Many papers have appeared during the past 10 years calling attention to the high incidence of cancer among nontoxic nodular goiters that are surgically treated. Different authors have presented their data in somewhat different ways. Most have included patients with known cancer, patients suspected of having cancer, and patients thought to have benign goiter in a single group and have given the percentage of cancer found in resected specimens for the entire group. Many have indicated that cancer was found much more commonly in thyroid glands with single nodules than in multinodular glands.1 Others have pointed out, however, that it is very difficult to determine clinically just how many nodules there are in a particular gland and that preoperative classification on this basis is of relatively little value.2 Some have stated that nodular goiters are more likely to be cancerous in men than in women.3 All authorities appear to agree that cancer is not common in toxic goiter.Many authors have made the mistake of applying surgical statistics to the population at large, without con¬ sidering whether their patients constituted a representa¬ tive sample of the population. They have assumed that, if 5% of nontoxic nodular thyroid glands that are sur¬ gically treated contain cancer, the same must be true of the glands that are not operated on. Such an assumption is incorrect, of course, since it implies a complete lack of selection of the patients referred for operation. Ac¬ tually, as was recently reemphasized by Crile and Dempsey,4 there is a great deal of selection, both by the patients themselves and by the various physicians who see them prior to surgery. Relatively few nontoxic goiters are operated on. These are selected by a process that statistically favors cancer and tends to exclude asympto¬ matic benign nodules. The extent of this selection-and hence the correction that must be used to make the surgical statistics applicable to the population at largeis not known for most communities. Crile and Dempsey estimated that at least a tenfold concentration of cancer had been achieved in the cases in which operation was done at the Cleveland Clinic.In toxic goiter, however, the situation is quite differ¬ ent. Until relatively recently, surgery was the only ther¬ apy available for hyperthyroidism and it is still the most commonly used treatment for this disease. The surgical experience with toxic goiter, therefore, includes the great majority of all cases in which this diagnosis was made.There is no reason to believe that exclusion of the mi¬ nority that was not operated on has affected the statistics on cancer incidence more than slightly. The surgical data may thus be accepted as valid for all toxic goiter in this regard. It is the purpose of this paper to review some of the American statistics on goiter and on thy¬ roid cancer, to call attention to some of the erroneous conclusions that have resulted from the study of un¬ representative samples, and to attempt to estimate the incidence of malignancy in an unselected gr...
