73: Segmentectomy for Primary Lung Cancer

Segmentectomy was initially described by Churchill and Belsey¹ in 1939 for the treatment of bronchiectasis. Although the operation is still used to treat suppurative and other nonmalignant processes (e.g., aspergilloma, pulmonary sequestration), other pulmonary infections, pulmonary abscesses, and benign tumors of the lung (hamartomas, papillomas), this chapter concerns its controversial use in early-stage lung cancer.² Until 1950, pneumonectomy was the standard of care for lung cancer. However, increasing awareness of the diminution of respiratory function caused by pneumonectomy soon led to interest in lobectomy and other lesser resections for tumors of amenable size and location. In 1973, Jensik et al.³ reported the first series of segmentectomies for early-stage lung cancer. Since then, limited resection for lung cancer has been a topic of much debate, and the controversy has been plagued by conflicting results between studies comparing segmentectomy and standard lobectomy or pneumonectomy.

Segmentectomy is an anatomic sublobar resection that involves the removal of functionally discrete units of the bronchovascular anatomy. The bronchovascular architecture is composed of a series of individual segments. Each segment has a pyramidal structure with its apex at the hilum and its base on the surface of the lung. Individually, the segments are supplied solely, with few collateral connections, by the following structures: (1) a segmental bronchus as a tertiary branch of the bronchial tree; (2) a segmental branch of the pulmonary artery (as well as the bronchial artery); and (3) a segmental (± intersegmental) branch of the pulmonary vein together with lymphatics. Each segment behaves as a discrete anatomical and functional unit that can be removed by segmental resection without affecting the functionality of the remaining lobe or adjoining bronchial segment. In properly selected patients with early-stage non–small-cell lung cancer (NSCLC), segmentectomy can achieve outcomes that are equivalent in overall survival to pneumonectomy and lobectomy.

A thorough knowledge of the human lung anatomy is mandatory for any surgeon undertaking this resection. There are 10 segments in the right lung (3 in upper lobe, 2 in middle lobe, and 5 in lower lobe) and 8 to 10 segments in the left lung (4–5 in upper lobe and 4–5 in lower lobe) (Fig. 73-1).

Several operations fall under the umbrella of sublobar pulmonary resection. These are the wedge, the segmentectomy, and the extended segmentectomy. It is important to note that a wedge resection, sometimes called an atypical segmentectomy because of its sublobar resection status, is not an anatomical resection. Although sometimes preferred because it is less technically demanding than segmentectomy, the wedge resection is associated with numerous pitfalls, including a difficulty in obtaining or identifying a tumor-free resection margin, limited extent of lymph node sampling and excision, high rate of local recurrence, and diminished overall survival. Segmentectomy, on the other hand, respects anatomic barriers and is associated with better overall and disease-free survival. Extended segmentectomy describes the technique whereby the parenchyma is divided lateral to the intersegmental plane to accommodate a wider resection margin.

Two groups of patients with NSCLC may benefit from segmentectomy, those who are able to tolerate lobectomy, but for whom curative resection is likely with a sublobar resection based on the small size of the tumor and negative nodes. The second group is those who are unable to tolerate lobectomy and for whom a lesser resection constitutes an alternative local cancer treatment. Patients with poor pulmonary reserve (forced expiratory volume in 1 second [FEV₁] less than 50% of predicted) or with multiple resectable lesions may also be considered for a segmental resection if complete resection is likely.

Conventional fractionated radiation therapy has been the alternative local treatment for medically inoperable patients with early-stage NSCLC and is associated with modestly prolonged survival compared to observation. Recently, several nonsurgical treatments have become available, including stereotactic body radiation therapy (SBRT) and percutaneous ablative therapy including radiofrequency, cryotherapy, and microwave (Chapter 85). Although these treatments appear to decrease the risk of respiratory failure, disability, and death, there is currently little evidence of efficacy compared to a parenchymal-sparing surgical resection. In the few trials that are available, all nonoperative modalities have higher recurrence rates over shorter intervals. Currently, two ongoing randomized trials are investigating the value of SBRT versus sublobar resection or lobectomy.

Preoperative workup and evaluation for thoracic surgery is discussed elsewhere in this book in detail (see Chapter 4). The same principles apply for segmentectomy.

The presence of stage I NSCLC should be ensured preoperatively. Combined PET/CT scan is the best radiographic and metabolic test available for ruling out extrathoracic disease, assessing the mediastinal and hilar lymph nodes, and excluding other suspicious nodules in the lung. Endobronchial ultrasound (EBUS) and endoscopic esophageal ultrasound (EUS) are first-line tools for mediastinal staging and should be used followed by videomediastinoscopy to rule out false-negative results when suspicious-looking lymph nodes are detected on radiographic imaging. MRI of the brain with contrast should be considered in all patients with headache or other neurologic symptoms, since a brain metastasis cannot be visualized by PET/CT.

Medical operability should be checked in patients with poor performance status if any surgical procedure is being planned. Algorithms for differentiating risk levels for patients being considered for a lung resection have recently been published.⁴ These European guidelines provide cutoff values for subjecting at-risk patients to additional assessment and threshold values to differentiate low-risk from high-risk patients. Cardiology risk stratification and pulmonary function testing including the diffusion capacity of carbon monoxide for the lung (D_LCO) are recommended in every patient undergoing pulmonary resection. Further exercise testing with oxygen consumption (VO₂) should be performed in patients with FEV₁ and/or D_LCO <80%. Patients with VO₂ max or a predicted postoperative (ppo) value below 10 mL/kg/min or 35% are not candidates for major anatomic resections. In patients with VO₂ max values between 10 and 15 mL/kg/min or between 35% and 75%, ppo FEV₁ and ppo D_LCO should be calculated. If both values are >30%, resection can be performed up to the calculated extent. Otherwise, ppoVO₂ max should be calculated.

The patient is placed in lateral decubitus position and intubated under general anesthesia with a double-lumen endotracheal tube. An epidural catheter is placed for pain. The chest is usually entered through the fifth intercostal space or one rib higher if there is a concern for pathology at the lung apex. The operation can be performed using open or video-assisted thoracic surgery (VATS) technique. For the open procedure described later, we routinely perform a muscle-sparing posterolateral thoracotomy with division of the latissimus dorsi and mobilization and retraction of the serratus anterior (see Chapter 2).

After entering the chest, the hemithorax is inspected and lung palpated to rule out evidence of advanced disease that would preclude segmental resection. If there is any uncertainty about the preoperative diagnosis, tissue is obtained for frozen-section analysis before proceeding. Central lesions may be sampled via needle biopsy. Frozen sections then are obtained of N1 and N2. The presence of metastatic disease in any lymph node constitutes an indication to proceed with lobectomy,⁵ provided the patient is surgically fit based on the preoperative assessment. Frozen section of the resection margin is also recommended.⁶

Segmentectomy

Although any bronchovascular segment can be removed, certain operations are more commonly performed. These include taking or sparing the superior segment for lower lobe cancers and taking or sparing the lingula for left upper lobe cancers. It is generally easier to remove the superior segment of the lower lobe (S6), the lingular segments (S4 + S5), and the basilar segments of the lower lobe (S7–S10), whereas the individual segments in the upper lobe (S1, S2, S3) and lower lobe (S7, S8, S9, S10) are more challenging. Although the spatial approach and the order of dividing the individual bronchovascular structures may vary depending on the individual segment(s), the overall principles remain the same.

Regardless of which segment is being resected, the fissure is opened first to reveal the pulmonary arterial branch. The appropriate segmental artery(ies) are identified and divided in the usual manner to expose the underlying segmental bronchus. Gentle traction on the segment with an atraumatic clamp may help to expose segmental branches that are hidden deep within the lung parenchyma (Fig. 73-2).

The S4, S5, and S6 segments each have an individual central vein that can be ligated or clipped, whereas the veins in other segments run close to the periphery and cannot be identified until dissection of the intersegmental plane has commenced and drainage into the superior or inferior venous trunks becomes visible. In some cases, early division of the individual central segmental vein can facilitate visualization of the pulmonary artery. However, these veins cannot be identified upon stapling of the parenchyma, so it is important to identify the anatomy of the venous trunks carefully before dividing the segmental vessels to avoid inadvertent ligation of veins draining blood from adjacent segments. This may result in venous thrombosis, lobar infarct, and potentially disastrous complications postoperatively.

Intraoperative bronchoscopy can be very helpful if there is confusion regarding the segmental anatomy or concern about potential compromise of the adjacent segmental bronchial orifice. A pediatric bronchoscope fits easily through the lumen of a double-lumen endotracheal tube to reveal the endobronchial view. In addition, the light on the scope illuminates the airway which can be visualized from the operative field.

Before dividing the parenchyma, it can be difficult to identify the appropriate plane, especially when VATS techniques are used. Observing the differential ventilation of the individual segments by clamping the segmental bronchus before (or after) full inflation can better delineate the intersegmental plane (Fig. 73-3). The segmental bronchus is then divided with a stapler or scalpel and closed in routine fashion. For suturing the bronchial stump, we prefer two over-and-over running 4-0 absorbable monofilament sutures after folding the membranous part inside the lumen of the bronchus and bringing the cartilaginous rings together according to the technique previously described by the Dutch surgeon Klinkenberg. Finally, the parenchyma between the involved and adjacent segments needs to be divided for which two techniques have been described: the open and stapled division.

For the open technique, the intersegmental plane is teased apart by using a clamp to place traction on the stump of the transected segmental bronchus, whereas the rest of the lung is well ventilated. Sharp and blunt finger dissection may also help to open the intersegmental plane (Fig. 73-4). Cautery, harmonic scalpel, or small vascular clips are used to ensure hemostasis. Small air leaks can be oversewn with fine sutures. In addition, the raw surface of the adjacent segment can be covered with pleural or pericardial fat pad. Care has to be taken to avoid compression or kinking of the remaining bronchovascular structures which can result in a nonfunctional lobe thereby eliminating the benefit of segmentectomy versus lobectomy. The advantage of the open technique is that reexpansion of the adjacent parenchyma is maximal, but it carries a higher risk of prolonged air leaks.

For the stapled technique, the virtual fissure is compressed and cut with the aid of a linear stapler. We prefer an endovascular linear cutting stapler. This technique results in better pneumostatic control in the remaining lung but it comes at the expense of volume loss as the visceral pleural layers are drawn together when closing the device. The remaining parenchyma is then somewhat trapped by the individual staplers, which blocks reexpansion of the lobe to its maximum volume (Fig. 73-5). The “extended” segmentectomy is accomplished by deploying the stapler lateral to the intersegmental plane so as to include the adjacent subsegments in the specimen.⁷

VATS segmentectomy is a safe and feasible operation, as confirmed by published reports from several high-volume centers. The benefits of using VATS with segmental resection seem to mirror the more robust data available for VATS lobectomy in terms of decreased morbidity, length of stay, postoperative pain, and other metrics. However, caution must be exercised as these resections can be quite challenging and one should always proceed as if performing an open procedure with careful individual dissection and ligation of the bronchial and vascular structures. Port placement can literally make or break VATS segmentectomy. In general, the standard port sites described by McKenna or Flores are the most versatile. McKenna routinely uses four incisions, whereas Flores⁸ describes only three for VATS lobectomies. We have found both to be appropriate for performing segmental resections. There are certainly times when it is quite easy to perform an operation through only three ports, while at other times, the fourth port significantly increases the options and ease of the resection. In general, we start with three and add a fourth port anteriorly, as described by McKenna, or at the scapular tip as needed.

Right Upper Lobe Segments

The apical segment dissection begins by incising the mediastinal pleura. The apical branch of the anterior arterial trunk is identified and divided. The apical segmental bronchus is then approached posteriorly and isolated after ligating the bronchial artery branches. For right upper lobe segmentectomy, the vein is usually encompassed in the staple line.

The anterior segment is approached from the medial aspect, beginning with incision of the mediastinal pleura. The anterior segmental artery is identified as it branches from the anterior arterial trunk. The anterior segmental vein is likewise identified and ligated, taking care not to compromise other segmental tributaries to the superior pulmonary vein. The horizontal fissure is completed, the anterior segmental bronchus is divided, and the segment is excised.

In performing a posterior segmentectomy, the major fissure is opened posteriorly and the pulmonary artery is traced proximally to identify the posterior segmental artery. Once identified, it is ligated. The posterior segmental bronchus is deep and lies slightly medial to the artery. Again, the vein is taken with the intersegmental plane. Alternatively, the artery branch may be ligated after dissection and transection of the segmental bronchus, which can be tracked posteriorly along the right upper lobe bronchus.

Superior Segment, Right Lower Lobe

If the major fissure is well developed, the pulmonary artery can be approached directly in the fissure. If the fissure is incomplete, it must first be developed to expose the artery. The pleura is opened posteriorly at the bifurcation of the right upper lobe bronchus and the bronchus intermedius and continued anteriorly to identify the recurrent artery to the posterior segment of the upper lobe and the superior segmental artery to the lower lobe. With these bifurcations visualized, the posterior portion of the major fissure can be developed. A stapler can be used to divide this parenchyma, taking care to exclude the arteries and bronchi from the jaws of the stapler. Sometimes it is helpful to place a ¼ in Penrose drain through the “window” that has been created and feed the lower jaw of the stapler into the drain. The drain can then be used to guide the stapler through the window without risk of falsely passing the stapler and injuring or dividing a vessel or bronchus. The posterior fissure can be a difficult spot to manipulate the stapler. Passing the stapler through the future chest tube site may afford a more benevolent angle of attack.

Once the posterior portion of the major fissure is opened, the artery can be isolated and divided. This will provide exposure to the bronchus, which runs deep to the artery. The superior segmental vein can usually be identified posteriorly as a separate tributary running into the inferior vein. Again, care must be taken not to compromise the remainder of the vein.

Basal Segment, Right Lower Lobe

This operation begins in the major fissure with dissection of the pulmonary artery to identify the right middle lobe artery, the superior segmental artery, and the basilar segmental artery. Circumferential dissection of the artery allows it to be divided. The basal segmental bronchus lies deep to artery. The inferior pulmonary vein is identified by first taking down the inferior pulmonary ligament. The anterior and posterior hilum is dissected to complete the inferior pulmonary vein isolation. Further dissection into the lung parenchyma permits visualization of the confluence of the segmental pulmonary veins. The bronchus is then dissected and divided. This allows demonstration of the veins deep to the bronchus. The superior segmental vein is identified and the intersegmental plane is completed using this vein as a guide, making sure that it is not compromised.

Left Upper Lobe Upper Division (Lingular Sparing)

Proximal control of the pulmonary artery may facilitate left upper lobe segmentectomy, but it is not mandatory. First, the anterior mediastinal pleura is opened to permit identification of the superior pulmonary vein. The lingular vein is identified and preserved. We begin by dividing the upper division segmental vein. A lymph node is normally encountered at this bifurcation and sent for frozen section. If positive, the segmentectomy should be abandoned in favor of lobectomy. Dividing this vein facilitates identification of the left main pulmonary artery. The dissection is continued distally along the main pulmonary artery to expose the segmental arteries, which are divided. At this point we find it beneficial to complete the oblique fissure. With this accomplished and the arterial branches divided, only the bronchus remains. The upper lobe bronchus can then be dissected up to the bifurcation of the upper division and lingula. This allows the upper division bronchus to be stapled and divided. Placing the stapler along the pulmonary vein from the lingula and heading toward the lingular bronchus, the intrasegmental fissure can be developed and completed.

Lingula

As with most segmental resections, the ease of dissection is a function of the completeness of the fissure. If the fissure is complete, it is quite easy to identify and isolate the artery and all its branches. If the fissure is incomplete, it must first be developed and this is best accomplished by tracing the pulmonary artery into the fissure. Commonly, a plane can be developed between the lobes using sharp dissection or a very low setting on the electrocautery to avoid significant air leaks. If raw parenchyma is encountered as the fissure is developed, we complete the fissure with a stapling device. The lingular arteries may arise separately as a common trunk from the ongoing pulmonary artery. The lingular bronchus will be found under the divided arteries. Care must be taken to avoid compromise of the superior divisional bronchus to the upper lobe when dividing the lingular bronchus. The lung is then retracted posteriorly, and the hilar pleura is incised to expose the superior pulmonary vein. The lingular branch is then identified and divided (Fig. 73-6). Finally, the parenchyma is divided as described for other segments. The other option for this procedure is to begin with identification and division of the lingular vein in the hilum. Attention then turns to the fissure for dissection and division of the appropriate branches of the pulmonary artery and bronchus.

Basal Segment, Left Lower Lobe

After releasing the inferior pulmonary ligament, attention is turned to the hilum. The inferior pulmonary vein is identified and circumferentially dissected. There is usually a lymph node that can be dissected from the posterior superior aspect of the inferior pulmonary vein, and this aids in completing the dissection. The subcarinal lymph nodes are dissected next. The branches of the inferior pulmonary vein are further delineated. Attention is turned to the pulmonary artery in the major fissure. The fissure is completed and the segmental artery to the basal segments is circumferentially dissected. It sits directly anterior the basal segmental bronchus. This artery (sometimes arteries) is divided. The basal segment bronchus is then dissected between the inferior pulmonary vein and the bronchus. The stapler then comes from anteriorly and the bronchus is stapled and divided. Care must be made to not compromise the superior segmental bronchus. With the basal segmental bronchus divided, the veins are easily identified and divided. The intersegmental fissure is completed using the superior segmental vein and artery as a guide.

Superior Segment, Left Lower Lobe

This resection is accomplished in a similar fashion to the right lower lobe superior segmentectomy. The oblique fissure is completed first which permits identification of the main pulmonary artery. The branch(es) of the pulmonary artery to the superior segment are then dissected and divided, allowing visualization of the segmental bronchus that lies immediately behind and inferior to the divided pulmonary artery. The segmental bronchus is then dissected out and the N1 lymph nodes are removed. Once this is done, the segmental bronchus is divided. If the segmental vein is identified, it can either be individually ligated or incorporated into the completion of the intersegmental fissure using a staple technique.

The postoperative care after segmentectomy is comparable to other patients undergoing pulmonary resection. Upon closure, we usually leave two 28 Fr drains in the chest, the first directed toward the apex and the second above the diaphragm. However, one author (Shamus Carr) only uses a single 24 Fr chest tube directed posteriorly toward the apex with additional holes cut in the tube to aid drainage near the diaphragm. We do not put the drains on suction routinely unless there is an increasing pneumothorax or subcutaneous emphysema due to a large air leak. The focus of postoperative care is directed toward adequate analgesia and pulmonary toilet. A dedicated pain team visits patients daily to monitor the patient-controlled analgesia pumps. Dedicated physiotherapists, early ambulation, regular portable chest x-rays, and a low threshold for bronchoscopy are all helpful to minimizing the risk of pulmonary complications.

Morbidity and mortality after segmentectomy are comparable to lobectomy in most reported series. It is unusual to have a residual space problem after segmentectomy compared with lobectomy. However, the postoperative course can be complicated by atelectasis in the remaining segments as a result of sputum impaction, disturbed bronchial anatomy, or by persistent air leaks, especially if stapling devices were not used to divide the intersegmental plane in an attempt to preserve maximum volume of the remaining lung.² A hematoma in the adjacent lung parenchyma can sometimes be seen on chest radiographs, but this process usually has begun to resolve prior to discharge and has completely resolved by the first outpatient follow-up visit. Signs of a progressive lobar infiltrate unique to the remaining lobe may be the first sign of lobar infarct. Reexploring the patient may be indicated after appropriate imaging and bronchoscopic workup. Not recognizing this complication may lead to sepsis and increased morbidity and mortality.

The last several decades have seen an increasing number of conflicting reports about the putative benefits of wedge resection versus anatomic lobectomy for the treatment of early-stage NSCLC (Table 73-1), with recent recommendations concerning the benefits of low-dose CT for lung cancer screening, more smaller-sized tumors are being visualized. This has led many surgeons to question the appropriateness of lobectomy for all cases of NSCLC.⁷

A randomized trial of lobectomy versus limited resection by Ginsberg and Rubinstein,⁹ on behalf of the Lung Cancer Study Group (LCSG), in patients with a clinical T1N0M0 NSCLC randomized to a lobectomy (n = 125) or lesser resection (n = 122; wedge resection 67% or segmentectomy 33%) demonstrated a three-fold decrease in locoregional recurrence (6.4% vs. 17.2%) and a reported statistically significant (p = 0.088) 5-year survival benefit (65% vs. 44%) in favor of lobectomy, based upon a one-sided p-value. This study published in 1995 established lobectomy with systematic nodal dissection as the “procedure of choice” relative to lesser resections in patients with early-stage NSCLC. Nevertheless, the LCSG study was criticized because wedge resection and segmentectomy were not separated and a high percentage (33%) of patients in the sublobar group underwent wedge resection.

A number of single-institutional retrospective reports, primarily from Japan,^10–12 have demonstrated that segmentectomy might be equivalent to lobectomy in terms of survival for carefully selected patients. Koike et al.¹⁰ compared survival and local recurrences in a retrospective study in patients with tumors smaller than 2 cm. Seventy-four patients underwent limited resection (14 wedge resections, remaining segmentectomies) and 159 patients had lobectomy. There was no difference in local recurrence or survival between these groups. Importantly, postoperative spirometry was significantly better in patients who underwent sublobar resections.

In 2005, Okada et al.¹² reported the results of a large retrospective study in 1272 patients. There was no significant difference in 5-year cancer-specific survival between patients undergoing segmentectomy versus lobectomy for tumors ≤3 cm. However, patients undergoing wedge resection for such tumors had significantly lower survival at 5 years (39.4% vs. 85.7% for lobectomy). On the other hand, for tumors greater than 30 mm in diameter, survivals were 81.3% after lobectomy, 62.9% after segmentectomy, and 0% after wedge resection.

In 2006, Okada et al.¹¹ reported on a multi-institutional, nonrandomized study over a 9-year period comparing survival and local recurrence for peripheral NSCLC ≤2 cm (90% adenocarcinoma) after sublobar resection in 305 patients (30 wedge resections, remaining segmentectomies) versus 262 patients after lobectomy. Overall 5-year survival, local recurrence, and distant recurrence were 89.6%, 4.9%, and 9.2% in the sublobar group versus 89.1%, 6.9%, and 10.3% in the lobectomy group, respectively.

In a study from the University of Pittsburgh (UPMC) authored by Schuchert et al.,¹³ 428 patients underwent resection for Stage IA and IB NSCLC. No significant difference in 5-year survival (80% vs. 83%) or local recurrence (4.9% vs. 7.7%) was found between lobectomy and segmentectomy, respectively. However, when the tumor margin to size ratio was less than 1, the local recurrence rate was higher. When the ratio was greater than 1, the local recurrence rate was 6.2% and not statistically different from lobectomy. However, when the ratio was under 1, the local recurrence rate after segmentectomy was 25%. Although this ratio was easier to achieve with smaller tumors, larger tumors in lower lobe basal segments enjoyed the same benefits.

The same group recently reported¹⁴ similar outcomes in 429 patients with pathologically proven stage 1A tumors only (5-year cancer specific survival 91% vs. 90% for T1a and 78% vs. 82% for T1b and recurrence rates 14.7% vs. 14.0%, respectively). This demonstrated that anatomic segmentectomy achieves equivalent recurrence and survival compared with lobectomy.

Despite favorable results in many studies, a report from Brigham and Women's hospital in Boston¹⁵ of 238 patients who underwent resection for a small (2 cm or less) NSCLC between 2000 and 2005, lobectomy (n = 84) was associated with longer overall (p = 0.0027) and recurrence-free (p = 0.0496) survival, but no statistically significant improvement in rate of local recurrence after sublobar resection (16% vs. 8%, p = 0.1117).

From all these nonrandomized studies mentioned earlier, it appears that for tumors less than 2 cm or when a margin to tumor ratio is greater than 1, anatomic segmental resection is justified with excellent results comparable to lobectomy. However, a recent SEER database analysis of 14,473 patients meeting the selection criteria, lobectomy was found to confer a significant survival advantage compared with segmentectomy (excluding wedge resections) for stage I NSCLC.¹⁶ This study has some inherent flaws and caution should be exercised with the interpretation of this study. First, it is a retrospective national database based upon population-based cancer registries where the type of surgery was self-reported. Therefore, it is not clear if all of the segmental resections were actually anatomic or if they represented just very large wedge resections.

Finally, a meta-analysis of published studies with 11,360 patients in total concluded that for stage I patients, sublobar resection causes lower survival than lobectomy, whereas the outcome after segmentectomy is comparable to that after lobectomy for stage IA patients with tumors ≤2 cm.¹⁷

In an effort to increase the level of evidence to support limited resection for these small (≤2 cm) tumors, two prospective randomized studies are currently ongoing, one in the United States and Canada with an expected enrollment of 1258 patients in 5 years (CALGB-140503) and one in Japan with an accrual of 1100 patients in 3 years (JCOG0802/WJOG4607L). Inclusion criteria in both trials are peripheral small (≤2 cm) NSCLC excluding noninvasive lung cancer on CT (so-called pure ground-glass opacity). In the North American trial, the investigational arm is segmentectomy or wedge resection, whereas in the Japanese trial only segmentectomy is allowed. The results of these trials are hoped to be available within 5 to 10 years from now.

Segmentectomy can also be performed via minimally invasive techniques. Reported series have shown that the outcome after VATS segmentectomy for small lung cancers is quite favorable in terms of perioperative course and intermediate survival.^18–20 VATS segmentectomy was compared to open segmentectomy in two retrospective studies, one from Duke University including 77 patients (48 VATS vs. 29 open)²⁰ and one from the University of Pittsburgh including 225 patients (104 VATS vs. 121 open).¹⁹ Both series reported a significantly reduced length of stay (4–5 days vs. 7 days), whereas a lower 30-day mortality (0% vs. 6.9%) was observed in the Duke series and a reduced rate of postoperative pulmonary complications (15.4% vs. 29.8%) was found in the Pittsburgh series in favor of the VATS approach. There were no differences in the operative time, estimated blood loss, recurrence, or survival between both groups in the latter series, whereas the survival was significantly (p = 0.0007) better in favor of VATS patients in the first series. One study reported by Swanson and colleagues²¹ compared segmentectomy (n = 31) versus lobectomy (n = 113) in patients operated for stage I NSCLC by VATS. Aside from a slightly increased number of patients with prolonged air leaks in the segmentectomy group, no major differences were seen in the length of stay (4 days), local recurrence (3.5%), and 3-year survival (80%). However, VATS segmentectomy can be quite challenging and may not be widely applicable at the current time.

Segmentectomy has been part of the thoracic surgeon's armamentarium for more than 70 years. In the last four decades, the indication for segmentectomy has shifted from a procedure for infectious etiologies (bronchiectasis, empyema, tuberculosis, and other infectious and nonmalignant lesions) to a surgical option for patients with peripheral stage IA NSCLC with a favorable size (≤2 cm) or with limited cardiopulmonary reserve. The only randomized study published demonstrates lower recurrence rates for lobectomy compared to a sublobar resection. These early published survival results after sublobar resection for lung cancer was tainted by the inclusion of nonanatomic wedge resections. Wedge resection resulted in a higher chance of leaving involved sump nodes behind and exposing the patient to a higher risk of local and regional recurrence. More recent retrospective and single-institutional reports suggest that the outcome in carefully selected patients with small (≤3 cm, and more likely ≤2 cm) peripheral tumors (mainly adenocarcinoma) is excellent after anatomic segmentectomy and comparable to lobectomy. Nevertheless, we must await results from ongoing randomized trials before segmentectomy can be embraced as the standard oncologic procedure for early-stage NSCLC. A VATS approach for sublobar resection has proved feasible, achieving excellent survival results comparable to open segmentectomy, but with shorter hospital stay and lower pulmonary morbidity. Proponents of this approach believe it may well become the future of surgery for lung cancer.²²

A recent article from Japan demonstrated the use of indocyanine green and infrared thoracoscope²³ to identify the intersegmental plane between segments after the segmental pulmonary artery has been ligated. In patients with severe COPD, collateral ventilation between adjacent segments may hinder identification of the intersegmental plane.

Dirk Van Raemdonck is a senior clinical investigator supported by the Fund for Research-Flanders (G.3C04.99).

The first “non lingular” segmentectomy was actually performed by my mentor, Dr. Richard Overholt, at the New England Deaconess hospital in Boston. His “posterior approach,” described in 1947 (Overholt RH, Langer L. A new technique for pulmonary resection; its application in the treatment of bronchiectasis. Surg Gynecol Obstet. 1947;84:257–268.) used a transcostal approach with dissection in the hilum. This “bronchus” first approach is still useful in patients after prior surgery or pulmonary/pleural infection or inflammation. Once the data from current randomized become available, a rational approach to subcentimeter lesions, found during lung cancer screening can be undertaken. The concept of lung-sparing surgery is paramount to treating patients with a long anticipated survival to minimize pulmonary morbidity. The margin-to-size ratio popularized by the Pittsburgh group promises to be an important determinant of adequacy of resection.