Microwave Ablation Charts Overpromise in Cirrhotic Livers, Study Finds

When doctors cook a liver tumor with microwave energy, they rely on a small but crucial piece of paperwork: the manufacturer’s ablation chart. Enter a power setting and a duration, and the chart predicts the volume of tissue that will be destroyed. A new study from a tertiary referral center in Bergamo, Italy, suggests those predictions may be systematically too optimistic in one of the patients who need them most. In a retrospective cohort published in CVIR Oncology, researchers led by Riccardo Muglia of the University of Milano-Bicocca and ASST Papa Giovanni XXIII Hospital found that the ablation zones actually measured on immediate post-procedural imaging were smaller than the chart-predicted volumes in 41 of 43 treated hepatocellular carcinoma lesions, a striking 95.3 percent of the time. The finding strikes at the heart of a procedure where success is defined in millimeters, and it adds fuel to a growing conversation about how thermal ablation should be planned in diseased liver tissue.

The clinical stakes are considerable. Hepatocellular carcinoma, the most common primary liver cancer, almost always arises in livers scarred by cirrhosis, and the same scarring that produced the tumor limits how much healthy tissue can be sacrificed during treatment. For patients who are not surgical candidates, percutaneous microwave ablation has become a mainstay: an antenna is threaded into the tumor, typically under ultrasound or CT guidance, and dielectric heating coagulates the lesion while sparing the surrounding parenchyma. Technical success depends on complete coverage of the tumor by the ablation zone, and most interventional radiologists aim for an ablative margin of at least five millimeters in every direction. Fall short of that margin, and residual tumor cells at the edge of the zone can seed a local recurrence. Manufacturers therefore publish charts that translate power and time into expected ablation dimensions based on in vivo liver models, and clinicians use those numbers to plan each burn.

The Bergamo team set out to test whether those chart numbers hold up in the real world of cirrhotic livers. They assembled a single-center retrospective cohort of patients with cirrhosis and BCLC stage 0, A, or B hepatocellular carcinoma who underwent microwave ablation between November 28, 2022, and June 30, 2025, with a database cutoff of March 31, 2026. After applying strict inclusion and exclusion criteria, the analytic cohort comprised 29 patients, all male by chance, who underwent 29 procedures treating 43 lesions with a single microwave platform, the Medtronic Emprint HP system. Every lesion had both an immediate post-procedural volume measurement and a one-month follow-up measurement, allowing the researchers to compare reality against prediction twice. Lesions were classified by size, small up to 25 millimeters or larger lesions up to 30 millimeters, and by location, either within one centimeter of the liver capsule or within one centimeter of a hepatic or portal vein branch.

The measurement technique was deliberately straightforward. Using a commercial workstation, a radiologist measured three orthogonal diameters of each ablation zone on contrast-enhanced CT or CT hepatic arteriography performed immediately after the procedure, and applied the ellipsoid formula to convert those diameters into a volume. Power settings were selected from the manufacturer’s in vivo liver chart according to lesion dimensions and the intended margin, with most lesions treated at 100 watts and a minority at 150 watts, and recorded durations ranging from one to ten minutes with a median of six. The primary comparison was the difference between the measured volume and the chart-predicted volume, so a negative number meant the actual ablation fell short of the promise. Because some patients contributed multiple lesions, the team used a cluster-aware statistical approach, a linear mixed-effects model with a patient-level random intercept, and validated the results with a sensitivity analysis restricted to one lesion per patient.

The numbers told a consistent and sobering story. Across all 43 lesions, the mean chart-predicted volume was 23.4 cubic centimeters, while the mean measured volume on immediate imaging was just 15.1 cubic centimeters, a shortfall of roughly 35 percent. The cluster-adjusted mean bias was minus 8.50 cubic centimeters, with a 95 percent confidence interval running from minus 10.28 to minus 6.71, and the limits of agreement stretched from minus 18.63 to plus 1.64 cubic centimeters, meaning individual lesions could deviate from prediction by amounts that would matter enormously at the margin. A Wilcoxon signed-rank test confirmed the difference was highly significant, and the one-lesion-per-patient sensitivity analysis reproduced the bias almost exactly at minus 8.44 cubic centimeters. Only two lesions, both subcapsular, bucked the trend and came in larger than their chart values.

The statistical texture of the disagreement is as important as its average. The researchers detected proportional bias, with a slope of 0.34 indicating that the measured-minus-predicted gap grew as the predicted volume increased, and they found heteroscedasticity, meaning the scatter of differences widened for larger ablations. Together, these findings rule out the tempting quick fix of a single uniform correction factor that could be applied to every chart entry. In other words, clinicians cannot simply assume the chart overestimates by a fixed percentage; the error behaves differently depending on the size of the planned ablation, which complicates procedural planning and underscores the need for careful post-procedural verification rather than blind trust in lookup tables.

Why would microwave energy underperform in cirrhotic tissue? The authors point to biologically plausible mechanisms rooted in the physics of dielectric heating. Microwave antennas deposit energy by oscillating water molecules, and the efficiency of that deposition depends on the tissue’s electrical conductivity and thermal properties. Experimental work has shown that these dielectric properties differ measurably between normal liver, cirrhotic liver, and tumor tissue, and the fibrosis, altered perfusion, and changed water content of a cirrhotic liver can impede the spread of heat. Notably, the shortfall in the Bergamo cohort remained relatively constant across lesion sizes and locations, which the authors interpret as a uniform impairment of thermal conduction in the cirrhotic substrate rather than a size- or position-dependent effect. Their series also included an unusual concentration of difficult anatomy: every lesion was either subcapsular or perivascular, reflecting the referral patterns of a tertiary center.

The two outliers that exceeded their chart predictions offer an intriguing counterpoint. Both were small subcapsular lesions that showed transient volumetric expansion on immediate imaging before shrinking substantially at one month. The authors speculate that a lower degree of fibrosis, or excess liver hydration from intra-arterial contrast administration during CT hepatic arteriography, may have improved microwave diffusion in these cases, since better-hydrated tissue generally conducts microwave energy more effectively. The observation, while anecdotal, reinforces the central physical insight: the ablation zone is not a fixed product of the device settings but an emergent result of the interaction between delivered energy and the specific tissue it encounters.

The study also tracked how ablation zones evolve over time, a dimension that has received little attention in cirrhotic patients. Mean volume fell from 15.1 to 12.66 cubic centimeters between the immediate and one-month imaging, a cluster-adjusted change of minus 2.34 cubic centimeters, with a median within-lesion shrinkage of 15.9 percent. Thirty-seven lesions decreased and six increased. Meanwhile, clinical outcomes were strong by conventional measures: immediate technical success, defined as visual inclusion of the tumor within the non-enhancing ablation zone with an intended five-millimeter margin, was achieved in all 43 lesions, and local tumor progression occurred in five lesions, or 11.6 percent, over a median follow-up of 13 months. Four minor complications, two pneumothoraces and two mild hemoperitoneums, occurred across 29 procedures and were managed conservatively.

The authors are careful about the limits of their evidence. The study was retrospective, single-center, and non-randomized, involved one microwave platform and an all-male cohort, and lacked a non-cirrhotic comparator group. Quantitative fibrosis measurements and detailed liver-disease variables were unavailable, precluding mechanistic analysis, and volumes were measured by a single radiologist without formal reproducibility testing, using an ellipsoid formula that may not capture irregular ablation zones. Still, the message for practice is clear enough that the authors argue for adjusted ablation parameters in cirrhotic patients, potentially increasing power or duration within safety limits to achieve more adequate margins, since larger margins are known to reduce local recurrence. Their longer-term vision is more ambitious: liver-specific ablation charts that incorporate tissue perfusion and stiffness data, so that the lookup table reflects the actual liver on the table rather than an idealized one. Until such charts exist, the study suggests, the immediate post-procedural scan remains the only honest referee of whether the burn was big enough.

Subject of Research: Agreement between manufacturer-predicted and measured ablation volumes after microwave ablation of hepatocellular carcinoma in cirrhotic livers

Article Title: Manufacturer-chart versus measured ablation volumes after microwave ablation of hepatocellular carcinoma in cirrhotic livers: a retrospective cohort study

Article References: Muglia, R., Martins de Mattos, K. D., Carbone, F. S., Bertuletti, M., Dulcetta, L., Gargiulo, C., Marra, P., & Sironi, S. (2026). Manufacturer-chart versus measured ablation volumes after microwave ablation of hepatocellular carcinoma in cirrhotic livers: a retrospective cohort study. CVIR Oncology, 2(1), Article 34. https://doi.org/10.1007/s44343-026-00066-y

Image Credits: AI Generated

DOI: 10.1007/s44343-026-00066-y

Keywords: microwave ablation, hepatocellular carcinoma, cirrhosis, ablation volume, manufacturer chart, ablative margin, interventional radiology, liver cancer, Bland-Altman analysis, thermal ablation, CVIR Oncology, retrospective cohort