Material may reach the GI tract directly by ingestion, by transfer from the respiratory tract as described above, or by transfer from other body organs. The GI tract model defined in ICRP Publication 30 Part 1 (ICRP, [28]) was used in ICRP Publications 67, 68, 69, 71, 72 and 78 [30, 32, 33, 36-38] to describe the behaviour of radionuclides in the GI tract, and to calculate doses from radionuclides in the contents of the GI tract. 

Recently the Publication 30 model of the gastrointestinal tract has been replaced by the Human Alimentary Tract Model (HATM) described in ICRP Publication 100 ([39]). The structure of the HATM is shown in Fig 13. Its main features can be summarised as follows:

• Inclusion of all alimentary tract regions. Doses are calculated for the oral cavity, oesophagus, stomach, small intestine, right colon, left colon and rectosigmoid (the sigmoid colon and rectum). Colon doses are combined as a mass-weighted mean to include the right colon, left colon and rectosigmoid.

• Gender-dependent parameter values for adults for dimensions and transit times of contents through the regions (age-dependent parameter values are also specified for use in future calculations of doses to members of the public).

• Transit times for food and liquids, as well as for total diet, for the mouth, oesophagus and stomach.

• A default assumption that absorption of an element and its radioisotopes to blood occurs exclusively in the small intestine, ie. The total fractional absorption, f_A coincides with the fractional absorption from the small intestine, f_SI.It is assumed there is no recycling from the wall to blood.

• Model structure to allow for absorption in other regions, where information is available.

• Model structure to allow for retention in the mucosal tissues of the walls of alimentary tract regions, and on teeth, where information is available.

• Explicit calculations of dose to target regions for cancer induction within each alimentary tract region, considering doses from radionuclides in the contents of the regions, and considering mucosal retention of radionuclides when this is taken into account.

The organs and fluids represented in Fig 13 by dashed boxes show connections between the HATM and the HRTM and systemic biokinetic models. First-order kinetics is assumed for all transfers in the HATM. This is a considerable simplification of the complex processes involved in transfer of material through the lumen of the alimentary tract but is expected to provide a reasonably accurate representation of the mean residence time of a radionuclide in each segment of the tract.

Mucus and associated materials cleared from the respiratory tract enter the oesophagus via the oropharynx. For ingested food and liquids, the HATM specifies two components of oesophageal transit representing relatively fast transfer of 90% (mean transit time of 7 seconds for total diet) of the swallowed material and relatively slow transit of the residual 10% (40 seconds for total diet). It is assumed that the slower oesophageal transit times apply to all material cleared from the respiratory tract.

The oral cavity and oesophagus will receive very low doses from radionuclides in transit because of their short transit times (ICRP, ([39]). However, these regions were included for completeness, because a specific w_T is assigned to the oesophagus (ICRP, [40]), and because retention in the mouth, on teeth for example, can result in a substantial increase in dose to the oral mucosa. In general, the alimentary tract regions of greatest importance in terms of doses and cancer risk are the stomach and particularly the colon. While the small intestine may receive greater doses than the stomach, it is not sensitive to radiation-induced cancer and is not assigned a specific w_T value. Doses are calculated separately for the right colon, left colon and rectosigmoid. This partitioning of the colon for the purposes of dose calculations is predicated on the availability of transit time data. The rectum is taken to be part of the rectosigmoid, primarily because of difficulties in determining transit times separately.  Mean transit times for the stomach and colon are about one-third greater in adult females than males. Slightly smaller masses in females (eg. 10% lower mass of colon tissue) will compound this gender difference.

In most cases, the values of f_A for ingestion will be the same as the f₁ values given previously for use with the Publication 30 model, since in most cases there is unlikely to be sufficient new information to warrant a revision in values. In addition, the general default assumption will be that absorption occurs solely from the small intestine, as in the Publication 30 model; that is, f_SI = f_A. However, the HATM allows absorption to be specified for other regions as well as the small intestine. As discussed in Publication 100 (ICRP ([39]) for the example of isotopes of iodine doses to alimentary tract regions and other tissues will in many cases be insensitive to assumptions regarding the site of absorption.

For inhaled particles reaching the alimentary tract after escalation from the respiratory tract, it is appropriate to take account of solubility in the lungs in specifying f_A values. For elements exhibiting a range in solubility according to their physicochemical form, there is evidence that the reduced solubility of Type M or S materials is also associated with reduced intestinal absorption. As discussed in Publication 71 (ICRP, ([33]), in many cases for which a single f_A value is specified for ingestion of an element, this is taken to apply to inhaled Type F materials and lower default values of 0.1 and 0.01 are applied to Types M and S. However, because of the need for realism in estimates of absorption for application to bioassay interpretation, attempts have been made wherever possible to use available data to specify f_A values for different forms rather than rely on defaults.  

An important development in the HATM is the methodology used to calculate doses in the various regions from non-penetrating alpha and electron radiations. Thus, while the Publication 30 approach was to assume that the dose to the wall was one half of that to contents of the region, with an additional factor of 0.01 included for alpha particles to allow for their short range, the HATM takes explicit account of the location of the target tissue in the mucosal layer of the wall of each region. The targets relating to cancer induction are taken in each case to be the epithelial stem cells, located in the basal layers of the stratified epithelia of the oral cavity and oesophagus and within the crypts that replenish the single cell layer epithelium of the stomach and small and large intestines.

As discussed in Publication 100 (ICRP ([39]), the HATM generally results in substantially lower estimates of doses to the colon from beta-emitting radionuclides than obtained using the Publication 30 model. This is because the HATM takes explicit account of dose to the target region throughout the length of the colon, and of loss of energy in the colon contents and the mucosal tissue overlying the target stem cells (at a depth of 280 - 300 mm). This reduces energy deposition in the target tissue for electrons and results in zero dose in the target tissue from alpha particles. In the absence of retention of radionuclides in the alimentary tract wall, doses from ingested alpha emitters to all regions of the alimentary tract will be solely due to their absorption to blood and subsequent irradiation from systemic activity in soft tissues. For the stomach, the HATM and Publication 30 approaches give more similar estimates of doses from electron-emitting nuclides.