Predictors of post-thymectomy long-term neurological remission in thymomatous myasthenia gravis: an analysis from a multi-institutional database Free

Abstract

Open in new tabDownload slide

OBJECTIVES

Thymectomy is the treatment of choice for thymomatous myasthenia gravis (MG) for both oncological and neurological aspects. However, only a few studies comprising small numbers of patients have investigated post-thymectomy neurological outcomes. We examined post-thymectomy long-term neurological outcomes and predictors of thymomatous MG using a multi-institutional database.

METHODS

In total, 193 patients (47.3 ± 12.0 years; male:female = 90:103) with surgically resected thymomatous MG between 2000 and 2013 were included. Complete stable remission (CSR) and composite neurological remission (CNR), defined as the achievement of CSR and pharmacological remission after thymectomy, were evaluated. Predictors for CSR and CNR were examined by Cox regression analysis.

RESULTS

The median duration between MG and thymectomy was 3.1 months. In addition, 161 patients (83.4%) had symptoms less than Myasthenia Gravis Foundation of America clinical classification III. All patients underwent an extended thymectomy; there were no perioperative deaths. The 10-year cumulative probability of CSR and CNR was 36.9% and 69.1%, respectively. Mild preoperative symptoms were a significant predictor for CSR (P = 0.040), and a large tumour was a predictor for CNR (P < 0.001). Patients with a large tumour were associated with early MG onset and no steroid treatment. Surgical methods, thymoma stage and histological subtypes were not associated with long-term neurological remission.

CONCLUSIONS

Large tumour size and preoperative mild symptoms were predictors for long-term neurological outcome in thymomatous MG. Considering that patients with early onset of MG and no immunosuppressive treatment tend to have large tumours, early surgical intervention for patients with thymomatous MG having mild symptoms might be beneficial for controlling neurological outcomes.

INTRODUCTION

Thymoma is an uncommon neoplasm derived from thymic epithelial cells and shows heterogeneity in malignant behaviour. Thymoma is associated with numerous types of autoimmune diseases, including myasthenia gravis (MG). MG occurs in ∼15% of patients with a thymoma, and thymoma occurs in ∼10–25% of patients with MG [1, 2]. MG is a relatively rare autoimmune disorder clinically characterized by fatigability and weakness of skeletal muscles, which deteriorate with activity and improve with rest. The clinical course of MG is variable, ranging from early remission to an acute exacerbation and even death. MG is caused by autoimmune antibodies against the acetylcholine receptor at the neuromuscular junction, and abnormalities in the thymic gland play a central role in the pathogenesis of MG [3].

Thymectomy has been the standard of care for treating thymoma [4]. The treatment options for MG may include the combination of thymectomy and various medications, such as anticholinesterase agents, immunosuppressive drugs and immunoglobulin therapy. Current guidelines and recommendations suggest that thymectomy is a valuable option for treating MG because it contributes to a dramatic improvement of the prognosis of MG over several decades along with advances in pharmacological development [3, 5, 6]. Some past studies reported long-term neurological outcomes and predictors in patients with MG after thymectomy [7–12]. However, information about post-thymectomy neurological outcomes and predictors in thymomatous MG is limited, and, given its rare incidence, the reported studies included small numbers of patients.

The purpose of this study was to identify predictive factors for achieving post-thymectomy long-term neurological remission in patients with thymomatous MG from a multi-institutional database.

MATERIALS AND METHODS

The study protocol was reviewed by the institutional review board and approved as a minimal-risk retrospective study (approval no. H-1905-122-1035; approval date: 28 May 2019) that did not require individual consent according to the institutional guidelines for consent waiver.

Patients

The Korean Association for Research on the Thymus recently established a multi-institutional database of patients who were surgically treated for thymic malignancies at 4 tertiary referral hospitals in South Korea.

We compiled 253 patients with thymomatous MG between January 2000 and December 2013 from the database. Among them, we excluded patients (i) who were followed up for <12 months, (ii) who underwent partial thymectomy or thymomectomy, (iii) who underwent thoracotomy for surgical resection of thymic malignancies and (iv) who did not have data on neurological status after the thymectomy. In total, 193 patients were included in the analysis, and their medical records were retrospectively reviewed. Preoperative clinical and demographic characteristics of the study population are summarized in Table 1. The mean age at onset of MG was 47.3 ± 12.0 years; the mean age at thymectomy was 47.5 ± 12.0 years. The study population included 90 male (46.6%) and 103 female (53.4%) patients. The median duration of MG before thymectomy was 3.1 months; and the median follow-up duration was 50.4 months.

Evaluation and management of myasthenia gravis

Diagnosis of MG was conducted by experienced neurologists from each hospital based on clinical criteria, electromyographic scans and serum acetylcholine receptor antibody levels. Preoperative and postoperative clinical assessment and management of MG were also conducted by neurologists from each hospital. The Myasthenia Gravis Foundation of America (MGFA) recommendations for clinical research standards classification [13] were used for the clinical classification and postintervention clinical status assessment. Preoperative MGFA clinical classifications are described in Table 1. Most of the patients had symptoms less than MGFA clinical classification III (n = 161, 83.4% for classes I and II). Twenty-one patients (10.9%) received steroid treatment preoperatively for MG symptoms.

According to the MGFA recommendations, the post-thymectomy neurological status was described as follows: complete stable remission (CSR) (patients who were asymptomatic without having any medication for at least 12 months) and pharmacological remission (PR) (same criteria for CSR except the patient continues to take some form of medication for MG, excluding cholinesterase inhibitors). We set the primary neurological end point as CSR after thymectomy. Because achieving either CSR or PR has been regarded as the neurological end point in previously published articles, we set composite neurological remission (CNR) as the secondary neurological end point; it was defined as an achievement of either CSR or PR after thymectomy.

The postoperative treatment included anticholinesterase, low-dose corticosteroids (10–20 mg/day prednisone) and the immunosuppressant azathioprine. If the patient was symptom free, he or she was progressively weaned off the medications; if not, azathioprine was added to the treatment regimen.

Statistical analyses

The cumulative probability rate of achieving CSR and CNR was estimated with the Kaplan–Meier method. The univariable analysis to potentially predict CSR and CNR was evaluated by the Cox proportional hazard model. Variables with P-values <0.20 were entered into the multivariable analysis. The proportional hazards assumption in the Cox model was assessed using the Schoenfeld residuals. Continuous variables were compared using the Student’s t-test and one-way analysis of variance, and the correlation analysis was conducted using Spearman’s method. The receiver operator curve analysis was used to detect the best threshold for the size of the tumour to predict CSR or CNR. All results were expressed as the mean ± standard deviation or the proportion. Statistical significance was considered with P-values <0.05. All statistical analyses were performed without correction for multiple testing. All statistical analyses were performed using the R software package, version 3.4.3 (http://www.R-project.org).

RESULTS

Operative and pathological data

Surgical resection was performed mostly by median sternotomy (n = 109, 56.5%), and a minimally invasive technique was used in 80 patients (41.4%). All patients underwent extended thymectomy; combined resection of adjacent organs was performed in 22 patients (11.4%) (Table 2).

The distribution of pathological Masaoka-Koga stages was as follows: stage I in 82 patients (42.5%), stage II in 88 patients (45.6%), stage III in 17 patients (8.8%) and stage IV in 6 patients (3.1%). The most common World Health Organization cell type was type B (n = 159, 82.4%); complete resection was achieved in 182 (94.3%) patients. The mean size of the tumour was 4.8 ± 2.5 cm (range 0.6–16.0 cm) (Table 2).

Early clinical and neurological outcome

There were no perioperative deaths. Postoperative complications occurred in 20 patients (10.4%) and included pneumonia (n = 1), bleeding (n = 3), arrhythmia (n = 3), chylothorax (n = 1), prolonged air leakage (n = 9), wound infection (n = 1), reintubation (n = 1) and others (n = 2). Myasthenic crisis was documented in 4 patients (2.1%). The neurological status 6 months postoperatively was symptom free in 73 patients (37.8%), showed improvement in 53 patients (27.5%), was unchanged in 55 patients (28.5%) and was worse or exacerbated in 9 patients (4.7%).

Long-term clinical and neurological outcomes

Seventy-five patients (38.9%) received adjuvant radiotherapy, and 4 patients (2.1%) received adjuvant concurrent chemoradiotherapy for thymoma. The recurrence of thymoma occurred in 11 patients (5.7%). During the follow-up period, 5 patients (2.6%) died of thymoma recurrence (n = 1), of deterioration of the underlying MG (n = 3) and of other medical conditions unrelated to thymoma or MG (n = 1).

The crude CSR and PR rates were 25.9% (50/193 patients) and 23.3% (45/193 patients), respectively; the crude CNR rate was 49.2% (95/193 patients). The cumulative probability of reaching CSR was 21.1%, 29.0% and 36.9% at 3, 5 and 10 years, respectively (Fig. 1A). Mild preoperative symptoms (MGFA clinical classification < III) were the only significant predictor for achieving CSR by multivariable Cox regression analysis. Furthermore, large tumour size was associated with CSR with marginal significance (Table 3). The cumulative probability of reaching CNR was 26.7%, 55.0% and 69.1% at 3, 5 and 10 years, respectively (Fig. 1B). Multivariable Cox regression analysis revealed that large tumour size was the only significant predictor for achieving CNR, and preoperative mild symptoms showed a marginal significance for achieving CNR (Table 4). We performed subgroup analysis for patients with generalized MG (MGFA clinical classification > I, N = 139) and found that MGFA clinical classification ≥III was the only significant negative predictor for CSR [hazard ratio 0.263, 95% confidence interval (CI) 0.081–0.856; P = 0.027] by univariable Cox regression analysis. Additionally, tumour size was a significant predictor for achieving CNR (hazard ratio 1.230, 95% CI 1.128–1.495; P < 0.001) by multivariable Cox analysis. MGFA clinical classification ≥III showed a negative association with CNR but it did not reach statistical significance (hazard ratio 0.543, 95% CI 0.283–1.041; P = 0.066).

Figure 1:

Cumulative probability of (A) complete stable remission and (B) composite neurological remission after thymectomy in patients with thymomatous myasthenia gravis.

Open in new tabDownload slide

We further evaluated the association between tumour size and other MG-related factors. Patients with early MG onset (≤40 years) (early onset 4.7 ± 2.2 cm vs late-onset 4.0 ± 1.8 cm; P = 0.024) and patients with no preoperative steroid therapy (steroid therapy 3.4 ± 1.6 cm vs no steroid therapy 4.3 ± 1.9 cm; P = 0.039) had significantly larger tumours (Fig. 2A and B). There was no association between tumour size and preoperative MGFA clinical classification (MGFA clinical classification I vs II vs >III = 4.5 ± 2.3 vs 4.0 ± 1.8 vs 4.4 ± 1.7; P = 0.21) (Fig. 2C). Because our study population mostly had a short duration of MG symptoms before thymectomy, we evaluated the association between tumour size and symptom duration in patients who experienced more than 12 months of MG symptoms. A significant negative correlation was noted between MG symptom duration and tumour size in this subset of the patients (rho = −0.528; P = 0.004) (Fig. 2D). Furthermore, the receiver operator curve analysis demonstrated that the size of 4.8 cm was identified as the optimal cut-off value for predicting both CSR and CNR.

Figure 2:

Association between tumour size and (A) early onset of myasthenia gravis (MG) (≤40 years); (B) preoperative steroid treatment; (C) preoperative severity of MG symptoms; and (D) duration of MG before thymectomy [subset: patients with prolonged duration of MG symptoms (>12 months)]. MGFA: Myasthenia Gravis Foundation of America.

Open in new tabDownload slide

DISCUSSION

The present study demonstrated 3 major findings. First, there were no operative deaths, and 2.1% of patients experienced an MG crisis after extended thymectomy for thymomatous MG. Second, 65.3% of the study population had symptomatic improvement after extended thymectomy in the early postoperative period. Finally, the 10-year CSR and CNR rates were 36.9% and 69.1%, respectively. The mild preoperative symptoms were identified as a significant predictor for achieving CSR, and large tumour size was a significant predictor for achieving CNR.

The role of thymectomy in patients with MG is based on several lines of evidence that support the role for the thymus in the pathogenesis of the disease, and the clinical benefit of thymectomy has been demonstrated in recent prospective randomized controlled studies [3, 14]. Especially in patients with thymomatous MG, thymectomy should be undertaken as an oncological intervention in addition to treating MG. Some old studies reported MG as a poor prognostic factor for early clinical outcome in patients with a thymoma due to increased surgical risk and risk of death from the underlying MG [15]. However, recent studies demonstrated favourable early prognosis after thymectomy in patients with thymomatous MG, probably due to the improvements in the medical therapy of MG and an earlier diagnosis of a less advanced stage of thymoma. The early clinical outcomes of our study also showed no deaths and a 10.4% complication rate with 2.1% myasthenic crisis rate, which were similar to previous observations [7–11, 16].

CSR has been widely selected as the primary neurological outcome, because it reflects a crucial modification of the natural history of MG. The reported CSR rate after thymectomy in patients with MG exhibits a wide range (10–50%) at the time points of 5–10 years across the studies [7–12, 16]. Regarding thymectomy in MG, thymoma has been considered a poor prognostic factor. Maggi et al. [11] reported a crude CSR rate of 9.6% at 90 months of follow-up, and López-Cano et al. [9] reported a remission rate of 16% for the 10-year follow-up period. Park et al. [7] also reported statistically worse long-term CSR of thymomatous MG than non-thymomatous MG. Our results demonstrated 5- and 10-year CSR rates of 29.0% and 36.9%, respectively. We suggested CNR as another end point in this study to examine long-term neurological remission without symptoms after thymectomy; the cumulative CNR rates were 55.0% and 69.1% at 5 and 10 years, respectively. Similar to results in previous studies, our results showed a delayed response to thymectomy, with increasing CSR and CNR rates with prolonged follow-up. The higher remission rate reported in our study might be attributed to the fact that our study population underwent thymectomy in a more recent period, reflecting improvements in medical therapy for MG and a better understanding of the disease over the decades. However, this possibility must be confirmed.

Our study demonstrated that patients with thymomatous MG with mild symptoms (MGFA clinical classification under III) are more likely to achieve CSR and CNR after thymectomy. This finding was consistent with the recent systematic review by Mao et al. [12]. Of note, most of the thymoma-associated factors, including invasiveness represented by the Masaoka-Koga stage, histological subtype and surgical factors, did not affect either CSR or CNR. This result highlights the importance for long-term neurological remission of an early diagnosis with early surgical intervention of MG when patients exhibit mild symptoms. In addition, the duration of MG was not a predictor of long-term neurological remission. Some authors reported that patients with a shorter duration of symptoms before thymectomy had a better neurological outcome, likely due to prolonged disease duration provoking increasing damage to the neuromuscular plate [17, 18]. The relatively short duration of disease (median 3.1 months) in our study reflects the fact that most of our study population benefitted from early intervention, perhaps profiting from previously published experiences with early thymectomy.

Notably, tumour size was a significant predictor of CNR in our study. Two previous studies evaluated the effect of tumour size on CSR. López-Cano et al. [9] demonstrated that a large tumour was a poor predictive factor, and Lucchi et al. [10] demonstrated no significant effect of tumour size on neurological outcome. In our study, patients with early MG onset (≤40 years), no preoperative steroid therapy and ocular MG tended to have a large tumour (Fig. 2A–C). Moreover, patients with large tumours tended to have a short duration of MG symptoms before thymectomy (Fig. 2D). Late-onset MG, preoperative steroid therapy, more severe preoperative symptoms and prolonged duration of MG symptoms before thymectomy are the poor prognostic factors for neurological remission reported in previous studies [7, 12, 17, 18]. We assume that the patients who had large thymomas were identified at an early age with mild symptoms that did not require immunosuppressive therapy, thereby showing better prognosis with early surgical intervention.

Limitations

Several limitations in our study should be addressed. This study is limited by its retrospective nature without randomization. As a multi-institutional database study, interinstitutional differences in the management of thymomatous MG in terms of surgical indications, operative methods and treatment strategy might exist. Additionally, the considerable time interval covered in the study implicitly includes patients treated with different diagnostic and therapeutic strategies not delineated in our analysis. The recall bias by the patients may exist in each follow-up visit, though CSR and PR have a clear definition by MGFA postintervention status to minimize interobserver discrepancies.

CONCLUSION

We concluded that thymectomy for thymomatous MG is a safe procedure, and long-term neurological remission is favourable. Mild preoperative symptoms were significant predictors of long-term neurological remission, and most of the thymoma-associated factors were not predictors of neurological remission. We also identified a large tumour size as a significant predictor for long-term neurological remission. The trend of earlier age of MG onset and the lack of preoperative steroid treatment in patients with large tumours might indicate that these patients tend to benefit from thymectomy in the early stage of MG.

ACKNOWLEDGEMENTS

The KART comprises the following Korean institutes (representatives): Asan Medical Center, University of Ulsan College of Medicine (D.K.K.); Samsung Medical Center, Sungkyunkwan University School of Medicine (H.K.K.); Seoul National University Hospital, Seoul National University College of Medicine (C.H.K.); and Severance Hospital, Yonsei University College of Medicine (C.Y.L.).

Conflict of interest: none declared.

Author contributions

Kwon Joong Na: Conceptualization; Formal analysis; Methodology; Writing—original draft; Writing—review & editing. Kwanyong Hyun: Conceptualization; Data curation; Formal analysis; Methodology; Writing—original draft. Chang Hyun Kang: Conceptualization; Data curation; Methodology; Writing—review & editing. Samina Park: Conceptualization; Data curation; Writing—review & editing. Hyun Joo Lee: Data curation; Writing—review & editing. In Kyu Park: Data curation; Supervision; Writing—review & editing. Young Tae Kim: Data curation; Methodology; Supervision; Writing—review & editing. Geun Dong Lee: Data curation; Methodology. Hyeong Ryul Kim: Data curation; Methodology; Writing—review & editing. Se Hoon Choi: Data curation; Writing—review & editing. Yong-Hee Kim: Data curation; Supervision; Writing—review & editing. Dong Kwan Kim: Conceptualization; Data curation; Supervision; Writing—review & editing. Seung-Il Park: Data curation; Methodology; Writing—review & editing. Sumin Shin: Data curation; Writing—review & editing. Jong Ho Cho: Data curation; Methodology; Writing—review & editing. Hong Kwan Kim: Conceptualization; Data curation; Writing—review & editing. Yong Soo Choi: Conceptualization; Data curation; Methodology; Writing—review & editing. Jhingook Kim: Data curation; Methodology; Writing—review & editing. Jae Il Zo: Data curation; Supervision; Writing—review & editing. Young Mog Shim: Data curation; Supervision; Writing—review & editing. Chang Young Lee: Data curation; Methodology; Writing—review & editing. Jin Gu Lee: Data curation; Writing—review & editing. Dae Joon Kim: Data curation; Writing—review & editing. Hyo Chae Paik: Data curation; Writing—review & editing. Kyung Young Chung: Conceptualization; Data curation; Supervision; Writing—review & editing.

REFERENCES

ABBREVIATIONS

ABBREVIATIONS
 
• CI

  Confidence interval

 
• CNR

  Composite neurological remission

 
• CSR

  Complete stable remission

 
• MG

  Myasthenia gravis

 
• MGFA

  Myasthenia Gravis Foundation of America

 
• PR

  Pharmacological remission

Author notes