ORIGINAL 
Annals of Nuclear Medicine Vol. 5, No. 2, 65-68, 1991 
Limited value of delayed radiothallium image in the diagnosis of nodular goiter 
Takashi MISAKI,* Kanji YAGI,* Shigeharu DOKOH,* Toshikiyo KOH** and Shin-ichiro SHIMBO** 
Departments of *Radiology and **Internal Medicine, Kyoto City Hospital, Kyoto 
To test whether delayed thallium image can improve the diagnostic accuracy of thyroid tumor, we have reviewed 35 patients with cold nodules of proven histopathology. Early and delayed images were taken 10-20 min and 3-4 hr postinjection, respectively. In this study, a thyroid nodule was interpreted as malignant if it had normal or increased uptake on early scan and more residual radioactivity than paranodular tissue on delayed scan. In the eighteen patients who had carcinoma, a disappointing 44 % false negative rate was observed. No significant difference was seen between retention-positive and negative cases in tumor size or histological type of cancer. Nor could any demographic bias explain the low sensitivity. Of the six false negatives, macroscopic cystic degeneration was seen in two cases, but tumors in the other four were grossly solid. Considering the histological heterogeneity often seen within a thyroid tumor, the portion with increased retention of radiothallium may be too small to be detected in the false negative cases. Furthermore, 3 false positive readings were obtained in 17 patients with benign conditions. We conclude that the contribution of the delayed thallium image was quite limited in predicting or ruling out malignancy in nodular goiters. 
Key words : thallium-201, tumor scintigraphy, thyroid cancer, thyroid adenoma, delayed image

INTRODUCTION 
THALLIUM-201 chloride is one of the most widely used radiopharmaceuticals, serving both in myocardial imaging and in tumor scintigraphy. In the diagnosis of nodular goiter, many authors have reported the accumulation of this radionuclide in thyroid cancers and their lymphnode metastasis.1-7 Meanwhile, since a significant fraction of adenomas or other benign conditions also appear positive in the scintigram, a single radiothallium image could not be a satisfactory test to use in estimating the 
Received February 4, 1991, revision accepted April 3, 1991. For reprints contact : Takashi Misaki, Department of Radiology, Kyoto City Hospital, Mibuhigashitakada, Nakagyoku, Kyoto 604, JAPAN. 
histological nature of a given struma nodosa. The only general consensus in the literature appears to be that a thyroid nodule photopenic in a thallium study is most likely to be a benign cystic one. To overcome this poor specificity, some researchers have claimed that a delayed thallium image taken three to five hours after i.v. can improve the differentiating accuracy,8 whereas others have been more pessimistic about the value of this additional scan.9,10 In the past, while routinely taking both early and delayed thallium scintigrams for thyroid nodules, we have gradually become doubtful of the usefulness of this dual procedure. We have therefore retrospectively reviewed the cases with proven histological diagnosis to address this issue statistically. The consistency between radionuclear findings and histopathology was closely examined with particular attention to the contribution of the delayed image. 

MATERIALS AND METHODS 
Patients 
Thirty-five cases of untreated "cold nodules" in the thyroid were analysed in the present study. The tumorous lesions all appeared as trapping defects in the iodine-123 and technetium-99m scintigrams. Patients were excluded from the study if they had a history of previous thyroid surgery, or scintigraphically diffuse goiter, even if they showed nodularity on palpation. Histological diagnoses were made by two expert pathologists on surgical specimens. The whole patient group consisted of 6 men and 29 women, aged 20-85 (mean 50.2) years. 
Thallium scintigraphy 
Early and delayed images were taken 10-20 min and 3-4 hr, respectively, after i.v. injection of 74 MBq (2 mCi) of thallium-201 chloride (Nihon Mediphysics, Takarazuka, Hyogo). The scinticamera used was a Searl Pho/Gamma LFOV equiped with a pinhole collimator, with a window setting of 80 KeV+-15 % and a preset count of 300 K. Interpretation of obtained scintigrams was done visually by a panel of three experienced specialists in nuclear medicine, and was unanimous on all occasions. To interpret early and delayed scintigrms combained, we followed the method of Ochi and colleagues:8 A thyroid nodule was consldered malignant if it had normal or in-creased uptake on early scan and more residual radioactivity than paranodular tissue on delayed scan; a nodule without these findings was regarded as scintigraphically benign. 

RESULTS 
Table 1 summarizes the correlation between the diagnosis based on dual thallium imaging and final histopathology in the 35 subjects examined. In the eighteen patients who had carcinoma, a 44% false negative rate was observed. No demographic bias could explain this low sensitivity, since there was little difference between scintigraphically "benign" and "malignant" groups as to age and sex. The same applies to the size of nodular lesions in these two groups. As for the histological type of malignant tumors, 4 out of 6 cases of follicular adenocarcinoma (FAC) were read as positive in the scintigrams, whereas the odds were 6 out of 12 for papillary adenocarcinoma (PAC). No other type of malignant goiter was observed in the present series. Of the six false negatives, macroscopic cystic degeneration was seen in two cases, one with FAC and another with PAC, but the other four were grossly solid. Histological diagnoses in the three false positives were hyperplasia (adenomatous goiter) for two and follicular adenoma for one; in the 14 true negatives, follicular adenoma for 4, adenomatous goiter for 8, and Hashimoto thyroiditis for 2. Representative cases of false positive and false negative are shown in Figs. I and 2, respectively. There were six cases in which combined reading of early and delayed scans differed from that of the early image only. The former gave correct radionuclear diagnoses in a false negative and two false positive cases of the latter, but was wrong in a true negative and two true positives. Hence, by adding delayed scan, specificity was slightly improved (from 76 % to 82 %) at the expense of sensitivity (from 61 % to 56 %), and overall accuracy remained the same at 69 %.

DISCUSSION 
Contrary to some enthusiastic earlier reports, our present study showed discouragingly low sensitivity of the delayed radiothallium scan in differentiating malignant thyroid tumors from benign conditions. The false negative findings were solely attributable to neither cystic necrosis of large tumors nor to microinvasive FAC. In fact, there were 5 cases of PAC without major cystic degeneration. We postulate that in those patients, partial vascular encasement by the tumors might have caused sufficient regional mild hypoperfusion to decrease the accumulation of radiothallium but still insufficient to necrotize the neoplastic tissue. Alternatively, as for the ability to accumulate thallium chloride, there may be two or more subsets within the "PAC" category. Concerning this point, it should be remembered that in current histopathology, thyroid carcinoma with both papillary and follicular components is classified as PAC.11,12 Furthermore, there often exists histological heterogeneity within a given thyroid tumor so that most of the mass appears cytologically benign while a small fraction of it is judged as malignant by pathologists. In these circumstances, it is no wonder that the nodule in question cannot retain enough radiothallium to be "positive" in a delayed scan, since the portion with increased ability to concentrate the radionuclide may be too small to be detected. 
Considering these facts, dynamic studies with thallium-201 13,14 seemed unlikely to be a fundamental solution for the low sensitivity. From the present analysis, we conclude that the contribution of the delayed thallium image was quite limited in forecasting malignancy in nodular goiters, and judgements based on its findings were sometimes even erroneous. For radionuclear evaluation of lymphnode metastasis, intrathoracic extension and cystic changes, early thallium image alone may be enough.6,7 Instead of an additional scintigram with inconvenient timing for patients and decreased patient throughput, other means, especially ultrasonography7 and fine needle aspiration biopsy,15 should be included in the initial diagnostic workup for nodular thyroid lesions, at least for now. Hopefully, future advances in tumor-specific radioimmunoscintigraphy or other nuclear medicine techniques will provide better presurgical prediction of the histology of thyroid nodules. 

ACKNOWLEDGEMENTS 
The authors are very grateful to the pathologists in our hospital, Drs. K. Takasu and Y. Kim, for their invaluable, cooperation and helpful discussion concerning the histopathology of nodular goiter; to Drs. J. Konishi. K. Kasagi. Y. Iida, and A. Hidaka of the Department of Nuclear Medicine. Kyoto University School of Medicine for their critical ccmments and warm encouragement.
 
REFERENCES 
1. Fukuchi M, Tachibana K, Kuwata K, et al : Thallium-201 imaging in thyroid carcinoma-appearance of a lymph node metastasis. J Nucl Med 19 : 195-196, 1978 
2. Tonami N, Bunko H, Michigishi T, et al: Clinical application of 201Tl scintigraphy in patients with cold thyroid nodules. Clin Nucl Med 3 : 217-221, 1978 
3. Palermo F, Bruniera F, Caldato L, et al: 201Tl in the scintigraphic evaluation of the "cold" thyroid areas. Eur J Nucl Med 4 : 43-48, 1979 
4. Harada T, Ito Y, Shimaoka K, et al: Clinical evaluation of 201Thallium chloride scan for thyroid nodule. Eur J Nucl Med 5: 125-130, 1980 
5. Hdschl R. Murray IPC, McLean RG, et al: Radio-thallium scintigraphy in solitary nonfunctioning thyroid nodules. World J Surg 8 : 956-962, 1984 
6. Makimoto K, Ohmura M, Tamada A, et al: Combined scintiscans in the diagnosis of thyroid carcinomas. Acta Otolaryngol Suppl 419: 189-194, 1985 
7. Ikekubo K, Higa T, Hirasa M, et al: Evaluation of radionuclide imaging and echography in the diagnosis of thyroid nodules. Clin Nucl Med 11 : 145-149, 1986 
8. Ochi H. Sawa H, Fukuda T, et al: Thallium-201-chloride thyroid scintigraphy to evaluate benign and/ or malignant nodules : Usefulness of the delayed scan. Cancer 50 : 236-240, 1982 
9. Henze E, Roth J, Boerer H, et al: Diagnostic value of early and delayed 201Tl thyroid scintigraphy in the evaluation of cold nodules for malignancy. Eur J Nucl Med 11 : 413-416 1986 
10. Bleichrcdt RP, Vermey A, Piers DA, et al: Early and delayed thallium 201 imaging: Diagnosis of patients with cold thyroid nodules. Cancer 60: 2621-2623, 1987 l 1. The Japan Society of Thyroid Surgery: General rules for the description of thyroid cancer, Tokyo, Kane-hara & Co., 1983 
12. LiVolsi VA: Pathology of thyroid disease. In Thyroid disease : Endocrinology, surgery, nuclear medicine, and radiotherapy, Falk SA, (ed.), New York, Raven Press, pp. 127-175, 1990 
13. Tennvall J, Palmer J, Cederquist E,, et al: Scintigraphic evaluation and dynamic studies with thallium-201 in thyroid lesions with suspected cancer. Eur J Nucl Med 6: 295-300, 1981 
14. Palermo F. Saitta B, Coghetto F, et al : Computerized study with 201Tl of the cold thyroid node. Nucl Med 21 : 23-29, 1982 
15. Hamburger JI: Getting the most out of fine needle biopsy. In Thyroid disease : Endocrinology, surgery, nuclear medicine, and radiotherapy, Falk SA, (ed.), New York, Raven Press, pp. 361-368, 1990