Dissolvable Gelatin Plug Triggers Liver Regrowth Before Major Cancer Surgery

For decades, surgeons preparing patients for the removal of large portions of the liver have relied on a paradoxically permanent solution: deliberately blocking the portal vein branches feeding the tumour-bearing side of the liver, forcing the remaining liver tissue to enlarge so that it can carry the metabolic load after surgery. The embolic agents used for this procedure, known as portal vein embolization, are designed never to dissolve, and that permanence has always carried a shadow risk. If the embolic material migrates to the wrong side of the liver or if thrombosis spreads into the main portal or mesenteric veins, the regenerative stimulus can turn into a disaster, sometimes ruling out curative surgery altogether. Now, results from the first prospective multicentre trial of a fully reversible version of the procedure suggest that the blockage does not need to last for the liver to respond.

The EMBORES trial, a phase 2 study conducted at six French hepatobiliary centres and published in eClinicalMedicine, tested whether a temporary embolization using powdered absorbable gelatin sponge could induce enough growth of the future liver remnant to permit major hepatectomy. Between May 2018 and May 2021, the investigators enrolled 34 patients with primary or secondary liver malignancies who required preoperative portal vein embolization before resection of three or more liver segments. Two patients were excluded from the efficacy analysis, one because of consent withdrawal and one because of technical difficulty in catheterizing the portal branches, leaving 32 patients who underwent the temporary procedure and could be evaluated with CT volumetry.

The technique itself is elegantly simple in concept. Interventional radiologists dissolved powdered absorbable gelatin, marketed as Curaspon, in a mixture of 80 percent iodixanol contrast agent and 20 percent saline, producing a homogeneous, radiopaque suspension that could be seen fluoroscopically as it was injected. The mixture was selectively delivered into the portal vein branches supplying the diseased liver until complete angiographic occlusion was achieved, while flow to the future liver remnant was carefully preserved. Roughly four grams of gelatin powder were typically needed for a right portal vein embolization, and the amount was tailored to each patient’s portal anatomy. No coils or permanent agents were used at any point.

The results were striking. After a mean follow-up interval of 30 days, CT scans showed that the portal veins had completely recanalized in 17 of 32 patients, or 53 percent, and partially recanalized in another nine, or 28 percent. In other words, in more than four out of five patients, blood flow through the embolized vessels had returned. Yet the liver had grown anyway. Mean future liver remnant volume rose from 580 millilitres before the procedure to 690 millilitres afterwards, corresponding to a mean hypertrophy ratio of 9.5 percent, with a median kinetic growth rate of 1.94 percent per week. The comparison with the historical benchmark of 13 percent hypertrophy after conventional permanent embolization yielded a p value of 0.059, just short of conventional statistical significance but well within the range of what the trial was powered to detect.

This recanalization paradox is the scientific heart of the study. Conventional wisdom holds that sustained portal occlusion is necessary to drive compensatory hypertrophy, on the logic that the deprived liver tissue must remain starved of portal blood for the growth signal to persist. The EMBORES findings, consistent with earlier experimental work in nonhuman primates and swine by the same French group, suggest instead that a transient period of portal deprivation is enough to switch on the regenerative programme, which then continues even after blood flow is restored. The authors note that the kinetics of recanalization, rather than its mere occurrence, may be the critical variable, since experimental models showed progressive restoration of portal flow during the first two weeks after embolization while regeneration continued unabated.

Safety data were reassuring. Intra-procedural complications occurred in six patients, or about 19 percent, and consisted mainly of abdominal pain, nausea, or transient fever. No patient developed liver failure or sepsis after embolization. A single clinically significant complication, a 3.1 percent rate, involved one patient with bilobar colorectal metastases and a history of multiple prior liver procedures who developed portal vein thrombosis extending to the contralateral branch and the superior mesenteric vein; the persistent thrombosis and inadequate hypertrophy ultimately precluded surgery, and the patient was shifted to palliative chemotherapy. For context, major complications after conventional permanent portal vein embolization are reported in roughly 2 to 5 percent of patients, and some of those complications stem precisely from the irreversibility of the occlusion.

Surgery followed in 23 of the 32 patients, a resection rate of 71.8 percent that matches figures reported after conventional embolization. Hepatectomy took place a median of 50 days after embolization, with a mean operative time of 357 minutes and median blood loss of 600 millilitres. Notably, adhesions around the portal pedicle were absent or filmy in 74 percent of operated patients, a finding the authors interpret as evidence that temporary occlusion provokes far less inflammation than permanent agents, which have been associated in prior reports with dense inflammatory scarring that complicates dissection. Ninety-day morbidity affected 11 of the 23 surgical patients, including two cases of post-hepatectomy liver failure defined by the 50-50 criterion of prothrombin time below 50 percent and bilirubin above 50 micromoles per litre. There were no deaths within 90 days.

Not every patient responded equally well. Five patients, or 15.6 percent, required a second embolization using conventional permanent materials because their livers did not grow enough, and four of those five ultimately underwent surgery without developing liver failure. More importantly, the hypertrophic response was significantly blunted in cirrhotic patients, whose mean hypertrophy ratio of 5.7 percent fell statistically short of the 13 percent reference value, whereas non-cirrhotic patients achieved 10.7 percent with no significant difference from the benchmark. This is consistent with the well-documented impairment of regenerative capacity in cirrhotic livers, and the authors caution that the absorbable gelatin approach should be used carefully in this population until more evidence accumulates.

The study has honest limitations that temper enthusiasm. Recruitment fell short of the planned 44 evaluable patients because of funding constraints, reducing statistical power, and the trial lacked a control group receiving conventional permanent embolization, so no direct comparison was possible. Recanalization was assessed at a single time point roughly one month after the procedure, leaving the precise timing and kinetics of vessel reopening unknown. Functional liver reserve was not measured, an omission highlighted by the two cases of post-hepatectomy liver failure, one of which occurred in a patient whose liver had grown briskly. The authors argue that future trials should combine volumetry with functional imaging such as technetium-99m mebrofenin hepatobiliary scintigraphy with SPECT/CT.

Even with those caveats, the conceptual implications are considerable. A reversible method of modulating portal inflow could reshape preoperative liver conditioning, particularly for patients who need only a modest increase in remnant volume, and could preserve therapeutic options for those whose disease progression ultimately rules out resection. It might even find applications beyond surgery, wherever transient portal deprivation could be useful without committing the patient to permanent vascular change. The investigators are careful to frame EMBORES as hypothesis-generating rather than practice-changing, and properly so: adequately powered comparative trials against permanent embolization are the necessary next step. But the core message is already provocative. The liver, it appears, does not need the door locked shut to be told to grow; a brief, dissolvable signal may be enough to set the regenerative machinery in motion.

Subject of Research: Temporary portal vein embolization with absorbable gelatin sponge to induce future liver remnant hypertrophy before major hepatectomy

Article Title: Efficacy and safety of temporary selective portal vein embolization before major hepatectomy (EMBORES): a phase 2 trial

Article References: Tranchart, H., Maitre, S., Girard, E., Ghelfi, J., Gelli, M., Tselikas, L., Barat, M., Fuks, D., Sarran, A., Benoist, S., Meyrignac, O., Sepulveda, A., Ronot, M., Pou, M., Chirica, M., Dagher, I., Sengel, C., Brouquet, A., Gayet, B., … Debry, J.-B. (2026). Efficacy and safety of temporary selective portal vein embolization before major hepatectomy (EMBORES): a phase 2 trial. eClinicalMedicine, 101, Article 104262. https://doi.org/10.1016/j.eclinm.2026.104262

Image Credits: AI Generated

DOI: 10.1016/j.eclinm.2026.104262

Keywords: portal vein embolization, liver regeneration, hepatectomy, absorbable gelatin sponge, future liver remnant, EMBORES trial, phase 2 study, liver surgery, interventional radiology, post-hepatectomy liver failure, cirrhosis, eClinicalMedicine