Interferon alpha (IFN) is used for the adjuvant treatment of melanoma as high-dose therapy (according to Kirkwood, et al.1), low-dose therapy (3 × 3 million international units subcutaneously per week2) and intermediate-dose therapy.3 The results of clinical trials are inconsistent with regard to a possible survival benefit associated with this therapy, in particular for low-dose and intermediate doses.1, 3, 4 Only a minority of patients might benefit, and markers to identify such patients are missing. It is well known that autoimmune diseases can be triggered by IFN therapy, such as lupus erythematosus,5 vitiligo,6 autoimmune thyroid disease (AIT),7 autoimmune thrombocytopenia,8, 9 antiphospholipid syndrome,10 polymyositis,11 diabetes mellitus,12 autoimmune hepatitis13 and rheumatoid arthritis.14 A recent study demonstrated that melanoma patients developing autoantibodies or clinical manifestations of autoimmunity during adjuvant high-dose IFN treatment had a significant survival benefit.14 Thus, the development of autoimmune phenomena might present a marker to identify patients who benefit from adjuvant IFN in melanoma. We hypothesized that autoimmune diseases are also a marker to identify melanoma patients who benefit from adjuvant low-dose IFN. Therefore, we retrospectively reviewed 134 consecutive melanoma patients from our Department who received an adjuvant low-dose IFN therapy. AIT represented the autoimmune disease most commonly triggered by low-dose IFN, it occurred in around 15% of patients. Patients developing AIT had an improved recurrence free survival (RFS) and overall survival (OS). Thus, we provide evidence that patients who develop autoimmune disease during adjuvant treatment with low-dose IFN have a prognostic benefit.
Interferon alpha is used for the adjuvant treatment of malignant melanoma at different dosages (high-, intermediate-, low-dose therapy). Only a minority of patients might benefit from this therapy, and markers to identify such patients are missing. A recent study suggested that melanoma patients developing autoantibodies or clinical manifestations of autoimmunity during adjuvant high-dose interferon alpha treatment had a significant survival benefit. We retrospectively reviewed 134 melanoma patients from our institution treated with adjuvant low-dose interferon alpha therapy and correlated the development of autoimmune diseases with prognosis. Interferon (IFN) therapy was routinely monitored by history, physical examination and laboratory tests before, after the first month and then after every 3 months of therapy. During a median follow up of 46.0 months (8.5–79.0 months) 28 patients (20.9%) suffered from recurrences and melanoma related deaths occurred in 16 patients (11.9%). In 20 patients (14.9%) autoimmune thyroiditis (AIT) was diagnosed during IFN therapy, one of these 20 patients developed rheumatoid arthritis later while continuing IFN therapy. Other autoimmune diseases were not observed. In 2 patients (one with AIT and one with arthritis) the autoimmune disease led to discontinuation of IFN therapy, in the other patients AIT remained subclinical or responded well to treatment while IFN therapy was continued. Kaplan–Meier analyses revealed a significant better recurrence free survival and a trend for a better overall survival for patients with AIT. Thus, autoimmunity triggered by low-dose IFN therapy appears to indicate an improved prognosis and should encourage continuation of IFN therapy. © 2007 Wiley-Liss, Inc.
Patients and methods
We retrospectively analyzed patients from our Department who received adjuvant IFN treatment for malignant melanoma. Patients had to match the following criteria:
- 1Surgical removal of primary melanomas was performed by wide local excision with a safety margin of 2 cm. In addition, sentinel lymph node (SLN) biopsy was performed after detection of the SLN by lymphoscintigraphy and patent blue dye as described previously.15
- 2At the time of melanoma diagnosis, there was no clinical or radiological evidence of melanoma metastasis (initial staging was performed by chest X-ray, ultrasound of abdomen and lymph nodes and S-100 melanoma marker in the serum, further investigations were performed only to clarify unclear reports or in case of specific complaints). In case of positive SLN, completion lymph node dissection (CLND) of the involved region(s) was recommended.
- 3There were no contraindications for IFN therapy, such as history of autoimmune diseases, severe cardiac disease (e.g. NYHA functional class III or IV, myocardial infarction within 6 months, ventricular tachyarrhythmias, unstable angina), severe liver or renal disease, seizure disorders requiring anticonvulsant therapy, pregnancy or lactation.
- 4The age at initiation of therapy ranged between 18 and 75 years.
- 5IFN therapy was initiated between January 2000 and December 2005.
Adjuvant therapy and follow up
Adjuvant therapy with IFN at low-dose (3 × 3 million international units subcutaneously per week with either IFN alpha 2a or IFN alpha 2b) was initiated within 2 months after the last surgery, i.e., after melanoma excision and SLN procedure or after CLND. Therapy was offered to patients with primary melanoma thickness >1.5 mm and/or positive SLN, since it is licensed in Germany for these patients. Some of the patients were enrolled in clinical trials, and depending on these trials the projected duration of treatment was 18–24 months.
Twenty-one patients received IFN alpha 2b while participating in the EADO 2001/CMII trial (Randomized, multicenter Phase III trial comparing adjuvant treatment with PegIntron over 36 months vs. reference treatment with IntronA over 18-months in cutaneous melanoma patients AJCC stage II). This trial recruited patients with a tumor thickness of 1.5 mm or higher without macrometastases regardless of micrometastases in the SLN (stage II of the AJCC staging system for cutaneous melanoma, version 1992). Nine patients received IFN alpha 2a while participating in the DeCOG trial “Randomized, multicenter, open label study to compare the efficacy and tolerability of pegylated IFN-alpha-2a in patients with malignant melanoma in stages IIA (T3b)-IIIB (AJCC-2002).”
IFN therapy was routinely monitored by history, physical examination and laboratory tests [including differential blood count, kidney function tests (creatinine, blood urea), liver function tests (GOT, GPT, gammaGT), thyroid function tests (fT3, fT4, TSH), creatine kinase (CK), alkaline phosphatase (AP), bilirubine, lactate dehydrogenase (LDH), S-100] before, after one month and then after every 3 months of therapy. In addition, radiological follow-up investigations were performed according to the guidelines of the DeCOG of the German Society of Dermatology,16i.e., chest X-ray and ultrasonography of lymph nodes and abdomen every 6 months. Further investigations were performed only in case of specific complaints or suspicious laboratory or radiologic results.
Follow up examinations and laboratory tests for patients participating in the clinical trials mentioned earlier were similar apart from the first 3 months of therapy, where control intervals were more frequent for patients in the EADO 2001/CMII trial (4 weeks, 8 weeks) and the DeCOG trial (2 weeks, 4 weeks, 8 weeks). Thus, the consistency of follow up is similar in patients treated within trials and ad hoc, apart from the first 3 months of treatment.
Thyroid function tests were performed using the Immulite system (DPC Biermann, Bad Nauheim, Germany), with normal ranges for free T3 (fT3) from 1.8 to 4.2 ng/l, for free T4 (fT4) from 8.0 to 19.0 ng/l, and for thyroid stimulating hormone (TSH) from 0.4 to 4.0 mIU/l. In case of abnormalities of thyroid function tests, patients were referred to specialists (nuclear medicine or internal medicine) for further diagnosis and treatment.
For determination of melanoma tumor marker S-100B, the Sangtec® S-100B ELISA was used (DiaSorin, Dietzenbach, Germany). This ELISA has a detection limit of 0.03 ug/l, values ≥ 0.15 ug/l were regarded as increased.
Kaplan–Meier and Levene's analyses were performed with the software SPSS 13.0. In case of a p-value < 0.05, the result was considered as statistically significant.
Characterization of study population
One hundred eighty-one consecutive melanoma patients from our Department received adjuvant IFN therapy and met the criteria stated in Patients and methods. Fifteen patients received high-dose IFN therapy, 28 patients received therapy with pegylated IFNs and 138 patients received low-dose IFN therapy. Of the latter 138 patients, 4 patients did not adhere to the follow up recommendations and were therefore excluded from further evaluation. Details of the remaining 134 patients treated with low-dose IFN who were available for this study are given in Table I. In 71 patients, the SLN was positive by histology or immunohistology, CLND of the involved region(s) was performed in 68 patients, 3 patients refused this operation.
|Type||Superficial spreading||51 (38.1)|
|Breslow thickness||Mean||2.7 mm|
|Clark Level||III||44 (32.8)|
|Sentinel lymph node/AJCC classification/S-100 tumor marker|
|SLN status||Negative||63 (47.0)|
|AJCC classification (Ref.17)||IB||20 (14.9)|
|Initial S-1001||Mean||0.088 ug/l|
|Patients with S-100 ≥ 0.15 ug/l||16 (12.3)|
|Follow up||Mean||44.2 months|
|Recurrence||No recurrence||111 (79.1)|
|First recurrence site||Locoregional cutaneous||14 (50.0)|
|Regional lymph nodes||3 (10.7)|
Performance and side effects of interferon therapy
Treatment was performed either with IFN alpha 2a (n = 63) or IFN alpha 2b (n = 71). In 18 patients, the dosage of IFN was reduced from 3 × 3 million international units to 3 × 1.5 million international units due to side effects (in 5 patients because of increasing liver function tests, in 13 patients due to ongoing subjective complaints such as flu-like symptoms, myalgia, fatigue and depression). Duration of treatment was planned for 18–24 months. The median treatment duration was 20 months (1–54 months). In 38 patients, IFN therapy was stopped earlier than planned because of melanoma recurrence during therapy (23 patients), side effects (subjective complaints in 7 patients, thyrotoxicosis, increasing liver function tests, apoplex, absolute arrhythmia, seronegative rheumatoid arthritis, ocular central vein thrombosis and depression in one patient, respectively) or other reasons (one patient died during an operation).
With regard to autoimmune diseases, in 20 patients (14.9%) AIT was diagnosed during therapy with IFN (detailed in Table II). Of these 20 patients, 15 were female and 5 male. Elevation (patients 1–7) or suppression (patients 8–20) of the TSH level was usually the first sign of thyroid disease noted during the follow up and leading to further laboratory studies of autoantibody titers. In most patients, antibodies directed against thyroglobulin and/or thyroid peroxidase were detected (Table II). Only 3/20 patients had antithyroid receptor antibodies, in addition to antithyroglobulin and antithyroid peroxidase antibodies (patients 12, 17, and 19; Table II).
|Patient number||Sex/age (years)||Type IFN||Duration therapy||fT3||fT4||TSH||TPO-Ab (MAK)||TAK||TSH-R-Ab (TRAK)||Clinical course|
|1||F/49||a||6||n||n||6.22||↑||↑||n||Euthyreot while IFN was continued for a total of 32 months, antibody titres decreased to normal within 12 months after cessation of therapy|
|2||M/75||a||6||n||0.8||8.22||↑||n.d.||n||Euthyreot with 75ug/day L-thyroxine while IFN was continued for a total of 18 months|
|3||M/45||a||13||n||n||10.53||n||↑||n||Decrease of TSH and TAK antibody within 11 months while interferon treatment was continued|
|4||F/23||a||14||n||n||4.90||↑||n||n||Euthyreot while IFN was continued for a total of 54 months|
|5||F/47||b||4||n||n||41.80||↑||↑||n||While IFN was continued for a total of 21 months, hypothyreosis developed and was treated with L-thyroxine|
|6||F/51||b||6||n||n||6.48||↑||n||n||Euthyreot while IFN was continued for a total of 24 months|
|7||M/61||a||1||n||n||5.40||↑||↑||n||Euthyreot while IFN was continued for a total of 24 months|
|8||F/38||a||5||n||n||0.12||↑||↑||n||Euthyreot while IFN was continued for a total of 24 months|
|9||F/52||b||1||n||n||0.15||↑||n.d.||n||Euthyreot while IFN was continued for a total of 24 months|
|10||F/55||b||2||n||n||<0.01||n||↑||n||Euthyreot while IFN was continued for a total of 6 months (then discontinued due to alopecia and depression)|
|11||F/70||a||1||n||n||0.16||↑||↑||n||Euthyreot while IFN was continued for a total of 24 months|
|12||F/60||b||1||4.7||23.8||<0.01||↑||↑||↑||Carbimazol normalized thyroid function within 4 months, IFN was continued for 18 months while autoantibody titres decreased, no endocrine orbithopathia developed|
|13||F/46||a||4||n||n||0.02||↑||↑||n||T4 below normal after 7 months of IFN, L-thyroxine normalized thyroid function tests, but autoantibody titres further increased during therapy|
|14||F/35||a||17||n||n||<0.01||↑||↑||n||Euthyreot while IFN was continued for a total of 24 months|
|15||F/45||b||2||7.2||30.4||<0.01||n||↑||n||Thiamazol normalized thyroid function, discontinuation of IFN after 17 months due to development of seronegative rheumatoid arthritis|
|16||F/24||b||6||n||n||<0.01||n||↑||n||Euthyreot while IFN was continued for a total of 24 months|
|17||M/60||b||1||13.7||73.0||<0.01||↑||n.d.||↑||Discontinuation of IFN after 2 months due to thyrotoxicosis, normalization of thyroid function by Thiamazol within 4 weeks|
|18||F/44||a||3||n||n||<0.01||↑||↑||n||Euthyreot while IFN was continued for a total of 24 months|
|19||M/69||a||5||n||n||0.16||↑||↑||↑||Euthyreot while IFN was continued for a total of 18 months|
|20||F/65||b||2||n||n||0.17||n||↑||n||Euthyreot while IFN was continued for a total of 18 months|
In the majority of patients, AIT was subclinical at initial diagnosis and remained subclinical while IFN therapy was continued (Table II). Overt thyroid disease was noted in 4 patients at the initial diagnosis of AIT (hypothyroidism in patient 2, hyperthyroidism in patients 12, 15, 17). Two additional patients (patients 5 and 13) developed hypothyroidism during the follow-up. Overt thyroid disease responded well to treatment, and IFN therapy was continued in all patients but one patient (patient 17; Table II). In 17/20 patients AIT occurred during the first 6 months of IFN therapy.
In one of the 20 patients with AIT a seronegative rheumatoid arthritis developed later while continuing IFN therapy (patient 15; Table II). Other autoimmune diseases such as vitiligo or lupus erythematosus were not observed.
Kaplan–Meier analyses revealed a significantly better RFS and a clear trend for a better OS for patients developing autoimmune disease while treated with IFN (Fig. 1). In contrast, the type of IFN alpha used (2a vs. 2b) did not correlate with prognosis (data not shown).
Comparing the 2 groups of patients with and without AIT with regard to prognostic parameters at the initial melanoma diagnosis, patients with AIT showed a trend toward a higher initial S-100 tumor marker value. The mean S-100 value of the AIT group was 0.102 ug/l, 3/19 (15.8%) patients had elevated S-100 levels ≥0.15 ug/l. In the non-AIT group the mean S-100 was 0.086 ug/l (p = 0.079 compared to the AIT group, Levene's test), and 13/111 (11.7%) patients had elevated S-100 levels. Increased initial S-100 (≥0.15 ug/l) was a significant prognostic marker for the whole group of patients [4/16 (25.0%) patients with increased initial S-100 and 10/114 (8.8%) with normal S-100 died of melanoma, p = 0.022, Kaplan–Meier analysis]. However, the 3 patients with increased S-100 in the AIT group (patients 3, 11, 17; Table II) did not suffer a melanoma recurrence, whereas melanoma relapses occurred in 5/13 patients with increased S-100 in the non-AIT group.
No significant differences between the 20 patients with AIT compared to the 114 patients without AIT were found with regard to AJCC classification, and type of IFN used (IFN alpha 2a vs. 2b). The 20 patients with AIT did not participate in the clinical studies referenced in Patients and methods, whereas none of the 30 patients treated within clinical trials developed AIT.
Spontaneous autoimmune disorders, in particular vitiligo, have been associated with an improved prognosis in melanoma patients.18, 19 Autoimmunity induced by therapy with interleukin 2 and/or IFN alpha has also been associated with an improved prognosis in metastatic melanoma20 and metastatic renal cell cancer.21
IFN has been licensed for the adjuvant treatment in melanoma for almost 10 years in several countries, since a (minor) subgroup of patients might benefit from this therapy. A recent study showed that patients who develop autoimmune phenomena or clinically manifest autoimmune diseases during adjuvant high-dose IFN therapy have an improved prognosis and suggested autoimmunity as a predictive parameter to select patients who might benefit from high-dose IFN therapy.14 In the present study, we show that autoimmune diseases triggered by adjuvant low-dose IFN therapy might also indicate an improved prognosis. 20/134 (14.9%) of melanoma patients in our study developed autoimmune diseases, primarily AIT. These patients had a significantly prolonged RFS and a trend toward a better OS as compared to the 114 patients without autoimmunity. Thus, our study extends the findings by Gogas et al.,14 indicating that autoimmunity might be a factor to identify melanoma patients benefiting from adjuvant high-dose as well as low-dose IFN therapy.
Our study showed that AIT is the most common autoimmune disease triggered by adjuvant low-dose IFN in melanoma patients, whereas other autoimmune diseases were rare (one case with rheumatoid arthritis). AIT is a well recognized and the most common autoimmune side effect of IFN therapy in different diseases such as haematologic malignancies,22, 23 viral hepatitis,7, 24 melanoma14 and metastatic renal cell cancer.21 Other autoimmune diseases have been reported less frequently, often sporadically as case reports or small case series.
In accordance with studies in patients with hepatitis C, AIT triggered by IFN in our study was either clinically silent or responded well to therapy. Therefore, IFN therapy could be continued in almost all patients.25
However, the conclusions that can be drawn from our study are limited.
First, in our retrospective study we did not perform an autoantibody-screen before IFN therapy, thus we cannot exclude that preexisting subclinical AIT was exacerbated by IFN therapy. Preexisting antithyroid antibodies were detected in up to 11.3% of the white adult population without known thyroid disease26 and preexisting antithyroid antibodies have been described as major risk factor for AIT in patients treated with IFN for hepatitis C.27 Furthermore, we cannot exclude that other, clinically silent autoimmune diseases were triggered by IFN, such as lupus erythematosus.
Second, AIT developed throughout the period of IFN therapy, mainly within the first 6 months. However, the first recurrences were also noted within the first few months of IFN therapy and led to discontinuation of IFN therapy. Thus, patients with autoimmune disease might represent a selection of patients with a better prognosis, since patients with an early relapse might discontinue therapy before they develop IFN related autoimmunity. However, in our group of patients we did not find evidence that the AIT patients represent a selection of patients with a better survival based on initial S-100 levels and AJCC classification status.
Therefore, larger prospective studies are needed to support our findings. Evaluations in this line are currently performed for EORTC 18891 and EORTC 189523 trials, that investigated the use of pegylated IFN and intermediated dose IFN, respectively, in the adjuvant therapy of melanoma patients. At present, we conclude that melanoma patients who develop an AIT during adjuvant low-dose IFN therapy appear to have an improved prognosis. On the basis of these data, AIT should encourage the continuation of IFN therapy, in particular since it is often clinically silent or responds well to therapy.
Authors thank Mr. Heinz Geerlings for his support in the statistical analysis. A.H. got study grants from Schering-Plough and Roche. A.H. acted as an advisor/consultant for Schering-Plough and Roche. R.G. received research grants from Essex Pharma and Roche. R.G. served as paid speaker for Essex Pharma and Roche.