Simon Hoelzer1, Dagmar Steiner2, Richard Bauer2, Christoph Reiners3, Jamshid Farahati3, Scott ..... Grant CS, Hay ID, Gough IR, Bergstralh EJ, Goellner JR,.
Original article Current practice of radioiodine treatment in the management of differentiated thyroid cancer in Germany Simon Hoelzer1, Dagmar Steiner2, Richard Bauer2, Christoph Reiners3, Jamshid Farahati3, Scott A. Hundahl4, Joachim Dudeck1 1 Institute
of Medical Informatics, Justus-Liebig-University Giessen, Germany for Nuclear Medicine, Justus-Liebig-University Giessen, Germany 3 Clinic and Policlinic for Nuclear Medicine, University of Würzburg, Germany 4 Commission on Cancer, American College of Surgeons, Chicago, USA 2 Clinic
Received 15 April and in revised form 23 June 2000 / Published online: 17 August 2000 © Springer-Verlag 2000
Abstract. This prospective, observational study of a cohort of thyroid cancer patients in Germany focusses on the “real-world” practice in the management of thyroid cancer patients. This report includes data from 2376 patients with primary differentiated thyroid carcinoma first diagnosed in the year 1996. The study reveals considerable differences in actual practice concerning surgery and radioiodine treatment. The results indicate that consensus is lacking with respect to the multimodality treatment approach for differentiated thyroid carcinoma. Our analysis represents the most current and comprehensive national assessment of presenting patient characteristics, diagnostic tests, treatment and complications for thyroid cancer. Key words: Differentiated thyroid cancer – Guidelines – Radioiodine – Treatment Eur J Nucl Med (2000) 27:1465–1472 DOI 10.1007/s002590000333
Introduction Because of the rarity and the comparatively low mortality of differentiated thyroid carcinoma, with a 10-year overall survival rate of more than 85% [1, 2], analyses of outcomes in controlled, randomised, clinical trials are impractical and consequently unavailable. Recommendations for diagnostic and therapeutic interventions in differentiated thyroid cancer are mainly based on the analysis of retrospective, non-controlled studies [3, 4, 5, 6, 7, 8, 9]. Such studies often address patient cohorts of a certain geographic region treated in the distant past. This Correspondence to: S. Hoelzer, Institute of Medical Informatics, Heinrich-Buff-Ring 44, 35392 Gießen, Germany
European Journal of Nuclear Medicine Vol. 27, No. 10, October 2000
focus on specialised clinical settings and atypical or insufficiently diverse patient populations means that the findings are of limited external validity. Moreover, comparing a contemporary intervention with such historically treated groups may overestimate the effectiveness of the former because of the improvements in outcome associated with general improvements in medical care. Given the rapid evolution of medical knowledge and technology, the need for up-to-date information is obvious. Contested areas in the management of thyroid carcinoma are especially the extent of the applied surgery and the most appropriate use of radioactive iodine. Consensus and expert-based guidelines for the treatment of differentiated thyroid cancer are available in Germany, but the actual practice, e.g. the surgical approach, the proportion of patients treated with radioiodine, the activity of radioiodine administered and the percentage of shortterm adverse effects, remains unknown. In order to answer some of these questions, a Patient Care Evaluation Study (PCE study) has been carried out since the year 1996. PCE studies are prospective observational studies in oncology that have been developed by the Commission on Cancer (CoC) of the American College of Surgeons (ACoS) in the United States. One of the major functions of this organisation is to develop criteria for the periodic evaluation of the care (diagnosis, treatment and follow-up) of patients with specific cancer diseases. PCE studies are based on patient-oriented, treatment-accompanying tumour documentation that is compiled in all medical facilities involved in cancer care. In this way, detailed information is made available on patient characteristics, extent of disease and course of disease with reference to the applied medical interventions. This allows a statistical, mostly explorative analysis of factors that influence the clinical decision-making process, the use of available diagnostic and therapeutic modalities for certain medical conditions and the consequences of cur-
1466
rent practice. In the year 1996, two parallel PCE studies of thyroid cancer were started in Germany and the US. This project addresses the present concerns with respect to the diagnosis and treatment of these patients. Both studies are still ongoing, but sufficient information is now available to characterise the first course of treatment and proxy outcomes. In this paper we focus on the current role of radioiodine treatment in the management of papillary and follicular thyroid carcinoma. Materials and methods In a prospective, multicentre, observational study with free choice of treatment, patients with primary thyroid cancer were entered free of selection between 1 January 1996 and 31 December 1996 in Germany. The methods of the PCE studies have been described previously [10, 11, 12, 13, 14, 15]. The data on patient and disease characteristics, applied diagnostics, treatment and follow-up were collected prospectively by the involved physicians in the fields of surgery, pathology, internal medicine, endocrinology, radiology, radio-oncology, and nuclear medicine under the co-ordination of regional cancer centres and tumour registries. The used structured data form was specially designed for thyroid cancer. This questionnaire was developed by a subcommittee of the CoC’s National Cancer Data Committee. Patient questionnaires were not used for this PCE study. In order to fit the questionnaire to the requirements and demands of the interdisciplinary expert panel in Germany, extensions of the data form were integrated by members of the involved medical societies. More detailed information was requested concerning the diagnostic approach, the chronological sequence of the treatment modalities, the radioiodine uptake test, the activity of radioiodine administered and the complications of the different treatment options. Since these extra data are not included in the US data form, we cannot report any comparative results of this parallel study. A comprehensive report including a comparison of the patterns of care in the US and Germany is available (www.patientcare.de). This report, generated from the German PCE study, summarises complex and disparate clinical data as recorded in different hospitals across the country. The interpretation should take into account the complex nature of the data classification and coding and the circumstances characterising the data sources. These are descriptive survey data and were not collected to prove the efficacy of clinical interventions, as do randomised, controlled clinical studies. We exclusively performed the kind of explorative analysis that the data reasonably support. Thus we carried out a straightforward, stratified and conservative assessment of the data patterns presented. In this context, significance testing (e.g. Pearson chisquare) is only used to support some of the findings. We consider these data suitable for benchmarking, for describing patterns of care and for generating new hypotheses. For stratification, the PCE study concept uses the internationally accepted coding and staging schema, e.g. the ICD-O-2 of the WHO and the TNM classification of the UICC/AJCC (Union International Contre le Cancer/American Joint Committee on Cancer) [16, 17, 18, 19, 20]. For analytical purposes, a “combined” UICC stage grouping is used for tabular summaries which includes the pathological stage (pTNM), where documented, augmented by the clinical stage (cTNM) when the pathological stage has not been adequately recorded (lymph node involvement and
presence of distant metastases). Problems with stage grouping according to the UICC system have been identified. For example, patients with papillary or follicular carcinoma under the age of 45 years without distant metastases are grouped in stage I regardless of tumour size and lymph node involvement. Whereas the UICC stage groups I–IV, which take into account the TNM classification and the patient’s age (under or over 45 years at the time of diagnosis), are seen to have prognostic significance for thyroid cancer, clinical decisions on the most appropriate therapeutic approach rely mainly on the distinct TNM stage, i.e. the size of the primary tumour, the lymph node involvement and the presence of distant metastases. This results in a distribution with at least 12 distinct TNM subgroups (see Table 3). For decision analysis, this grouping of dissimilar cases was addressed by creating alternative groupings of the TNM classifications – T1N0M0 (early localised disease), T2/3N0M0 (advanced localised disease), T1–3N1M0 (regional metastases), T4anyNM0 (advanced primary tumour), TxNxM0, and anyTanyNM1 – taking into account the important clinical parameters on which the treatment decisions are potentially based. The German PCE studies are designed as an external quality assurance programme with local-level performance improvement in mind. Detailed, customised reports are generated of the entire cohort, which subsequently can be used to compare the patterns of care and outcomes of the participating hospitals with the national standard. The following analysis covers the use of nuclear medicine in the management of primary differentiated cancer of the thyroid gland, classified as papillary or follicular carcinoma.
Results Data on the first course of treatment of 1685 patients with papillary carcinoma (66.4%), 691 with follicular carcinoma (27.2%) and 161 with medullary or anaplastic carcinoma (6.4%) have been submitted. The following analysis includes 2376 patients with primary differentiated carcinoma diagnosed in the year 1996. This number accounts for about 80% of all cases in this year in Germany. Patient characteristics and staging The median age at the time of diagnosis was 50 years for papillary and 55 years for follicular carcinoma. Of the papillary carcinomas, 77.7% occurred in females and 22.3% in males (non-standardised ratio of 3.5). The female to male ratio in patients with follicular thyroid cancer was lower than 3:1 (Table 1). Most of the patients with papillary thyroid cancer (59.6%) were diagnosed with UICC stage I disease. Of the patients with follicular thyroid cancer, 38.7% had stage I disease and and 37.2% stage II disease. In 11.3% and 14.3% of the cases of papillary and follicular cancer, respectively, the collected data were not sufficient to permit staging according to the UICC groupings (Table 2). Table 3 shows the distribution of patients among the 12 distinct TNM subgroups for both histologies. The
European Journal of Nuclear Medicine Vol. 27, No. 10, October 2000
1467 Table 4. Distribution of patients among the alternative (combined) TNM subgroups, according to the type of cancer
Table 1. Sex distribution of papillary and follicular cancer Gender
Papillary
Follicular TNM groupings
Male Female Total Median age
No.
%
No.
%
375 1310 1685 50 years
22.3 77.7 100
184 507 691 55 years
26.6 73.4 100
Table 2. Stage distribution of papillary and follicular cancer UICC stage
I II III IV Subtotal known Unknown stage Total a
Papillary cancer
Follicular cancer
No.
%
No.
%
891 320 235 48 1494 191 1685
59.6 21.4 15.7 3.2 100.0 11.3a
229 220 69 74 592 99 691
38.7 37.2 11.7 12.5 100.0 14.3a
Denominator is total cases
Table 3. Distribution of patients among the TNM subgroups, according to the type of cancer TNM stage
T1aN0M0 T1bN0M0 T 1N 0M 0 T 1N 1M 0 T 2N 0M 0 T 2N 1M 0 T 3N 0M 0 T 3N 1M 0 T 4N 0M 0 T 4N 1M 0 T xN xM 0 anyTanyNM1 TNM unknown Total
T 1N 0M 0 T2/3N0M0 T1–3N1M0 T4anyNM0 T xN xM 0 anyTanyNM1 TNM unknown Total
Papillary cancer
Follicular cancer
No.
%
No.
%
397 513 162 265 22 68 258 1685
23.6 30.4 9.6 15.7 1.3 4.0 15.3 100.0
48 325 22 64 13 81 138 691
6.9 47.0 3.2 9.3 1.9 11.7 20.0 100.0
agnostic subgroup was again T2N0M0 disease (30.2%), followed by T3N0M0 disease with 16.8%. Only a small proportion of patients (6.9%) had a tumour size smaller than 1 cm. Distant metastases were shown to be present in 11.7% of the follicular cases overall. In one-fifth of the patients the combined TNM stage remained unknown. For further analysis and stratification, some TNM subgroups were combined into alternative groupings as outlined in the previous section. The frequencies are reported in Table 4.
Papillary cancer
Follicular cancer
Diagnostic procedures
No.
%
No.
%
154 79 164 47 445 106 68 9 136 129 22 68 258 1685
9.1 4.7 9.7 2.8 26.4 6.3 4.0 0.5 8.1 7.7 1.3 4.0 15.3 100.0
21 4 23 4 209 13 116 5 45 19 13 81 138 691
3.0 0.6 3.3 0.6 30.2 1.9 16.8 0.7 6.5 2.7 1.9 11.7 20.0 100.0
Ultrasound of the thyroid and thyroid scan are the most frequently used diagnostic imaging techniques for thyroid cancer in Germany. The results of the diagnostic test performed were reported as either normal, pathological but not suggestive of cancer, suggestive of cancer or unknown. The results were based on the subjective estimations of the investigators. No objective criteria for pathological test results were defined in the study protocol. Table 5 reports the results of thyroid scanning for the investigated histologies. A very high rate (50.5% and 44.7%) of non-specific, pathological test results (abnormal, not suggestive of cancer) is apparent. The thyroid scan seems to be more specific in the detection of follicular cancer but one has to take into account the fact that patients with follicular thyroid cancer are diagnosed at more advanced tumour stages than patients with papillary thyroid cancer, as shown in Tables 2, 3 and 4. Of those patients in whom thyroid ultrasound was suggestive of cancer, 77.0% also had a pathological scan result. On the other hand, 66.1% of patients in whom the thyroid scan was suggestive of cancer also had a positive, suggestive ultrasound result. In more than 70% of the cases, both tests (thyroid scan and ultrasound) yielded identical results (normal, pathological and not suggestive of cancer or pathological and suggestive of cancer). In 10.9% of cases, thyroid scan revealed a pathological
most frequent diagnostic classification of patients with papillary thyroid cancer was T2N0M0 disease (26.4%), followed by T1N0M0 disease (9.7%). More than 15% of the patients with papillary thyroid cancer had a tumour extending beyond the thyroid capsule. A positive lymph node status was often seen in patients with T4 tumours whereas the presence of distant metastases was reported in only 4% of the patients with papillary thyroid cancer. In patients with follicular cancers the most frequent di-
European Journal of Nuclear Medicine Vol. 27, No. 10, October 2000
1468 Table 5. Results of thyroid scan, according to type of cancer Test results
Papillary cancer
Follicular cancer
No.
No.
Test not done 101 Abnormal, 519 suggestive of cancer Abnormal, 666 not suggestive of cancer Normal 32 Subtotal known 1318 Unknown 367 Total 1685 a
%
%
7.7 39.4
38 247
7.1 46.4
50.5
238
44.7
2.4 100.0 21.8a
9 532 159 691
1.7 100.0 23.0a
result whereas ultrasound images were non-specific or normal. By comparison, in only 6.2% of cases did ultrasound permit tumour detection while thyroid scan was negative. The overall specificity of both tests was found to be rather low in the diagnosis of differentiated thyroid cancer. Fine-needle aspiration biopsy was performed in only 624 patients (26.3%). Test results suggesting the presence of malignant cells were reported in 405 cases (65%). Cancer-directed surgery The surgical procedures are described to the extent that they are relevant to the subsequent radioiodine treatment. Table 6 illustrates the type of cancer-directed surgery for papillary thyroid cancer. Only a few patients were treated with a local excision or lobectomy, even if
Denominator is total cases
Table 6. Type of surgery for papillary thyroid cancer Type of surgery
TNM subgroups T 1N 0M 0
T2/3N0M0
T1–3N1M0
T4anyNM0
T xN xM 0
anyTanyNM1
TNM unknown
No.
No.
No.
No.
No.
No.
No.
%
%
%
%
%
%
%
No cancer-directed surgery
4
1.0
3
0.6
0
0
1
0.4
2
9.1
1
1.5
2
0.8
Local excision
14
3.5
0
0
0
0
1
0.4
2
9.1
1
1.5
8
3.1
9
2.3
7
1.4
5
3.1
2
0.8
2
9.1
0
0
10
3.9
Near-total thyroidectomy
142
35.8
108
21.1
19
11.7
38
14.3
7
31.8
10
14.7
56
21.7
Total thyroidectomy without LND
110
27.7
162
31.6
7
4.3
44
16.6
3
13.6
17
25.0
133
51.6
Total thyroidectomy with limited LND
42
10.6
110
21.4
42
25.9
63
23.8
2
9.1
13
19.1
15
5.8
Total thyroidectomy with radical LND
59
14.9
90
17.5
82
50.6
103
38.9
3
13.6
24
35.3
10
3.9
Thyroidectomy, NOS
5
1.3
32
6.2
4
2.5
7
2.6
1
4.5
1
1.5
18
7.0
Surgery for regional/distant sites
1
0.3
1
0.2
2
1.2
1
0.4
0
0
1
1.5
0
0
11
2.8
0
0
1
0.6
5
1.9
0
0
0
0
6
2.3
397
100.0
513
162
100.0
265
100.0
22
100.0
68
100.0
258
100.0
Lobectomy
Surgery, NOS Total
100.0
LND, Lymph node dissection; NOS, not otherwise specified
European Journal of Nuclear Medicine Vol. 27, No. 10, October 2000
1469 Table 7. Frequency with which different therapeutic combinations were used for the TNM subgroups of papillary thyroid cancer
Type of multimodality treatment
TNM subgroups T 1N 0M 0
Surgery, radioiodine and hormone therapy Surgery, radioiodine, ext. beam radiation and hormone therapy Surgery and hormone therapy Surgery alone Others Total
they were diagnosed in the early localised stage (
