Keywords
pulmonary metastasectomy - sarcoma - survival - prognostic factor
Introduction
Sarcomas comprise 1% of adult malignancies and the lungs are the most common site
of metastasis. The incidence of isolated pulmonary metastases is approximately 20%
in sarcoma patients.[1] Pulmonary metastases cannot be treated sufficiently by chemo- or radiotherapy due
to their drug-resistant characteristics. Therefore, pulmonary metastasectomy (PM)
is widely accepted as standard therapy in isolated pulmonary metastases and is associated
with long-term survival.[2]
[3] This survival rate is superior to that obtained with chemotherapy which is induced
when PM is not proper.[4]
Although the outcomes and benefits of PM have been evaluated by several retrospective
studies, the effectiveness of this treatment has not been disclosed because of the
heterogeneity of the sarcomas.[2] Prognostic factors associated with improved survival after PM are suggested as the
histological type of primary tumor, the number and size of pulmonary nodules, disease-free
interval (DFI), tumor doubling time, and age of patient.[1]
[2]
[3] The 5-year survival rate varying between 15 and 52% and recurrence rate of approximately
40% after PM were reported in the literature.[5]
[6]
[7]
In this study, we aimed to analyze clinical properties of patients who underwent PM
for osteogenic and soft tissue sarcomas (STSs) and to investigate the prognostic factors
associated with overall survival (OS) and disease-free survival (DFS) after PM.
Materials and Methods
Data of 69 patients who underwent PM for osteogenic and STS metastases between January
1999 and December 2017 were evaluated retrospectively, according to patients characteristics
as age and gender, primary tumor characteristics as histological type, presentation
age, localization and local recurrence, and pulmonary metastases characteristics as
presentation time after sarcoma diagnosis, number, size, localization, laterality,
and recurrence.
The primary tumor was histopathologically diagnosed first and curative surgical resection
was performed after neoadjuvant chemotherapy in all patients by the orthopedic team.
Complete local control of primary tumor and isolated pulmonary metastases were validated
by radiological findings on computed tomography (CT) and/or positron emission tomography/CT
(PET/CT) scan before PM. PET/CT scan was started to be used after the year 2004 in
our center, so all 62 patients who were operated after 2004 had a PET/CT scan before
the operation. There were no lymph node metastases detected in these patients by PET/CT
scan before the operation. PM was performed with the aim of complete radical resection
of metastases. Mediastinal or hilar lymph node dissection or sampling was not performed
routinely for this study population with sarcomas.
Video-assisted thoracic surgery (VATS) was used in patients having less than three
ipsilateral metastasis detected by a highly sensitive multislice CT scan if available
for VATS preoperatively or perioperatively. Other patients were operated by thoracotomy.
A anthracycline-based neoadjuvant chemotherapy was applied to adult patients with
osteosarcoma for 3 or 4 cycles and T-10 protocol (Methotrexate, Adriamycin, Cyclophosphamide,
Cis-Platin, Actinomycin-D, Bleomycin) was used as adjuvant chemotherapy for 8 cycles
to 1 year depending to the response of the neoadjuvant chemotherapy. For Ewing's sarcoma
and other STS, VAC/IE protocol (Vincristine, Adriamycin, Cyclophosphamide/Ifosfamide,
Etoposide) and/or radiotherapy was used as neoadjuvant and adjuvant chemotherapy for
4 cycles preoperatively and 8 cycles to 1 year depending on the response.
In early years, Mayo Pilot II protocol (Methotrexate, Cis-Platin, Doxorubicin, Ifosfamide)
and in recent years EURAMOS protocol (Methotrexate, Cis-Platin, Doxorubicin, Ifosfamide,
Etoposide) were applied as neoadjuvant chemotherapy for 3 or 4 cycles to pediatric
patients with osteosarcoma and high-dose methotrexate was used as adjuvant chemotherapy
for 6 to 12 months depending to the response. For Ewing's sarcoma, EICESS-92 protocol
(Vincristine, Doxorubicin, Ifosfamide, Dactinomycin) was used as neoadjuvant chemotherapy
for 4 cycles and no adjuvant chemotherapy was applied. The administration of chemotherapy
and long-term follow-up of the patients was done by medical and pediatric oncology
departments.
Patients were divided into two groups according to the diagnosis as “osteosarcoma
group” and “nonosteosarcoma group.” These two groups were compared in terms of age,
primary sarcoma localization, DFI, diameter, volume, number, localization and laterality
of nodules, ratio of surgical margin to nodule size, and operation type for the first
PM.
The largest nodule was taken as reference for the determination of size and ratio
of surgical margin to nodule size. Surgical margin was calculated as the distance
from the staple line to the edge of the nodule. Histopathological findings were used
for the determination of these parameters.
The OS was calculated from the time of first PM to the time of death or last follow-up,
and DFS was measured from the time of first PM to the recurrence of pulmonary metastases
or death. DFI was defined as the time between the first diagnosis of the primary tumor
and the first PM. The relation between the prognostic risk factors and OS and DFS
were evaluated.
Analyses were conducted via IBM SPSS Statistics 23. Kolmogorov–Smirnov test was used
for assessment of data distribution. Continuous variables were summarized as mean ± standard
deviation. Categorical data were shown as percentages and compared using the Fisher's
exact test for 2 groups and the Pearson's chi-squared test for > 2 groups. Student's
t-test and Mann–Whitney U-test were performed for data with normal and abnormal distribution, respectively.
OS and DFS, the effects of the factors on mortality, were examined using the Kaplan–Meier
and Mantel–Cox (log rank) analysis. Cox regression analysis was performed to determine
the effects of factors on DFS and OS. The cut-off values for the ratio of surgical
margin to nodule size were calculated using receiver operating characteristic curve
analysis and logistic regression models were performed for significance. Statistical
level of significance was taken as p-value of < 0.05 for all tests.
The study was approved by the institutional ethics review panel. Patient consent for
data collection was not required.
Results
Mean age of patients (43 males, 26 females) was 25.32 ± 14.30 years (range, 2–70 years)
at the time of primary tumor diagnosis and 30.13 ± 14.00 years at the time of first
PM (range, 11–75 years). Osteosarcoma was the most common histologic type (36.2%)
followed by Ewing's sarcoma (24.6%) and synovial sarcoma (18.8%). Primary site of
the sarcoma was lower extremity in 47 (68.1%) followed by upper extremity in 15 (21.7%)
patients. A total of 102 operations were performed. Pulmonary wedge resection via
thoracotomy (44.12%) was the most common type of operation followed by video thoracoscopy
(41.18%).
In the analysis of first PM, number of nodules resected in a session ranged between
1 (48 patients, 69.60%) and 9 (1 patient, 1.4%). Mean size of the largest nodules
detected was 23.61 ± 22.61 mm (range, 2–140 mm). Nodules were bilateral in 12 (17.6%)
patients. Segmentectomy or lobectomy was the largest anatomic resection type for the
first PM with a rate of 7.24% (5 patients). The surgical approach was thoracotomy
in 38 (55.1%) and VATS in 31 (44.9%) patients. Complete resection could not be achieved
in 4 (5.8%) patients due to the enlargement of tumor. The mean of the DFI was 28.32 ± 29.78
months (range, 5–134 months) ([Table 1]). All patients received chemotherapy after the PM.
Table 1
Patient characteristics
|
n (%)
|
Gender (male/female)
|
43 (62.3)/26 (37.7)
|
Age at first PM (mean ± SD, range) (y)
|
30.13 ± 14.00 (11–75)
|
Histology
|
Osteosarcoma
|
25 (36.2)
|
Ewing's sarcoma
|
17 (24.6)
|
Synovial sarcoma
|
13 (18.8)
|
Others
|
14 (20.4)
|
Primary site
|
Lower extremity
|
47 (68.1)
|
Upper extremity
|
15 (21.7)
|
Vertebrae
|
5 (7.2)
|
Chest wall
|
2 (2.9)
|
No of nodules (1/2/3/4/5/6/8/9)
|
48 (69.6)/8 (11.6)/6 (8.7)/2 (2.9)/
2 (2.9)/1 (1.4)/1 (1.4)/1 (1.4)
|
Size of nodule (mean ± SD, range) (mm)
|
23.61 ± 22.61 (2–140)
|
Surgical margin/nodule size (mean ± SD)
|
0.75 ± 0.69
|
Laterality (unilateral/bilateral)
|
57 (82.6%)/12 (17.4%)
|
Surgical procedure
|
Pulmonary wedge resection
|
62 (89.8)
|
Segment/lobe/pneumonectomy
|
7 (10.2)
|
Complete resection
|
65 (94.2)
|
DFI (mean ± SD, range) (mo)
|
39.13 ± 36.32 (4–221)
|
Recurrence
|
38 (55.1)
|
Abbreviations: DFI, disease-free interval; PM, pulmonary metastasectomy; SD, standard
deviation.
Under the age of 20 years, osteosarcoma was the most common histologic type and 21
of 25 cases were seen in this age group (p < 0.001). All of the osteosarcoma cases were younger than 40 years of age (p < 0.001). In spite of these, nearly all of the others group (13 of 14 cases) including
malign fibrous histiocytoma, liposarcoma, chondrosarcoma, and pleomorphic cell sarcoma
was observed over the age of 20 years (p < 0.001). Also, osteosarcoma was the histologic type that had been resected with
the highest ratio of surgical margin to nodule size (p = 0.018). There was no significant relation between histologic type and DFI.
The comparison of groups showed significant difference according to age (p < 0.001 and p = 0.002), size of nodule (p = 0.033), ratio of surgical margin to nodule size (p = 0.007), and DFI (p = 0.039). There was no significant difference according to gender, site of primary
sarcoma, volume, number, localization and laterality of nodule, recurrence, and operation
type for the first PM ([Table 2]).
Table 2
Comparison of osteosarcoma and nonosteosarcoma patients according to prognostic factors
|
Osteosarcoma
|
Nonosteosarcoma
|
p
|
Age
|
< 20
|
21 (30.4%)
|
16 (23.2%)
|
< 0.001
|
≥ 20
|
4 (5.8%)
|
28 (40.6%)
|
< 40
|
25 (36.2%)
|
30 (43.5%)
|
0.002
|
≥ 40
|
0
|
14 (20.3%)
|
Gender
|
Male
|
15 (21.7%)
|
28 (40.6%)
|
0.764
|
Female
|
10 (14.5%)
|
16 (23.2%)
|
Site of primary sarcoma
|
Lower extremity
|
18 (26.1%)
|
29 (42.0%)
|
0.727
|
Upper extremity
|
5 (7.2%)
|
10 (14.5%)
|
Vertebrae
|
2 (2.9%)
|
3 (4.3%)
|
Chest wall
|
0
|
2 (2.9%)
|
Size of nodule (mm)
|
< 20
|
18 (26.1%)
|
20 (29.0%)
|
0.033
|
≥ 20
|
7 (10.1%)
|
24 (34.8%)
|
No. of nodule
|
≤ 2
|
20 (29.0%)
|
36 (52.2%)
|
0.853
|
> 2
|
5 (7.2%)
|
8 (11.6%)
|
Surgical margin/size of nodule
|
< 1
|
12 (%17.4)
|
35 (50.7%)
|
0.007
|
≥ 1
|
13 (18.8%)
|
9 (13%)
|
Laterality
|
Unilateral
|
20 (29.0%)
|
37 (53.6%)
|
0.667
|
Bilateral
|
5 (7.2%)
|
7 (10.1%)
|
Operation type
|
Thoracotomy
|
16 (23.2%)
|
22 (31.9%)
|
0.261
|
VATS
|
9 (13.0%)
|
22 (31.9%)
|
DFI
|
< 12 mo
|
4 (5.8%)
|
1 (1.4%)
|
0.039
|
≥ 12 to < 24 mo
|
8 (11.6%)
|
13 (18.8%)
|
≥ 24 to < 36 mo
|
9 (13.0%)
|
11 (15.9%)
|
≥ 36 mo
|
4 (5.8%)
|
19 (27.5%)
|
Recurrence
|
Yes
|
13 (18.8%)
|
25 (36.2%)
|
0.699
|
No
|
12 (17.4%)
|
19 (27.5%)
|
Abbreviations: DFI, disease-free interval; VATS, video-assisted thoracic surgery.
The median follow-up time after the first PM was 35 months (range, 2–194 months).
The estimated 5-year survival was 48% with a median of 43 months and 5-year DFS was
38% with a median of 22 months ([Fig. 1]).
Fig. 1 Disease-free (A) and overall (B) survival of patients after pulmonary metastasectomy.
The cut-off values of ratio of surgical margin to nodule size for DFS and OS were
calculated as 0.94 for both with odds ratios of 4.727 and 6.587, respectively. The
ratio was grouped as ≥ 1 and < 1 according to this cut-off value.
The univariate analysis showed that the number of nodules (p = 0.008), ratio of surgical margin to nodule size (p = 0.001), and localization of the nodule (p = 0.039) were the significant factors associated with DFS ([Fig. 2]). Also, nodule size (p = 0.042), number of nodules (p = 0.003), ratio of surgical margin to nodule size (p < 0.001), and laterality (p = 0.027) were significant prognostic factors associated with OS ([Fig. 3]) ([Table 3]). The multivariate analyses demonstrated that the ratio of surgical margin to nodule
size was an independent prognostic factor for DFS ([Table 4]), while the number of nodules and ratio of surgical margin to nodule size were both
independent prognostic factors for OS ([Table 5]). Also, the logistic regression analysis determined the ratio of surgical to nodule
size as the common significant risk factor for DFS and OS ([Table 6]).
Table 3
Univariate analyzes of factors associated with DFS and OS
|
Mean (median) DFS
|
p
|
Mean (median) OS
|
p
|
Gender
|
Male
|
71.9 (32)
|
0.189
|
77.1 (32)
|
0.101
|
Female
|
90.4 (77)
|
112.5 (103)
|
Age
|
≤ 20
|
89.1 (37)
|
0.388
|
105.2 (104)
|
0.457
|
> 20
|
64.7 (40)
|
71.9 (43)
|
Histology
|
Osteosarcoma
|
95.5 (51)
|
0.205
|
110.2 ()
|
0.262
|
Nonosteosarcoma
|
66.9 (40)
|
75.5 (40)
|
DFI
|
≤ 24 mo
|
82.1 (27)
|
0.539
|
86.8 (27)
|
0.470
|
> 24 mo
|
78.5 (43)
|
92.7 (51)
|
Nodule size
|
≤ 20 mm
|
87.7 (56)
|
0.081
|
114.3 (124)
|
0.022
|
> 20 mm
|
67.3 (31)
|
68 (31)
|
No of nodule
|
≤ 2
|
86.4 (56)
|
0.027
|
102.7 (103)
|
0.003
|
> 2
|
30.5 (18)
|
30.5 (18)
|
Surgical margin/nodule size
|
< 1
|
56.5 (27)
|
0.001
|
62.1 (31)
|
< 0.001
|
≥ 1
|
126.7 (156)
|
155.8 ()
|
Nodule localization
|
Limited to one lobe
|
87.5 (51)
|
0.041
|
99.8 (103)
|
0.088
|
> 1 lobe
|
51.9 (24)
|
60.0 (24)
|
Laterality
|
Unilateral
|
85.5 (56)
|
0.183
|
101.6 (103)
|
0.027
|
Bilateral
|
39.5 (23)
|
39.5 (23)
|
Operation type
|
Thoracotomy
|
75.4 (37)
|
0.561
|
87.4 (37)
|
0.423
|
VATS
|
75.3 (56)
|
84.0 (103)
|
Abbreviations: DFI, disease-free interval; DFS, disease-free survival; OS, overall
survival; VATS, video-assisted thoracic surgery.
Table 4
Multivariate analysis of the factors associated with DFS rates
Variables
|
Hazard ratio
|
95% CI
|
p
|
No of nodule
|
0.617
|
0.222–1.713
|
0.354
|
Ratio of surgical margin to nodule size
|
3.349
|
1.595–7.034
|
0.001
|
Nodule localization
|
0.659
|
0.216–2.013
|
0.464
|
Abbreviations: CI, confidence interval; DFS, disease-free survival.
Table 5
Multivariate analysis of the factors associated with overall survival rates
Variables
|
Hazard ratio
|
95% CI
|
p
|
Nodule size
|
0.939
|
0.431–2.046
|
0.874
|
No of nodule
|
0.395
|
0.165–0.945
|
0.037
|
Ratio of surgical margin to nodule size
|
4.531
|
1.534–13.384
|
0.006
|
Laterality
|
1.895
|
0.785–4.570
|
0.155
|
Abbreviation: CI, confidence interval.
Table 6
Classification tables for ratio of survival margin to nodule size on the logistic
regression model
|
Predicted
|
Recurrence
|
|
Observed
|
(–)
|
(+)
|
Percentage correct
|
Recurrence
|
(–)
|
13
|
11
|
54.2
|
(+)
|
9
|
36
|
80.0
|
Overall percentage
|
|
|
71.0
|
|
Mortality
|
|
Alive
|
Exitus
|
Mortality
|
Alive
|
17
|
16
|
51.5
|
Exitus
|
5
|
31
|
86.1
|
Overall percentage
|
|
|
69.6
|
Fig. 2 Significant risk factors for disease-free survival: (A) number of nodules, (B) ratio of surgical margin to nodule size, and (C) localization of nodule.
Fig. 3 Significant risk factors for overall survival: (A) nodule size, (B) number of nodules, (C) ratio of surgical margin to nodule size, and (D) laterality.
Discussion
Especially, pulmonary metastasis is a poor prognostic factor for STS, but the survival
can be extended with PM in selected patients. Therefore, PM remains a cornerstone
in the treatment of STS since neither chemotherapy nor radiotherapy has shown any
effective results.[8]
[9]
In this study, overall 5-year survival was 48% with a median of 43 months. This was
a similar and favorable survival outcome when compared with the recent studies.[2]
[3]
[6] Beside, the 5-year DFS (43% with a median of 40 months) was significantly higher
in our highly selected patient group when compared with the same studies. This can
be a result of patient selection which affects the risk factors of survival like histologic
type, tumor grade and aggressiveness, number of metastasis, localization, and laterality.
Also, all of the patients in this study received chemotherapy before and after the
resection of the primary tumor and after PM. The approach type with a VATS rate of
45% was another mark of patient selection.
Histologic type was considered as a prognostic factor in some recent studies,[1]
[10]
[11] but we did not find any difference in comparing the most frequent histologic types.
We grouped patients as osteosarcoma and nonosteosarcoma and compared these two groups
according to patient characteristics. Osteosarcoma was seen especially before the
age of 20 years and none of them was diagnosed after the age of 40 years. The nodule
size and ratio of the surgical margin to nodule size was significantly different in
the osteosarcoma group. These differences may be due to the nature of the sarcoma
and a closer follow-up of pediatric patients for metastatic diseases is mandated.
Although we expected a difference because of these factors, there was no significant
difference in terms of survival between the two groups. On the other hand, DFI was
significantly shorter in the osteosarcoma group and we thought this as a reason of
absence of difference.
Most of the recent studies in the literature mentioned DFI as a very important prognostic
factor for the survival of sarcoma patients with pulmonary metastasis. Especially
DFI longer than 12 months was indicated as a strong prognostic factor for DFS and
OS.[3]
[12]
[13] In our study, there was no correlation between DFI and survival, despite different
grouping variations according to interval. This can be due to the number of patient
included and distribution of histologic types in our series. Other studies similar
to that of ours also reported similar findings, that is, no difference.[14]
[15]
Number of nodules, nodule size, laterality, and complete resection were reported as
the main risk factors in previously published studies involving large number of patients.[3]
[7]
[8]
[12]
[13] Number of nodules, ratio of surgical margin to nodule size, and localization for
DFS, nodule size, nodule volume, number of nodules, ratio of surgical margin to nodule
size, and laterality for OS were the significant risk factors in univariate analysis.
The common risk factors for DFS and OS were number of nodules and ratio of surgical
margin to nodule size. After the multivariate analysis, the ratio was the common significant
risk factor with p-values of 0.001 and 0.006 for DFS and OS, respectively. The cut-off values were determined
as 0.94 for this factor and showed that the patients could be grouped as < 1 and ≥
1 according to the ratio. This ratio was also considered as a significant prognostic
factor by the results of logistic regression analysis. We did not evaluate the effect
of complete resection because of the results already known. Instead, we examined the
ratio of surgical margin to nodule size which was thought as more meaningful. This
analysis showed the importance of length of tumor-free surgical margin for recurrence
and survival.
Hilar and mediastinal lymph node involvement is very rare especially in sarcomas,
so the role of lymph node dissection in sarcoma patients remain contradictory.[16]
[17] Because of this contradiction, some centers perform no lymph node sampling or dissection
whereas others favor lymph node sampling or radical dissection.[18] We do not perform lymph node sampling or dissection routinely especially in patients
with negative CT and PET/CT scan, that is, if there is no significant finding of lymph
node involvement perioperatively.
Type of surgical approach is a common issue today. Number of nodules, localization,
and laterality are main determinants in selecting the surgical approach. Two studies,
one of them including only sarcoma patients[19] and the other including colorectal cancer patients,[20] applied VATS (with nodules less than 3 and peripheral localization, etc.) and yielded
similar survival benefits. We also preferred VATS for the nodules which were few and
peripheral and found no difference in DFS and OS. In addition, parenchymal protective
surgery is an important point of PM, and therefore the surgeon must consider this
when selecting the kind of approach.
Limitations
There are some limitations of this study that should be kept in view when commentating.
This is a retrospective and single center-based study and hence the methodology used
cannot be generalized to other centers. Also, because of the retrospective design
and the missing data the grade of the primary sarcoma which has an effect on survival
could not be analyzed.
Comment
PM seems to be the best choice in selected sarcoma patients with pulmonary metastasis.
The efficacy of chemo-/radiotherapy on survival is limited according to PM. Size,
number and localization of nodules, DFI, complete resection, and laterality are reported
as the main risk factors of survival. Although this prognostic factor should be supported
by randomized controlled studies with large patient numbers and meta-analysis, this
study shows that the ratio of surgical margin to nodule size ≥ 1 must be taken as
a common risk factor for DFS and OS. Therefore, the resection of the nodules with
the possible widest surgical margin is an important point of PM.