Hepatocellular carcinoma, the most common form of primary liver cancer, is a disease deeply entangled with metabolism. Most patients develop it on a backdrop of cirrhosis, chronic inflammation, and often obesity or diabetes, all of which reshape body composition in ways that standard measures like body mass index cannot capture. Over the past decade, computed tomography has emerged as the gold standard for assessing body composition in cancer care, because the same scans used to stage tumors and monitor response also allow precise quantification of skeletal muscle and adipose tissue at a single anatomical level, typically the third lumbar vertebra. Abnormal patterns on these scans, particularly muscle depletion and distorted fat distribution, have increasingly been linked to worse outcomes across many cancer types. What has remained unclear is whether these measurements, which capture a moving target, remain meaningful as a patient’s body wastes away during months of therapy.

The research team, led by Salvatore Corallo of the University of Pavia and Fondazione IRCCS Policlinico San Matteo together with Margherita Rimini of IRCCS San Raffaele Hospital and Andrea Casadei-Gardini, set out to answer two questions at once. First, how does body composition at the start of treatment influence survival and the risk of dose-limiting toxicities in patients with advanced hepatocellular carcinoma? Second, how does body composition change over the course of treatment, and do those changes alter the predictive value of the baseline measurements? Their retrospective analysis included 113 patients with advanced disease who received first-line systemic therapy, roughly seventy percent treated with tyrosine kinase inhibitors and thirty percent with immune checkpoint inhibitors, primarily the widely used combination of atezolizumab and bevacizumab that has become the standard of care for unresectable disease.

The technical heart of the study lies in quantitative CT analysis performed at two time points: baseline, before treatment began, and the moment of best radiologic response. At the level of the third lumbar vertebra, the researchers measured the lumbar skeletal muscle index, a cross-sectional muscle area normalized for height, alongside separate indices for visceral and subcutaneous adipose tissue. From the two fat measurements they derived the visceral-to-subcutaneous adipose tissue ratio, abbreviated VSR, a dimensionless number that expresses the balance between metabolically active internal fat and the more inert peripheral fat depot. This distinction is biologically important. Visceral adipose tissue is a hormonally active organ that secretes inflammatory cytokines, adipokines, and free fatty acids directly into the portal circulation that drains into the liver, fueling inflammation, insulin resistance, and in the context of cirrhosis, a hostile metabolic environment. Subcutaneous fat, by contrast, has a comparatively quiescent secretory profile, and a high proportion of visceral relative to subcutaneous fat is considered a marker of metabolically adverse adiposity.

The statistical analysis relied on Cox proportional hazards regression to model the relationship between body composition variables and survival outcomes, complemented by generalized linear models for the toxicity endpoint. In univariable analyses, two features stood out: a high visceral-to-subcutaneous adipose tissue ratio and muscle depletion, both of which predicted poor survival. When the investigators moved to bivariable models adjusted for established prognostic factors in advanced hepatocellular carcinoma, the association between elevated VSR and reduced survival held firm, indicating that the ratio carries prognostic information that is not simply a proxy for tumor burden, liver function, or other clinical variables. The strength of the effect was striking. Patients with a higher ratio, meaning a greater share of their fat packed into the visceral compartment, fared substantially worse than those whose fat was distributed more toward the subcutaneous depot.

Perhaps the most clinically consequential result concerns toxicity. Elevated VSR was associated with a significantly increased incidence of dose-limiting toxicities, with a hazard ratio of 1.16 per unit increase in the ratio, a ninety-five percent confidence interval of 1.08 to 1.26, and a p-value below 0.001. Dose-limiting toxicities are the severe adverse events that force clinicians to interrupt treatment, reduce doses, or abandon a regimen altogether, and they are a major reason why patients with advanced liver cancer fail to derive full benefit from otherwise effective drugs. A pre-treatment scan that identifies patients at high risk of such events could in principle guide closer monitoring, prophylactic supportive care, nutritional intervention, or earlier consideration of alternative regimens. The magnitude of the hazard ratio may appear modest per unit, but because VSR is a continuous variable spanning a wide range across patients, the cumulative difference in risk between the lowest and highest ratios is substantial.

The longitudinal component of the study delivered an equally important insight. Over the course of treatment, most patients lost both muscle and fat, as expected in advanced cancer, where cachexia and the metabolic side effects of therapy progressively erode body stores. Yet when the researchers recalculated the visceral-to-subcutaneous ratio at the time of best radiologic response, they found that it had remained essentially stable even as the absolute tissue quantities declined. This stability proved decisive for the ratio’s predictive value. At the follow-up time point, high VSR remained associated with poorer outcomes, with hazard ratios of 1.37 for progression-free survival and 1.12 for overall survival, both statistically significant at p below 0.001. In other words, the ratio behaved as a time-stable trait, a fixed characteristic of an individual’s metabolic constitution rather than a fluctuating state, which is precisely the quality that makes a biomarker trustworthy for repeated clinical use.

This property sets VSR apart from other body composition metrics that the field has pursued. Absolute muscle area and fat area shrink under the combined pressures of malignancy, inflammation, reduced intake, and drug toxicity, so measurements taken at different times are difficult to compare and their prognostic meaning drifts. A ratio, by contrast, normalizes one compartment against another, largely canceling out the generalized wasting that afflicts nearly all these patients. The study’s authors emphasize that this time stability is what allows VSR to function as a robust, repeatable predictor, and they frame it as a candidate for integration into the nutritional and prognostic assessment of patients with advanced hepatocellular carcinoma, pending validation in larger and more diverse cohorts.

The findings also resonate with a growing literature on the paradoxical role of adiposity in liver cancer prognosis. Prior work has shown that adipose tissue distribution predicts the prognosis of cirrhotic patients undergoing surgical resection for hepatocellular carcinoma, that different body composition patterns affect outcomes under immunotherapy in ways that confound expectations based on body mass index alone, and that clinical-adipometric nomograms can help predict tumor response to transarterial chemoembolization. The new study adds a unifying thread to this scattered evidence by proposing a single, stable, easily computed index that captures the metabolically relevant dimension of fat biology. The putative mechanisms are plausible: visceral fat’s portal drainage delivers inflammatory and immunomodulatory mediators directly to the diseased liver, potentially promoting tumor progression, undermining hepatic drug metabolism, and amplifying the systemic inflammation that heightens vulnerability to treatment-related injury.

The researchers are careful about the limits of their work. The study was retrospective, single-country, and modest in sample size, and it spanned two different classes of systemic therapy, which introduces heterogeneity even though the results were adjusted for relevant prognostic factors. The authors state explicitly that validation in larger studies is needed before VSR can be recommended for routine clinical decision-making. Nevertheless, the appeal of the approach is hard to overstate. Every patient with hepatocellular carcinoma undergoes serial abdominal CT as part of standard care, and the software to segment visceral and subcutaneous fat at the lumbar level is already available in most radiology departments. No additional radiation, cost, blood draw, or biopsy is required. A measurement that is essentially free, stable over time, and predictive of both the length of life and the quality of that life under treatment represents exactly the kind of low-hanging fruit that oncology nutrition research has long promised.

If the results are confirmed, the implications could extend beyond hepatocellular carcinoma to other cancers in which body composition shapes treatment tolerance, and they could reframe how clinicians think about obesity in oncology. Two patients with identical body mass index may carry radically different metabolic risk depending on the distribution of their fat, and the visceral-to-subcutaneous ratio offers a way to see that difference on a scan that would otherwise be read only for tumor response. For now, the message for the field is that the information needed to anticipate who will struggle with advanced liver cancer therapy may already be sitting, unexamined, on the picture archiving systems of cancer centers around the world.