Roles and outcomes of thoracoscopic anatomic lung subsegmentectomy for lung cancer

Abstract

 

OBJECTIVES

We performed sublobar resections, including thoracoscopic segmentectomy and subsegmentectomy for small lung cancers, and analysed the results of indications and outcomes of thoracoscopic subsegmentectomy.

METHODS

Between March 2005 and May 2020, 357 consecutive patients underwent thoracoscopic anatomic sublobar resections for lung cancer, including 68 patients undergoing subsegmentectomy. These patients were compared with 289 patients who underwent segmentectomy during the same period.

RESULTS

Subsegmentectomies included mono-/bi-/tri-subsegmentectomies for 34/23/11 of 68 patients, respectively. The median tumour size was 13.5 mm, significantly smaller than tumours in patients who underwent a segmentectomy (P < 0.001). Tumours obtained by mono-subsegmentectomy (11.0 mm) were significantly smaller than bi-/tri-subsegmentectomy (P = 0.028). The proportion of ground-glass opacity-dominant tumours obtained by subsegmentectomy (85.3%) was higher than that obtained by segmentectomy. The proportion of intentional cases satisfying the criteria for sublobar resection was higher than that of segmentectomy cases. Although tumour locations in 40 patients were not identified during surgery, tumours were correctly resected in 39 patients without tumour markers. The median operative time and blood loss were 167 min and 13 ml, significantly shorter and less, respectively, in subsegmentectomy than in segmentectomy patients (P = 0.005, P = 0.006). Duration of drainage and hospitalization were 1 and 5 days, respectively, for subsegmentectomy patients; complications occurred in 6 (8.8%). Outcomes were similar to those of the segmentectomy patients. Although 4 subsegmentectomy patients died of other diseases, none showed cancer recurrence during a mean follow-up of 50 months.

CONCLUSIONS

Thoracoscopic subsegmentectomy can be used for patients with ground-glass opacity-dominant lung cancers <1.5 cm if adequate margins can be secured.

The number of sublobar resections performed for small lung cancers has increased.

INTRODUCTION

The number of sublobar resections performed for small lung cancers has increased. This trend might be attributed to the increasing detection rates for small lung nodules by the recent widespread use of computed tomography (CT) screening. Small nodules containing ground-glass opacities (GGOs) especially can only be detected by CT. Small GGO-dominant lung cancers have been shown to be associated with better prognosis than solid-dominant lung cancers [1, 2]. Sublobar resections have therefore become preferred for patients with small GGO-dominant lung cancers, although the standard procedure for lung cancers has traditionally been considered to be lobectomy plus lymph node dissection [3]. Wedge resection or segmentectomy is generally the procedure used for sublobar resections. Furthermore, the thoracoscopic approach has been preferred over open thoracotomy since thoracoscopic surgery is minimally invasive, with the advantages of reduced pain, rapid recovery and shortened period of hospitalization [4, 5].

Therefore, we have performed thoracoscopic segmentectomies for small GGO-dominant lung cancers since 2004 and also used thoracoscopic subsegmentectomies for patients with the following indications: adequate surgical margin possible and tumour located where a wedge resection would be inappropriate. However, the selection criteria between subsegmentectomy and segmentectomy were actually unclear because we had selected the procedure according to a team discussion in the preoperative conference based on preoperative CT findings. Although previously we mainly reported on thoracoscopic subsegmentectomy guided by 3-dimensional (3D) CT simulation, with short-term surgical outcomes [6], we could not elucidate the definitive selection criteria and report on the types of patients who could undergo thoracoscopic subsegmentectomy because of our limited number of patients.

The aim of this study was to perform a retrospective investigation comparing the tumour characteristics of patients who underwent thoracoscopic subsegmentectomy with those of the tumours of patients who underwent thoracoscopic segmentectomy during the same period of time. We also evaluated the results of indications for the procedure and outcomes of the patients who underwent thoracoscopic subsegmentectomy.

MATERIALS AND METHODS

This retrospective study was approved by the Institutional Ethics Committee (# 2020-220), and the requirement to obtain written informed consent from each patient was waived by the hospital’s Institutional Review Board.

Between March 2005 and May 2020, 461 consecutive patients underwent thoracoscopic anatomic sublobar resections at our institution. The indications for thoracoscopic anatomic sublobar resections were as follows: (i) patients with adequate pulmonary function able to tolerate lobectomy with a lung tumour <2 cm plus >80% GGO on high-resolution CT, (ii) patients with indeterminate lung nodule <1 cm and (iii) patients considered poor candidates for lobectomy because of limited cardiopulmonary function or other organ failure. Wedge resection was considered to be inappropriate in all patients because the targeted nodules were located in the deep parenchyma from the visceral pleura.

Of the 461 patients who underwent thoracoscopic sublobar resections, 90 underwent thoracoscopic subsegmentectomy. Among the 90 patients, 22 were excluded for the diagnosis of metastatic lung tumours, etc., and remaining 68 were diagnosed with lung cancer postoperatively and were enrolled in this study. During the same time, 311 and 60 patients underwent segmentectomy and segmentectomy combined with adjacent subsegmentectomy, respectively. Among the 371 patients, 82 were excluded for the diagnosis of metastatic lung tumours, etc., and remaining 289 were diagnosed with lung cancer. Finally, among the patients whose tumours were histopathologically diagnosed with lung cancer during the same period, the tumour characteristics and surgical outcomes of the 68 patients who underwent subsegmentectomy were compared with the 289 patients who underwent thoracoscopic segmentectomy or segmentectomy combined with adjacent subsegmentectomy to elucidate the indications for subsegmentectomy (Fig. 1). The characteristics of the tumour and the surgical indications (‘intentional’ or ‘compromised’ cases) were evaluated in these patients, and the surgical outcomes were analysed. ‘Intentional’ cases were defined as those patients who satisfied our indications for sublobar resection, and the ‘compromised’ cases were defined as those patients who were considered poor candidates for lobectomy because of limited cardiopulmonary function or other organ failure. Finally, the long-term survival of patients who underwent thoracoscopic subsegmentectomy were evaluated.

Figure 1:

Flowchart to determine the study eligibility of patients who underwent thoracoscopic anatomical lung resections in our institution.

The numbers and symbols used to denote pulmonary segments or subsegments were based on the methods described in Yamashita’s Roentgenologic Anatomy of the Lung [7]. Subsegmentectomy was defined as a pulmonary segmentectomy at the subsegmental (third order) level of the arterial and bronchial branches.

Three-dimensional computed tomography simulation

Multidetector CT scanning was preoperatively performed according to the Digital Imaging and Communications in Medicine Standard, and the data were saved to a computer. Workstations or a client viewer (Synapse Vincent system; Fujifilm Corporation, Tokyo, Japan, and AquariusNET; TeraRecon, Inc., San Mateo, CA, USA) were used for image analysis.

A 3D volume-rendering method was used for pulmonary arterio-venous and bronchial reconstructions and was performed by the operating surgeon instead of a technician as we previously reported [6, 8]. Simulation images that were based on the CT reconstruction were used to identify intrasubsegmental arteries (arterial branches entering the affected subsegment) and intersubsegmental veins (venous branches surrounding the affected subsegment) (Fig. 2A and B).

Figure 2:

(A and B) Thin section computed tomography (CT) and 3-dimensional CT image of a patient who underwent thoracoscopic right S2b + S3a subsegmentectomy. The intersubsegmental veins were identified as the veins running between the adjacent subsegments.

Surgical methods

Using a four-port approach, the trocar placements were performed as our previous reports [6, 8]. A single-port approach was introduced in 2019 that used an incision ∼3–4 cm long at the fourth or fifth intercostal space along the mid-axillary line that was protected by a wound retractor (Alexis; Applied Medical, Rancho Santa Margarita, CA, USA).

First, the hilar parenchyma was dissected along the intersubsegmental veins by electrocautery or an energy device. Second, the arterial branches were dissected, the hilar lymph nodes were subsequently resected and frozen sections were made from the resected nodes. Only hilar lymph nodes were dissected; mediastinal lymph node dissection was omitted when no hilar lymph node metastasis was observed in the frozen sections. Third, the bronchus was threaded and ligated by a monofilament suture in a slip knot, based on a previous report [9]. After the bilateral lungs were inflated, the bronchus was then divided by a stapler or ligated by a silk thread. The parenchyma was further dissected towards peripheral site along the intersubsegmental veins by either electrocautery or an energy device, and the venous branches running into the affected subsegment were divided. Staplers were used to divide the parenchyma in the peripheral lung. The subsegmental line was used as the resection line to obtain a sufficient surgical margin around the tumour. A sufficient surgical margin was defined as the distance from the tumour to the cut margin and was 2 cm or the equivalent to the tumour. If subsegmentectomy alone was thought insufficient for providing an adequate surgical margin, subsegmentectomy combined with adjacent subsegmentectomy (bi- or tri-subsegmentectomy) was performed. The intersubsegmental veins were divided or preserved on the intersegmental planes of remaining subsegments (Fig. 3). The surgical margin between the tumour and cut surface was measured at the surgeon’s discretion based on the macroscopic status of the resected specimens (Video 1).

Figure 3:

Surgical view of an S1 + 2c subsegmentectomy. Division along the intersubsegmental vein and the inflation–deflation line.

Statistical methods

Data were analysed by JMP statistical software, version 14.0 (SAS institute Inc., Cary, NC, USA). Continuous data are expressed as the median with interquartile ranges (IQRs) and compared between each group with the Mann–Whitney or Kruskal–Wallis test. Comparisons of categorical variables were performed by the χ² test or Fisher’s exact test. Overall and cancer-specific recurrence-free survival were estimated using the Kaplan–Meier method. The overall survival was assessed from the date of subsegmentectomy to the date of death from any cause or the last visit. And the cancer-specific recurrence-free survival was assessed from the date of subsegmentectomy to the first date of recurrence or the last visit, non-cancer-related deaths were censored.

RESULTS

Patient characteristics

Table 1 shows the characteristics of patients. Among the 68 patients who underwent subsegmentectomy, most tumours were diagnosed as adenocarcinomas; 1 patient with a small cell carcinoma underwent a completion lobectomy and lymph node dissection 7 days after the initial subsegmentectomy. The median tumour size in the subsegmentectomy group, as measured on CT, was 13.5 mm (IQR 9.3–17 mm) in diameter and was significantly smaller than the median size of the tumours measured on CT of the segmentectomy group (median 16 mm, IQR 12–21 mm; P < 0.001). The proportion of patients with GGO-dominant tumours (partially solid and pure GGO) obtained by subsegmentectomy [58/68 patients (85.3%)] was significantly higher than the proportion obtained by segmentectomy [196/289 patients (67.8%); P < 0.001]. Regarding the surgical indications, 76.5% of patients who underwent subsegmentectomy were ‘intentional’ compared to 57.4% of patients who underwent segmentectomy (P = 0.004). More patients were ‘intentional’ than ‘compromised’ within each group and the proportion of ‘intentional’ patients in the subsegmentectomy group was higher than in the segmentectomy group. The median tumour size measured on CT in the ‘intentional’ patients who underwent a subsegmentectomy was 13 mm (IQR 10–17.75 mm) in diameter and was significantly smaller than the median size of the tumours (15 mm) in the ‘intentional’ segmentectomy patients (IQR 12–18.25 mm; P = 0.024).

Table 1:

Patient characteristics

	Subsegmentectomy (n = 68)	Segmentectomya (n = 289)	P-value
Age (years)	68 (61–75)	72 (64–77)	0.02
Male/female	28/40	155/134	0.065
Preoperative diagnosis			0.063
Lung cancer	64 (94.1)	283 (97.9)
MLT	4 (5.9)	4 (1.4)
Inflammatory tumour	0 (0)	2 (0.7)
CT findings
Tumour size (mm)	13.5 (9.3–17.0)	16.0 (12.0–21.0)	<0.001
Tumour characteristics			<0.001
Solid	10 (14.7)	93 (32.2)
Partially solid	30 (44.1)	143 (49.5)
Pure GGO	28 (41.2)	53 (18.3)
c-Stage (TNM 8th)			<0.001
0	28 (41.2)	53 (18.3)
1A1	35 (51.4)	104 (36.0)
1A2	3 (4.4)	95 (33.0)
1A3	1 (1.5)	27 (9.3)
1B	1 (1.5)	5 (1.7)
2A or 2B	0 (0)	5 (1.7)
Pathological findings
Tumour size (mm)	11.0 (8.0–15.0)	15.0 (11.0–20.0)	<0.001
Postoperative diagnosis			0.149
Adenocarcinoma	63 (92.6)	236 (81.7)
Squamous cell carcinoma	4 (5.9)	44 (15.2)
Small cell carcinoma	1 (1.5)	5 (1.7)
Others	0 (0)	4 (1.4)
p-Stage (TNM 8th)			<0.001
0	35 (51.5)	57 (19.7)
1A1	20 (29.4)	88 (30.4)
1A2	11 (16.2)	89 (30.8)
1A3	2 (2.9)	35 (12.1)
1B	0 (0)	11 (3.8)
2Aor2B	0 (0)	7 (2.4)
3A	0 (0)	2 (0.7)
Surgical indications			0.004
Intentional caseb	52 (76.5)	166 (57.4)
Compromised casec	16 (23.5)	123 (42.6)
Tumour size on CT for each indication (mm)
Intentional caseb	13.0 (10.0–17.75)	15.0 (12.0–18.25)	0.024
Compromised casec	14.0 (8.25–16.75)	17.0 (13.0–23.0)	0.023

Data are expressed as n (%) or median (IQR).

^aIncluding segmentectomy combined with adjacent subsegmentectomy.

^bDefined as patients who met our indications for sublobar resection.

^cDefined as patients who were considered poor candidates for lobectomy due to limited cardiopulmonary function or other organ failure.

CT: computed tomography; GGO: ground-glass opacity; IQR: interquartile range; MLT: metastatic lung tumour; TNM: tumour, node and metastasis.

Types of subsegmentectomy

Table 2 summarizes the resected subsegments. Of the 68 patients, 34 underwent mono-subsegmentectomy, 23 underwent bi-subsegmentectomy and 11 underwent tri-subsegmentectomy. The types of subsegmentectomy performed varied, and accordingly, the approach needed to obtain adequate surgical margins also varied.

Table 2:

Procedure types and resected lung subsegments

Procedure type		n	Right	n	Left	n
Subsegmentectomy	Mono-	34	S1a	2	S1 + 2a	2
			S1b	4	S1 + 2c	5
			S2b	1	S3aii	1
			S3a	1	S4a	2
			S3b	3	S6c	1
			S4a	1	S8a	2
			S6b	1	S9a	1
			S6c	1
			S8a	4
			S8b	1
			S*b	1
	Bi-	23	S1a + S2a	1	S1 + 2ab	5
			S1b + S3a	2	S1 + 2a + S3c	1
			S2b + S3a	4	S1 + 2c + S3a	1
			S6b + S8a	1	S3ab	1
			S8b + S9b	1	S3a + S4a	1
					S3b + S4b	1
					S6b + S8a	2
					S8b + S9b	1
					S10ac	1
	Tri-	11	S1a + S2aii + S2bi	1	S1 + 2a + S3bc	1
			S2b + S3a + S3bi	1	S1 + 2ab + S3c	3
			S6b + S8a + S9a	1	S1 + 2bc + S4a	1
			S6bc + S10a	1	S1 + 2c + S3a + S4a	1
					S1 + 2ab + S3c	1
Total		68		33		35

(a) Apical or posterior subsegment; (b) anterior subsegment; and (c) posterior subsegment.

S1: apical; S2: posterior; S1 + 2: apical posterior; S3: anterior; S4: superior; S6: superior; S8: anterior basal; S9: lateral basal; S10: posterior basal; S*: subsuperior; S: segment.

Tumour characteristics and surgical outcomes of each type of subsegmentectomy

Table 3 presents the tumour characteristics and surgical outcomes of each type of subsegmentectomy. The mono-subsegmentectomy tumours ranged from 8 to 15 mm (median 11 mm) in diameter, as measured by CT, and the median size was significantly smaller than the median sizes of bi- and tri-subsegmentectomy tumours [16 mm (IQR 10–20 mm) and 16 mm (IQR 10–24 mm), respectively; P = 0.028].

Table 3:

Details of the subsegmentectomy types

	Subsegmentectomy type			P-value
	Mono-	Bi-	Tri-
	(n) 34	23	11
CT findings
Tumour size (mm)	11.0 (8.0–15.0)	16.0 (10.0–21.0)	16.0 (10.0–24.0)	0.028
Tumour characteristics				0.75
Solid	5 (14.7)	4 (17.4)	1 (9.1)
Partially solid	13 (38.2)	12 (52.2)	5 (45.45)
Pure GGO	16 (47.1)	7 (30.4)	5 (45.45)
Preoperative diagnosis				0.12
Lung cancer	30 (88.2)	23 (100)	11 (100)
MLT	4 (11.8)	0 (0)	0 (0)
c-Stage (TNM 8th)				0.72
0	16 (47.1)	7 (30.4)	5 (45.5)
1A1	15 (44.1)	14 (60.9)	6 (54.5)
1A2	2 (5.9)	1 (4.3)	0 (0)
1A3	1 (2.9)	0 (0)	0 (0)
1B	0 (0)	1 (4.3)	0 (0)
Pathological findings
Tumour size (mm)	10.0 (7.5–13.5)	12.0 (10.0–15.0)	15.0 (10.0–21.0)	0.032
Postoperative diagnosis				0.73
Adenocarcinoma	31 (91.2)	21 (91.3)	11 (100)
Squamous cell carcinoma	2 (5.9)	2 (8.7)	0 (0)
Small cell carcinoma	1 (2.9)	0 (0)	0 (0)
p-Stage (TNM 8th)				0.22
0	22 (64.7)	8 (34.8)	5 (45.4)
1A1	8 (23.5)	9 (39.1)	3 (27.3)
1A2	3 (8.8)	6 (26.1)	2 (18.2)
1A3	1 (2.9)	0 (0)	1 (9.1)
Operative findings
Undetectable tumour	22 (64.7)	13 (56.5)	5 (45.5)	0.51
Operative outcomes
Operative time (min)	161.5 (126.25–184.75)	169 (128–192)	170 (151–195)	0.58
Bleeding volume (ml)	12.5 (0–74.25)	10 (0–101)	50 (0–85)	0.86
Duration of drainage (days)	1 (1–1)	1 (1–2)	1 (1–2)	0.48
Postoperative hospital stay (days)	5 (4–7.25)	5 (4–6)	5 (4–6)	0.55
Conversion to thoracotomy	0 (0)	1 (4.3)	0 (0)	0.37
Complete resection	33 (97.1)	23 (100)	11 (100)	0.6
Conversion to segmentectomy	1 (2.9)	0 (0)	0 (0)	0.6
Complications	3 (8.8)	3 (13.0)	0 (0)	0.46

Data are expressed as n (%) or median (IQR).

CT: computed tomography; GGO: ground-glass opacity; MLT: metastatic lung tumour; IQR: interquartile range; TNM: tumour, node and metastasis.

Regarding the operative findings, although more than half the tumours were undetectable by visualization or palpation during surgery, most tumours could be completely resected with sufficient surgical margins, except in 1 patient. That patient was scheduled to undergo a left thoracoscopic S8a subsegmentectomy and was converted to a S8 segmentectomy because the initially resected specimens did not contain targeted tumour because the incorrect bronchus had been mistakenly targeted. In all other cases, the targeted tumours were palpable in the resected specimens and were diagnosed as lung cancer with frozen sections intraoperatively after removal of the specimen. The median operative time was ∼160–170 min. Although the volume of blood loss from each subsegmentectomy was small, 1 patient who underwent bi-subsegmentectomy required a blood transfusion. That patient required conversion to small thoracotomy (∼7 cm) because of bleeding from a torn pulmonary artery. Postoperative complications occurred in a total of 6 patients who underwent mono- and bi-subsegmentectomy.

Surgical parameters of subsegmentectomy versus those of segmentectomy

Surgical outcomes are shown in Table 4. The median operative time and bleeding volume of subsegmentectomies were 167 min (IQR 130.25–191.75 min) and 13 ml (IQR 0–79.5 ml), respectively, which were significantly less than those parameters of segmentectomies [median operative time, 178 min (IQR 151–212 min) and bleeding volume, 53 ml (IQR 3.25–134.25 ml), P = 0.005 and P = 0.006, respectively]. The differences of other parameters between subsegmentectomy and segmentectomy were not significant. Tumour recurrences after subsegmentectomy were not observed during the follow-up period, which ranged from 3 to 140 months (median 50 months).

Table 4:

Surgical outcome: subsegmentectomy versus segmentectomy

Variables	Subsegmentectomy (n = 68)	Segmentectomya (n = 289)	P-value
Operative time (min)	167 (130.25–191.75)	178 (151–212)	0.005
Bleeding volume (ml)	13 (0–79.5)	53 (3.25–134.25)	0.006
Duration of drainage (days)	1 (1–1)	1 (1–2)	0.063
Postoperative hospital stay (days)	5 (4–6.75)	6 (4–8)	0.090
Conversion to thoracotomy	1 (1.5)	10 (3.4)	0.70
Complications	6 (8.8)	47 (16.3)	0.12
Prolonged air leakage	4 (5.9)	30 (10.4)
Arrhythmia	1 (1.5)	1 (0.3)
Delirium	1 (1.5)	2 (0.7)
Pneumonia	0 (0)	5 (1.7)
Respiratory failure	0 (0)	2 (0.7)
Other	0 (0)	7 (2.4)
Recurrence	0 (0)	2 (0.7)	0.49

Data are expressed as n (%) or median (IQR).

^aIncluding segmentectomy combined with adjacent subsegmentectomy.

IQR: interquartile range.

Long-term outcomes of subsegmentectomy

Figure 4 shows the long-term survival of patients who underwent subsegmentectomy. The 5-year overall and cancer-specific recurrence-free survival rates were 92.9% [95% confidence interval (CI) 79–98%] and 100% (95% CI 100–100%), respectively. Although 4 patients died of other diseases during the 5-year follow-up period, there were no recurrences.

Figure 4:

Five-year overall survival and cancer-specific recurrence-free survival.

DISCUSSION

With the clinical reports showing that the outcomes of sublobar resections versus lobectomy for small lung cancers are equivalent [10, 11], randomized trials of patients with small peripheral lung cancers, including CALBG 140503 and JCOG0802/WJOG4607L, are in progress, and the validated non-inferiority of sublobar resections was recently reported [12, 13].

Most sublobar resections are generally performed by wedge resection or segmentectomy. Subsegmentectomy, which is another procedure used for sublobar resection, has also been performed at our institution. In 2013, we reported on the thoracoscopic subsegmentectomy techniques based on 3D CT; there have recently been other reports on subsegmentectomy [6, 14, 15]. Our previous report, however, did not discuss the appropriate indications for subsegmentectomy because the number of relevant cases was too small. We have since accumulated a number of subsegmentectomy cases; however, the selection criteria for subsegmentectomy were not established before this study.

In this study, it was elucidated that thoracoscopic subsegmentectomy was performed for patients with significantly smaller tumours and tumours with a higher proportion of GGO than the tumours of patients undergoing segmentectomy. The tumour size and tumour characteristics were 1.5 cm less than and GGO-dominant tumours. Specifically, the median tumour size on CT was 13.5 mm in the subsegmentectomy group. Moreover, among the 68 subsegmentectomies, the median tumour size of the patients who underwent a mono-subsegmentectomy was 11 mm, while the median tumour size of the patients who underwent bi- and tri-subsegmentectomy was both 16 mm. Although the median tumour size of 13.5 mm in the subsegmentectomy group may appear slightly large, the tumour size may have reflected the patients who underwent bi- and tri-subsegmentectomies, and the median size of ‘intentional’ cases was smaller. The tumour characteristics on CT might play a role in procedure selection. And the favourable long-term outcomes in this series might be accounted for by the following reasons: the ‘intentional’ cases underwent subsegmentectomy based on our strict criteria for sublobar resections and a higher proportion of subsegmentectomy than segmentectomy patients consisted of ‘intentional’ cases, although segmentectomy had also sometimes been considered applicable to ‘compromised’ cases with poor cardiopulmonary function and contraindications for a lobectomy for lung cancer.

The important issue for the surgical resection of small lung tumours is the localization of the tumour. Palpation or visualization of small GGO-dominant tumours might be difficult during surgery. Tumour markers are sometimes required for wedge resections. Various methods for detecting tumours intraoperatively have been reported [16, 17]. CT-guided hookwire tumour localization is widely performed. Improved methods such as the image-guided video-assisted thoracoscopic surgery method have been reported by Gill et al. as an alternative to the CT-guided hookwire method, which can occasionally lead to serious complications such as air embolism caused by the introduction of air into a pulmonary vein after accidental puncture [18, 19]. Our subsegmentectomy can avoid such a complication because, as we previously reported on the roles of segmentectomy for undetectable tumours, subsegmentectomy does not require any tumour markers [8]. In this series of 68 patients who underwent subsegmentectomy, the tumours of 40 (58.8%) patients were undetectable during surgery. Our excellent technical results for these undetectable tumours might be attributed to the use of 3D CT simulation, as we have previously reported on the usefulness of 3D CT simulation for thoracoscopic segmentectomy and subsegmentectomy [6, 10, 20]. In this study, various subsegments could be resected by referring to the 3D CT reconstructions of the pulmonary vasculature and bronchi, as shown in Table 2. The usefulness of 3D CT simulation was reaffirmed because 3D CT simulation provides a more detailed understanding of the subsegmental anatomy than that required for segmentectomy.

With the exception of 3D CT simulation, the following technical aspects might also contribute to precise resection. We divided the intersubsegmental plane along the intersubsegmental veins with reference to both the 3D CT images and the inflation–deflation lines because the intersubsegmental veins run along the intersubsegmental plane and the inflation to targeted subsegment becomes a landmark of the intersubsegmental plane. Although the injection of indocyanine green has been reported to be a useful method for identifying the intersegmental plane, indocyanine green injection requires effort and the use of expensive instruments, such as those used for infra-red thoracoscopy [21]. The inflation–deflation line can be easily used for segmentectomy and provides adequate time for identifying the intersegmental plane. The inflation–deflation line, however, may not be apparent to surgeons when attempting to detect the intersubsegmental plane in an emphysematous lung. In this series, most patients with GGO-dominant tumours had normal lung parenchyma. Therefore, the inflation–deflation line was sufficient to create a clear description of the intersubsegmental plane. Moreover, we first dissected the pulmonary artery or intersubsegmental veins of the hilar site, always confirmed the targeted artery, veins and bronchus when dissecting these structures from the hilar site, and rechecked whether the dissected structure was correct while confirming the surrounding structure and referring to the 3D CT image. Our intersubsegmental division was performed by electrocautery or an energy device in the hilar site and a stapler in the peripheral site. The combined methods might depend on the reduction of complications, such as pulmonary air leakage; there were no problems with using a stapler. As a result, the precise resection could be completed for targeted tumours, even if the tumour was non-palpable or not visualized. Regarding the technical skills, although the technique of thoracoscopic subsegmentectomy may be more complex than wedge resection and segmentectomy, the learning curve is thought to be similar to that of the thoracoscopic segmentectomy previously reported because the steps of thoracoscopic subsegmentectomy are nearly the same as segmentectomy [22].

Our operative outcomes were comparable to those from previous studies of subsegmentectomy [14, 15]. In our series, although the thoracoscopic approach was mainly performed via 4 ports, a single-port approach was also used for some patients beginning in 2019. Recently, Chang et al. [15] reported that single-port subsegmentectomy obtained a mean operative time of 160 min and blood loss of 13.2 ml. Our operative time was 167 min and the blood loss volume was 13 ml. The results were remarkably similar, although it was difficult to simply compare the findings to the findings in the previous paper because the number of single-port approaches was only 4. And compared to segmentectomy performed during the same period, the operative time for our subsegmentectomy was shorter with less blood loss. These results might be attributed to the smaller extent of surgery performed by subsegmentectomy than by segmentectomy.

Wedge resection is a relatively simple procedure and is generally performed for small peripheral tumours. However, tumours not located in the peripheral parenchyma are problematic for wedge resection in 2 ways: tumour markers are difficult to place deeply in the parenchyma and adequate surgical margins are difficult to obtain, especially at the bottom of the cut margin. Suzuki et al. [23] reported that the probability of failed nodule detection is high for tumours located >5 mm from the visceral pleura and for tumours <10 mm in diameter. In our series, most tumours were located >5 mm from the visceral pleura (data not shown) and the deeply situated tumours were resected with sufficient surgical margins by subsegmentectomy. Thus, subsegmentectomy is assumed to play a useful role for securing a sufficient surgical margin.

Anatomic subsegmentectomy might provide other oncological benefits. The procedure allows for the evaluation of lymph nodes and can be easily converted to curative lobectomy if a metastatic lymph node is found during the procedure. Our subsegmentectomy omitted the mediastinal lymph node dissection if metastasis to the hilar lymph node was negative on intraoperative frozen section because GGO-dominant small-sized lung cancers are less likely to have lymph node metastases. Indeed, a recent paper also reported that there were no lymph node metastases in GGO-dominant lung cancers [24].

In this study, 68 patients underwent a thoracoscopic subsegmentectomy and were enrolled among a total of 357 patients who underwent thoracoscopic anatomic sublobar resections for lung cancer. The number of subsegmentectomies was not smaller compared to previous reports [6, 14, 15]. The indications for subsegmentectomy might be based on tumour characteristics. If the tumour size is <1.5 cm and the tumour characteristics exhibit GGO dominance, thoracoscopic subsegmentectomy can be an alternative curative procedure among sublobar resections for lung cancer.

Limitations

This study has limitations; it was a single-centre study with retrospective design and lack of analysis of pulmonary function. No adjustments were made for confounders. It might be difficult to demonstrate the superiority of subsegmentectomy under these circumstances. Future studies that enrol more patients are needed to clarify the outcomes after anatomic subsegmentectomy.

CONCLUSION

Thoracoscopic anatomic subsegmentectomy based on 3D CT simulation can be safely performed and can be curative for a GGO-dominant lung tumour <1.5 cm.

Conflict of interest: none declared.

 

Author contributions

Hirohisa Kato: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Project administration; Resources; Software; Supervision; Validation; Visualization; Writing—original draft; Writing—review & editing. Hiroyuki Oizumi: Validation. Jun Suzuki: Surgeon. Katsuyuki Suzuki: Surgeon. Satoshi Takamori: Surgeon.

Reviewer information

Interactive CardioVascular and Thoracic Surgery thanks Dominique Gossot, Madhuri Rao and the other, anonymous reviewer(s) for their contribution to the peer review process of this article.

ABBREVIATIONS

3D

3-Dimensional

CI

Confidence interval

CT

Computed tomography

GGO

Ground-glass opacity

IQR

Interquartile range

 

Presented at the 34th Annual Meeting of the European Association for Cardio-Thoracic Surgery, Barcelona, Spain, 8–10 October 2020.