Methodologic Issues In Determining Doseeffect Relationships

In radiobiologic studies of tumor, dose-response curves are frequently constructed based on the conditions of the study. For in vitro studies of cell survival this is not a difficult task, but in human studies of these radiopharmaceuticals (called "unsealed sources" of radiation by the U.S. Nuclear Regulatory Commission) for pain relief, the production of a dose-response relationship is difficult. Tumor and marrow doses are difficult to calculate for several reasons. Intraoss-eous trabecular distribution and thickness are both variable in and around tumors. The thicker the trabeculae, as in osteoblastic lesions, the greater the chance of beta particle absorption, reducing the average distance that the particle can travel. The anatomic distribution of tumor throughout the bone, especially trabecular bone where the largest proportion of any radiopharmaceutical is deposited, is also variable. The biologic half-life of Sr-89, Sm-153 lexidronam, Re-186 etidronate, and Sn-117m pentetate, for which data are available, and by extrapolation for P-32, is prolonged on reactive versus normal bone, and inevitably a mixture of these will occur in any osseous metastatic lesion. The target bone-to-nontarget ratios will differ between different metastatic sites in the same patient. Thus, a number of assumptions are required before radiation dose estimation, and these can lead to quite different results. For example, the published range of dose estimates from Sr-89 to bone metastases has been reported to range between 59 and 611 mGy/MBq (220-2260 rads/mCi). These questions raise the issue of the need for individual lesion dosimetry, if a dose-response curve could be established.

The other part of this conundrum, measurement of response, is perhaps more difficult, as it is rooted in human subjective perceptions of pain. Pain thresholds differ between individuals, making generalizations about a radiation dose which relieves pain not easily reproducible. The crucial question of objective pain measurement has been approached in many ways. Semiquantitative scales, patient diaries, the use of descriptor adjectives, and the visual analog scale (VAS) (a 10-cm line on which the patient is asked to point to where his degree of pain lies, between 0 cm for no pain and 10 cm for the worst pain imaginable) have all been tried. The VAS has a reproducibility of + 20%, but has nevertheless proven to be a reliable tool. The amelioration of pain may occur if the patient reduces his activities of daily living (ADL), for example, going to bed, or if the patient increases his analgesic dose. Therefore measurements of the efficacy of a drug in reducing pain must also take into account changes in ADL and medication. In most studies employing the radiopharmaceuticals of Table 1, this important methodologic approach has not been taken. The Utrecht group has demonstrated the value of a tripartite scale which requires an examination not only of pain reduction, but also ADL, and medication changes. This approach should be applied to all future research in the area.

0 0

Post a comment