Absorption refers to the movement of an agent from a peripheral site into the systemic circulation. Because oral administration is the most important nonintravenous route of anticancer drug administration, it is important to consider limitations produced by the gastrointestinal tract. After oral administration, drugs may be degraded by acids in the stomach, adsorbed by food or other medications, metabolized by enzymes in gut luminal cells, extracted by the liver in first-pass metabolism before a systemic effect can be produced, or simply not absorbed because of physicochemical characteristics of the drug molecule. The rate and extent of drug absorption after oral compared with intravenous administration is referred to as bioavailability. Before a drug can be incorporated into anticancer treatment regimen, its bioavailability must be understood. The most reliable way to study bioavailability is to compare plasma AUCs after oral and intravenous administration in the same patient; bioavailability is usually expressed as the fraction of the AUC after intravenous administration produced by oral administration.

The two agents most commonly administered by mouth in childhood anticancer therapy are methotrexate and mercaptopurine (MP), which are used in nearly all maintenance regimens for ALL. Both these agents exhibit considerable variation in bioavailability (58,59,60,61,62,63). Concomitant food administration diminishes the absorption of both agents (64,65). There may also be diurnal variation in the
absorption or elimination of these drugs, as suggested by the observation that children who take oral methotrexate and MP in the evening appear to have lower risk of relapse than those who take them in the morning (66,67,68,69).

Bioavailability of methotrexate at low doses (7.5 to 40 mg per m²) is quite variable. In one study, absorption ranged from 23% to 95% (58). Furthermore, methotrexate absorption appears to be saturable, with a plateau in AUC observed at doses of 30 mg per m² (58,70,71). Therefore, the relationship between dose and AUC is not linear for oral methotrexate. To overcome these limitations, intramuscular administration is sometimes used instead of oral administration of methotrexate. In one study of children with ALL, intramuscular injection of methotrexate resulted in approximately twice the bioavailability of oral administration, and at doses exceeding 40 mg per m² intramuscular absorption did not show the same nonlinearity as oral administration (71).

Oral administration of MP also results in variable systemic exposure. MP undergoes extensive first-pass metabolism in the liver and gut mucosa by the enzyme xanthine oxidase (63,72). In addition, there is considerable interpatient variability of plasma MP concentrations even when the dose is administered under fasting conditions (61,62,63). Intrapatient variability has also been observed with repeated monitoring of individual patients (73).

Other agents may have more predictable bioavailability after oral administration. The absorption of prednisone and dexamethasone is greater than 80% (74,75), although variability in the absorption of prednisone has been reported in children (76). Oral busulfan is rapidly absorbed, with a bioavailability of 70% in children (77,78). Interestingly, there appears to be circadian variation in busulfan plasma concentrations, with the highest troughs occurring in the early morning (79).

Distribution of a drug refers to its transfer among various tissues, blood, tumor, and any other compartments. Although it is possible to measure drug concentrations in different tissues under laboratory conditions, it is not feasible to do so in most clinical settings. Drug concentrations in various compartments can be estimated using pharmacokinetic modeling techniques, but the compartments referred to in models do not usually correlate directly with physiologic compartments. However, the use of pharmacokinetic models, in addition to preclinical studies, permits some conclusions to be drawn about the distribution of most anticancer agents. For example, the pharmacokinetic behavior of both the anthracyclines and the Vinca alkaloids is consistent with extensive tissue binding. Disappearance of these drugs from plasma drug during the distributive phase is extremely rapid, with a distribution half-life of less than 10 minutes. These agents also exhibit volumes of distribution that are much larger than the circulatory volume (several hundred L per m²) as well as a prolonged terminal half-lives (80,81). In addition, tissue levels of anthracyclines, which bind extensively to DNA, are 10- to 500-fold higher than plasma concentrations of the drugs (13).

Methotrexate distributes widely in total body water (82). This is important in relationship to drug toxicity because methotrexate distributes freely into extravascular fluid collections such as pleural effusions or ascites. These fluid collections can contain substantial amounts of methotrexate after high-dose administration and can act as depots that release drug slowly back to other issues, resulting in prolonged systemic drug exposure and increased toxicity (83). Thus, methotrexate should be administered with caution to patients with ascites or pleural effusions, and consideration should be given to draining such fluid collections prior to methotrexate therapy.

Distribution of drugs into the central nervous system (CNS) represents a special challenge in anticancer therapy. Both primary and metastatic CNS tumors are very common in children, and drugs must cross the blood–brain barrier (BBB) to gain access to these tumors. In many tumors the BBB is at least partially disrupted by the tumor itself (84). In other cases, the BBB is relatively intact. Drugs that are poorly soluble in lipids, undergo significant ionization in plasma, or are highly protein bound are unlikely to penetrate extensively across the BBB (85,86,87). Because it is very difficult to measure drug penetration into tumor or brain tissue, the concentration of drug in the cerebrospinal fluid (CSF) is often used as a surrogate marker for CNS penetration. CNS drug penetration is most accurately described expressed as the ratio of drug AUC in CSF to drug AUC in plasma. The majority of anticancer drugs in clinical use do not cross the intact BBB to any significant extent (Table 44.2). CSF:plasma ratios for the Vinca alkaloids, anthracyclines, and epipodophyllotoxins
are less than 10%. A few agents, such as thiotepa and the nitrosoureas, penetrate well. Several useful agents, such as cytarabine and methotrexate, do not penetrate well but can be administered in such high systemic doses that even though the CSF:plasma ratio is low, the CSF concentration is still cytotoxic (88,89,90,91,92,93,94). However, such high-dose systemic approaches produce significant systemic toxicity. The most common approach to overcome this problem is the direct intrathecal administration of anticancer agents. Methotrexate and cytarabine remain the mainstay of intrathecal therapy, although other agents, such as thiotepa (95), DepoCyt (liposomally encapsulated cytarabine) (96,97,98,99), topotecan (100), gemcitabine (101,102), and busulfan (103) have also been studied.

Biotransformation, or metabolism, of anticancer drugs can result in production of active agents or inactivation of the agents. Some commonly administered anticancer agents are really prodrugs that require metabolic transformation before they have antitumor activity (Table 44.3). Other drugs are active in the form administered but undergo metabolism to additional active species that may contribute importantly to activity or toxicity. For these agents, all active moieties must be considered in evaluating the relationship between pharmacokinetic parameters like AUC or half-life and pharmacodynamic parameters like response or toxicity. Drug activation most commonly occurs in the liver, where the bulk of drug metabolism takes place. Some drugs, especially antimetabolites, may be activated in their target tissues as they are incorporated into DNA, RNA, or other macromolecules. A few drugs, such as alkylating agents, undergo spontaneous chemical decomposition in solution to cytotoxic reactive intermediates. Interpatient variability in metabolic activation may be an important part of interpatient variability in drug activity or especially toxicity at a given dose. Saturation of drug-metabolizing enzymes at high dose levels may lead to nonlinear dose-response curves. Interaction or competition with other drugs for a metabolic pathway, or even a drug’s induction of its own metabolic pathways, could also result in unexpected alterations in exposure to active species. In designing plans for regional therapy, such as intrathecal or intra-arterial drug administration, it is important to remember that agents requiring activation at a distant site will not be useful.

As mentioned, most antimetabolites are activated at their target sties. For example, after entering cells, cytarabine is converted to the active nucleotide arabinosylcytosine triphosphate (Ara-CTP) in three phosphorylation steps catalyzed by deoxycytidine kinase, deoxycytidylate kinase, and nucleoside diphosphate kinase (104,105). Ara-CTP then acts as a fraudulent substrate for DNA replicative and repair enzymes. The pharmacokinetic parameters of the active intracellular metabolites are more predictive of response than plasma levels of the parent drug. For example, studies have shown little correlation between the pharmacokinetic parameters of cytarabine in plasma and leukemic cell concentrations of Ara-CTP. Ara-CTP retention in blasts correlates with response better than plasma drug concentrations (106,107,108). Furthermore, decreased deoxycytidine kinase activity, resulting in decreased accumulation of Ara-CTP, is a primary mechanism of Ara-C resistance (109,110,111,112).