Scaling Functional Liver Mass in Chronic Disease

Why Liver Size Is Not Liver Capacity

A chronic liver disease model may need to represent two changes that do not necessarily move together: the anatomical size of the liver and the amount of viable hepatocyte tissue within it. A liver can be enlarged while containing a reduced proportion of viable parenchyma, or it can be smaller with substantial loss of viable tissue. Therefore, multiplying a drug-metabolizing capacity by the liver-volume ratio alone can misrepresent the biological change.

The relevant PBPK input for a mass-scaled whole-organ capacity is an estimate of functional liver mass: the amount of viable hepatocyte tissue available to contribute to that capacity. This is not necessarily a directly observed mass. It is often an operational model quantity formed from liver size and a measure or estimate of viable hepatocyte content. It should be kept separate from liver compartment volume, which represents anatomical volume for distribution and tissue partitioning, and from hepatic blood flow, which represents delivery. The earlier material on cirrhosis-related hepatic blood flow and portosystemic shunting addresses the latter inputs.

The quantity \(f_H\) may be an estimated fraction of tissue that is viable hepatocyte parenchyma or a normalized index derived from an appropriate measurement. Its definition and units must be consistent between reference and disease values. For a normalized index, \(f_{H,ref}=1.00\) is a reference value by convention; it does not mean that every part of the liver consists of hepatocytes. If density or the measurement scale differs materially between conditions, the simple product \(V_L f_H\) is not sufficient unless that difference is incorporated into the estimate.

From Functional Mass to a PBPK Input

The scaling argument concerns the total amount of capacity distributed across viable hepatocytes. Suppose a whole-liver capacity \(C_H\) is proportional to viable hepatocyte mass and the capacity per unit viable mass is unchanged by disease. Then \(C_H=M_H c_H\), where \(c_H\) is capacity per unit functional mass. Taking the ratio of disease to reference conditions gives \(C_{H,d}/C_{H,ref}=M_{H,d}/M_{H,ref}=R_M\). Thus a reference whole-liver capacity can be multiplied by \(R_M\), but only when the assumptions behind proportional scaling are reasonable.

The capacity \(C_H\) must be a whole-organ quantity, or a parameter that the PBPK implementation explicitly scales as a whole-organ quantity. For example, if the model represents a total liver enzyme amount or a whole-liver intrinsic capacity, mass scaling may be applied to that amount. Do not multiply a per-cell or per-mass concentration by \(R_M\) and then treat the result as a total-organ amount without checking how the model uses the parameter; that can apply the scaling twice. Nor should \(R_M\) automatically be applied to liver volume or blood flow. Those are distinct physiological inputs.

The simple scaling assumes that viable hepatocytes have comparable capacity per unit mass in disease and reference conditions, that the volume and viable-content measurements refer to compatible populations and definitions, and that the capacity is adequately represented as proportional to total viable mass. It also assumes that the chosen disease value represents the modeled individual or population. These assumptions are modeling choices, not consequences of observing a liver-volume change. Disease-related changes in the capacity of each viable hepatocyte require a separate parameterization rather than being silently absorbed into the mass factor.

When Liver Volume Is Preserved or Increased

A volume-only approach can be especially misleading when fibrosis, inflammation, congestion, or other structural changes preserve or increase total organ size while viable hepatocyte content falls. The mass factor combines the two pieces of information; it does not presume that a larger liver has more metabolically active tissue. Before calculating it, determine whether the available “liver size” measurement is total anatomical volume, a tissue-specific volume, or another imaging-derived quantity. The denominator and numerator must refer to comparable quantities.

Direct Mass Estimates and the Scaling Justification

If a validated estimate of viable hepatocyte mass is available in the relevant population, direct mass ratios avoid the additional assumptions needed to infer mass from organ volume and a content index. In practice, direct measurement may be unavailable, so the product formula is a transparent approximation. Its value depends on how well the volume estimate and the viable-content estimate represent the same organ and the same disease state.

The proportionality can be justified by summing the contribution of viable tissue. If each small viable mass element \(dM\) contributes \(c_H dM\), with constant capacity per unit mass \(c_H\), total capacity is the integral over viable tissue: \(C_H=\int c_H\,dM=c_H M_H\). Dividing disease capacity by reference capacity cancels \(c_H\), giving \(C_{H,d}/C_{H,ref}=M_{H,d}/M_{H,ref}\). If capacity per unit mass differs across regions or between disease and reference, that cancellation does not hold, and a single mass factor cannot fully represent the change.

Implementation Checks and Failure Modes

A careful implementation records the provenance, definition, and uncertainty of each input. Liver volume may come from imaging, while viable content may be estimated from histology or another source. A local biopsy can be unrepresentative of the whole organ, and histological area fractions do not automatically equal whole-organ mass fractions. If a biopsy-derived estimate is used as a proxy, state the representativeness assumption rather than treating it as an exact organ-wide measurement.

A value of \(R_M\) greater than 1 is not inherently invalid: it could arise from increased viable mass or from the measurement and reference definitions. But it should prompt a check that a larger anatomical volume has not been mistaken for more viable tissue. Likewise, a very low or high factor may expose incompatible scales, a transcription error, or a nonrepresentative sample. Because the factor is a product, proportional uncertainty in either component affects the result; uncertainty in both can be consequential even when each estimate appears moderately precise.

Alternatives, Limits, and a Common Mistake

There are several possible representations, and the choice should follow the available evidence. A direct viable-mass estimate is conceptually closest to the target, but may be difficult to obtain. The volume-times-content approach is usable when the two inputs are compatible and the assumptions are stated. A liver-volume-only adjustment is simpler but does not account for changes in the viable fraction. A functional score or clinical severity category may help classify disease, but it is not automatically a measurement of hepatocyte mass; converting it into \(R_M\) requires an explicitly supported mapping. Treating all disease effects as one global capacity multiplier is also limited when different processes or regions do not scale together.

A common mistake is to reason, “The liver is 20% larger, so its metabolic capacity is 20% higher.” That conclusion confuses anatomical volume with viable hepatocyte mass and assumes unchanged viable content and unchanged capacity per unit mass without checking either. The correction is to represent liver compartment volume and functional mass separately, estimate \(R_M\) using compatible inputs, and scale only a whole-organ capacity under the stated proportionality assumptions.

Mass scaling is therefore a structured approximation, not a universal description of chronic liver disease. It cannot by itself capture regional heterogeneity, different capacities among viable cells, or disease effects that alter drug disposition through mechanisms other than the quantity of viable tissue. The model should preserve these distinctions so subsequent evaluation can identify whether a prediction is sensitive to mass assumptions, flow, or other disease inputs rather than hiding them inside one number.