resection margins, pathological necrosis, and overall survival. Twenty-three patients (7 women, 16 men, mean age 62 ± 10 years) met inclusion for the study. Only 14/23 (61%) patients had underlying liver disease due to one or more of: alcohol (n=2), hepatitis C (n=5), hepatitis B (n=1), and non-alcoholic fatty liver disease (n=6). Tumors included HCC (n=15), ICC (n=6), and metastatic disease (n=2). ECOG performance status was 0 (n=21) or 1 (n=3). Treated tumors (average size 6.5 ± 4.3 cm) were in the left (n = 4) or right hepatic lobe (n = 20). Y-90 treatment was lobar (n=9), segmental (n=2), or lobar with segmental boost (n=12). The average lobar dose administered was 195 ± 77 Gy and the average segmental dose was 195 ± 94 Gy. Average time to resection post Y-90 treatment was 145 ± 98 days). All patients underwent successful transplant (n=9), right (n=7), extended right (n=6), or extended left (n=1) hepatectomies. Only one patient had a positive resection margin near the hepatic vein margin, otherwise all others were R0 resections. 11/23 (48%) patients had 95-100% necrosis, 13/23 (56%) patients had at least 90% necrosis on explant pathology. No mortality occurred after transplant. Mortality after surgical resection at 30-days was 1/14 (7%), and 90-days was 3/14 (21%). Overall survival after surgical resection or transplant was 4/23 (17%) over a cumulative follow-up time of 37 patient years. No major Y-90 procedure complications or biochemical toxicities occurred. Hepatic resection and transplant after Y-90 can be safe and effective with high long-term overall survival. This was seen in a population of large tumors (average size 6.5 ±4.3 cm) and high lobar Y-90 dose (average 195 Gy), often with a boost dose to the tumor (average 195 Gy). Substantial tumor necrosis rates were seen with this Y-90 dosing.

To examine recent trends in national utilization and outcomes of transjugular intrahepatic portosystemic shunt (TIPS) creation. The National Inpatient Sample, the largest capture of hospitalizations in the United States, was queried across the most recently available study years 2012 to 2018. TIPS creation and primary diagnoses were identified using International Classification of Diseases (versions 9 and 10) codes. Patient sociodemographics, hospital characteristics, and mortality were extracted from the database. Utilization over time and by indication was assessed using linear regression. Categorical variables were examined using chi-square analysis. TIPS creation did not change in annual volume: 30845 overall, 4805 in 2012 vs. 5080 in 2018. Fewer Black patients underwent TIPS over time (280, 6.2% in 2012 vs. 170, 3.4% in 2018; P< 0.05). Most TIPS were created at urban teaching hospitals, the relative proportion of which increased from 81% in 2012 to 90% in 2018 (P< 0.001). Ascites was the most common indication (18115, 59% overall), the proportional prevalence of which increased over the study period (53% in 2012 vs. 63% in 2018). Variceal bleeding was the second most common indication (9030, 29% overall) and its prevalence did not change over time (31% in 2012 vs. 32% in 2018). However, a greater proportion underwent TIPS in planned fashion for non-bleeding varices (20% in 2012 vs. 27% in 2018; P< 0.001). TIPS for portal vein thrombosis also slightly increased over time (480, 10% in 2012 vs. 685, 13% in 2018; P< 0.001). Inpatient mortality did not change over time (9.6% in 2012 vs. 7.0% in 2018; P = 0.50), was highest for variceal bleeding (15%) and lowest for the elective non-bleeding subset (3.3%). Annual volume of TIPS creation remains stable, with ascites and variceal bleeding being the most common indications. A greater proportion of TIPS for variceal bleeding are occurring in planned fashion for non-bleeding varices, for which hospital mortality is significantly lower than those with bleeding varices. Urban teaching hospitals provide the vast majority of TIPS creation, the use of which is seen with lowering frequency at urban non-teaching and rural hospitals over time.

To investigate outcomes of patients who received modified radiation lobectomy for hepatocellular carcinoma (HCC) and intrahepatic cholangiocarcinoma (ICC). Institutional review board approved this retrospective study of patients who underwent modified radiation lobectomy with yttrium-90 (90Y) glass microspheres for HCC or ICC with intent to grow the future liver remanent (FLR) between 2019-2022. Outcome measures included tumor response according to modified Response Evaluation Criteria in Solid Tumors (mRECIST), FLR kinetics, tumor progression, and surgical resection or liver transplantation. Adverse events were assessed according to the Society of Interventional Radiology classification. Descriptive statistics were used to assess data. Fifteen patients (7 women, 8 men; mean age 66 ± 11 years) with HCC (n = 7) and ICC (n = 8) underwent modified radiation lobectomy of left (n = 1) or right (n = 14) sided tumors. Mean index tumor size was 7.8 ± 3.7 cm (range, 2.4-16.2). HCC patients were intermediate stage BCLC B (n = 6), or advanced stage BCLC C (n = 1) due to portal vein tumor invasion. Six patients had cirrhosis. Neoadjuvant chemotherapy was concurrently given to 4/8 (50%) of the ICC patients. The average lobar dose was 198 ± 58 Gy and the average segmental boost dose was 186 ± 81 Gy, with an average of 1.6 segmental doses per patient. No severe adverse event from 90Y treatment occurred. Mean bilirubin was unchanged at 30-days. All patients had objective response at 30 days, including 6/15 (40%) with complete response and 9/15 (60%) with partial response. FLR increased by an average of 16.8% ± 14.6% at 30 days (n = 15), and 42.1% ± 22.8% at 90 days (n = 9). A total of 12/15 (80%) were eligible for surgical resection after treatment. One patient refused surgery and another developed concurrent breast cancer, thus 10/15 (67%) went on to hepatectomy (n = 9) or transplant (n = 1) with negative margins at a mean 122 ± 77 days post-treatment. Surgical resection included right (n = 4), extended right (n = 4), or extended left (n = 1) hepatectomies. The remaining 3 patients were ineligible for resection due to new disease progression in the contralateral liver lobe (n = 2) or to lungs (n = 1) at 1-, 2- and 3-months post-treatment. This study demonstrates that modified radiation lobectomy is safe and effective for treatment of HCC and ICC, with a high number of patients going on to definitive surgical management. This is an effective method to attain liver remnant hypertrophy while maintaining control over tumor progression.