metastases, is well documented (41,87,117,120,125,126,129,130,254-256, 260). In only one case (No. 72) was there evidence of cervical lymph node metastases from a clinically occult papillary carcinoma (see Figure 4). Lymph node metastases note in Table 4 were all limited to regional cerviThese has been no evidence of distant metastasis in any of the cal nodes. Marshallese, nor has there been local recurrence of any resected carcinoma. Other forms of thyroid malignancies, such as medullary, squamous, anaplastic, or sarcomatous, have not been observed. Table 5 summarizes the distribution of these tumors among exposure The prevalence of these lesions within the study population subsets groups. has been discussed in Section III.1.b. No specific type of histologic lesion appeared directly attributable to radiation, since they occurred in similar relative proportions in both exposed and unexposed groups, but histologic and cytologic changes consistent with radiation injury were observed in some of the exposed individuals. These included the postirradiation interstitial fibrosis, lymphocytic thyroiditis, and oxyphilic changes noted by other investigators (257,258). In a number of instances, more than one diagnosis applied to a single pa- tient. Most of the specific lesions, such as papillary carcinomas, occurred in a local milieu of adenomatous nodularity, but the latter was recorded as a separate diagnosis only when it was judged to be a significant contributor to the palpable nodularity. When carcinoma was found, it was not always the clinically suspicious nodule that originally prompted surgery, a circumstance experienced by a number of other investigators (148,153,257). Favus et al. (153) reported the frequent coexistence of adenomatous hyperplasia and radiogenic papillary or mixed carcinomas. As noted in Table 4, the present study included numerous cases in which prominently hyperplastic foci, often with papillary configurations, occurred within adenomatous nod- In two of these cases (Nos. 8 and 2221), focal atypia was also noted. ules. Otherwise, there was no histologic evidence to suggest that such hyperplastic areas might eventually progress to carcinoma, especially since they were also The reasons commonly observed among the nonirradiated control populations. for such highly localized, exuberant responses to TSH stimulation remain unclear, however, and radiogenic injury cannot be ruled out as one potential cause. Studies of children with thyroid exposures of <1000 rads led Spitalnik and Straus to suggest that focal hyperplasia in such cases might represent a premalignant alteration (258). Many of the references cited indicate that radiation-induced papillary or mixed carcinomas of the thyroid usually represent low-grade malignancies with relatively little or no clinical significance in terms of patient morbidity or mortality. Factors that appear to influence prognosis adversely in- clude nonpapillary histology, size of the primary, local extent, local recurrence rate, and distant metastases, but not regional lymph node metastases. The single most important adverse prognostic indicator may be onset at more than 40 to 45 years of age (41~44,45,118,119,121-123,125). Of the fourteen Marshallese with either occult or overt papillary carcinomas, five were in the Of the five with cerviolder age group, three of whom were radiation exposed. cal node metastases, one (No. 3042) was in the unexposed comparison population from Utirik. Since the preponderance of cases among the Marshallese was in the younger age group, and all were papillary carcinomas devoid of local recurrence or - 67-