Liver Transplant Costs and Activity After United Network for Organ Sharing Allocation Policy Changes.

Importance: A new liver allocation policy was implemented by United Network for Organ Sharing (UNOS) in February 2020 with the stated intent of improving access to liver transplant (LT). There are growing concerns nationally regarding the implications this new system may have on LT costs, as well as access to a chance for LT, which have not been captured at a multicenter level. Objective: To characterize LT volume and cost changes across the US and within specific center groups and demographics after the policy implementation. Design, Setting, and Participants: This cross-sectional study collected and reviewed LT volume from multiple centers across the US and cost data with attention to 8 specific center demographics. Two separate 12-month eras were compared, before and after the new UNOS allocation policy: March 4, 2019, to March 4, 2020, and March 5, 2020, to March 5, 2021. Data analysis was performed from May to December 2022. Main Outcomes and Measures: Center volume, changes in cost. Results: A total of 22 of 68 centers responded comparing 1948 LTs before the policy change and 1837 LTs postpolicy, resulting in a 6% volume decrease. Transplants using local donations after brain death decreased 54% (P < .001) while imported donations after brain death increased 133% (P = .003). Imported fly-outs and dry runs increased 163% (median, 19; range, 1-75, vs 50, range, 2-91; P = .009) and 33% (median, 3; range, 0-16, vs 7, range, 0-24; P = .02). Overall hospital costs increased 10.9% to a total of $46 360 176 (P = .94) for participating centers. There was a 77% fly-out cost increase postpolicy ($10 600 234; P = .03). On subanalysis, centers with decreased LT volume postpolicy observed higher overall hospital costs ($41 720 365; P = .048), and specifically, a 122% cost increase for liver imports ($6 508 480; P = .002). Transplant centers from low-income states showed a significant increase in hospital (12%) and import (94%) costs. Centers serving populations with larger proportions of racial and ethnic minority candidates and specifically Black candidates significantly increased costs by more than 90% for imported livers, fly-outs, and dry runs despite lower LT volume. Similarly, costs increased significantly (>100%) for fly-outs and dry runs in centers from worse-performing health systems. Conclusions and Relevance: Based on this large multicenter effort and contrary to current assumptions, the new liver distribution system appears to place a disproportionate burden on populations of the current LT community who already experience disparities in health care. The continuous allocation policies being promoted by UNOS could make the situation even worse.

Living Donor Liver Transplantation for Hepatocellular Carcinoma Within and Outside Traditional Selection Criteria: A Multicentric North American Experience.

OBJECTIVE: To evaluate long-term oncologic outcomes of patients post-living donor liver transplantation (LDLT) within and outside standard transplantation selection criteria and the added value of the incorporation of the New York-California (NYCA) score. BACKGROUND: LDLT offers an opportunity to decrease the liver transplantation waitlist, reduce waitlist mortality, and expand selection criteria for patients with hepatocellular carcinoma (HCC). METHODS: Primary adult LDLT recipients between October 1999 and August 2019 were identified from a multicenter cohort of 12 North American centers. Posttransplantation and recurrence-free survival were evaluated using the Kaplan-Meier method. RESULTS: Three hundred sixty LDLTs were identified. Patients within Milan criteria (MC) at transplantation had a 1, 5, and 10-year posttransplantation survival of 90.9%, 78.5%, and 64.1% versus outside MC 90.4%, 68.6%, and 57.7% ( P = 0.20), respectively. For patients within the University of California San Francisco (UCSF) criteria, respective posttransplantation survival was 90.6%, 77.8%, and 65.0%, versus outside UCSF 92.1%, 63.8%, and 45.8% ( P = 0.08). Fifty-three (83%) patients classified as outside MC at transplantation would have been classified as either low or acceptable risk with the NYCA score. These patients had a 5-year overall survival of 72.2%. Similarly, 28(80%) patients classified as outside UCSF at transplantation would have been classified as a low or acceptable risk with a 5-year overall survival of 65.3%. CONCLUSIONS: Long-term survival is excellent for patients with HCC undergoing LDLT within and outside selection criteria, exceeding the minimum recommended 5-year rate of 60% proposed by consensus guidelines. The NYCA categorization offers insight into identifying a substantial proportion of patients with HCC outside the MC and the UCSF criteria who still achieve similar post-LDLT outcomes as patients within the criteria.

Predicting tumor recurrence on baseline MR imaging in patients with early-stage hepatocellular carcinoma using deep machine learning.

Tumor recurrence affects up to 70% of early-stage hepatocellular carcinoma (HCC) patients, depending on treatment option. Deep learning algorithms allow in-depth exploration of imaging data to discover imaging features that may be predictive of recurrence. This study explored the use of convolutional neural networks (CNN) to predict HCC recurrence in patients with early-stage HCC from pre-treatment magnetic resonance (MR) images. This retrospective study included 120 patients with early-stage HCC. Pre-treatment MR images were fed into a machine learning pipeline (VGG16 and XGBoost) to predict recurrence within six different time frames (range 1-6 years). Model performance was evaluated with the area under the receiver operating characteristic curves (AUC-ROC). After prediction, the model's clinical relevance was evaluated using Kaplan-Meier analysis with recurrence-free survival (RFS) as the endpoint. Of 120 patients, 44 had disease recurrence after therapy. Six different models performed with AUC values between 0.71 to 0.85. In Kaplan-Meier analysis, five of six models obtained statistical significance when predicting RFS (log-rank p < 0.05). Our proof-of-concept study indicates that deep learning algorithms can be utilized to predict early-stage HCC recurrence. Successful identification of high-risk recurrence candidates may help optimize follow-up imaging and improve long-term outcomes post-treatment.

A narrative review of the current and future role of robotic surgery in liver surgery and transplantation.

Background: Minimally invasive surgery (MIS) is the technique of choice in selected patients for the treatment of liver tumors. The robotic approach is considered today the natural evolution of MIS. The application of the robotic technique in liver transplantation (LT) has been recently evaluated, especially in the living donation. The aim of this paper is to review the current role of the MIS and robotic donor hepatectomy in the literature and to evaluate the possible future implication in the transplant field. Methods: We conducted a narrative review using PubMed and Google Scholar for reports published so far, using the following keywords: minimally invasive liver surgery, laparoscopic liver surgery, robotic liver surgery, robotic living donation, laparoscopic donor hepatectomy and robotic donor hepatectomy. Results: Several advantages have been claimed in favor of robotic surgery: three-dimensional (3-D) imaging with stable and high-definition view; a more rapid learning curve than the laparoscopic one; the lack of hand tremors and the freedom of movements. Compared to open surgery, the benefits showed in the studies evaluating the robotic approach in the living donation are: less postoperative pain, the shorter period before returning to normal activity despite sustaining longer operation time. Furthermore, the 3-D and magnification view makes the technique excellent in distinguishing the right plane of transection, vascular and biliary anatomy, associated with high precision of the movements and a better bleeding control (essential for donor safety) and lower rate of vascular injury. Conclusions: The current literature does not fully support the superiority of the robotic approach versus laparoscopic or open method in living donor hepatectomy. Robotic donor hepatectomy performed by teams with high expertise and in properly selected living donors is safe and feasible. However, further data are necessary to evaluate properly the role of robotic surgery in the field of living donation.

Living donor liver transplant candidate and donor selection and engagement: Meeting report from the living donor liver transplant consensus conference.

INTRODUCTION: Living donor liver transplantation (LDLT) is a promising option for mitigating the deceased donor organ shortage and reducing waitlist mortality. Despite excellent outcomes and data supporting expanding candidate indications for LDLT, broader uptake throughout the United States has yet to occur. METHODS: In response to this, the American Society of Transplantation hosted a virtual consensus conference (October 18-19, 2021), bringing together relevant experts with the aim of identifying barriers to broader implementation and making recommendations regarding strategies to address these barriers. In this report, we summarize the findings relevant to the selection and engagement of both the LDLT candidate and living donor. Utilizing a modified Delphi approach, barrier and strategy statements were developed, refined, and voted on for overall barrier importance and potential impact and feasibility of the strategy to address said barrier. RESULTS: Barriers identified fell into three general categories: 1) awareness, acceptance, and engagement across patients (potential candidates and donors), providers, and institutions, 2) data gaps and lack of standardization in candidate and donor selection, and 3) data gaps regarding post-living liver donation outcomes and resource needs. CONCLUSIONS: Strategies to address barriers included efforts toward education and engagement across populations, rigorous and collaborative research, and institutional commitment and resources.

Honoring the gift: The transformative potential of transplant-declined human organs.

For decades, transplantation has been a life-saving treatment for those fortunate enough to gain access. Nevertheless, many patients die waiting for an organ and countless more never make it onto the waitlist because of a shortage of donor organs. Concurrently, thousands of donated organs are declined for transplant each year because of concerns about poor outcomes post-transplant. The decline of any donated organ-even if medically justified-is tragic for both the donor family and potential recipients. In this Personal Viewpoint, we discuss the need for a new mindset in how we honor the gift of organ donation. We believe that the use of transplant-declined human organs in translational research has the potential to hasten breakthrough discoveries in a multitude of scientific and medical areas. More importantly, such breakthroughs will allow us to properly value every donated organ. We further discuss the many practical challenges that such research presents and offer some possible solutions based on experiences in our own research laboratories. Finally, we share our perspective on what we believe are the necessary next steps to ensure a future where every donated organ realizes its full potential to impact the lives of current and future patients.

Perceptions and Early Outcomes of the Acuity Circles Allocation Policy Among Liver Transplant Centers in the United States.

Recently, a new liver allocation policy called the acuity circles (AC) framework was implemented to decrease geographic disparities in transplant metrics across donor service areas. Early analyses have examined the changes in outcomes because of the AC policy. However, perceptions among transplant surgeons and staff regarding the new policy remain unknown. Methods: A 28-item survey was sent to division chiefs and surgical directors of liver transplantation across the United States. Questions assessed the respondents' perceptions regarding center-level metrics and staff satisfaction. We used Organ Procurement and Transplantation Network data to study differences in allocation between the pre-AC implementation period (2019) and the post-AC implementation period (2020-2021). Results: A total of 40 participants completed this ongoing survey study. Most responses were from region 8 (13%), region 10 (15%), and region 11 (13%). Sixty-three percent of respondents stated that the wait time for a suitable offer for recipients with model of end-stage liver disease score <30 has decreased, whereas 50% stated that wait time for a suitable offer for recipients with model of end-stage liver disease score >30 has increased. However, most respondents (75%) felt that the average cost per transplant had increased and that the rate of surgical complications and 1-y graft survival had remained the same. In most states, an observable decrease in in-state liver transplantations occurred each year between 2019 and 2021. In addition, most allocation regions reported an increase in donations after circulatory deaths between 2019 and 2021. Conclusions: Perceptions of the new AC policy among liver transplant surgeons in the United States remain mixed, highlighting the potential strengths and concerns regarding its future impact. Further studies should assess the effects of the AC policy on clinical outcomes and liver transplantation access.

Epidemiology and outcomes of surgical site infections among pediatric liver transplant recipients.

INTRODUCTION: Surgical site infections (SSI) are a significant cause of morbidity in liver transplant recipients, and the current data in the pediatric population are limited. The goal of this study was to identify the incidence, classification, risk factors, and outcomes of SSIs among children undergoing liver transplantation (LT). METHODS: A single-center, retrospective descriptive analysis was performed of patients age ≤18 years undergoing LT between September 2007 and April 2017. SSI identified within the first 30 days were analyzed. Primary endpoints included incidence, classification, risk factors, and outcomes associated with SSIs. RESULTS: We included 86 patients, eight patients (9.3%) developed SSIs. Among segmental grafts (SG) recipients, 7/61 (11.4%) developed SSI. Among whole grafts recipients, 1/25 (4%) developed SSI. SSIs were associated with the presence of biliary complications (35% vs. 3%, p < .01; odds ratios 24, 95% CI: 3.41-487.37, p<.01). There were no differences in long term graft or patient survival associated with SSI. Patients who developed SSI were more likely to undergo reoperation (50% vs. 16.7%, p = .045) and had an increased total number of hospital days in the first 60 days post-transplant (30.5 vs. 12.5 days, p = .001). CONCLUSIONS: SSIs after pediatric LT was less frequent than what has been previously reported in literature. SSIs were associated with the presence of biliary complications without an increase in mortality. SG had an increased rate of biliary complications without an association to SSIs but, considering its positive impact on organ shortage barriers, should not be a deterrent to the utilization of SGs.

Sustained Improvement in Patient Experience by Optimizing Patient Flow in Ambulatory Settings.

Patient experience has become a priority for healthcare institutions as it affects clinical quality of care, financial reimbursement, provider, and patient satisfaction. We report our experience of improving patient experience measured by Press Ganey surveys in a busy multidisciplinary clinic over 65 months. We optimized patient flow in the clinic by technology-facilitated communication among the clinic staff and by a modest space redesign. We noted a significant improvement in "clinic visit" scores from baseline of 82.1 to 84.6 at year 1, 86.1 at year 2, 88.7 at year 3, and 88.9 at year 4 (P < .001). In comparison with previous short-term studies, we were able to sustain improvement in patient experience scores over 4 years due to optimized patient flow and monitoring of clinic operations. A similar approach can be implemented in other ambulatory settings and is likely to cause a long-term positive impact on patient experience.

North American Practice-Based Recommendations for Transjugular Intrahepatic Portosystemic Shunts in Portal Hypertension.

Complications of portal hypertension, including ascites, gastrointestinal bleeding, hepatic hydrothorax, and hepatic encephalopathy, are associated with significant morbidity and mortality. Despite few high-quality randomized controlled trials to guide therapeutic decisions, transjugular intrahepatic portosystemic shunt (TIPS) creation has emerged as a crucial therapeutic option to treat complications of portal hypertension. In North America, the decision to perform TIPS involves gastroenterologists, hepatologists, and interventional radiologists, but TIPS creation is performed by interventional radiologists. This is in contrast to other parts of the world where TIPS creation is performed primarily by hepatologists. Thus, the successful use of TIPS in North America is dependent on a multidisciplinary approach and technical expertise, so as to optimize outcomes. Recently, new procedural techniques, TIPS stent technology, and indications for TIPS have emerged. As a result, practices and outcomes vary greatly across institutions and significant knowledge gaps exist. In this consensus statement, the Advancing Liver Therapeutic Approaches group critically reviews the application of TIPS in the management of portal hypertension. Advancing Liver Therapeutic Approaches convened a multidisciplinary group of North American experts from hepatology, interventional radiology, transplant surgery, nephrology, cardiology, pulmonology, and hematology to critically review existing literature and develop practice-based recommendations for the use of TIPS in patients with any cause of portal hypertension in terms of candidate selection, procedural best practices and, post-TIPS management; and to develop areas of consensus for TIPS indications and the prevention of complications. Finally, future research directions are identified related to TIPS for the management of portal hypertension.

Inactive status is an independent predictor of liver transplant waitlist mortality and is associated with a transplant centers median meld at transplant.

BACKGROUND: Approximately 30% of patients on the liver transplant waitlist experience at least one inactive status change which makes them temporarily ineligible to receive a deceased donor transplant. We hypothesized that inactive status would be associated with higher mortality which may differ on a transplant centers' or donor service areas' (DSA) Median MELD at Transplant (MMaT). METHODS: Multi-state models were constructed (OPTN database;06/18/2013-06/08/2018) using DSA-level and transplant center-level data where MMaT were numerically ranked and categorized into tertiles. Hazards ratios were calculated between DSA and transplant center tertiles, stratified by MELD score, to determine differences in inactive to active transition probabilities. RESULTS: 7,625 (30.2% of sample registrants;25,216 total) experienced at least one inactive status change in the DSA-level cohort and 7,623 experienced at least one inactive status change in the transplant-center level cohort (30.2% of sample registrants;25,211 total). Inactive patients with MELD≤34 had a higher probability of becoming re-activated if they were waitlisted in a low or medium MMaT transplant center or DSA. Transplant rates were higher and lower re-activation probability was associated with higher mortality for the MELD 26-34 group in the high MMaT tertile. There were no significant differences in re-activation, transplant probability, or waitlist mortality for inactivated patients with MELD≥35 regardless of a DSA's or center's MMaT. CONCLUSION: This study shows that an inactive status change is independently associated with waitlist mortality. This association differs by a centers' and a DSAs' MMaT. Prioritization through care coordination to resolve issues of inactivity is fundamental to improving access.

Developing simultaneous liver-kidney transplant medical eligibility criteria while providing a safety net: A 2-year review of the OPTN's allocation policy.

The OPTN's simultaneous liver-kidney (SLK) allocation policy, implemented August 10, 2017, established medical eligibility criteria for adult SLK candidates and created Safety Net kidney allocation priority for liver-alone recipients with new/continued renal impairment. OPTN SLK and kidney after liver (KAL) data were analyzed (registrations as of December 31, 2019, transplants pre-policy [March 20, 2015-August 9, 2017] vs. post-policy [August 10, 2017-December 31, 2019]). Ninety-four percent of SLK registrations met eligibility criteria (99% CKD: 50% dialysis, 50% eGFR). SLK transplant volume decreased from a record 740 (2017) to 676 (2018, -9%), with a subsequent increase to 728 (2019, 1.6% below 2017 volume). For KAL listings within 1 year of liver transplant, waitlist mortality rates declined post-policy versus pre-policy (27 [95% CI = 20.6-34.7] vs. 16 [11.7-20.5]) while transplant rates increased fourfold (46 [32.2-60.0] vs. 197 [171.6-224.7]). There were 234 KAL transplants post-policy (94% Safety Net priority eligible), and no significant difference in 1-year patient/graft survival vs. kidney-alone (patient: 95.9% KAL, 97.0% kidney-alone [p = .39]; graft: 94.2% KAL, 94.6% kidney-alone [p = .81]). From pre- to post-policy, the proportion of all deceased donor kidney and liver transplants that were SLK decreased (kidney: 5.1% to 4.3%; liver: 9.7% to 8.7%). SLK policy implementation interrupted the longstanding rise in SLK transplants, while Safety Net priority directed kidneys to liver recipients in need with thus far minimal impact to posttransplant outcomes.

Bile Microbiota in Liver Transplantation: Proof of Concept Using Gene Amplification in a Heterogeneous Clinical Scenario.

Objective: Historically, bile in the biliary tract has been considered sterile. Most of the series are based on patients with biliary tract diseases or the bile has been obtained with procedures susceptible to contamination. Methods: We evaluated the bile in a heterogeneous cohort of liver donors and recipient patients, with samples obtained in a sterile way, directly from the gallbladder and the common bile duct. Results: We assessed the bile microbiota in six liver donors and in six liver recipients after whole or split liver procedures in adult or pediatric recipients. Bile samples were studied using PCR sequencing of the 16S ribosomal RNA gene amplification (rDNA). Conclusions: We demonstrated that the bile is sterile, thereby ruling this out as a source of contamination following transplant.

Impact of the new heart allocation policy on patients with restrictive, hypertrophic, or congenital cardiomyopathies.

BACKGROUND: Patients with restrictive or hypertrophic cardiomyopathy (RCM/HCM) and congenital heart disease (CHD) do not derive clinical benefit from inotropes and mechanical circulatory support. Concerns were expressed that the new heart allocation system implemented in October 2018 would disadvantage these patients. This paper aimed to examine the impact of the new adult heart allocation system on transplantation and outcomes among patients with RCM/HCM/CHD. METHODS: We identified adult patients with RCM/HCM/CHD in the United Network for Organ Sharing (UNOS) database who were listed for or received a cardiac transplant from April 2017-June 2020. The cohort was separated into those listed before and after allocation system changes. Demographics and recipient characteristics, donor characteristics, waitlist survival, and post-transplantation outcomes were analyzed. RESULTS: The number of patients listed for RCM/HCM/CHD increased after the allocation system change from 429 to 517. Prior to the change, the majority RCM/HCM/CHD patients were Status 1A at time of transplantation; afterwards, most were Status 2. Wait times decreased significantly for all: RCM (41 days vs 27 days; P<0.05), HCM (55 days vs 38 days; P<0.05), CHD (81 days vs 49 days; P<0.05). Distance traveled increased for all: RCM (76 mi. vs 261 mi, P<0.001), HCM (88 mi. vs 231 mi. P<0.001), CHD (114 mi vs 199 mi, P<0.05). Rates of transplantation were higher for RCM and CHD (P<0.01), whereas post-transplant survival remained unchanged. CONCLUSIONS: The new allocation system has had a positive impact on time to transplantation of patients with RCM, HCM, and CHD without negatively influencing survival.

Changes in Use of Left Ventricular Assist Devices as Bridge to Transplantation With New Heart Allocation Policy.

OBJECTIVES: The goal of this study was to describe outcomes of patients with bridge to heart transplantation (BTT) after changes were made to the donor heart allocation system. BACKGROUND: Left ventricular assist devices (LVADs) have been used as a BTT. On October 18, 2018, the donor heart allocation system in the United States was updated. METHODS: This study identified adults in the United Network for Organ Sharing database with durable, continuous-flow LVAD at listing or implanted while listed between April 2017 and April 2020. Baseline recipient and donor characteristics, waitlist survival, and post-transplantation outcomes were compared pre- and post-allocation system change. RESULTS: A total of 1,794 patients met inclusion criteria: 983 in the pre-change period and 814 afterward. The number of patients listed with LVAD decreased nationally over time from 102 in April 2017 to 12 in April 2020 (p < 0.001). The proportion of patients with LVAD at time of transplant decreased from 47% to 14%. Before the change, the majority were Status 1A (75.8%) at transplantation; afterward, most were Status 2/3 (67.8%). Transplantation rates were not different (85.4% vs. 83.6%; p = 0.225), but waitlist time decreased in the post period (82 vs. 65 days; p = 0.004). Donors were more likely to be high risk (39.0% vs. 32.2%; p = 0.005), and both ischemic times and distance traveled increased (3.4 h vs. 3.1 h; p < 0.001; 199 miles vs. 82 miles; p < 0.001). Waitlist survival did not change, but post-transplantation survival was worse in patients with BTT post-change (p < 0.001). CONCLUSIONS: The number of patients with BTT on the transplant list decreased steadily and dramatically after the allocation system change. Although time to transplant decreased, there was an increase in post-transplant mortality. These data suggest that the risks and benefits of LVAD implantation as a BTT have changed under the new allocation system and that the appropriate indication for this treatment strategy warrants a re-evaluation.

Current status: meeting the regulatory goals of your liver transplant program.

PURPOSE OF REVIEW: The regulatory framework set by the Organ Procurement and Transplantation Network (OPTN) and Center for Medicare and Medicaid Services (CMS) for practice of liver transplantation in US is periodically updated and risk adjusted. Therefore, it is prudent for transplant centers to know the rules of engagement as it pertains to their practice. RECENT FINDINGS: OPTN besides providing the regulatory oversight for safe and continued practice of transplant centers, provides necessary tools like: advanced statistical models and technological platforms to aid, and guide transplant centers including the necessary safeguards for high-quality transplant care.CMS regulations although had different thresholds to flag underperformance, often covered common grounds similar to the OPTN, therefore considered duplicative and unnecessary. But with much deliberation and consideration CMS undertook a major overhaul to the final rule for re-approval applications, a giant leap in the positive direction for transplant innovation and growth. SUMMARY: The duplicative regulatory framework of OPTN and CMS has although achieved the goal of improving 1-year patient outcomes, it has proven costly in terms of slowing innovation, increasing organ discard and stunting growth of transplant volume. But the new updates in effect and also in the pipeline are a long-awaited opportunity for waiting transplant patients.

American Association for the Study of Liver Diseases Expert Panel Consensus Statement: Vaccines to Prevent Coronavirus Disease 2019 Infection in Patients With Liver Disease.

The aim of this document is to provide a concise scientific review of the currently available COVID-19 vaccines and those in development, including mRNA, adenoviral vectors, and recombinant protein approaches. The anticipated use of COVID-19 vaccines in patients with chronic liver disease (CLD) and liver transplant (LT) recipients is reviewed and practical guidance is provided for health care providers involved in the care of patients with liver disease and LT about vaccine prioritization and administration. The Pfizer and Moderna mRNA COVID-19 vaccines are associated with a 94%-95% vaccine efficacy compared to placebo against COVID-19. Local site reactions of pain and tenderness were reported in 70%-90% of clinical trial participants, and systemic reactions of fever and fatigue were reported in 40%-70% of participants, but these reactions were generally mild and self-limited and occurred more frequently in younger persons. Severe hypersensitivity reactions related to the mRNA COVID-19 vaccines are rare and more commonly observed in women and persons with a history of previous drug reactions for unclear reasons. Because patients with advanced liver disease and immunosuppressed patients were excluded from the vaccine licensing trials, additional data regarding the safety and efficacy of COVID-19 vaccines are eagerly awaited in these and other subgroups. Remarkably safe and highly effective mRNA COVID-19 vaccines are now available for widespread use and should be given to all adult patients with CLD and LT recipients. The online companion document located at https://www.aasld.org/about-aasld/covid-19-resources will be updated as additional data become available regarding the safety and efficacy of other COVID-19 vaccines in development.

Deep learning-assisted differentiation of pathologically proven atypical and typical hepatocellular carcinoma (HCC) versus non-HCC on contrast-enhanced MRI of the liver.

OBJECTIVES: To train a deep learning model to differentiate between pathologically proven hepatocellular carcinoma (HCC) and non-HCC lesions including lesions with atypical imaging features on MRI. METHODS: This IRB-approved retrospective study included 118 patients with 150 lesions (93 (62%) HCC and 57 (38%) non-HCC) pathologically confirmed through biopsies (n = 72), resections (n = 29), liver transplants (n = 46), and autopsies (n = 3). Forty-seven percent of HCC lesions showed atypical imaging features (not meeting Liver Imaging Reporting and Data System [LI-RADS] criteria for definitive HCC/LR5). A 3D convolutional neural network (CNN) was trained on 140 lesions and tested for its ability to classify the 10 remaining lesions (5 HCC/5 non-HCC). Performance of the model was averaged over 150 runs with random sub-sampling to provide class-balanced test sets. A lesion grading system was developed to demonstrate the similarity between atypical HCC and non-HCC lesions prone to misclassification by the CNN. RESULTS: The CNN demonstrated an overall accuracy of 87.3%. Sensitivities/specificities for HCC and non-HCC lesions were 92.7%/82.0% and 82.0%/92.7%, respectively. The area under the receiver operating curve was 0.912. CNN's performance was correlated with the lesion grading system, becoming less accurate the more atypical imaging features the lesions showed. CONCLUSION: This study provides proof-of-concept for CNN-based classification of both typical- and atypical-appearing HCC lesions on multi-phasic MRI, utilizing pathologically confirmed lesions as "ground truth." KEY POINTS: • A CNN trained on atypical appearing pathologically proven HCC lesions not meeting LI-RADS criteria for definitive HCC (LR5) can correctly differentiate HCC lesions from other liver malignancies, potentially expanding the role of image-based diagnosis in primary liver cancer with atypical features. • The trained CNN demonstrated an overall accuracy of 87.3% and a computational time of < 3 ms which paves the way for clinical application as a decision support instrument.