Transxiphoid Bilateral Palpation in Video-Assisted Thoracoscopic Lung Metastasectomy FREE

SURGICAL metastasectomy has become a standard therapy in selected patients with lung tumor metastases.¹ Complete resection proved to be the most important prognostic factor in these patients, who often underwent aggressive and iterative procedures to achieve this purpose.² The frequent discovery of unexpected metastases at intraoperative manual palpation³ has provoked question as to whether video-assisted thoracoscopy (VAT) is adequate in this setting, since it does not allow bilateral manual palpation.⁴ Indeed, VAT has been proposed as a minimally invasive approach for the resection of unilateral metastases,⁵ but the advantages of the procedure may be frustrated by the inaccuracy in detecting nodules. To overcome this limitation, we recently developed a transxiphoid approach through which one can reach both hemithoraces in one operation without performing sternotomy.⁶ The low invasiveness and safety of this approach allowed us to routinely carry out bilateral manual palpation of the lung even in patients with radiologically unilateral disease.

The aim of this experience was to better define the indications and feasibility of the transxiphoid VAT approach in pulmonary metastasectomy, and to investigate the efficacy of transxiphoid manual palpation of both lungs in discovering occult metastasis.

The program of transxiphoid VAT pulmonary metastasectomy started in December 1995. Informed consent was obtained from all patients who were given the fundamental information about the characteristics and the possible risks of the procedure. Until September 1999, a total of 45 patients underwent surgery for pulmonary metastasectomy. Sixteen were judged unfit for the transxiphoid VAT procedure for various reasons that are reported in Table 1. Thus 29 patients comprised the study group: 21 men and 8 women, ranging in age from 19 to 80 years (mean ± SD age, 60.51 ± 15.01 years). Primary tumors originated from carcinoma of the colon-rectum (n = 13), kidney (n = 4), uterus (n = 2), larynx (n = 2), breast (n = 1), prostate gland (n = 1), ovary (n = 1), limb osteosarcoma (n = 3), back fibrosarcoma (n = 1), and skin melanoma (n = 1).

Most of the patients were asymptomatic (n = 19), and the presence of metastasis was discovered during routine follow-up radiological examination. Major symptoms were dyspnea (n = 5) and weight loss (n = 5). The mean ± SD disease-free interval was 32.6 ± 40.1 months (range, 0-204 months). In 2 cases, the metastases were synchronous with the primary tumor; in another 4 patients, pulmonary metastases were discovered at the same time as hepatic ones. In 6 cases, the disease-free interval was longer than 35 months, and in 2 of those cases, longer than 100 months. Because of the nature of the study, we did not plan a sample size a priori.

After the discovery of lung metastases, all patients underwent further examination (total-body computed tomography [CT] scan, ultrasonography, bone scan) to exclude primitive recurrence or extrapulmonary relapse. The lung lesions were also restaged immediately before planned surgery by means of a helical chest CT scan (Tomoscan SR 7000; Philips Medical Systems, Eindhoven, the Netherlands), and without intravenous contrast medium. This also allowed determination of the stability of pulmonary lesions. The mean interval between helical CT scan and surgical exploration was 8 days (range, 4-15 days). The scanning protocol included a tube voltage of 120 kilovolts (peak), a tube current of 250 mA, a slice thickness of 5 mm, and a table increment of 5 mm per rotation. Reconstruction was performed at 5-mm intervals, and images were printed at a window center of −530 Henrys (H), and a window width of 1500 H.

Complete control of the primary tumor, and absent or resectable extrapulmonary metastases were considered the 2 mandatory prerequisites. Other exclusion criteria and their occurrence in the patients excluded from the procedure are presented in Table 1. Previous abdominal surgery did not in itself represent a contraindication to this operation. In 4 instances transxiphoid incision was made along the middle trait of the scar of the previous subcostal incision.

Surgical steps of the procedure can be summarized as follows. Anesthesia was induced through a double-lumen tube, with the patient placed in a 60° off-center position. Exploration started from the side that was deemed more technically demanding either because of the greater number of lesions or a more difficult resection. A thoracoscopic trocar was inserted in the fourth intercostal space between midclavicular and anterior axillary lines to verify the presence of adhesions that might contraindicate the procedure. Another two thoracoscopic ports were created in the fifth and seventh intercostal spaces along the posterior and midaxillary lines, respectively. Thereafter, a midtransverse arcuate skin incision was performed just along the inferior margin of the thoracic cage. The rectus abdominis muscles were longitudinally incised along the linea alba for approximately 8 cm to facilitate hand introduction; thus avoiding traumatic disinsertion from the rib cage. The xiphoid appendix was resected without entering the peritoneal cavity (Figure 1, A), and the lower margin of the ribcage was carefully regularized by a rongeur forceps to avoid injury during hand insertion. Under thoracoscope visualization, the assistant introduced one hand below the sternum and bluntly dissected the retrosternal areolar tissue. Mediastinal pleura were then incised with thoracoscopic scissors, allowing the whole hand to enter the cavity (Figure 1, B). The whole lung (both partially inflated and totally deflated) was carefully palpated between thumb and forefinger under thoracoscopic visualization (Figure 1, C). After this complete preliminary examination, all regions predicted on helical CT scan to be sites of metastases were specifically explored until the predicted lesions could be identified and evaluated for resection. All palpated nodules were excised with minimal resection by endostapler (Figure 1, D) or laser beam, as described in our previous report³; taking care to spare the surrounding healthy parenchyma. During these maneuvers, it was helpful to have one hand inside the hemithorax to present the lesions to be resected or to protect the healthy parenchyma. Lesions were placed in a sterile glove or in a commercially produced retrieval bag and extracted from the chest through the transxiphoid incision. At the end of the procedure, 1 or 2 chest tubes were inserted. Size and location of the resected nodules were immediately recorded.

The same surgical maneuvers were usually repeated on the opposite side using the same subxiphoid incision and entering the hemithorax with the same technique. Although only 10 patients had radiological evidence of bilateral disease, bilateral exploration and manual palpation of both hemithoraces were performed on a routine basis and accomplished in as many as 23 patients. The only reason to refrain from bilateral exploration was the presence of pleural adhesions in the side determined radiologically free of metastases.

Evaluation of acute postoperative pain was assessed using a visual analog scale (VAS) that graded from 0 (no pain) to 10 (most severe imaginable pain) as an index of the patient's perception of pain. The VAS was explained to patients preoperatively and consisted of patients marking a grade of pain on a graded ruler. Pain was quantified both at basal level and during particular movements. Assessments were made 3 times daily for the first 3 days, then once daily until the discharge. Analgesic requirement and modification of pain perception after analgesic administration were recorded. Postdischarge follow-up visits occurred at 1, 3, and 6 months after operation, and pain perception was determined according to the same VAS.

All patients with metastatic osteosarcoma had adjuvant chemotherapy consisting of 100 mg/m² of cisplatinum, 60 mg/m² of doxorubicin, and high-dose methotrexate. One patient with laryngeal carcinoma also had postoperative chemotherapy, which consisted of 1 mg/m² of vincristine, 150 mg/m²of bleomycin sulfate, and 20 mg/m² of methotrexate. Patients underwent total-body CT scans twice per year for the first 2 years; ultrasonography, a bone scan, and a laboratory examination were also performed when indicated.

There was neither operative mortality nor major morbidity. All patients had a fast recovery. Acute and chronic postoperative pain was compared with the pain provoked by other VAT procedures that did not imply a transxiphoid approach, and by bilateral metastasectomy performed via sternotomy in the same period and with the same analgesic protocol (ie, epidural administration until postoperative day 2, followed by continuous intravenous drug delivery). No significant difference by the Mann-Whitney test was found in connection with the other VAT procedures in any of the acute postoperative and postdischarge determinations. Compared with median sternotomy, pain was less significant at day 1 (mean basal pain score, 1.5 vs 2.7; P = .01), day 5 (mean basal pain score, 0.9 vs 1.8; P = .002), 1 month (mean at movement, 0.2 vs 1.5; P = .09), 3 months (mean at movement, 0.2 vs 1.3; P<.001), and 6 months (mean at movement, 0.1 vs 1.0; P<.001). Chest drainage was removed after a mean ± SD of 2.86 ± 1.19 days (range, 2-7 days); mean ± SD hospital stay was 4.37 ± 1.78 days (range, 2-10 days), which is significantly shorter than that of patients who underwent single wedge resection through a thoracotomy (6.1 ± 2.02 days) or metastasectomy through a sternotomy (10.5 ± 1.59 days) in the same period. Sixty-nine nodules (mean ± SD diameter, 11.05 ± 9.3 mm) were palpated and resected, of which only 54 had been detected on CT scans. Bilateral palpation was possible in 23 patients. In the remaining 6 patients, exploration was restricted to the side with visible metastases, because of pleural adhesions in the contralateral hemithorax.

Twenty-eight nodules were located in the upper lobes (14 right, 14 left), 32 were found in the lower lobes (15 right, 17 left), and 9 were detected in the middle lobe. Seven lesions were located in the visceral pleura. The maximum number of resected nodules per patient was 6. The mean ± SD operative time for bilateral metastasectomy was 117.3 ± 24.4 minutes (range, 95-170 minutes). In 3 cases, we preferred to perform a rib-spreading mini-thoracotomy of 6 to 8 cm through which it was possible to better circumscribe the lesion and separate it from adjacent major vessels (this was due to the presence of deeply located metastasis). This incision was usually accomplished by prolonging a trocar port. Twenty-nine lesions were resected by laser, and 40 by stapler.

Sixty lesions were identified as metastatic at histological examination. Bilateral palpation allowed the discovery of 11 unexpected metastases in 9 patients (Table 2). Five patients who were believed to have unilateral metastases at CT scan were proven to have bilateral metastases. The diameter of occult lesions ranged between 3 and 12 mm (mean ± SD size, 6.06 ± 2.78 mm), and 10 of these were smaller than 6 mm in diameter. There were 5 false-positives with a mean ± SD diameter of 8.00 ± 4.47 mm. Three of them were smaller than 6 mm in diameter. According to histological examination, these nodules were fibrotic tissue (n = 2) and benign chondromas (n = 2). One more nodule diagnosed using the CT scan was not detected at palpation, and it was interpreted as a small blood vessel.

According to the prognostic classification formulated by the International Registry of Lung Metastases² 1 patient was classified as group I, 16 as group II, and 12 as group III.

After a mean ± SD follow-up of 22.89 ± 10.87 months (range, 9-60 months), no patient showed signs of subxiphoid incisional hernia. Four patients died of the disease at 15, 16, 22, and 38 months following the metastasectomy. Thirteen patients relapsed in various sites; however, only 6 patients developed pulmonary metastases after a mean interval of 12.8 months. Three of these 6 patients' tumors metastasized in the unexplored lung after intervals of 6, 9, and 12 months, respectively. No recurrence was found in the transxiphoid incision, along the thoracoscopic ports, or adjacent to previous resection margins. Data on relapse and second free intervals are presented in Table 3. Although selection biases do not allow statistical comparison, overall recurrence rate (6/29 vs 5/16) in the lungs was lower than that of the remaining 16 patients operated on in the same period; this was likely attributable to later findings in the nonexplored lung.

Radical resection is considered worldwide the aim of surgery for selected patients with lung metastases, thus encouraging aggressive^7- 9 and iterative operations.^10- 11 This aim was basically limited by the relative inaccuracy of preoperative radiological studies in detecting all pulmonary metastases¹² and by the difficulty of approaching both lungs with a single operation.⁴ The use of CT scanning produced considerable diagnostic progress, but small-sized metastases still remained difficult to identify.⁶ The frequent discovery of occult metastases at manual palpation convinced many medical centers to propose intentional bilateral exploration¹ also in patients with radiologically unilateral disease.¹³ According to this purpose, routine exploration of both lungs via median sternotomy has been proposed for sarcoma¹⁴ even in the absence of demonstrated bilateral metastases. Other centers voiced some criticism toward such an aggressive approach, which has a too high a cost in terms of morbidity compared with the global benefit.¹⁵

At the same time, the advent of VAT has produced new waves of interest and criticism. Indeed, VAT seems to be an ideal tool for the surgery of pulmonary metastases, which most of the time are peripheral and responsive to excision via wedge resection.⁵ Both diagnostic and therapeutic aims can be pursued with significantly decreased postoperative pain, functional impairment, and hospital stay.¹⁶ Furthermore, minimally invasive procedures seem preferable as the first approach in patients who are often predetermined to receive iterative surgery. On the other hand, VAT inherits the theoretical defects of a unilateral approach, but its major limit is the impossibility to perform the thumb-forefinger palpation that still remains the most accurate method of detection of metastases.^3- 4 A prospective study undertaken to compare its value with that of CT and open thoracotomy was prematurely closed to accrual, showing a 50% probability to miss metastases by VAT⁴; this is likely owing to the impossibility of performing manual palpation. Even the introduction of helical CT techniques proved less effective than manual palpation in detecting the smallest metastases.^17- 18

Although it is generally believed that complete resection of pulmonary metastases confers both a disease-free and overall survival advantage (as proven by the International Registry of Lung Metastases on a population of 5206 patients), the prognostic significance of these smaller lesions that were undetected on helical CT scan is controversial. For this purpose, a specific trial named CALBG 39804, comparing VAT vs open resection, is currently in progress.¹⁹

We developed a new transxiphoid approach without using sternotomy or clamshell incision, for allowing manual palpation of both lungs in VAT metastasectomy.²⁰ The approach proved technically feasible and secure. Postoperative adhesions are relatively scant, and repeated resection is possible. The theoretical risk of cardiac tamponade or arrhythmia during lung palpation was deemed tolerable in light of the advantages of the procedure: decreased postoperative pain, shorter in-hospital stay, minimal surgical trauma with no bone separation, and easier patient acceptance. All lung regions can be easily reached and palpated. Resection is facilitated with one hand inside the thorax, thus avoiding tedious maneuvers for appropriate endostapler positioning. Moreover, a transxiphoid incision may also represent another site for endostapler introduction and a less painful site for chest tube insertion.

Nonetheless, in this study, we focused our attention on the potential oncological benefits of the procedure, which are usually the more criticized aspects of VAT approaches. Despite the use of the helical CT scan, the most accurate device in detecting metastases, in our selected population manual palpation was able to identify 15 (22%) of 69 supplementary unexpected lesions. Eleven (73%) of them, which were found in one third (9/29) of the patients, had metastases at histological examination. On the other hand, another 5 (10%) of the 54 lesions deemed metastatic on the CT scan were benign at histological examination, and 3 patients were falsely diagnosed as having a bilateral disease. This seems to us to be another important reason to explore.

The low sensitivity and subsequent incomplete resection may reasonably have an effect on recurrence rate and long-term survival. This is valid for either an unexplored contralateral lung or a nonpalpated ipsilateral lung, especially in patients operated on for metastatic sarcoma. Although median sternotomy or thoracotomy approaches proved non–statistically different benefits on long-term survival,¹³ we believe that recurrence rates may be influenced by the pattern of exploration. The uncontrolled character of our study does not authorize definitive conclusions, but it was our observation that one half of the nonpalpated lungs relapsed, and this frequency was consistently higher than those (5%) found amongst the palpated lungs (Table 3). Also, recurrence time was shorter in the subset of patients with nonpalpated lungs (9 vs 16 months) and this may likely be owing to occult metastases situated on a nonaccessible side. Once again, patients with sarcoma revealed those patients with the highest probability of overlooked lesions and false unilateral disease (Table 3). However, in contrast with previous statements,²¹ we also found multiple and unexpected metastases within the carcinoma group in a consistent percentage of cases. All these findings suggest that whenever possible, and if achievable with the same morbidity rate, bilateral palpation represents the best way to perform radical surgery in these patients. This maneuver, once possible only through aggressive incisions, can be easily and satisfactorily performed through the transxiphoid VAT approach. Limits of the procedure may be its use on centrally located metastases or lesions requiring lobectomy, cardiomegaly, diffuse pleural adhesions, or previous sternotomy. However, the procedure is feasible in the majority of patients selected for lung metastasectomy. In our series, despite the wide exclusion criteria, as many as 64.4% of the patients who underwent pulmonary metastasectomy were able to undergo this approach.

To conclude, despite the introduction of helical CT, the relevance of occult metastases is still not negligible and has led us to prefer routine bilateral palpation of the lung, which can be easily accomplished with the transxiphoid approach. We recommend that surgeons use this approach whenever feasible, as it seems to conciliate 2 apparently incompatible issues: low invasiveness and radicality.

This work has been carried out within the research fellowship program, Dottorato di Ricerca in Oncologia Toracica, appointed by Tor Vergata University of Rome, and was supported by a grant from Ministero dell' Universitá e della Ricerca Scientifica e Tecnologica, Rome (60%).

Corresponding author and reprints: Tommaso Claudio Mineo, MD, Cattedra di Chirurgia Toracica, Università Tor Vergata, Policlinico Tor Vergata, Viale Oxford 81, 00133 Rome, Italy (e-mail: mineo@med.uniroma2.it).