Natural history of portal vein embolization before liver resection: a 23-year analysis of intention-to-treat results Fernando A. Alvarez a , Denis Castaing a,b,c , Rodrigo Figueroa a , Marc Antoine Allard a,b,d , Nicolas Golse a,b,c , Gabriella Pittau a , Oriana Ciacio a , Antonio Sa Cunha a,b,d , Daniel Cherqui a,b,c , Daniel Azoulay e , René Adam a,b,d , and Eric Vibert a,b,c, * a AP-HP, Hôpital Paul Brousse, Centre Hépato-Biliaire, Villejuif, France b Université Paris-Sud, Villejuif, France c Inserm, Unité 1193, Villejuif, France d Inserm, Unité 776, Villejuif, France e Departement of Hepatopancreaticobiliary Surgery and Liver Transplantation, Hôpital Henri Mondor, AP-HP, Créteil, France ARTICLE INFO Article history: Accepted 22 December 2017 A B ST R AC T Background. Portal vein embolization (PVE) use is nowadays debated due to the risk of technical or bi- ological unresectability after the period of time needed to achieve future liver remnant (FLR) hypertrophy. We evaluated the safety and efficacy of PVE in a single high-volume hepatobiliary center, with empha- sis in the feasibility to achieve tumor resection. Methods. Patients undergoing PVE before major hepatectomy at our institution between 1993 and 2015 were retrospectively analyzed. Results. A total of 431 patients formed the study population. Morbidity and mortality rates of PVE were 16.7% and 0.2% respectively. Morbidity was similar between percutaneous and ileocolic approaches or between histoacryl and ethanol as embolization materials (P > 0.05). On the contrary, the percutaneous ipsilateral approach was associated with significantly less complications than the contralateral ap- proach (10.3% vs 19.4%; P = 0.024). Almost all patients (96%) achieved sufficient FLR volume after embolization, but only 66% finally underwent planned liver resection. Disease progression was the most common cause of unresectability (67%). Patients with extrahepatic biliary tumors experienced signifi- cantly higher unresectability rates compared to other entities (45.1% vs 31.4%; P = 0.019). Conclusion. PVE was not followed by hepatectomy in 34% of our patients. Biliary tumors displayed the higher dropout rates after PVE and the higher chances of tumor progression preventing curative resec- tion. Although PVE may be performed with acceptable morbidity, PVE-related complications prevented curative resection in 5% of patients. Careful multidisciplinary selection is crucial to avoid PVE overuse in technically resectable patients who will experience a not negligible risk of futile use and non- therapeutic laparotomy. © 2018 Elsevier Inc. All rights reserved. One of the main limitations to achieve complete tumor removal in patients bearing technically resectable liver tumors is the pres- ence of a small future liver remnant (FLR), insufficient to sustain liver function after resection. Preoperative portal vein emboliza- tion (PVE) was first described by Makuuchi et al 1 in 1984 before extended hepatectomy for bile duct carcinoma. This strategy has become the gold standard modality to increase the volume of normal liver tissue and thereby allow major liver resections with reduced risk of hepatic insufficiency and death in patients with an insuffi- cient FLR. 2,3 The occlusion of the right or left portal vein branches causes redistribution of blood flow and results in atrophy of the ip- silateral embolized lobe and hypertrophy of the contralateral nonembolized lobe. Even though there is no strict consensus for a safe volume of FLR, PVE is generally considered when the FLR volume is <20–25% of total liver volume (TLV) or <0.5% of body weight (BW) in normal livers. 4-6 We have reported previously how patients who underwent resection after PVE may present similar outcomes com- pared to those without preoperative PVE. 7,8 Nevertheless, resection rates after PVE vary among different studies and patients may become ineligible for surgery due to multiple causes. Even though it is well known that PVE may directly induce tumor progression within the ipsilateral or contralateral lobe, 9-11 the exact causes of * Corresponding author. AP-HP, Hôpital Paul Brousse, Centre Hépato-Biliaire, 14 Ave Paul Vaillant Couturier, Villejuif, 94800, France. E-mail address: eric.vibert@aphp.fr (E. Vibert). https://doi.org/10.1016/j.surg.2017.12.027 0039-6060/© 2018 Elsevier Inc. All rights reserved. Surgery ■■ (2018) ■■–■■ ARTICLE IN PRESS Please cite this article in press as: Fernando A. Alvarez, et al., Natural history of portal vein embolization before liver resection: a 23-year analysis of intention-to-treat results, Surgery (2018), doi: 10.1016/j.surg.2017.12.027 Contents lists available at ScienceDirect Surgery journal homepage: www.elsevier.com/locate/ymsy