Insulin-like growth factor-1 receptor and phosphorylated AKT-serine 473 expression in 132 resected thymomas and thymic carcinomas


  • This article is US Government work and, as such, is in the public domain in the United States of America.



Thymic malignancies are rare tumors. The insulin-like growth factor-1 (IGF-1)/IGF-1 receptor (IGF-1R) system is involved in the development of the thymus. IGF-1R expression in thymic epithelial malignancies is unknown.


The authors investigated the expression of IGF-1R and phosphorylated AKT serine 473 (p-AKT) by using immunohistochemistry and examined the clinicopathologic correlations in a retrospective, single-institution surgical series of 132 patients with thymic epithelial malignancies.


Earlier disease stage, less aggressive histologic types, and complete resection were significant positive prognostic factors for disease-related survival and progression-free survival, and being a woman was a better prognostic factor for disease-related survival. IGF-1R and p-AKT protein levels were expressed in 20% and 36% of thymic tumors, respectively. Both markers were expressed more commonly in recurrent disease than in primary tumors, in more aggressive subtypes, and in more advanced disease stages. There was a trend toward better survival and progression-free survival in patients who were negative for IGF-1R or p-AKT expression in the whole series. When only the 91 primary tumors, IGF1R expression was associated with worse progression-free survival (P < .001).


The current retrospective analysis demonstrated that disease stage, tumor histology, sex, and resection type were major prognostic factors in the survival of patients with thymic malignancies. The expression levels of IGF-1R and p-AKT in thymic tumors suggested that IGF-1R is a potential target for treatment. Cancer 2010. Published 2010 by the American Cancer Society.

Thymic epithelial malignancies (TEMs) are rare tumors with an overall annual incidence of 0.15 per 100,000 population, yet they are the most common anterior mediastinal malignancies in adults, representing 50% of anterior mediastinal masses.1, 2 TEMs can invade through the capsule and infiltrate the surrounding organs and great vessels and, albeit rarely, they can metastasize to distant organs. The epithelial cells represent the tumor cells, whereas the lymphocytes are considered benign infiltrating cells. The World Health Organization (WHO) classifies TEMs into 6 histologic subtypes (A, AB, B1, B2, B3, and C).3 Surgery is the mainstay of localized tumors. However, late recurrences are not uncommon, especially when resection is not complete or in more aggressive histologic types. The major prognostic factors, in patients with TEMs, are disease stage, completeness of resection, and histologic classification.4-6 There is a recognized major difference in prognosis and clinical behavior between thymomas, which are relatively indolent tumors, and thymic carcinomas, which have much more aggressive behavior.

The biology of these rare tumors largely remains unknown, and systemic therapy for advanced TEMs basically has not changed in the last 10 to 20 years and includes platinum-based chemotherapy.7 Chemotherapy, although active, is not curative in patients with advanced-stage disease, and targeted therapies to date have not been successful.6, 8, 9 The development of novel agents is important, especially for the more aggressive tumor types, in which chemotherapy is less effective.

Insulin-like growth factor-1 receptor (IGF-1R) is a tyrosine kinase receptor and a transmembrane, heterotetrameric protein that is encoded by the IGF-1R gene located on chromosome 15q25-q26. IGF-1R reportedly plays roles in promoting oncogenic transformation, growth, and cancer cell survival.10-13 IGF-1R is expressed on multiple immune cell types, including bone marrow lymphocyte precursors, thymocytes, thymic epithelial cells, and mature lymphocytes. By using T-cell–specific IGF-1R knock-out mice, Chu et al demonstrated that IGF-1 enhances thymopoiesis mainly through an expansion of thymic epithelial cells.14 These observations suggest an important role for the IGF/IGF-R system in the development and maturation of the thymus. IGF-1R transduces proliferative and antiapoptotic stimuli to the cell mainly through activation of the mitogen-activated protein kinase (MAPK) and AKT pathways. Phosphorylation of AKT is a downstream event of the activation of many tyrosine kinase membrane receptors, such as epidermal growth factor receptor (EGFR), human epidermal growth factor receptor 2 (HER2), and c-KIT, and it represents an indirect index of IGF-1R activation, because it has been suggested that 80% of activated AKT is caused by IR/IGF-1R activation.15

Recently, both monoclonal antibodies and small-molecule kinase inhibitors have been developed against IGF-1R,16 and promising anticancer activity has been reported in Ewing sarcoma and lung cancer. It is noteworthy that, in phase 1 trials of CP-751,871 and IMC-A12, 2 monoclonal antibodies against IGF-1R, prolonged stabilization for >1 year have been reported in 2 patients with metastatic thymomas.17, 18 To our knowledge, no data exist on IGF-1R expression and related phosphatidylinositol-3 kinase (PI3K)-AKT pathway activation in TEMs. The objective of the current study was to evaluate the protein expression of IGF-1R and phosphorylated AKT-serine (Ser) 473 (p-AKT) in tumor samples in relation to clinical characteristics, survival, and known prognostic factors in a relatively large series of resected TEMs treated at a single institution.


Patients and Samples

Patients underwent surgery consecutively at the Istituto Clinico Humanitas (Rozzano, Milan, Italy) between 1996 and 2008 for either a primary tumor or a local recurrence. For patients who had recurrent tumors, the date of the first surgery was taken into consideration, extending the observation period from 1976 to 2008. The formalin-fixed, paraffin-embedded surgical specimens were retrieved in each patient. Patient selection was based on a diagnosis of thymoma or thymic carcinoma and the availability of tumor tissue. Tissue slides were reviewed and classified according to the WHO classification.3 Tumor staging was determined according to the revised Masaoka system.19 The completeness of resection was classified as complete (R0), microscopic residual disease infiltrating resection margins (R1), and macroscopic residual disease (R2), as described previously.20 This study was conducted in agreement with the Declaration of Helsinki and was approved by the institutional ethical review boards (National Clinical Trials [NCT] identification number NCT00965627).

Construction of Tissue Microarrays

Formalin-fixed, paraffin-embedded tumor specimens were assessed for quality and adequacy of fixation and storage. A tissue microarray (TMA) block that contained 132 TEMs was generated. In brief, 3 0.36-mm2 punches (0.6 mm in greatest dimension) were taken from different intratumoral areas in each tumor sample and arranged in the recipient tissue array block. A pathologist (H.S.L.) verified the presence of tumor tissue on a hematoxylin and eosin-stained TMA slide. Cases were considered adequate if tumor occupied 1 or more cores of 3 punches.


IGF-1R and p-AKT expression levels were analyzed by immunohistochemistry (IHC). Formalin-fixed, paraffin-embedded TMAs were cut at 4 μm, deparaffinized with xylene, and rehydrated in graded ethanol. Antigen retrieval was performed by heating the slides at 95°C for 20 minutes in 10 mM sodium citrate buffer, pH 6.0, for both antibodies. Endogenous peroxidase blocking solution (EnVision+ System horseradish peroxidase-3,3′-diaminobenzidine [HRP-DAB]; Dako, Carpinteria, Calif) was applied to the tissues for 10 minutes followed by incubation in protein-free T20 (Tris-buffered saline [TBS]) blocking buffer (Thermo Scientific, Rockford, Ill) for 1 hour. Samples were then incubated with either rabbit polyclonal antibody IGF-1 receptor β (dilution, 1:100; Cell Signaling Technology, Danvers, Mass) or rabbit monoclonal anti-p-AKT (Ser-473) (dilution, 1:50; Cell Signaling Technology) overnight in a humid chamber at 4°C. After several washes with TBS-Tween 20 buffer, the slides were incubated for 30 minutes with labeled polymer-HRP as a secondary antibody; then, and immune reactions were observed with DAB as the chromogen (EnVision+ System HRP-DAB; Dako). The slides were counterstained with hematoxylin, dehydrated, and mounted. During all runs, negative control specimens were included by replacing the primary antibodies with buffer solution. For positive controls, specimens of human prostatic adenocarcinoma and formalin-fixed, paraffin-embedded xenografts from a Rh4 rhabdomyosarcoma cell line with high levels of IGF-1R were used for p-AKT and IGF-1R, respectively, according to the vendor's instructions and as reported previously.21

The sections were reviewed and scored by a pathologist who was blinded to patients' information and clinical outcomes. Analyses of specific protein staining (IGF-1R, membranous and cytoplasmic staining; p-AKT, cytoplasmic and nuclear staining) took into consideration both the percentage and the staining intensity of positive tumor cells.22 Staining intensity was scored from 0 to 3 as negative (0), weak (1), moderate (2), or strong (3). The percentage of positive cells was scored from 0 to 4 as follows: 0% to 5% positive cells (0), 6% to 26% positive cells (1), 26% to 50% positive cells (2), 51% to 75% positive cells (3), and 76% to 100% positive cells (4). The IHC score was obtained by multiplying the staining intensity by the percentage of positive cells. For IGF-1R, samples were graded from 0 to 3 according to the IHC score as follows: grade 0 (negative), IHC scores from 0 to 2; grade 1, IHC scores from 3 to 5; grade 2, IHC scores from 6 to 8; and grade 3, IHC scores from 9 to 12. For p-AKT, samples also were graded from 0 to 3 as follows: grade 0 (negative) samples, IHC scores of 0 and 1; grade 1, IHC scores of 2 and 3; grade 2, IHC scores of 3 and 4; and grade 3, IHC scores from 6 to 12.

Statistical Analysis

Clinical and biologic characteristics were compared using Fisher exact tests or chi-square tests when appropriate. Correlations were considered significant at the P < .05 level (2-tailed). Survival curves were generated using the Kaplan-Meier method, and differences between curves were analyzed using the log-rank test. Survival was calculated from the date of primary surgery to the date if death. Overall survival (OS), disease-related survival (DRS), and progression-free survival (PFS) were calculated. DRS was calculated with the exclusion of those patients who died of causes other than progression of TEM. A Cox regression multivariate analysis of prognostic factors was performed. All tests were done using the SPSS software package (version 17; SPSS Inc. Chicago, Ill).


Patient Outcome

Data from 132 patients were collected. The main patient and tumor characteristics are summarized in Table 1. Most samples were collected at primary surgery. The median patient age was 55 years, and the ratio of men to women was of 1:1. Most patients had early stage disease (stage I and II, 59%) and underwent radical resection (R0, 59.8%). Survival was assessable in all patients, and the median follow-up was 7 years (84.5 months; 95% confidence interval [CI], 72.1-96.9 months). Thirty-one patients (23.5%) were reported as dead at the update performed in June 2009. The median OS was 31.5 years (377.8 months; 95% CI, 112.5-634.2 months), and the 10-year OS rate was 70% (Fig. 1A). Only 13 patients (9.8%) died of disease progression, 2 patients died of myasthenia gravis, and 4 patients died of secondary cancers (3 hepatocellular carcinomas and 1 nonsmall cell lung carcinoma). The median DRS was not reached after 32 years of follow-up. The 10-year DRS rate was 87% (Fig. 1B). In total, 131 patients were evaluable for PFS; for 1 patient, no updated information on the date of disease progression was available. Thirty-six patients developed disease progression during the study period. The median PFS was 12.3 years (147 months; 95%CI, 121.6-172.4 months), and the 10-year PFS rate was 69% (Fig. 1C). Of the potential clinical prognostic factors that were analyzed, disease stage, histology, and resection type had significant influence on DRS (Fig. 1D) and PFS (Table 2), but not on OS, in univariate analysis. Patients with thymic carcinomas had worse DRS and PFS than patients with thymomas (log-rank test, P = .002 and P = .021, respectively). The 10-year DRS rate was 91% for thymomas and 60% for thymic carcinomas, whereas the 10-year PFS rates were 72% and 43%, respectively. Sex was a significant prognostic factor for DRS (P = .046), and women had more favorable outcomes. Patients aged >55 years had a worse OS than patients aged ≤55 years, but this difference was not reflected by a worse DRS or PFS. Information on the presence of paraneoplastic syndromes was available for 125 patients: Thirty-two patients (25.6%) developed myasthenia gravis during the course of their disease, 1 patient developed autoimmune encephalopathy, and 1 patient developed autoimmune glomerulonephritis. Paraneoplastic syndromes did not have a significant impact on survival.

Figure 1.

These Kaplan-Meier survival curves illustrate estimates for (A) overall survival, (B) disease-related survival, (C) progression-free survival, and (D) disease-related survival among patients with stage I and II disease versus patients with stage III and IV disease.

Table 1. Patient Characteristics (n = 132)
CharacteristicNo. of Patients (%)
  • WHO indicates World Health Organization; NA, not assessable.

  • a

    One micronodular thymoma and 1 cystic thymoma.

  • b

    For 18 patients, the disease stage at the diagnosis was not assessable.

  • c

    For 20 patients, there was insufficient information to assess the completeness of resection of the first surgery.

Median age [range], y55 [20-86]
 Primary108 (81.8)
 At recurrence24 (18.2)
Paraneoplastic syndromes
 Myasthenia gravis32 (25.6)
 Other syndromes2 (1.6)
WHO histotype
 A15 (11.4)
 AB28 (21.2)
 B124 (18.2)
 B1, B26 (4.5)
 B28 (6.1)
 B2, B311 (8.3)
 B324 (18.2)
 C14 (10.6)
 Othera2 (1.5)
Disease stage
 I35 (26.5)
 IIA26 (19.7)
 IIB17 (12.9)
 IIIA16 (12.1)
 IIIB3 (2.3)
 IVA5 (3.8)
 IVB12 (9.1)
 NAb18 (13.6)
Type of resection
 R079 (59.8)
 R123 (17.4)
 R210 (7.6)
 NAc20 (15.1)
Table 2. Univariate Analysis of Potential Prognostic Factors
 10-Year OSa10-Year DRSa10-Year PFSa
Parameter% of Patients (No. Alive)bSELog- Rank P% of Patients (No. Alive)bSELog- Rank P% of Patients (No. Alive)bSELog- Rank P
  • OS indicates overall survival, SE, standard error; DRS, disease-related survival; PFS, progression-free survival; NR, not reached; IGF-1R, insulin-like growth factor- receptor; p-AKT, phosphorylated AKT-serine 472.

  • a

    The 5-year OS rate was 71% for type A and 100% for type B1,B2; the 5-year DRS rate was 100% for type A and 100% for type B1,B2; and the 5-year PFS rate was 100% for type A and 78% for B1,B2.

  • b

    Numbers in parentheses indicate the numbers of patients alive or without disease progression at 10 years.

Sex  .544  .046  .151
 Men68 (16)0.07 80 (16)0.06 65 (13)0.07 
 Women73 (11)0.07 95 (11)0.04 73 (7)0.07 
Stage I/II vs III/IV  .087  .017  .003
 I/II77 (13)0.06 94 (13)0.04 85 (11)0.05 
 III/IV55 (2)0.1 73 (2)0.09 53 (1)0.1 
Histology  .326  .049  .005
 ANR  NR  NR  
 AB84 (7)0.09 100 (6)0 94 (6)0.06 
 B171 (5)0.11 93 (5)0.06 80 (3)0.1 
 B1, B2NR  NR  NR  
 B271 (4)0.17 85 (4)0.14 57 (3)0.19 
 B2, B386 (2)0.13 86 (2)0.13 73 (2)0.16 
 B360 (8)0.11 78 (8)0.1 40 (4)0.11 
 C52 (2)0.15 60 (2)0.15 43 (2)0.15 
Resection type  .439  .049  .028
 R073 (11)0.06 89 (11)0.04 83 (10)0.05 
 R167 (2)0.12 92 (2)0.07 66 (1)0.12 
 R247 (1)0.18 57 (1)0.19 47 (1)0.18 
Paraneoplastic syndromes  .203  .714  .421
 Yes79 (17)0.08 85 (17)0.07 67 (12)0.09 
 No68 (9)0.06 90 (9)0.04 74 (8)0.06 
Age ≤55 y vs >55 y  .007  .801  .289
 ≤55 y79 (19)0.6 85 (19)0.05 60 (12)0.07 
 >55 y61 (8)0.7 90 (8)0.05 79 (8)0.06 
Age <40 y vs 40-65 y vs >65 y  .15  .735  .110
 <40 y71 (7)0.11 76 (7)0.01 57 (6)0.12 
 40-65 y79 (18)0.06 88 (18)0.05 65 (12)0.07 
 >65 y52 (2)0.1 94 (2)0.05 89 (2)0.07 
IGF-1R  .762  .310  .083
 Positive73 (4)0.11 79 (4)0.11 49 (4)0.12 
 Negative71 (19)0.06 93 (19)0.03 78 (14)0.05 
Cytoplasmic p-AKT  .606  .106  .335
 Positive74 (15)0.7 83 (15)0.06 67 (14)0.08 
 Negative69 (7)0.8 97 (7)0.03 75 (3)0.08 
Nuclear p-AKT  .845  .245  .915
 Positive73 (12)0.7 85 (12)0.06 68 (10)0.08 
 Negative65 (9)0.7 94 (9)0.04 74 (6)0.02 

Of 109 patients who underwent resection for their primary tumor who had information on further treatment available, only 12 patients received chemotherapy either as adjuvant treatment (3 patients) or at disease recurrence (9 patients). No apparent impact of chemotherapy could be observed.



In total, 111 patients were evaluable for IGF-1R IHC staining; the remaining 21 patients did not have adequate material on TMA. Twenty-two patients (19.8%) had samples that were positive for IGF-1R staining (Fig. 2A,B), including 19.8% of the 91 primary tumors and 20% of the 20 recurrent tumors (Table 3). There were 16 grade 1 IGF-1R–positive samples, 3 grade 2 IGF-1R–positive samples, and 3 grade 3 IGF-1R–positive samples.

Figure 2.

Insulin-like growth factor-1 receptor (IGF-1R) and phosphorylated AKT serine 473 (p-AKT) immunohistochemical staining is shown (A) in a sample with positive IGF-1R membrane staining and (B) in a sample with negative IGF-1R membrane staining. (C) Kaplan-Meier curves for disease-related survival among patients with thymic epithelial malignancies (TEMs) who had either positive or negative IGF-1R expression are shown. (D) Kaplan-Meier curves for progression-free survival among patients with TEMs who had either positive or negative IGF-1R expression are shown. (E) A sample with positive p-AKT cytoplasmic staining is shown. (F) A sample with negative p-AKT cytoplasmic staining is shown. (G) Kaplan-Meier curves of disease-related survival among patients with TEMs who had either positive or negative p-AKT expression are shown. (H) Kaplan-Meier curves of progression-free survival among patients with TEMs who had either positive or negative p-AKT expression are shown.

Table 3. Results of Immunohistochemistry for Insulin-Like Growth Factor-Receptor and Phosphorylated AKT-Serine 473
 IGF-1R PositiveCytoplasmic p-AKT TR Positive
VariableNo./Total (%)G1G2G3PNo./Total (%)G1G2G3P
  • IGF-1R indicates insulin-like growth factor- receptor; p-AKT, phosphorylated AKT-serine 472; G, grade; WHO, World Health Organization.

  • a

    Comparison between A, AB, and B1 versus B1/B2, B2, B2/B3, B3, and C.

  • b

    Comparison between A, AB, B1, B1/B2, B2, B2/B3, and B3 versus C.

  • c

    One micronodular thymoma and 1 cystic thymoma.

  • d

    Comparison between stages I and II versus stages III and IV.

  • e

    For 15 patients, the disease stage at diagnosis was not assessable.

  • f

    Comparison between R0 versus R1 and R2.

  • g

    For 16 patients, there was insufficient information to assess the completeness of resection at the first surgery.

Total22/111 (19.8)    51/105 (48.6)    
Tumor type    .589    .015
 Primary18/91 (19.8)1233 36/86 (41.9)15165 
 Recurrent4/20 (20)400 15/19 (78.9)582 
Sex    .683    .958
 Women12/59 (20.3)822 29/58 (50)12143 
 Men10/52 (19.2)811 22/47 (46.8)8104 
WHO histotype
 A0/8 (0)000 2/8 (25)110 
 AB2/26 (7.7)200 10/26 (38.5)550 
 B10/24 (0)000.009a6/22 (27.3)330.001a
 B1/B20/6 (0)000 2/6 (33.3)200 
 B20/8 (0)000 7/8 (87.5)241 
 B2/B33/10 (30)120 9/9 (100)252 
 B39/19 (47.4)810 9/17 (52.9)342 
 C7/8 (87.5)403.005b6/8 (75)222.234b
 Otherc1/2 (50)100 0/1 (0)000 
Disease stage    .058d    .0013d
 I5/28 (17.9)311 7/25 (28)430 
 IIA2/24 (8.3)200 9/25 (36)540 
 IIB0/16 (0)000 6/13 (46.2)321 
 IIIA5/14 (35.7)221 6/14 (42.9)150 
 IIIB1/3 (33)100 3/3 (100)111 
 IVA2/3 (66)200 2/3 (66)011 
 IVB3/8 (37.5)201 6/7 (85.7)222 
 NAe4/15 (26.7)400 12/15 (80)462 
Completeness of resection    .365f    .287f
 R07/66 (10.6)412 25/62 (40.3)11113 
 R16/21 (28.6)321 8/20 (40)260 
 R25/8 (62.5)500 6/7 (85.7)312 
 NAg4/16 (25)400 12/16 (75)462 

IGF-1R expression was significantly less common (Fisher exact test; P < .0001) in the histotype group that included WHO subtypes A, AB, and B1, in which it was expressed in 2 of 58 samples (3.4%), compared with the more aggressive histotypes B1/B2, B2, B2/B3, B3, and C, in which it was expressed in 19 of 51 samples (37.2%). All grade 2 and grade 3 samples were in the more aggressive histotype group. With this subdivision, 6 of 51 samples and 0 of 58 samples were grade 2 and 3 in the more aggressive group and in the less aggressive group, respectively (Fisher exact test; P = .009). When the histologic groups were divided into thymic carcinomas and thymomas, 3 of 8 samples and 3 of 103 samples were grade 2 and 3 in thymic carcinomas and in thymomas, respectively (Fisher exact test; P = .005). IGF-1R expression was significantly more common in patients with stage III and IV disease (11 of 28 patients; 39.3%) than in patients with stage I and II disease (7 of 68 patients; 10.3%; chi-square test, 10.942; P = .001). Four of 6 grade 2 and 3 samples were associated with a more advanced disease stage. Positive IGF-1R expression also was significantly more common in R2 resection specimens than in R0 resection specimens (62.5% vs 10.6%, respectively; Fisher test for the 3 variables; P = .002). Although there was a trend for patients with IGF-1R expression to have a worse DRS than patients without IGF-1R expression, this difference was not statistically significant (P = .301) (Table 3, Fig. 2C). Considering only the primary tumors or only the recurrent tumors, there also was no statistically significant difference in terms of DRS between patients with positive and negative IGF-1R expression (P = .178 and P = .960, respectively). The estimated median PFS was 10.2 years (122.5 months; 95% CI, 13.2-231.8 months) for the IGF-1R–positive group and 12.9 years (147 months; 95% CI, 119.1-174.9 months) for the IGF-1R–negative group (log-rank test, P = .083; Breslow test, P = .003) (Fig. 2D). When only patients with primary tumors were considered, the 5-year PFS rate was 96% for IGF-1R–positive patients and 65% for IGF-1R–negative patients (log-rank test; P < .001). No statistically significant difference in PFS was observed in patients who had recurrent disease (log-rank test; P = .609): The 5-year PFS rate was 63% for IGF-1R–positive patients and 50% for IGF-1R–negative patients.


Staining for p-AKT was observed in both the cytoplasmic cellular compartment and the nuclear cellular compartment. Nuclear staining for p-AKT was present in 51 of 105 evaluable samples (48.6%) (Fig. 2E,F), including 36 of 86 primary tumors (41.9%) and 15 of 19 recurrent tumors (78.9%) (Table 3). Calculated on the basis of IHC scores, among the samples that had p-AKT-positive nuclear staining, 20 samples were grade 1, 24 samples were grade 2, and 7 samples were grade 3. For further analysis, we selectively analyzed the samples that had cytoplasmic p-AKT–positive cells only, because the main function of p-AKT is in this compartment. In the thymoma group that included histologic types A, AB, and B1, 18 of 56 tumor specimens had positive cytoplasmic p-AKT staining (32.1%) versus 33 of 48 specimens of the more aggressive histologic types B1/B2, B2, B2/B3, B3, and C (68.7%), and this difference was statistically significant (chi-square test, 13.860; P < .0001). According to this subdivision, 22 of 48 specimens and 9 of 56 specimens were grade 2 and 3 in the more aggressive group and in the less aggressive group, respectively (chi-square test, 10.942; P = .001), suggesting that higher IHC scores are associated with more aggressive histotypes. When the histologic groups were divided into thymic carcinomas and thymomas, grade 2 and 3 staining was not significantly higher in thymic carcinomas (4 of 8 tumors) than in thymomas (27 of 97 tumors; Fisher exact test; P = .231). p-AKT expression was higher in patients with stage III and IV disease (17 of 27 patients; 62.9%) than in patients with stage I and II disease (22 of 63 patients; 34.9%; chi-square test, 6.053; P = .014). The score distribution according to tumor grade for high-stage (III-IV) disease was 41 grade 0 tumors, 12 grade 1 tumors, 9 grade 2 tumors, and 1 grade 3 tumor; whereas, the distribution for low-stage (I-II) disease was 10 grade 0 tumors, 4 grade 1 tumors, 9 grade 2 tumors, and 4 grade 3 tumors. p-AKT expression was more frequent in R2 and R1 specimens than in R0 specimens (Fisher test; P < .007). Although DRS was worse in patients who had positive p-AKT expression compared with patients who negative p-AKT expression, this difference was not statistically significant (P = .106) (Table 2, Fig. 2G). The difference was also not significant when primary tumors and recurrent tumors were considered separately (P = .207 and P = .388, respectively). There also was no statistically significant difference in PFS between patients with or without p-AKT expression (P = .355) (Fig. 2H). The tumor specimens from 104 patients were evaluable for both IGF-1R expression and p-AKT expression. Of 51 p-AKT–positive samples, 16 samples also were positive for IGF-1R (chi-square test, 0.005). In multivariate analysis, no independent prognostic factor was identified for DRS, whereas WHO classification was an independent prognostic factor for PFS (P = .008).


The current analysis included 132 patients with TEMs who underwent surgery at a single institution and had a long follow-up (7 years). Most patients had early stage disease and underwent complete resection, as reflected by an overall median survival of 31.5 years. In most surgical series, the Masaoka stage and the completeness of resection are the most important prognostic factors.6 Thymoma is 1 of the few tumors in which even tumor debulking can improve survival.4, 6 Surgery appears to have a much higher rate of failure in higher WHO types, such as in thymic carcinomas (WHO type C),23 and this also was confirmed in our study. However, the WHO classification has not always been confirmed as an accurate prognostic factor.24 Because of the more common occurrence of paraneoplastic syndromes, some of which may prove fatal, and secondary cancers, DRS and PFS usually are more informative than OS in these patients who otherwise have a very long life expectancy. When considering DRS, survival was better for women that for men, a phenomenon that has been observed in other tumors, such as lung cancer and lymphomas, but that has not been reported previously in TEM.25

The current results demonstrated that protein expression of IGF-1R and p-AKT is common in all histologic subtypes of TEM. It is noteworthy that the expression of both IGF-1R and p-AKT was correlated significantly with more aggressive histologic subtypes. Completeness of resection was strongly related to disease stage at diagnosis, as evidenced in our multivariate analysis. We identified greater IGF-1R and p-AKT expression in more samples from patients with advanced disease stages at diagnosis and in incomplete resection specimens. Also, samples of recurrent tumors had greater expression than primary tumor samples. All of these findings strongly suggest that this pathway may be important for tumor progression and potentially may be targeted for treatment, especially in patients with advanced-stage disease and aggressive histologic types. p-AKT expression can be considered a marker for activated IGF-1R signaling: It has been suggested that 80% of activated AKT is caused by IR/IGF-1R pathway activation.15 In patients with breast cancer, IGF-1R expression and activation have been associated with disease progression, increased resistance to radiotherapy, and a poor prognosis.26 In patients with lung cancer, a statistically significant association was observed between high coexpression of both IGF-1R and EGFR and worse DFS in early disease,27 but not between IGF-1R alone and DFS or OS. In our series of patients with thymic tumors, we observed a trend toward worse survival for patients with IGF-1R–positive tumors, but this trend did not reach statistical significance. PFS, however, was significantly worse in patients who had IGF-1R–positive primary tumors.

A better understanding of the biology of this rare tumor may allow us to identify potential prognostic markers and drug targets. The IGF system of ligands, receptors, and soluble binding proteins has been implicated in local and distant tumor spread. The demonstration of IGF-1R involvement in malignant transformation28, 29 and the frequent detection of IGF-1R expression in human cancers have contributed in part to the efforts to develop IGF-1R–targeted therapy. Frequent loss of heterozygosity of 6q26, which contains IGF-2R, has been reported in TEM,30 and it is possible that IGF-1R may be up-regulated in such malignancies to compensate for the loss of IGF-2R. Indirectly, the loss of IGF-2R may lead to an increase in IGF and stimulation of IGF-1R. Furthermore, in a recent report, IGF-2BP3, a translational activator of IGF-II, was highly overexpressed in thymic carcinomas.31

Drugs targeted at growth factor receptors have been investigated in advanced TEMs, and phase 2 trials failed to demonstrate any activity of gefitinib (an EGFR inhibitor) or imatinib (a c-KIT inhibitor).32-34 The activity of these agents is limited mainly to tumors that bear sensitizing mutations in the adenosine triphosphate-binding domain of these tyrosine kinases, and most TEMs do not have mutations of c-KIT or EGFR.35-38 Conversely, continuous activation of the PI3K-AKT pathway is considered an important mechanism of tumor growth and resistance to anti-EGFR therapies. Preclinical data on cell lines suggest that IGF-1R mediates resistance to anti-EGFR therapy through continued activation of the PI3K-AKT pathway.39, 40

In conclusion, the expression of IGF-1R was observed in a significant percentage of TEMs and was associated with more aggressive tumors, suggesting that IGF-1R may be a potential target of treatment in thymic tumors. Furthermore, this retrospective analysis confirmed the importance of disease stage, completeness of resection, and WHO histologic type as prognostic factors. On the basis of our current findings and the prolonged stabilization observed in 2 phase 1 studies of IGF-1R monoclonal antibodies in 2 patients with metastatic thymoma,17, 18 a phase 2 trial with the anti-IGF-1R antibody IMC-A12 recently has started accruing at the National Cancer Institute in patients with thymoma and thymic carcinoma.


This work was funded by the National Institutes of Health/National Cancer Institute intramural program.