Regional GI Compartments and Anatomical Boundaries

Why Regional Boundaries Matter

A gastrointestinal PBPK model assigns properties to places along a continuous organ system. The regions are therefore not just labels: their boundaries determine which measurements belong to each input, which quantities are added to make a total, and what a reported regional value means. If one source defines “ileum” as the entire distal half of the small intestine while another uses an anatomical segment ending at the ileocecal valve, their values cannot be compared as though they described the same region.

For a useful regional representation, define boundaries before entering measurements into a parameter table. Then record the quantity, units, source, and any transformation for each region. This is consistent with the earlier material on parameter provenance: an input is interpretable only when its meaning and applicability are documented. The aim here is to define the anatomical map and the kinds of regional quantities that need a place on that map, not to assign universal values to them.

Definition: A GI region is a named anatomical interval used as a unit for organizing PBPK inputs. Its boundary is a stated landmark or, where anatomy has no sharp transition, an explicit operational rule. A regional input describes a property assigned to that interval; it does not imply that the property is uniform throughout the tissue or lumen.

Anatomical Regions and Their Boundaries

A common coarse-grained map includes the stomach, duodenum, jejunum, ileum, and colon. It follows the direction of luminal passage and uses major anatomical landmarks where they are available. The map is a modeling choice rather than a claim that physiology changes abruptly at every boundary.

Stomach: begins at the gastroesophageal junction and extends to the pylorus, the outlet into the duodenum. If a model includes the esophagus or separates parts of the stomach, it should state those choices rather than silently including them in the gastric input.
Duodenum: extends from the pylorus to the duodenojejunal flexure, near the ligament of Treitz. These landmarks give the proximal and distal boundaries of the duodenal region.
Jejunum: begins at the duodenojejunal flexure and continues into the small intestine. There is no single sharp gross-anatomical landmark that consistently marks the jejunum-to-ileum transition.
Ileum: is the distal small intestine and ends at the ileocecal valve, where the small intestine joins the cecum. In a model with separate jejunal and ileal regions, their shared boundary therefore needs an operational definition.
Colon: begins at the cecum, beyond the ileocecal valve, and includes the ascending, transverse, descending, and sigmoid colon. The rectum and anal canal are anatomically distinct; state explicitly whether a model excludes them, includes them in a distal large-intestine region, or represents them separately.

The stomach, duodenum, and ileocecal junction have useful landmarks, but the jejunum-to-ileum boundary is not a crisp line that can always be located in an individual. A model may operationalize it using a documented fraction of small-intestinal length, a specified segment scheme, or another anatomy-informed rule. The selected rule should be consistent across the regional lengths, volumes, tissue quantities, and measurements that use that boundary. Do not present an operational division as though it were a universally visible anatomical landmark.

Worked Example: Assigning Measurements to Regions

Suppose a hypothetical source reports a measurement from the duodenojejunal flexure to the ileocecal valve, while a PBPK input table has separate jejunum and ileum rows. The table’s documented operational boundary places the jejunum-to-ileum division at 40% of the measured interval, counted from the flexure. How should the measurement be mapped?

First identify the anatomical extent of the source measurement: it spans the jejunum and ileum, not the duodenum or colon. The source has not measured those two regions separately. If the model uses the stated 40% rule for length and the measurement is a quantity that can legitimately be apportioned by length, the assignments are:

L_{\mathrm{jejunum}}=0.40L_{\mathrm{SI}}, \qquad L_{\mathrm{ileum}}=0.60L_{\mathrm{SI}}.

For a hypothetical measured interval of 500 cm, the assigned lengths are \(0.40(500)=200\) cm for jejunum and \(0.60(500)=300\) cm for ileum. The two assigned lengths sum to 500 cm, so the partition preserves the measured total. The 40% rule is a modeling assumption in this example, not a universal anatomical fact.

This allocation is appropriate only for the interval length under the stated rule. It does not justify splitting a measured tissue volume, surface area, or physiological measurement in the same proportions unless the relevant relationship is supported. If the source reports only a regional total for a quantity that cannot be defensibly apportioned, retain it as a combined small-intestinal value or document a separate allocation assumption.

What a Regional Input Record Should Represent

A region record should make it possible to tell both where an input applies and what kind of quantity it is. At a minimum, store the region name, proximal and distal boundaries, the boundary rule where needed, units, source, and any conversion or allocation. Beyond those identifiers, the required quantities depend on the model question and level of detail. Common categories include:

Geometry: regional length, a stated diameter or radius, wall thickness, and any geometric assumptions used to derive a volume or area.
Tissue amount: regional wall or tissue volume, mass, or another explicitly defined measure of tissue quantity. State which tissue layers are included.
Luminal space: a regional lumen volume or other defined measure of the space in which luminal material is represented. Keep this distinct from tissue volume.
Surface representation: anatomical surface area and, if used, an effective area for a specified model process. These are not automatically interchangeable; the latter may depend on how the model represents the tissue.
Regional physiological properties: any region-dependent values required by the model, with their conditions and measurement definitions. Examples include luminal pH, transit information, fluid availability, and bile-related quantities; each requires its own appropriately defined evidence and assumptions.
Model mapping: which state or process uses the value, whether it is measured or derived, and whether the value applies to the full region or only a subregion.

This list is a checklist, not a claim that every model needs every quantity. A model concerned with one question may need a coarse regional map and a small set of inputs; another may require subdivisions or additional measurements. In either case, do not mix geometry with physiology. For example, a regional length is not a residence time, and a lumen volume is not tissue volume. Each has its own definition, units, and evidence.

Modeling principle: A regional parameter is meaningful only together with its region definition, measurement basis, and units. When a boundary is operational, record the rule. When a value is derived or apportioned, record that transformation and its assumptions.

Deriving a Consistent Aggregate

When two adjacent regions are combined, an overall value must be calculated according to what the value represents. For an intensive property defined per unit tissue volume, a simple arithmetic mean is generally not the correct aggregate if the regions have different tissue volumes. The aggregate must preserve the total amount implied by the regional values.

Let \(q_i\) be an intensive property in region \(i\), and let \(V_i\) be the tissue volume to which that property applies. The amount represented in region \(i\) is \(q_iV_i\). After combining regions, the total represented amount is \(\sum_i q_iV_i\), and the total tissue volume is \(\sum_i V_i\). Therefore the combined intensive value is:

q_{\mathrm{combined}} = \frac{\sum_i q_iV_i}{\sum_i V_i}.

This expression is a volume-weighted mean: multiplying the combined value by the combined volume recovers the sum of the regional amounts. If instead the property is defined per unit surface area, the analogous area-weighted expression uses \(A_i\), not \(V_i\). The weighting basis must match the definition of the quantity.

Worked Example: Combining Two Tissue-Based Inputs

A hypothetical model combines two adjacent small-intestinal regions for an analysis. Region A has a tissue volume of 120 mL and a property \(q_A=2.0\) units/mL. Region B has a tissue volume of 280 mL and \(q_B=3.5\) units/mL. Find the combined property.

The represented amounts are \(q_AV_A=2.0(120)=240\) units in Region A and \(q_BV_B=3.5(280)=980\) units in Region B. The combined property is:

q_{\mathrm{combined}} = \frac{(2.0\ \mathrm{units/mL})(120\ \mathrm{mL})+ (3.5\ \mathrm{units/mL})(280\ \mathrm{mL})} {120\ \mathrm{mL}+280\ \mathrm{mL}} = \frac{1220\ \mathrm{units}}{400\ \mathrm{mL}} = 3.05\ \mathrm{units/mL}.

The result preserves the total represented amount: \(3.05\ \mathrm{units/mL}\times400\ \mathrm{mL}=1220\) units. The unweighted mean, \((2.0+3.5)/2=2.75\) units/mL, would not preserve that amount because the regions have unequal volumes. This aggregation is appropriate only if the stated property is defined per unit of the tissue volume used in the calculation.

Choosing the Level of Regional Detail

A single “gut” compartment is simple, but it removes location-specific distinctions. A five-region scheme of stomach, duodenum, jejunum, ileum, and colon makes those locations explicit while remaining manageable. More detailed schemes can divide the stomach, small intestine, or colon further. Greater detail is not automatically better: each added compartment requires a defensible boundary and inputs that correspond to it.

Choose the coarsest segmentation that can represent the model question without combining regions whose relevant inputs are meaningfully different for that question. If a source reports only a combined small-intestinal value, do not create precise jejunal and ileal values from it without a documented allocation method. Conversely, if a model needs to distinguish regions but the evidence only supports a combined input, that limitation should remain visible rather than being hidden by a detailed-looking table.

Worked Example: Auditing a Regional Input Table

A hypothetical PBPK table has these entries: “small intestine length: 520 cm,” “jejunum length: 240 cm,” and “ileum length: 310 cm.” The model defines the small intestine as the combined jejunum and ileum, with no separately counted segment. Are the entries consistent?

Under that definition, the regional lengths should satisfy:

L_{\mathrm{small\ intestine}} = L_{\mathrm{jejunum}}+L_{\mathrm{ileum}}.

The two subregion values sum to \(240+310=550\) cm, which differs from the reported total of 520 cm by 30 cm. The table is therefore inconsistent as written. The discrepancy could arise from different sources, different boundary conventions, or a transcription error; the numbers alone do not establish which explanation is correct.

The audit should check that all three entries refer to compatible anatomical extents and measurement definitions. If they do, the modeler must resolve the discrepancy or select a documented, applicable value. If the measurements are not comparable, retain the distinction and explain why they should not be forced to sum. Simply adjusting one entry to make the arithmetic work would create a derived input without a defensible basis.

Diagnostics, Alternatives, and Common Pitfalls

Several checks help reveal a regional map that is becoming unreliable. Confirm that each region has an unambiguous proximal and distal boundary; that adjacent regions meet without unexplained gaps or overlaps; and that regional totals agree with a whole-organ total when they are intended to partition that total. Check units and definitions before comparing values. Also check whether an input applies to a full region, a sampled location, or a model-defined subregion. A disagreement may indicate a genuine difference in source populations or measurement methods rather than a simple data error.

A whole-gut compartment is an alternative when regional distinctions are unnecessary or unsupported, but it cannot retain region-specific inputs. A coarse standard-region model offers a practical compromise. A finer segmentation can represent more local variation, but it also raises the burden of evidence: boundaries, regional values, and aggregation rules must all be specified. These approaches are not interchangeable merely because they cover the same organ system.

Common mistake: treating the jejunum-to-ileum boundary as a fixed visible landmark. That transition is not sharply demarcated in the same way as the pylorus or ileocecal valve. State the operational rule used in the model.
Common mistake: using an unweighted average for regions of unequal size. For a value defined per unit volume or area, use the matching volume or area weights when combining regions.
Common mistake: assuming every regional quantity follows the same partition. A length-based split does not automatically justify a split of tissue volume, surface area, or a physiological measurement.
Common mistake: counting the rectum as colon without saying so. State whether rectum and anal canal are excluded, combined into a distal region, or separately represented.

Regional detail is useful only when its meaning is stable enough to support comparison and modeling. A boundary definition that is explicit, a distinction between tissue and lumen quantities, and transparent aggregation rules are often more valuable than adding compartments without compatible evidence.

Check Your Understanding

Answer each question, then compare your reasoning with the solutions.

What landmarks bound the duodenum in the regional scheme described here?
Why should a model state an operational rule for the jejunum-to-ileum boundary?
Two regions have values defined per unit tissue volume. What weighting basis should be used to combine them, and why?
A source reports a value for the entire small intestine, but the model has separate jejunum and ileum inputs. What must be established before splitting the source value?
Does the colon necessarily include the rectum in a regional PBPK table? What should the table specify?

Solutions. (1) The duodenum extends from the pylorus to the duodenojejunal flexure near the ligament of Treitz. (2) The jejunum-to-ileum transition has no single sharp gross-anatomical landmark, so the model needs a reproducible boundary rule. (3) Use tissue-volume weighting, because multiplying each per-volume value by its tissue volume gives the represented amount; the combined value must preserve the sum of those amounts. (4) Establish the source’s anatomical extent and measurement definition, then justify an allocation rule suitable for that particular quantity; otherwise retain the combined value or document the limitation. (5) No. The table should say whether the rectum and anal canal are excluded, included in a distal large-intestine region, or represented separately.