Early and mid-term outcomes of trans-sternal and video-assisted thoracoscopic surgery for thymoma^†

Abstract

OBJECTIVES

Video-assisted thoracoscopic surgery (VATS) for thymoma has uncertain safety and effectiveness in comparison with trans-sternal resection. This feasibility study compared short- and mid-term outcomes for patients undergoing these two procedures, highlights weaknesses in current research and makes recommendations for long-term technological evaluations in this field.

METHODS

Consecutive thymoma cases between 2004 and 2010 were identified. Patients were divided into two groups according to surgical approach (Group I trans-sternal; Group II VATS) and comparisons were made between groups. The primary outcome was overall survival. Secondary outcomes included operative morbidity and mortality, hospital stay, recurrence rate and disease-free survival.

RESULTS

Thirty-nine patients were included (Group I: n = 22 vs Group II: n = 17). There were no differences between groups at baseline for all measured covariates. No deaths occurred within 30 days of surgery. More patients in Group I developed complications (Group I: n = 10 vs Group II: n = 3; P = 0.093), while hospital stay was shorter in Group II (Group I: 6.4 ± 4.6 days vs Group II: 4.4 ± 1.8 days; P = 0.030). Five-year overall survival (Group I: 93.8 ± 6.1% vs Group II: 83.3 ± 11.2%; P = 0.425), 5-year disease-free survival (Group I: 71.0 ± 15.3% vs Group II: 83.3 ± 11.2%; P = 0.827) and recurrence rates at final follow-up (Group I: n = 2 vs Group II: n = 1; P = 0.363) were similar between the groups.

CONCLUSION

VATS thymectomy for thymoma is feasible, safe and has comparable mid-term oncological outcomes to trans-sternal thymectomy. Future research is required to evaluate long-term oncological outcomes of VATS thymectomy for thymoma in national registries and randomized, controlled trials.

INTRODUCTION

Thymomas are uncommon tumours of the thymus gland, with an estimated incidence of 15 per 100 000 people in the UK [1]. The histological diagnosis of thymoma is broad and includes a spectrum ranging from tumours that are slow growing and benign to invasive thymic carcinoma [2]. Thymomas are usually an incidental radiological finding; however, patients may also present de novo with insidious symptoms of cough, dyspnoea and dysphagia due to mass effect or paraneoplastic features, including myasthenia gravis [3]. Owing to the relatively benign nature of thymomas, the time frame from the onset of disease to diagnosis is often quite protracted, while recurrence can occur up to 10 years after initial management. Complete surgical resection provides the best hope of cure, while patients may also be offered neoadjuvant chemotherapy or adjuvant radiotherapy to improve survival, when appropriate. Because of the relative scarcity of this tumour, there is currently a lack of high-quality evidence to inform surgical decision-making.

Trans-sternal thymectomy is currently considered the gold standard approach to thymoma, based on the perception that large tumours and local invasion can be identified and managed intraoperatively using this technique. In the last 20 years, video-assisted thoracoscopic surgery (VATS) for non-thymomatous myasthenia gravis has become widespread, and technical success rates have improved [4] This has led to growing interest in the role that VATS can play in the management of thymoma. Amidst some controversy, a number of studies have recently published good oncological outcomes for VATS thymectomy for thymoma [5–8]. These studies were, however, limited by the sole inclusion of early-stage tumours, short duration of follow-up and failure to make comparisons between VATS and trans-sternal techniques [9–13]. Indeed, a recent systematic review highlighted that VATS thymectomy for thymoma was technically safe, but there was a lack of oncological data to determine effectiveness, particularly in the long term [14]. Consequently, the current feasibility study aimed to compare early and mid-term outcomes of trans-sternal and VATS thymectomy for thymoma and to make recommendations to guide future long-term outcome research.

PATIENTS AND METHODS

This study comprised a retrospective case series of consecutive patients undergoing thymectomy for thymoma at a single UK institution between September 2004 and September 2010. Although retrospective in design, all data were collected prospectively at the time of events occurring as part of routine hospital records, thus limiting effects of recall bias. The study start date was based on the date of first VATS thymectomy for thymoma at our institution to ensure comparable duration of follow-up and hospital-related factors (e.g. quality and availability of perioperative care). Early outcomes were defined as those occurring within 30 days of surgery, while mid-term outcomes were considered to be those occurring between 30 days and 5 years after surgery. The long duration of mid-term follow-up was deemed appropriate because of the slow-growing nature of thymomas and their tendency to recur up to 10 years after initial treatment.

Patient selection

It is mandatory in the UK for all patients with suspected malignancy to be discussed by a multidisciplinary team (MDT) to obtain consensus on the most appropriate staging investigations and subsequent treatment. Patients were included in the study if the MDT recommended thymectomy on the basis of either histologically proven thymoma or computed tomography (CT) imaging highly suggestive of thymoma (later confirmed histologically). Treatment with neoadjuvant chemotherapy or adjuvant radiotherapy was also permitted, provided patients underwent surgery. All tumour stages were included as defined at final postoperative histology by both World Health Organization (WHO) nomenclature and Masaoka staging [15, 16]. Patients were excluded if they underwent thymectomy for non-thymomatous myasthenia gravis or other mediastinal mass; if they were aged <18 years at the time of surgery; or if the thymoma was deemed unresectable or the patient was considered unfit for surgery by the MDT. Patients were divided into two groups according to whether thymectomy was planned a priori through a trans-sternal approach (Group I) or using VATS (Group II). Decisions on surgical approach were individualized according to each surgeon's experience with VATS and the radiological findings.

Preoperative investigations

Preoperative work-up for patients with presumed thymoma included chest radiography and high-resolution contrast CT of the chest, abdomen, head and neck to determine the extent of local and distant involvement. Additional staging with positron emission tomography (PET) was not used routinely in patients included in this study. Myasthenia gravis was identified through a combination of neurological history, physical examination, isolation of acetylcholine receptor antibodies (AChR MG) or muscle-specific tyrosine kinase antibodies (MuSK MG), repetitive nerve stimulation or tensilon testing. CT-guided needle biopsy of the tumour was not routinely undertaken preoperatively owing to concerns over the potential for tumour seeding. Neoadjuvant chemotherapy was offered to all patients with bulky tumours or clear evidence of extracapsular invasion on CT in order to shrink the tumour and permit R0 resection.

Operative technique

Decisions regarding the operative approach (trans-sternal vs VATS) were made by the thoracic MDT, which comprised thoracic surgeons, radiologists, respiratory physicians, oncologists and pathologists. Surgical approach was based on tumour size, the degree of extracapsular spread and surgeon preference. The VATS approach was usually but not exclusively confined to lesions <5 cm in diameter without suspected capsular breach on CT.

The principle of extended thymectomy was applied to all resections regardless of whether performed through the trans-sternal or VATS approach. This involved removal of all fatty tissue from phrenic nerve to phrenic nerve, and from thyroid cartilage to diaphragm. All fat superficial to the pericardium was also excised en bloc with the thymus. If there was involvement of one of the phrenic nerves, this was sacrificed, while spread to any adjacent structures was managed as appropriate. The decision to perform diaphragmatic plication when the phrenic nerve is divided was taken on a case-by-case basis. When the tumour could not be removed due to extensive involvement of adjacent structures, a debulking procedure was performed, because this has been shown to improve long-term survival [17]. Extended thymectomy was carried out even if intraoperative frozen section biopsies showed complete excision of the thymoma.

VATS thymectomy evolved little over the study period with the exception that progressively larger tumours were excised as experience developed. Technique differed slightly according to surgeon preference; however, VATS thymectomy was generally performed through either right or left approaches depending on the predominant laterality of the tumour. The patient was positioned supine on the operating table with 45° elevation on side of thoracic access, with a towel roll under the shoulders. Single-lung ventilation was used for all VATS procedures to permit better visualization of the mediastinum. Carbon dioxide insufflation was necessary in some cases to facilitate lung collapse. Three ports were used routinely: a superior 10 mm port at the third intercostal space entering beneath pectoralis major muscle, a second port in the fifth intercostal space mid-clavicular line below the nipple and a third port midway between the two for the 30° camera. The thymus was removed en bloc through the widest port in a polythene bag. No attempt was made to fractionate the specimen for ease of extraction. Conversion to open sternotomy was performed if the thymus could not be technically resected thoracoscopically due to tumour size, invasion or risk of damage to adjacent structures. A single apical chest drain was inserted via the most inferior of the access ports, and removed 24 h after surgery.

Follow-up

In the postoperative period, adjuvant radiotherapy was considered following discussion with the MDT in cases where resection margins were involved or if there was evidence of capsular breach (Masaoka grade II or higher). Patients were subsequently followed-up at 4 weeks after surgery and then annually for 10 years. CT imaging was performed routinely to detect signs of disease recurrence at 12 and 24 months postoperatively, as per the institution protocol. Disease recurrence can occur much later than this in patients with thymoma, and if symptoms or signs of recurrence were identified at any time, further CT imaging was requested. This protocol was adopted to ensure early detection of recurrence while minimizing radiation exposure and maximizing appropriate use of limited resources.

Data collection and outcomes

Data were collected from prospectively maintained hospital databases. Details extracted included patient demographics (age, gender, body mass index, smoking, comorbidities and performance status); investigation findings, including AChR-Ab MG and MuSK MG antibody results; operative characteristics (indication, nature and duration of surgical procedure); early postoperative events (mortality, complications and length of hospital stay); histology (tumour stage, resection margins and capsular breach); MDT proceedings regarding the need for adjuvant therapy and the follow-up (duration, recurrence rate and survival).

The primary outcome was overall survival, which was defined as time to death from any cause or time to censorship. Patients lost to follow-up were censored at the date of last contact. Secondary outcomes included operative mortality (<30 days from surgery), complications, open conversion rate for VATS procedures, hospital stay, resection margin involvement, local and distant disease recurrence and disease-free survival. Disease-free survival was defined as death from any cause or disease recurrence (whichever came first) from the day of surgery to final follow-up.

Statistical analyses

Statistical analyses were performed using Stata version 11 (StataCorp, College Station, TX, USA). Categorical data were expressed as frequency and percentages, while continuous variables were expressed as means with standard deviations or as medians with interquartile range (IQR), where appropriate. Categorical variables were compared using the χ² or Fisher's exact tests. Continuous variables were compared using the Student's unpaired t-test or Mann–Whitney U-test. P-values of <0.05 were considered significant. Survival was expressed as percentage ± standard error and Kaplan–Meier analysis was used to plot survival curves. The log-rank test was used to compare survival rates between Groups I and II.

RESULTS

Patients

Sixty-three patients underwent thymic surgery at our institution between September 2004 and September 2010. Twenty-four patients were excluded because they did not have histologically confirmed thymoma, and 39 patients were consequently included in the study (Table 1). Twenty-two patients underwent trans-sternal thymectomy (Group I), while 17 were treated with VATS thymectomy (Group II). There were 19 men and 20 women with similar proportions in each group (P = 0.200). The mean age of the entire sample was 64.4 ± 13.1 years (65.4 years in Group I vs 63.1 years in Group II; P = 0.854). In 17 patients, the thymoma was an incidental radiological finding (n = 13 in Group I vs n = 4 in Group II; P = 0.200), 12 patients presented with thymomatous myasthenia gravis (n = 4 in Group I vs n = 8 in Group II; P = 0.082) and 10 patients presented with symptoms of cough, dyspnoea and dysphagia (n = 5 in Group I vs n = 5 in Group II; P = 0.721). Preoperative body mass index, comorbidities and performance status were similar between the groups.

Operative outcomes

Table 2 summarises the operative details and early postoperative outcomes. The mean operative time was similar between the groups (151.7 ± 63.3 min in Group I vs 177.1 ± 70.2 min in Group II; P = 0.156). Two (11.8%) VATS procedures required conversion to open sternotomy, while one (4.5%) patient in Group I required re-exploration for postoperative bleeding. Reasons for open conversion were both related to tumour adherence to major thoracic vascular structures. All admissions to the intensive care unit (ICU) were planned a priori for observation of myasthenic patients. The mean ICU stay was 1.5 days (2.0 days in Group I vs 1.1 days in Group II; P = 0.540) and only 1 patient required a protracted ICU stay (>2 days) for a slow respiratory wean. There was a trend towards more complications in Group I, although this difference did not reach statistical significance (P = 0.093). Two (11.8%) patients in Group II suffered phrenic nerve injury compared with none (0.0%) in Group I. In both cases, nerve injury was iatrogenic and caused intraoperatively by the removal of the tumour from the perineural tissue. Nerve sacrifice was necessary to increase the likelihood of R0 resection. Diaphragmatic plication was not required in either case because the neural deficit was unilateral and ventilation parameters were satisfactory. There was no operative mortality (<30 days) in either group. The length of hospital stay was significantly shorter in the VATS group (6.4 ± 4.6 days in Group I vs 4.4 ± 1.8 days in Group II; P = 0.030).

Pathological and oncological outcomes

Pathological results, adjuvant therapy and late postoperative outcomes are summarized in Table 3. Data on the pathological tumour size were unavailable for 4 patients (n = 2 in Group I vs n = 2 in Group II). The proportion of patients with postoperative pathology demonstrating that the tumour diameter exceeded 5 cm was similar between groups [Group I: n = 19 (95.0%), Group II: n = 13 (86.7%); P = 0.390]. This is despite tumours with a size of >5 cm on CT usually being recommended by the MDT for treatment with trans-sternal thymectomy. There was, however, a trend towards greater mean tumour volume in patients undergoing trans-sternal thymectomy (417 ± 638 cm³) when compared with VATS (191 ± 206 cm³), although this difference was again not statistically significant (P = 0.199). Capsular breach was observed in 14 patients (63.6%) in Group I and 11 patients (64.7%) in Group II (P = 1.000). Resection margin involvement was evident in a high proportion of patients in both groups [n = 9 (40.9%) in Group I vs n = 8 (47.1%) in Group II; P = 0.756]. Fifteen (88.2%) patients with R1/2 resection demonstrated extracapsular invasion compared with 10 (45.5%) patients who underwent R0 resection. In total, 25 patients had either neoadjuvant chemotherapy or adjuvant radiotherapy, of which 5 had both. Among patients with R1/2 resection, a high proportion (n = 15, 88.2%) received (neo)adjuvant therapy, while only 10 (47.6%) patients with R0 resection required (neo)adjuvant therapy. Specifically, in Group I (trans-sternal), there were 14 (63.6%) cases of extracapsular invasion, 9 (40.9%) had R1 resections, 9 (40.9%) patients had neoadjuvant chemotherapy and 4 (18.2%) had adjuvant radiotherapy. In Group II, there were 11 (64.7%) cases of extracapsular invasion, 8 (47.1%) had R1 resections, 6 (35.3%) patients underwent neoadjuvant chemotherapy and 2 (11.8%) had adjuvant radiotherapy. Pathological tumour volume was not shown to be associated with resection margin involvement in VATS (P = 0.132), trans-sternal (P = 0.758) or all patients combined (P = 0.545). The final WHO and Masaoka staging were similar between groups.

The follow-up was complete at a mean of 33 ± 17.8 months (Group I 35.0 ± 20.2 months versus Group II 30.5 ± 14.3 months; P = 0.44). A total of 3 (7.7%) patients underwent debulking (R2) resection and had evidence of disease progression at the final follow-up (n = 2 in Group I versus n = 1 in Group II; P = 0.363). Two cases of progression occurred in the pleural cavity in Group I, while there was one mediastinal progression in Group II. There were three mid-term deaths occurring between 30 days and 5 years after surgery (n = 1 in Group I versus n = 2 in Group II; P = 0.570), one of which was attributable to disease progression in Group II. Other deaths were due to colorectal cancer and T-cell lymphoblastic lymphoma. Overall survival (Fig. 1) at 5 years was 93.8 ± 6.1 and 83.3 ± 11.2% in Groups I and II, respectively (log-rank P = 0.425). Disease-free survival (Fig. 2) at 5 years was 71.0 ± 15.3 and 83.3 ± 11.2% in Groups I and II, respectively (log-rank P = 0.827).

DISCUSSION

This is one of the first feasibility studies to directly compare outcomes of trans-sternal and VATS thymectomy for thymoma. We have shown that patients undergoing VATS thymectomy have a shorter hospital stay and potentially fewer complications compared with trans-sternal thymectomy. In addition, mid-term follow-up revealed that VATS thymectomy had comparable oncological outcomes in terms of adjuvant therapy received, disease progression and survival. This is in a population of patients that were similar at baseline for all measured covariates. There was a trend towards larger tumours in the trans-sternal group as would be expected from the CT selection criteria for this approach; however, this difference was not statistically significant perhaps due to the small sample size. In addition, a high proportion (>85%) of patients in both groups had tumours in excess of 5 cm in at least one dimension at final pathological examination. This suggests that the CT imaging was inaccurate, the time delay between CT staging and surgery was protracted enough to permit tumour growth or that surgeons undertook resections of progressively larger tumours as experience developed. It was not possible in the current study to determine which of these factors was most important. The final histological WHO and Masoaka staging of tumours were also remarkably similar between the groups despite preoperative selection of patients for trans-sternal surgery on the basis of CT evidence of capsular breach. Indeed, the rates of capsular breach were comparable between groups. VATS thymectomy may, therefore, be undertaken safely in the mid-term in certain patients and should be routinely considered during preoperative MDT discussions. This study falls some way short of being able to recommend VATS thymectomy as a gold standard and raises a number of issues that need to be addressed in order to guide future long-term research into the surgical management of thymoma.

The first issue concerns the small number of patients included in the present study, which is unsurprising given the relative scarcity of thymoma. Indeed, our tertiary referral hospital receives thoracic cases from eight peripheral hospitals, while all thymoma patients treated over a 7-year period were included. We deliberately excluded patients who underwent trans-sternal thymectomy prior to the date of the first VATS procedure to ensure comparability of follow-up. Jurado et al. [18] reported early and mid-term outcomes of trans-sternal and VATS thymectomy in the largest study of its kind to date (n = 263 patients). In this series, only a quarter (n = 72, 27.4%) of patients had histologically proven thymoma, and just 10 of these were treated with VATS. This compares to 39 patients in our study, of which 17 underwent VATS thymectomy for thymoma. Other studies have included similar numbers of patients undergoing VATS thymectomy (Odaka et al. n = 40 [10], Takeo et al. n = 35 [11], Pennathur et al. n = 18, Marulli et al. n = 79) [19], while few have directly compared VATS and trans-sternal surgery [18, 19]. Several problems arise from small sample sizes such as these. Although there were a few, if any, baseline differences between patients undergoing each surgical procedure in either our study or those aforementioned, it is possible that minor differences between the groups may not have been detected. Furthermore, rare but catastrophic complications, such as tumour spillage into the pleural cavity using VATS, were not observed. In addition, comparisons by tumour characteristics, such as disease stage, are unreliable. These factors highlight the need for large-scale studies of the treatment of thymoma and we recommend the development of national or even international registries of patients that are compulsory and record standardized outcomes to facilitate data synthesis and cross-centre comparisons. This would help capture small but clinically important differences between outcomes in patients undergoing each procedure. Such registries would represent a major milestone in the treatment of thymoma and would ensure that patients receive optimal evidence-based care in the future.

The second issue in the current study relates to the duration of the follow-up. We reported a mean follow-up of 33 ± 17.8 months, with slightly longer follow-up in the trans-sternal group (35.0 ± 20.2 months) compared with VATS patients (30.5 ± 14.3 months). Thymoma recurrence tends to occur late in comparison with other tumours, and our survival and recurrence estimates only extend to 5 years, while thymomas are known to recur long after this period. Although numerous studies have reported longer follow-up periods for trans-sternal thymectomy, few have done so for VATS. Indeed, Odaka et al. (22 months) [10], Takeo et al. (65 months) [11], Jurado et al. (24 months) [18], Pennathur et al. (27 months) and Marulli et al. (40 months) [19] reported similar median follow-up periods for VATS patients [20]. The two series comparing trans-sternal and VATS thymectomy [18, 19] both reported similar rates of overall survival at final follow-up in trans-sternal and VATS patients with thymoma. These were also comparable with those observed in the current study. We believe that it is important to report mid-term outcomes for VATS thymectomy and to draw comparisons with trans-sternal surgery at interim analysis to ensure that patients do not undergo oncologically inferior procedures for many years before concerns are identified. Despite this, we recommend that future studies should include oncological follow-up for a minimum of 10 years.

A third issue relates to the selection of patients for each surgical approach. Although there were no measureable differences between the two groups at baseline in our study, the MDT clearly recommended trans-sternal access in some cases and VATS in others. A possible reason for this was that patients with suspected capsular breach, local invasion or larger tumours on CT were probably recommended for open resection, especially at the outset of the study when VATS thymectomy was in its infancy and operator experience with the technique was rightly reserved for those in whom it was expected to be successful [20, 21]. Recent evidence from a systematic review and meta-analysis has, however, shown that capsular breach has no effect on prognosis in thymoma [22], and this should no longer be a selection criterion for open surgery. It is noteworthy that such selection bias may be addressed with randomized, controlled trials (RCTs) in this field. An RCT comparing trans-sternal and VATS thymectomy would be methodologically difficult, however, due to the scarcity of the tumour, the need to power the study on an appropriate patient-centred outcome, the protracted time period between initial management and potential recurrence and possible inconsistency in operator experience with VATS thymectomy at multiple centres. Despite this, there is growing evidence to support the effectiveness of VATS thymectomy for thymoma, and RCTs are recommended in technological evaluations of surgery when the surgical learning curve is stabilizing yet clinical equipoise still exists [23]. This is the current status of research into the definitive management of thymoma.

One detail to highlight from the current study is that the R1/2 resection rate in both groups was high. There are a number of potential reasons for this, which requires further explanation. First, two-stage resections were sometimes undertaken in VATS patients, which involved an initial excision biopsy and intraoperative histological examination of the frozen specimen, followed immediately (during the same operation) by extended thymectomy if the biopsy confirmed thymoma. Two-stage procedures were necessary when the diagnosis of thymoma was uncertain from preoperative CT imaging or the macroscopic appearance of the tumour during surgery. The thoracic MDT at our institution takes a very cautious approach to reporting resection margin involvement to ensure patients maximally benefit from adjuvant treatment. In cases where excision biopsies were performed, the tumour capsule was breached meaning that when the main specimen was removed and examined histologically, the resection margin was reported as R1 at the location of the excision biopsy. In the early stages of this case series, it proved difficult for the thoracic pathologists to orientate the excision biopsy and main specimens and therefore rule out true resection margin involvement. Close cooperation in interpreting the specimen and its orientation was, therefore, required between the surgeon and the pathologist and changes at our institution were made accordingly.

A second possible reason for the high rate of R1/2 resection is that some surgeons in our institution use debulking surgery followed by adjuvant therapy with curative intent. There is conflicting evidence with regard to this management policy. Two patients in the trans-sternal group underwent debulking procedures (R2 resections) and had positive resection margins, while there was 1 patient in the VATS group. Our institution has recently changed its policy and has adopted neoadjuvant chemotherapy for large tumours unlikely to be cleared by surgery, followed by surgery (even if only R1 or R2) and by adjuvant radiotherapy.

The surgical learning curve for VATS thymectomy may also have meant that patients treated with VATS in the early period of this study may have been more likely to undergo R1 resection. Although the small sample size precludes statistical analyses to confirm this, a qualitative review of our data set suggests that the number of R1/2 resections did not reduce over time in either group. Further to this, VATS thymectomy is not a novel procedure for our surgeons; they have practiced this method for non-thymatous myasthenia gravis for many years. The surgical learning curve is probably a minor contributory factor to the observed results.

Finally, there is empirical evidence of publication bias and selective outcome reporting in studies of a wide range of interventional procedures. It is plausible that outcomes previously reported in the literature were selected on the basis of their statistical significance or potential importance, irrespective of the scientific quality of the data presented. For example studies with favourable resection margin results may be more likely to be published. We firmly believe in reporting negative as well as positive findings, as this reduces bias and aids accurate data synthesis in meta-analyses.

In conclusion, the results of the current study have shown that VATS thymectomy is feasible, safe and associated with reduced hospital stay and potentially fewer major complications than trans-sternal thymectomy. In addition, the mid-term oncological outcomes in terms of recurrence rate and survival are not compromised. However, this study has a number of limitations that are representative of the entire field of thymoma surgical research. We, therefore, propose two recommendations: (i) development of a national or international registry of patients treated with thymectomy for thymoma with compulsory registration and standardized outcome reporting and (ii) consideration of a large-scale multinational RCT to compare trans-sternal and VATS thymectomy for thymoma. It is envisaged that this will inform decision-making and improve outcomes for patients with thymoma in the long term.

Conflict of interest: Robert N. Whistance holds an NIHR Doctoral Research Fellowship award. This report is an independent research arising from a Clinical Fellowship supported by the National Institute for Health Research. The views expressed in this publication are those of the authors and not necessarily those of the National Health Service, the National Institute for Health Research or the Department of Health.

REFERENCES

APPENDIX. CONFERENCE DISCUSSION

Dr G. Marulli(Padova, Italy): The topic of your work is one of the most controversial because the minimally invasive approach to early stage thymoma is still considered as experimental due to many concerns about technical feasibility, but most of all oncological safety. To date several studies have demonstrated that VATS thymectomy is technically safe and sound; however, there is still a lack of long-term follow-up to definitively affirm that the oncological value is comparable with the trans-sternal approach. Also your study has the same limitation regarding the length of the follow-up which is short.

I have a couple of questions. The first is about the selection criteria. Your selection criteria for VATS were the radiological evidence of an encapsulated thymoma with a diameter of less than 5 cm. However, in your series you have 86.7% of tumours bigger than 5 cm, 64.7% of capsular breach, and 17% of Masaoka stage III. Do you have an explanation for that? Did you change your selection criteria or your approach on radiological evaluation?

Dr Manoly: With regards to this, as mentioned, it is always a decision from the MDT, and we take opinion from our radiologists because we do not perform CT-guided biopsy anymore. So if the radiologist commented that the tumour was less than 5 cm and there is no capsular breach, at least on the CT, then these patients become eligible for VATS thymectomy.

We have conventional surgeons and also surgeons who perform both VATS as well as trans-sternal thymectomy. So we haven't changed our criteria so far; however, what we have seen is that resection margin or capsular breach was evident in both groups, so maybe we'll have to use a better radiological analysis. We could do that.

Dr Marulli: Secondly, in your series you report a rate of microscopic positive margin, R1, of 59% for trans-sternal and 52.9% for the VATS approach. This is a very big rate of potentially incomplete resections that led you to submit several patients to adjuvant chemo- or radiotherapy in Masaoka stage II that usually is not treated with adjuvant chemotherapy. Can you explain why you have this high rate of R1 resection in early stage thymoma? And given this data and the short follow-up, do you consider a VATS thymectomy really safe from an oncological point of view? I mean, I expect many reoccurrences in the coming years if these data are reliable.

Dr Manoly: That's one of the observations which we have made in our study. The resection margin was often involved in both the groups and we do not know the reason why. We have implemented the principle of extended thymectomy as described in the literature, even when the tumour is supposed to be encapsulated. When we have a doubt that there is evidence of capsular breach, we discuss it with the MDT and then with the patient, and if the patient chooses adjuvant therapy, we just go for it.

Author notes