Prolonged Tumor-free Survival Research Articles

Abstract 397: The antitumor activity of the investigational drug MLN9708 in genetically engineered mouse models of plasma cell malignancy

Abstract Introduction: The proteasome inhibitor VELCADE® (bortezomib) is an important component of the chemotherapeutic strategy in treating multiple myeloma, a plasma cell malignancy (PCM). While genetically engineered mouse models (GEMMs) of cancer often accurately recapitulate their human disease counterparts, their usage in drug discovery settings has been very limited. Here we describe the antitumor activity of a second generation proteasome inhibitor, the investigational drug MLN9708, in the iMycCα/Bcl-XL GEMMs of PCM, in which neoplastic plasma cell development is driven by enforced expression of the Myc and Bcl-XL transgenes. Materials: MLN9708 immediately hydrolyzes to MLN2238, the biologically active form, upon exposure to aqueous solutions or plasma; MLN2238 was used for all studies below. We previously demonstrated that double transgenic iMycCα/Bcl-XL (C57BL6/FVB) mice develop de novo PCM with short onset (135 days) and full penetrance (100%). We derived a plasma cell tumor (PCT) cell line, DP54, from the bone marrow of a syngeneic mouse previously inoculated with an iMycCα/Bcl-XL tumor. DP54 PCT cells were stably transfected with the firefly luciferase gene, clonally isolated, and designated as DP54-Luc cells. Methods: Nine-week-old iMycCα/Bcl-XL (C57BL6/FVB) mice were untreated or treated with bortezomib (1.2 mg/kg intravenously [IV] twice weekly [BIW]) or MLN2238 (18 mg/kg IV BIW) for 6 consecutive weeks and monitored for tumor-free survival for an additional 25 weeks. To establish disseminated and intraosseous mouse models of iMycCα/Bcl-XL PCM, freshly dissociated DP54-Luc cells were aseptically injected into the lateral tail veins and the bone marrow space of the upper right tibia, respectively, of immunocompromised mice. Once tumor growth was established, mice were randomized and treated with vehicle (5%HPbCD), bortezomib (0.8 mg/kg IV BIW) or MLN2238 (13 mg/kg IV BIW) for 3-4 consecutive weeks. The doses used represent the maximum tolerated dose for each drug in each mouse strain. Results: In the iMycCα/Bcl-XL GEMM of de novo PCM, treatment with bortezomib or MLN2238 significantly prolonged tumor-free survival (+27 and +36 days, respectively; p<0.0001) and decreased plasma immunoglobulin levels compared to untreated controls. In the models of iMycCα/Bcl-XL PCM, treatment with bortezomib and MLN2238 significantly reduced disease burden as measured by IVIS® bioluminescent imaging. Conclusion: GEMMs of cancer have often been viewed as promising alternatives to traditional subcutaneous xenograft models, yet data to support their wider use in drug discovery settings are sparse. Here we demonstrated that carefully implemented GEMM studies can be integrated as an important part of the drug-discovery paradigm. MLN9708 is currently in clinical development for both hematologic and solid tumor indications. Safety and efficacy have not been established. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 101st Annual Meeting of the American Association for Cancer Research; 2010 Apr 17-21; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2010;70(8 Suppl):Abstract nr 397.

Abstract 4755: Importance of an induction dose to the immunogenicity of cancer vaccines: A reverse translational study

Abstract We have described a human cancer vaccine consisting of autologous, irradiated melanoma cells modified with the hapten, dinitrophenyl (DNP), and mixed with BCG. Because of its immunopotentiating effect, low dose cyclophosphamide (CY) was administered 3 days before the first of a series of vaccine injections. The vaccine induced T cell immunity to both haptenized and unmodified autologous melanoma cells; clinical effectiveness was suggested by tumor regression in some stage IV patients and an unexpectedly high proportion of 5-year survivors in stage IIIB and IIIC patients. A serendipitous finding from these studies was the critical importance of what we came to call an “induction dose” of vaccine - a single injection of irradiated, autologous melanoma cells without BCG 5-7 days prior to CY. Patients who received a properly timed induction dose had more intense T cell responses (positive DTH to unmodified melanoma cells 73% vs. 20%, p<.01) and were more likely to survive 5 years (51% vs. 33%, p=.007) than those who did not. To support this finding in a controlled experimental system, we developed a mouse model using the mammary carcinoma, 410.4. After injection of 410.4 cells into the mammary fat pads, primary tumors developed and were excised. Then we assessed the anti-metastatic activity of low dose CY followed by a DNP-modified vaccine prepared and administered in a manner comparable to the human melanoma vaccine. The published study showed that this regimen significantly prolonged tumor-free survival in the mice while CY followed by a control vaccine (irradiated, non-haptenized cells) did not. To mimic the induction dose used in the clinical trial, a cohort of mice was injected with irradiated, DNP-modified tumor cells without BCG 3-7 days prior to CY. The relapse-free survival (RFS) of mice receiving the induction dose (N=41) was significantly longer (p < .01) than that of mice receiving DNP-vaccine without the induction dose (N=27) (e.g., RFS day 37: CY+irradiated vaccine=30.8%, CY+irradiated-haptenized vaccine=40.7%, CY+induction dose+ irradiated-haptenized vaccine=53.7%). In this reverse translational model, the murine experiments corroborated a serendipitous clinical observation. Taken together our results suggest a new strategy for increasing the effectiveness of cancer vaccines. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 101st Annual Meeting of the American Association for Cancer Research; 2010 Apr 17-21; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2010;70(8 Suppl):Abstract nr 4755.

The Novel Proteasome Inhibitor MLN9708 Demonstrates Efficacy in a Genetically-Engineered Mouse Model of DeNovo Plasma Cell Malignancy.

Abstract 3849Poster Board III-785 IntroductionThe first-in-class proteasome inhibitor VELCADË (bortezomib) is a critical component of chemotherapeutic strategy in treating multiple myeloma (MM), a plasma cell malignancy characterized by monoclonal gammopathy and osteolytic lesions. However, patients eventually develop bortezomib resistance, highlighting the continuous challenge to further improve myeloma therapy. We have recently identified a novel proteasome inhibitor MLN9708 that shows greater antitumor activity than bortezomib in a number of preclinical xenograft models. Here we describe antitumor activity of MLN9708 in a genetically engineered mouse model of human MM. In this model, neoplastic plasma cell development is driven by deregulated expression of Myc (c-myc, myelocytomatosis oncogene) and Bcl-x (official gene name: Bcl2l1; encodes the anti-apoptotic Bcl-XL oncoprotein). This is the first preclinical in vivo study of proteasome inhibitors employing a transgenic mouse model of human MM in which plasma cell neoplasms develop de novo. MaterialsUpon exposure to aqueous solutions or plasma, MLN9708 immediately hydrolyzes to MLN2238, the biologically active form. MLN2238 was used for all preclinical studies described below. MethodsWe have previously demonstrated that intercrossing C57BL/6 mice carrying the iMycCa transgene (insertion of Myc in the immunoglobulin heavy-chain locus, Igh) with FVB/N mice carrying the 3'KE-Bcl-XL transgene (enforced expression of Bcl-x driven by the immunoglobulin light-chain 3' k enhancer and Vk21 promoter) produces double transgenic F1 hybrid iMycCa/Bcl-XL mice that develop plasma cell malignances with short onset (135 days on average) and full penetrance (100% tumor incidence; J. Clin. Invest. 113:1763-1773, 2004). These double transgenic iMycCa/Bcl-XL mice develop hypergammaglobulinemia and tissue plasmacytosis by 6-8 weeks of age and rapidly succumb to malignant plasma cell tumors by less than 200 days of age. Preliminary studies based on 11 tumor-bearing iMycCa/Bcl-XL mice showed this form of plasma cell malignancy recapitulates important features of human MM, including serum paraproteins, infiltration of bone marrow with malignant plasma cells, osteolytic lesions and myeloma-like global gene expression profiles (Cancer Res. 67:4069-4078, 2007). Here we used the iMycCa/Bcl-XL mouse model of human MM to assess the antitumor activity of bortezomib and MLN9708 in a preclinical setting. ResultsUntreated double transgenic iMycCa/Bcl-XL mice (n=30) invariably developed plasma cell tumors with short onset (median tumor-free survival = 112 days) and showed marked elevations in plasma immunoglobulin IgM, IgG1, IgG2a and IgG2b levels. To assess antitumor activity of bortezomib and MLN2238, 9-week-old iMycCa/Bcl-XL mice were treated intravenously (IV) twice per week (BIW) with bortezomib (1.2 mg/kg) or MLN2238 (18 mg/kg) for 6 consecutive weeks. These doses represent the maximum tolerated dose (MTD) for each drug as previously determined in normal C57BL/6 x FVB/N F1 hybrid mice. Treatment of iMycCa/Bcl-XL mice with bortezomib (n=30) significantly prolonged tumor-free survival compared to untreated controls (median tumor-fee survival = 139 days; hazard ratio = 0.116; 95% CI = 0.057 to 0.235; p<0.0001). Importantly, treatment of iMycCa/Bcl-XL mice with MLN2238 (n=30) also significantly prolonged tumor-free survival relative to untreated controls (median tumor-free survival = 148 days; hazard ratio = 0.144; 95% CI = 0.071 to 0.289; p<0.0001). Immunoglobulin levels, overall tumor burden, osteolytic lesions (mCT) and biochemical parameters of myeloma bone disease after treatment with bortezomib and MLN2238 will be presented. ConclusionThe novel proteasome inhibitor MLN9708 demonstrates striking antitumor activity in a genetically engineered mouse model of human MM, significantly prolonging tumor-free survival of double transgenic iMycCa/Bcl-XL mice. MLN9708 is currently in human clinical development for both hematological and solid tumor indications. Disclosures:Janz:Millennium – The Takeda Oncology Company: Research Funding. Van Ness:Millennium: Research Funding; Scientific Advisory Board, International Myeloma Foundation: Membership on an entity's Board of Directors or advisory committees. Liu:Milllennium: Employment, Equity Ownership. Pickard:Milllennium: Employment. Terkelsen:Milllennium: Employment. Bradley:Milllennium: Employment, Equity Ownership, Research Funding. Kupperman:Milllennium: Employment. Manfredi:Milllennium: Employment. Lee:Milllennium: Employment, Equity Ownership.

Abstract 5832: Overexpression of ODC1 is associated with poor outcome in childhood neuroblastoma and represents an important therapeutic target

Abstract MYCN gene amplification is a powerful predictor of poor clinical outcome in neuroblastoma, the most common solid tumor of young children. The mechanism by which MYCN amplification influences the prognosis of this disease has not been well defined. MYCN is a known transcription factor and available evidence indicates that ornithine decarboxylase (ODC1) is a critical downstream target for this oncoprotein. ODC1 is the rate-limiting enzyme in the synthesis of polyamines and overexpression of this gene has been observed in a range of tumor cells, including neuroblastoma. In particular, a number of studies have provided evidence for the oncogenic and transforming abilities of ODC1. To investigate the role of ODC1 in neuroblastoma, we used Real-Time PCR to analyze ODC1 expression in a large cohort (n=209) of primary untreated neuroblastomas from patients enrolled on POG biology protocol 9047. Older age, advanced stage, and MYCN amplification were all predictive of poor outcome in this cohort. ODC1 expression was significantly higher in MYCN-amplified tumors by comparison with non-amplified tumors (p&amp;lt;0.0001). Kaplan-Meier survival analysis indicated that high levels of ODC1 expression were strongly predictive of both event-free-survival (p&amp;lt;0.001) and overall survival (p&amp;lt;0.001) in this cohort as well as in patients whose tumors lacked MYCN amplification (p=0.01). Furthermore, a functionally important single nucleotide polymorphism within the ODC1 promoter (G316A) that has been shown previously to influence gene transcription rates, was found to be significantly associated with outcome in patients whose tumors lacked MYCN amplification (p&amp;lt;0.05). These data suggest that inhibiting ODC1 activity would have therapeutic advantage in treating neuroblastoma. To test this hypothesis, MYCN transgenic mice that spontaneously develop neuroblastoma closely mimicking the human disease, were treated either with cisplatin alone or in combination with the ODC1 inhibitor, α-difluoromethylornithine (DFMO). Importantly, DFMO therapy, either concomitant with or following cisplatin treatment, resulted in prolonged tumor-free survival (p&amp;lt;0.01) in these mice, by comparison with cisplatin administered alone. Similar results were obtained with cyclophosphamide treatment in combination with DFMO. No overt toxicity attributable to DFMO therapy was noted in treated mice. Collectively, these findings provide strong evidence that ODC1 contributes to the malignant phenotype of neuroblastoma and suggest that targeting this oncogene for suppression of polyamine synthesis is potentially a valuable therapeutic approach for inhibiting neuroblastoma growth.

The NO TRAIL to YES TRAIL in cancer therapy (Review)

Treatment of cancer patients with conventional therapies (chemotherapy, hormonal therapy, immunotherapy and radiation) respond initially well and experience prolonged tumor-free survival. However, in many patients tumor recurrences and relapses occur and such tumors exhibit the resistant phenotype i.e. cross-resistance to various cytotoxic and apoptotic agents. Therefore, new therapeutic strategies are currently being explored and are based on a better understanding of the underlying biochemical and molecular mechanisms of tumor cell resistance. Hence, novel sensitizing agents that can modify the tumor dysregulated apoptotic gene products can reverse resistance when used in combination with subtoxic doses of cytotoxic reagents. Targeted anti-tumor therapies are the current choice in the treatment of resistant tumors. One such targeted therapy is the application of TRAIL or TRAIL agonist monoclonal antibodies (mAbs) (anti-DR4 and anti-DR5) because, unlike Fas ligand and TNF-alpha, they are not cytotoxic to normal tissues. TRAIL as monotherapy will only be effective against TRAIL sensitive tumors, however, most tumors are resistant to TRAIL and their sensitization can restore their sensitivity to TRAIL apoptosis. We present, herein, one potential novel sensitizing agent, namely, nitric oxide (NO) that has been shown to sensitize TRAIL-resistant tumor cells to TRAIL apoptosis via its inhibitory effect on the transcription factors NF-kappaB and Yin Yang 1 (YY1), concomitantly with upregulation of DR5. We propose the therapeutic application of NO donors as sensitizing agents used in combination with TRAIL/DR4 or DR5 mAbs in the treatment of TRAIL-resistant tumors.

Adjuvant Chemotherapy for Prevention of Recurrence of Invasive Hepatocellular Carcinoma After Orthotopic Liver Transplantation

Orthotopic liver transplantation (OLT) is the best treatment for nonresectable hepatocellular carcinoma (HCC), but tumor recurrence reduces long-term and medium-term survival. The effectiveness of adjuvant chemotherapy to prevent tumor recurrence has not been fully established. Methods Three hundred eighty-seven consecutive patients, including 43 with HCC superimposed on liver cirrhosis, underwent OLT. Twelve patients with one or more prognostic criteria for HCC recurrence were entered into a prospective prophylaxis protocol with monthly cycles of cisplatin (60 mg/m 2) and adriamycin (30 mg/m 2), beginning the fourth week post-OLT for a maximum of seven sessions. Results The 5-year survival of the non-HCC patients was 65.7% and that of the HCC patients was 60.46% ( P = NS). Chemotherapy was reasonably well tolerated, but the 9 patients with hepatitis C- or B-associated cirrhosis showed viral and histological recurrence of the primary disease. A high proportion of patients (7 of 12) developed tumor recurrence during the first year after OLT. Six of these patients died, all but one due to HCC relapse. Five patients remain healthy and tumor free at 58 to 130 months. Post-OLT adjuvant chemotherapy does not avoid tumor recurrence and its fatal consequences but may contribute to prolonged tumor-free survival among a significant proportion of patients with high-risk HCC. However, the uncertain implications on viral recurrence and the lack of control groups do not allow post-OLT chemotherapy to be recommended outside controlled clinical trials, which are clearly warranted.

Trastuzumab mimotope vaccination of HER-2/neu transgenic mice results in prolonged tumor-free survival and reduced tumor load

12510 Background: Passive immunotherapy with the humanized monoclonal antibody trastuzumab (Herceptin) to date is the most effective treatment for patients with HER-2/neu overexpressing breast cancer. In previous studies we could show that active immunization with peptide mimotopes, i.e. structural mimics of the epitope recognized by trastuzumab, leads to formation of antibodies again recognizing HER-2/neu in mice. Functional in vitro analyses of the induced antibodies demonstrated “trastuzumab-like” properties, such as receptor internalization, inhibition of signaling, and antibody-mediated cytotoxicity against HER-2/neu overexpressing breast cancer cells. The aim of the present study was to test the effects of trastuzumab mimotope vaccination in vivo, namely in a HER-2/neu transgenic mouse model. Methods: We used BALB-neuT mice, which carry the activated neu oncogene on the BALB/c background. These mice constitute the most aggressive animal model for HER-2/neu driven carcinogenesis, as all females uniformly develop mammary carcinomas at the age of 12 weeks. One group of mice was immunized with the previously described trastuzumab mimotope - KLH conjugate, a control group with the carrier protein KLH alone, and a further control group was left naïve. Mice were palpated weekly to monitor tumor development and size, and blood samples were taken at regular intervals to follow up the induced immune responses. Results: Trastuzumab mimotope immunizations lead to delayed tumor development and thus to an increase in tumor-free survival. When tumors occurred, they were smaller and grew more slowly than in the control mice. In contrast, control KLH immunizations did not affect tumor growth kinetics as compared to the naïve mice. Serum analysis demonstrated that all immunized animals had mounted an anti-KLH immune response, so we accredit the observed tumor-inhibitory effects in the mimotope group to the biologic properties of induced anti-HER-2/neu antibodies. Conclusion: These results indicate that epitope-specific vaccination with mimotopes elicits trastuzumab-like antibodies that are effective in vivo against HER-2/neu overexpressing tumor cells also in HER-2/neu expressing organisms. [Table: see text]

Antitumor effects of a xenogeneic survivin bone marrow derived dendritic cell vaccine against murine GL261 gliomas.

Survivin is a member of the inhibitor of apoptosis protein family. Gliomas and many other tumors express survivin at high levels; whereas, normal fully differentiated cells generally do not. Therefore, survivin represents a tumor-specific target for cancer vaccine therapy. It has been shown that it is possible to produce a MHC-I-restricted cellular immunologic response to survivin vaccines. To study differences in immunogenicity between murine and human survivin proteins, we vaccinated C57BL/6 mice with bone marrow dendritic cells (BMDC) transfected with expression vectors containing the murine and human survivin genes. Mice vaccinated with BMDCs expressing a truncated human survivin protein developed cytotoxic T lymphocyte to subcutaneous GL261 glioma cells and exhibited prolonged tumor-free survival compared to mice vaccinated with BMDCs transfected with vector alone (P<0.01). While mice challenged with intracerebral GL261 cells had increased survival, no cures were observed. In contrast, vaccinated mice that fully resisted subcutaneous tumor challenge were rendered resistant to intracerebral GL261 re-challenge. BMDCs transfected with the full-length human survivin molecule were significantly more effective at prolonging survival than BMDCs expressing the full-length murine survivin gene (P=0.0175). Therefore, xenogeneic differences between human and murine sequences might be exploited to develop more immunogenic tumor vaccines.

Malignant insulinoma

Malignant insulinoma occurs in a few patients with insulinoma. Due to the small sample of patients, there are little data regarding their clinical manifestation as well as the preferred treatment modalities. The aims of the current study were to summarize the National Institutes of Health experience during the last two decades and to conduct a critical review of the current literature. The authors identified 10 patients with metastatic insulinoma. The patients presented with four patterns of clinical behavior. First, four patients presented with lymph node metastasis and, after surgical excision, maintained a prolonged tumor-free survival. Second, four patients presented with metastatic disease to the liver, which appeared years after the initial diagnosis and presumed curative surgery. Third, one patient presented with a large alpha-fetoprotein-secreting liver mass. Finally, 9 of the 10 patients had a prolonged survival. Various treatment modalities were used to control hypoglycemia. Short-term benefits were most often achieved with embolization and diazoxide. Less successful modalities included radiofrequency ablation, radical debulking surgery, verapamil therapy, octreotide therapy, and chemotherapy. The current study, as well as others, suggested that metastatic insulinoma may have a variable natural history. After the initial surgical resection, the biology of the tumor, rather than any treatment modality, was most likely the major determinant of long-term survival.

Apoptotic, but not necrotic, tumor cell vaccines induce a potent immune response in vivo.

Prophylactic tumor vaccination against subsequent tumor challenge depends on effective cross-priming in vivo. Professional APCs process tumor antigens from whole tumor cells and present them to CD4(+) and CD8(+) T cells. Data suggest that dendritic cells process antigens more efficiently from necrotic cells than from apoptotic cells in vitro. We compared the effect of apoptosis vs. necrosis in vivo using different tumor models (CT26, RENCA, B16 and CT26-HA). Apoptosis was induced by gamma-irradiation prior to injection and verified in vivo. Apoptotic CT26-HA, CT26-wt or RENCA prevented tumor outgrowth in 100%, 75% and 100%, respectively, of mice for more than 30 days after challenge. In contrast, injection of necrotic tumor cells led to protection of no more than 0-30%. Prolonged tumor-free survival was also observed in mice after vaccination with irradiated B16 cells. In vivo protection experiments correlated very well with in vitro cytotoxicity assays. Immunohistochemical analysis of the vaccine site showed a strong CD4(+) and CD8(+) T-cell response after injection of apoptotic cells, which was accompanied by the presence of dendritic cells. In contrast, necrotic cell vaccines attracted a strong local macrophage response. Our data clearly demonstrate that only apoptotic tumor cell vaccines induce a potent antitumor immune response.

Fibrolamellar hepatocellular carcinoma: pre- and posttherapy evaluation with CT and MR imaging.

To determine the features of advanced hepatic and extrahepatic fibrolamellar hepatocellular carcinomas (HCCs) and their effects on immediate surgical management and tumor recurrence. Thirty-one patients with fibrolamellar HCC underwent pretherapy computed tomography (CT); 11 underwent pretherapy magnetic resonance (MR) imaging. All 40 patients underwent posttherapy CT; four, follow-up MR imaging. Imaging, surgical, and histopathologic findings were correlated. Twenty-five (81%) patients had solitary tumors (mean maximum diameter, 13 cm). Thirteen (42%) patients had intrahepatic biliary obstruction; 27 (87%) patients had involvement of the portal or hepatic veins. Thirteen (42%) had extrahepatic tumor spread, nine (29%) had distant metastases on pretherapy images, and 20 (65%) had lymphadenopathy. Thirty-two (80%) of 40 patients underwent exploration surgery; curative resection was attempted in 25 (62%), including four patients who underwent liver transplantation. Only 17 patients were considered to have had hepatic and extrahepatic tumors completely excised. Tumor recurred in all eight of the 17 patients who had extrahepatic disease at pretherapy CT and in four of the seven patients who seemed to have tumor limited to the liver. A combination of repeat tumor resection and adjuvant chemotherapy resulted in prolonged tumor-free survival in some cases. Fibrolamellar HCC frequently demonstrates aggressive local invasion and nodal and distant metastases. Pretherapy and follow-up imaging are important for staging, surveillance, and optimal management. Aggressive surgical resection may be helpful to control fibrolamellar HCC and to prolong survival in appropriately selected cases.

Reoperation for thymoma: Report of 23 cases

Twenty-three patients underwent reoperation for thymoma at intervals of 2 months to 17 years 10 months after the initial operation. There were no operative or hospital deaths. Myasthenia gravis occurred in 12 patients, but in only 2 was it a determinant for reoperation. The longest survival after reoperation is 12 years 9 months, and that patient is free from tumor. Four distinct surgical groups emerged, and their recognition provides an improved method of reporting and suggests a strategy for better overall management. Group 1 (n = 5) had completion of thymectomy (reoperation) after thymomectomy alone or after incomplete thymectomy. The interval was 2 months to 17 years 10 months. All 5 had myasthenia gravis. At reoperation, thymomas were found in 3 and a hyperplastic thymus in 2. Four are alive and tumor free 2 years to 8 years 2 months after reopetation. One died tumor free after 5 years. Group 2 (n = 8) had reoperation for recurrent thymoma after standard (presumably complete) resection. The interval was 2 years to 13 1 / 2 years. Four had myasthenia gravis. Four are alive 8 months to 5 years 8 months after reoperation, 3 without detectable tumor. Four died 3 years 3 months to 8 years 4 months after reoperation, 3 free from tumor. Group 3 (n = 8) underwent reoperation for initially unresectable thymoma after adjuvant treatment with chemotherapy, radiotherapy, or both. The interval was 3 months to 4 years 8 months. Three had myasthenia gravis. Six are alive 4 months to 4 years after reoperation, only 1 with tumor. Two died 1 year 5 months and 6 1 / 2 years after reoperation, both with tumor. Group 4 (n = 2) underwent reoperation for miscellaneous reasons. The interval was 2 months and 3 1 / 2 years. Neither had myasthenia gravis. One had thymoma at reoperation and is alive, tumor free, 12 years 9 months postoperatively. The other, reoperated on for a mediastinal mass at 3 1 / 2 years, had metastatic breast cancer, not thymoma. Reoperation for thymoma can be followed by prolonged tumor-free survival. Thymomectomy alone or incomplete thymectomy should be abandoned. Recurrent thymoma is most often limited to the thorax, thus facilitating reoperation. The response of initially inoperable thymoma to adjuvant therapy favors subsequent reoperation. Slow growth and late appearance of recurrence necessitate lifelong follow-up. Myasthenia gravis, when present, had no deleterious effect on operation.