Using MRI to Assess Microvascular Invasion in Hepatocellular Carcinoma

HomeRadiologyVol. 297, No. 3 PreviousNext Reviews and CommentaryFree AccessEditorialUsing MRI to Assess Microvascular Invasion in Hepatocellular CarcinomaAn Tang An Tang Author AffiliationsFrom the Department of Radiology, Radio-oncology and Nuclear Medicine, Université de Montréal, Montréal, Canada; Department of Radiology, Centre Hospitalier de l'Université de Montréal (CHUM), 1000 rue Saint-Denis, Montréal, QC, Canada H2X 0C2; and Centre de Recherche du Centre Hospitalier de l'Université de Montréal (CRCHUM), Montréal, Canada.Address correspondence to the author (e-mail: [email protected]).An Tang Published Online:Sep 29 2020https://doi.org/10.1148/radiol.2020203376MoreSectionsPDF ToolsImage ViewerAdd to favoritesCiteTrack CitationsPermissionsReprints ShareShare onFacebookTwitterLinked In See also the article by Min et al in this issue.Dr An Tang is an abdominal radiologist and clinical researcher at the Centre Hospitalier de l'Université de Montréal (CHUM) and professor of radiology at the Université de Montréal. His research interests focus on imaging-based biomarkers of chronic liver disease and liver cancer. Dr Tang is a member of the LI-RADS steering committee, current chair of its International Working Group, and principal investigator on a CIHR grant.Download as PowerPointOpen in Image Viewer Hepatocellular carcinoma (HCC) may spread by means of several routes. Tumor growth may occur by direct extension into the surrounding parenchyma and tumor dissemination, invasion of sinusoids or portal venules to form satellite nodules, lymphatic spread to regional lymph nodes, tumor spread inside a vein, or hematogenous spread to distant organs (1,2).HCC has a propensity to invade vascular structures. Macrovascular invasion refers to macroscopically visible tumor in vein, whereas microvascular invasion (MVI) refers to the presence of tumor cells inside portal or hepatic venous systems (3). Macrovascular invasion and MVI are indicators of an aggressive biologic behavior associated with a more advanced disease stage, a higher risk of tumor recurrence, and shorter long-term survival after surgery (4). Although macrovascular invasion (or tumor in vein) can be diagnosed with imaging techniques, the preoperative detection of MVI with imaging techniques is challenging because MVI occurs at a microscopic scale. Hence, histopathologic analysis of a resected specimen is currently required for accurate identification of MVI.The presence of macrovascular invasion typically precludes curative treatment by means of partial hepatectomy or liver transplant. Although MVI was previously underestimated, perceived as a mild or early form of vascular invasion, its prognostic role is increasingly recognized. MVI is fairly common, found in 15%–57% of patients undergoing partial hepatectomy or liver transplant for HCC, with a higher rate of MVI in larger tumors (5). Short-term tumor recurrence rates are higher for small HCCs (≤2 cm) when MVI is present (6). Although long-term survival of small HCCs does not appear to be affected by MVI, patients with large HCCs (>2 cm) have worse long-term survival outcomes after surgery when MVI is present (7).The objective of staging is to predict the prognosis and to link the stage with the most appropriate treatment. Currently, major HCC staging systems stratify tumor stage on the basis of tumor size, number of tumors, presence of macrovascular invasion, and presence of extrahepatic disease in lymph nodes or distant metastases (8). MVI is not currently incorporated in imaging-based HCC staging systems, although this may change in the future.The propensity of HCC for vascular invasion, the high incidence of MVI, the potential need for closer monitoring of recurrence or use of adjuvant therapy in patients with MVI, the prognostic implication of MVI on long-term survival, and the current reliance on histopathologic analysis of resected specimens all underscore the critical need for reproducible and accurate noninvasive tests to identify the presence of MVI preoperatively.Because MVI cannot be imaged directly before surgery, predictive scoring systems have been developed to predict the presence of MVI at histopathologic examination and the risk of developing extrahepatic metastases. Typically, predictive models incorporate demographic information (eg, patient age), serum markers (eg, α-fetoprotein or protein induced by vitamin K absence-II, or PIVKA-II, levels), or imaging characteristics (eg, tumor number, size, margins, or capsule) to predict the presence of MVI.In this issue of Radiology, Min et al (9) assessed the ability to detect MVI in HCC with gadoxetate disodium–enhanced MRI. This timely article in an area of active research is superbly illustrated and contains impeccable tables. The retrospective study meticulously reports the interobserver agreement and diagnostic performance of individual and combinations of imaging features seen at preoperative MRI.As secondary end points, Min et al also evaluated the impact of radiologists' experience (more experienced vs less experienced reviewers) and tumor size (≤3 cm vs > 3 cm) on agreement. The index test was gadoxetate disodium–enhanced MRI, including hepatobiliary phase images. The four imaging features assessed to determine the possibility of MVI were nonsmooth tumor margins, irregular rim-like enhancement in the arterial phase, peritumoral arterial phase hyperenhancement, and peritumoral hepatobiliary phase hypointensity. The reference standard was pathologic assessment of MVI as described in pathology reports. The study revealed fair to moderate interobserver agreement for the four imaging features. Interobserver agreement was significantly higher for more experienced than less experienced radiologists. The diagnostic performance of MRI in the prediction of MVI was modest, with areas under the receiver operating characteristic curves ranging from 0.60 to 0.74.This study is important because it captures the current state of the art in MRI-based preoperative MVI prediction. The moderate interobserver agreement and modest diagnostic performance indicate that the current approach is promising, but not yet ready for clinical prime time. Higher agreement and diagnostic performance will be required before imaging criteria of MVI can be incorporated into imaging-based HCC staging systems.The findings of Min et al (9) must be appreciated in the broader context of research on MVI. A systematic review of 20 observational studies assessing the impact of MVI after liver transplant or resection of HCC revealed substantial intra- and interobserver variability in reporting and grading of MVI at histopathologic analysis, with κ values of 0.19 to 0.50 (10). In light of the slight to moderate agreement on MVI at histopathologic analysis—which is considered the accepted reference standard—variability of MVI assessment at MRI should not be surprising. Therefore, moderate interobserver agreement and modest diagnostic performance at MRI found in this study should not deter future research in this area for imaging biomarkers that predict or directly visualize MVI.Unlike previous studies of MRI-based features of MVI that have been conducted in single centers with a few readers, the study by Min et al is a multicenter, multireader study more likely to reflect real-world performance. None of the four imaging features assessed actually represent MVI, a histopathologic finding that occurs at a microscopic scale, whereas MRI images pixels at a scale of millimeters. Higher spatial resolution will be required to directly image MVI.The study has some limitations. The end point was the presence or absence of MVI in histopathologic reports. Preoperative identification of MVI may be further refined by identifying its location and distance between the invaded vessels and the tumor margins. This could help guide treatment of patients with HCC in terms of selection of hepatectomy type and width of resection margins. In addition, this study focused on four imaging features to predict the presence of MVI. Other imaging features, such as the presence of a capsule appearance, have been considered as predictors of MVI at preoperative CT or MRI and deserve inclusion in future studies.This study was performed with gadoxetate disodium, a hepatobiliary contrast agent that requires a 20-minute delayed phase to assess signal intensity characteristics in the hepatobiliary phase. Additional studies are required to assess the interobserver agreement and performance of preoperative MRI features with extracellular agents or blood pool agents (both of which arguably provide higher contrast for vascular assessment) for the presence of MVI in HCC. Future studies may also include radiologic and pathologic correlation to improve understanding of imaging features that accompany MVI spatially. Localization of MVI is also relevant conceptually because small branches of the portal vein are more likely to be invaded than those of the hepatic veins. Development of a reproducible and accurate imaging technique for prediction of MVI or, eventually, direct visualization of MVI will be required before it can be incorporated in imaging-based HCC staging systems.Disclosures of Conflicts of Interest: Activities related to the present article: received a research scholarship from Fonds de recherche du Québec en Santé (FRQ-S) and Fondation de l'association des radiologistes du Québec. Activities not related to the present article: received a speaking honorarium from Eli Lilly; received operating grants from Onco-Tech, Institut de Valorisation de Donnees, and Canadian Institutes of Health Research. Other relationships: disclosed no relevant relationships.References1. Kitao A, Zen Y, Matsui O, Gabata T, Nakanuma Y. Hepatocarcinogenesis: multistep changes of drainage vessels at CT during arterial portography and hepatic arteriography—radiologic-pathologic correlation. Radiology 2009;252(2):605–614. Link, Google Scholar2. Sirlin CB. Liver Imaging Reporting and Data System version 2018 CT/MRI Manual. Chapter 6: Hepatocarcinogenesis. Page 6–8. American College of Radiology. https://www.acr.org/-/media/ACR/Files/Clinical-Resources/LIRADS/Chapter-6-Hepatocarcinogenesis.pdf?la=en&hash=692ADBDC797A2C7F0002AE87CEBFD2C5. Published 2018. Accessed August 7, 2020. Google Scholar3. Sumie S, Kuromatsu R, Okuda K, et al. Microvascular invasion in patients with hepatocellular carcinoma and its predictable clinicopathological factors. Ann Surg Oncol 2008;15(5):1375–1382. Crossref, Medline, Google Scholar4. Lauwers GY, Terris B, Balis UJ, et al. Prognostic histologic indicators of curatively resected hepatocellular carcinomas: a multi-institutional analysis of 425 patients with definition of a histologic prognostic index. Am J Surg Pathol 2002;26(1):25–34. Crossref, Medline, Google Scholar5. Rodríguez-Perálvarez M, Luong TV, Andreana L, Meyer T, Dhillon AP, Burroughs AK. A systematic review of microvascular invasion in hepatocellular carcinoma: diagnostic and prognostic variability. Ann Surg Oncol 2013;20(1):325–339. Crossref, Medline, Google Scholar6. Yamashita Y, Tsuijita E, Takeishi K, et al. Predictors for microinvasion of small hepatocellular carcinoma ≤ 2 cm. Ann Surg Oncol 2012;19(6):2027–2034. Crossref, Medline, Google Scholar7. Shindoh J, Andreou A, Aloia TA, et al. Microvascular invasion does not predict long-term survival in hepatocellular carcinoma up to 2 cm: reappraisal of the staging system for solitary tumors. Ann Surg Oncol 2013;20(4):1223–1229. Crossref, Medline, Google Scholar8. Tang A, Fowler KJ, Chernyak V, Chapman WC, Sirlin CB. LI-RADS and transplantation for hepatocellular carcinoma. Abdom Radiol (NY) 2018;43(1):193–202. Crossref, Medline, Google Scholar9. Min JH, Lee MW, Park HS, et al. Interobserver variability and diagnostic performance of gadoxetic acid–enhanced MRI for predicting microvascular invasion in hepatocellular carcinoma. Radiology 2020;297:573–581. Link, Google Scholar10. Fan L, Mac MT, Frishberg DP, et al. Interobserver and intraobserver variability in evaluating vascular invasion in hepatocellular carcinoma. J Gastroenterol Hepatol 2010;25(9):1556–1561. Crossref, Medline, Google ScholarArticle HistoryReceived: Aug 9 2020Revision requested: Aug 17 2020Revision received: Aug 18 2020Accepted: Aug 21 2020Published online: Sept 29 2020Published in print: Dec 2020 FiguresReferencesRelatedDetailsCited ByWorld Journal of Gastroenterology, Vol. 27, No. 32Journal of radiology and nuclear medicine, Vol. 102, No. 5Accompanying This ArticleInterobserver Variability and Diagnostic Performance of Gadoxetic Acid–enhanced MRI for Predicting Microvascular Invasion in Hepatocellular Carcinoma29 Sep 2020RadiologyRecommended Articles LI-RADS Tumor in Vein at CT and Hepatobiliary MRIRadiology2021Volume: 302Issue: 1pp. 107-115Proliferative versus Nonproliferative Hepatocellular Carcinoma: Clinical and Imaging ImplicationsRadiology2021Volume: 300Issue: 3pp. 583-585Differentiation of Hepatocellular Carcinoma from Other Hepatic Malignancies in Patients at Risk: Diagnostic Performance of the Liver Imaging Reporting and Data System Version 2014Radiology2017Volume: 286Issue: 1pp. 158-172Portal Vein Thrombosis in Patients with Hepatocellular Carcinoma: Diagnostic Accuracy of Gadoxetic Acid–enhanced MR ImagingRadiology2016Volume: 279Issue: 3pp. 773-783Interobserver Variability and Diagnostic Performance of Gadoxetic Acid–enhanced MRI for Predicting Microvascular Invasion in Hepatocellular CarcinomaRadiology2020Volume: 297Issue: 3pp. 573-581See More RSNA Education Exhibits Infiltrative Appearance HCC: Key Features for Detection and Diagnosis of the LI-RADS MisfitDigital Posters2018Decision Making When Using LI-RADS: Tips and Tricks to Increase ConfidenceDigital Posters2020The Fast and the Furious: Rapid Progression of HCC After Locoregional Therapy — A Tri-Institutional Collaboration on Imaging PatternsDigital Posters2018 RSNA Case Collection LI-RADS 5RSNA Case Collection2022LI-RADS 5RSNA Case Collection2021Pseudocirrhosis of Hepatic Breast MetastasesRSNA Case Collection2020 Vol. 297, No. 3 Metrics Downloaded 1,017 times Altmetric Score PDF download