Abstract
The objective of the present study was to retrospectively analyze a cohort of patients who underwent surgery for colorectal cancer pulmonary metastases during a 12-year period. Patients and Methods: The sample included 75 patients who were monitored in terms of overall survival (OS) and disease-free interval (DFI) in relation to patient's age, preoperative values of biomarkers, type of surgery, number and size of metastases, occurrence of complications and length of hospitalisation. Results: A total of 95 surgical interventions were performed and 133 metastases were removed. Out of these, 28% of patients were free of any signs of relapse or disease progression for 5 years after metastasectomy. Those with two or more metastases are 2.3-times more at risk of disease progression. Tissue polypeptide specific antigen (TPS) values above the 140 IU/l cut-off point increase the risk of progression 3.9-times. The five-year survival rate among the group was 45%. Patients with 2 or more metastases are 2.7-times more at risk of death. TPS values above the 140 IU/l cut-off increase the risk of death 5.5 times, and carbohydrate antigen CA19-9 values above the 28 IU/ml cut-off point increase the risk of death by 3.2 times. Conclusion: The number of metastases and the preoperative TPS values are decisive prognostic factors influencing both OS and DFI.
The lungs are the second most frequent location of tumour dissemination and 25-30% of all solid tumours, including colorectal cancer (CRC), disseminate to the lungs. Nevertheless, in a considerable number of patients the dissemination of metastases is stopped or slowed due to a pulmonary involvement and these patients may benefit from radical metastasectomy. Surgical treatment of CRC pulmonary metastases is viewed as potentially curative, unlike other treatment modalities and is nowadays a generally accepted procedure (1-3). The objective of the present study was to retrospectively analyze a group of patients who underwent surgery for CRC pulmonary metastases during a 12-year period.
Patients and Methods
Study population. Between 2001 and 2012 we surgically treated 75 patients with CRC pulmonary metastases. The study population consisted of 42 men and 33 women with the median age of 66 years (range 29-80 years). In the male group, the median age was 67 years (range 29-80 years), while in the female group, the median age was 65 years (range 32-77 years).
The median duration of the follow-up period for the study population was 27 months. The study population was monitored for overall survival (OS) and disease-free interval (DFI) after the primary tumour surgery and metastasectomy and their relation to the patient's age, preoperative values of biomarkers, type of surgery, number and size of metastases, occurrence of complications and length of hospitalisation.
Metachronous metastases were diagnosed in 68 patients (90.7%), while synchronous metastases were diagnosed in 7 (9.3%) patients. Solitary lesions were found in 47 patients (62.7%), while the remaining patients had multiple metastases and 15 patients (20.0%) had bilateral lesions. Five patients (6.7%) underwent surgery due to a relapse of metastatic lung involvement. The median time period from primary tumour surgery to metastasectomy was 25 months. Fifty-three (70.7%) of the 75 patients had an adjuvant anticancer intervention to radical surgical treatment of their primary tumour prior to the metastasectomy and 14 (18.7%) had a surgical intervention of secondary liver metastases. Seven patients in the group (9.3%) were simultaneously diagnosed with lung and liver secondary lesions and both types were surgically removed in two periods.
Diagnostic methods. Standard examination methods for surgical pulmonary diseases were employed in order to diagnose lung metastases, i.e., computed tomography (CT) alone or in combination with positron emission tomography (PET/CT), fibrobronchoscopy, etc. To assist further diagnosis, serum levels of the following biomarkers were established: carcinoembryonic antigen (CEA) and carbohydrate antigen 19-9 (CA19-9) by a chemiluminescence method using kits from Beckman (Beckman Coulter, Inc., Brea, CA, USA) on a DxI 800 device (Beckman Coulter, Inc., Brea, CA, USA), thymidine kinase proliferation biomarker (TK) by a radioenzyme method (REA; Beckman Coulter, Inc., Brea, CA, USA) and two soluble cytokeratin fragments of the tissue polypeptide specific antigen (TPS; IDL Biotech AB, Bromma, Sweden) and tissue polypeptide antigen (TPA; DiaSorin, Inc., Stillwater, MN, USA) by immunoradiometric assay (IRMA). The cut-off values for the respective biomarkers in the laboratory are as follows: CEA: 3 μg/l; CA19-9: 28 IU/l; TK 8: IU/l; TPS: 90 IU/l and TPA: 100 IU/l. The OS and DFI calculations involved optimal calculated cut-off values related to prognosis.
Indication criteria. The radical removal of the primary malignant tumour and absence of other extrapulmonary metastases were the indication criteria, except for liver metastases suitable for resection. In multiple lesions we did not a priori determine a maximum possible number of removable metastases, but made our decision on a case-to-case basis. Exclusion criteria included the presence of extrapulmonary metastases (except for liver lesions suitable for resection) and preoperative evidence of tumour lesions on intrathoracic lymphatic nodes (LN).
Surgical methods. The surgical approach was either a posterolateral thoracotomy or video-assisted thoracoscopic surgery (VATS). The extent of surgery was based primarily on the location and on the number of lesions to be resected. With regard to anti-cancer radicality, we considered, as sufficient, removal of a metastasis with a safety resection edge of 10 mm of healthy tissue around the focus. In case of open surgeries, the particular procedures included anatomical pulmonary resections that involved segmentectomy, lobectomy, bilobectomy and pneumectomy and also non-anatomical pulmonary resections, i.e., wedge or high-precision laser excisions using the Nd:YAG laser MY 40 1.3 (Gebrüder Martin GmbH & Co. KG, KLS Martin Group, Tuttlingen, Germany) with a laser beam wave length of 1318 nm. During the minimally invasive surgeries, we removed the metastatic lesions using the wedge resection by means of endoscopic staplers. The lesions that could not be removed using the radical surgery were treated with radio-frequency ablation (RFA) transcutaneously under CT guidance.
Statistical analysis. The statistical analysis was conducted using Statistica (StatSoft, Inc., Tulsa, OK, USA) and SAS (SAS Institute, Inc., Cary, NC, USA) software. Basic statistical data was calculated for the measured parameters in the entire population and in the respective groups. OS and DFI were calculated using the Kaplan-Meier survival curve. Influence of individual factors on OS and DFI was tested using the Log-rank test and the Cox regression model. Statistical significance was determined to the level of alpha (α)=5%.
Types and numbers of surgeries.
Results
Surgical treatment. All patients in the study population underwent surgical therapy for pulmonary metastases, which included 60 unilateral (80.0%) and 15 bilateral surgeries (20.0%). These surgeries were performed as one-stage surgery in 1 patient (1.3%) and as two-stage surgery in 14 patients (18.7%). Subsequently, 5 patients (6.7%) underwent another radical surgery due to a relapse of the ipsilateral metastatic pulmonary lesion. In total, 95 surgeries were performed on 75 patients. The surgical approach included three VATS (3.2%), 85 (89.5%) posterolateral thoracotomies and seven (7.4%) transparietal RFAs. Particular types and numbers of surgical procedures are shown in Table I. Anatomical lung resections included 4 segmentectomies, 22 lobectomies, 1 bilobectomy and 1 pneumonectomy. In total, 133 metastases were radically resected and an additional 7 metastases were treated with RFA. The highest number of lesions removed in one patient was 11. The size of metastases ranged between 2 and 75 mm with the median of 21 mm. The final postoperative histology confirmed the preoperative assumption of CRC metastasis in all removed deposits.
Sixty-two surgeries (65.3%) involved removal of the hilar and mediastinal lymph nodes, by means of sampling in 36 cases and in 26 patients by means of ipsilateral systematic mediastinal lymphadenectomy (SMLA). In six cases (9.7% of biopsies performed), metastases of the same histological type as in the pulmonary lesions were found in the removed lymph nodes. Three of these were in the hilar and three in the mediastinal lymph nodes.
No perioperative complications were reported and the postoperative morbidity reached 8.0% (6 patients). On two occasions (2.7%) we were forced to apply drainage to the pleural space due to complicating pneumothorax or pleural empyema. None of the patients died in relation to the surgery within thirty days of the procedure. The mean duration of the postoperative hospitalisation was 6.9 days. Fifty-one patients (68.0%) underwent additional cancer treatment after the pulmonary metastasectomy (PM).
Disease-free interval after metastasectomy.
Factors employed in prognosis. The preoperative CEA value increased in 25 patients (33.3%), CA19-9 in 9 patients (12.0%), TK in 16 patients (21.3%), TPS in 22 patients (29.3%) and TPA values in 3 patients (4.0%). Changes to biomarker levels correlate to the cut-off values used in the laboratory (see Patients and Methods).
Thirty-one patients (41.3%) are alive after the PM without a disease progression; the median DFI is 24 months. In total, 42.5% of the patients were free of any signs of relapse or disease progression for 3 years and 28.1% for 5 years (Figure 1). Those with two or more CRC metastases are 2.3 times more at risk of disease progression (statistically significant; Log-Rank test: p-value=0.0061; Cox Hazard model: p-value=0.0080; 95% Hazard Ratio CI: 1.2-4.2) (Figure 2), with 3 and more metastases the risk of disease progression is 2.7-times higher (statistically significant; Log-Rank test: p-value=0.0215; Cox Hazard model: p-value=0.0280; 95% Hazard Ratio CI: 1.1-6.4). The TPS values above the 140 IU/l cut-off increase the risk of disease progression 3.9 times (statistically significant; Log-Rank test: p-value=0.0015; Cox Hazard model: p-value=0.0034; 95% Hazard Ratio CI: 1.6-9.9) (Figure 3).
A total of 45 patients (60.0%) are surviving in the study population with a median survival period of 46 months. All the patients who passed away suffered disease progression. Three-year and five-year survival rates in the study population were 58% and 45%, respectively (Figure 4). Those with two or more CRC metastases are 2.7-times more at risk of death (statistically significant; Log-Rank test: p-value=0.0055; Cox Hazard model: p-value=0.0077; 95% Hazard Ratio CI: 1.3-5.8) (Figure 5). The TPS values above the 140 IU/l cut-off increase the risk of death 5.5-times (statistically significant; Log-Rank test: p-value=0.0013; Cox Hazard model: p-value=0.0040; 95% Hazard Ratio CI: 1.7-17.8) (Figure 6), CA19-9 values above the 28 IU/ml cut-off increase the risk of death 3.2 times (statistically significant; Log-Rank test: p-value=0.0129; Cox Hazard model: p-value=0.0188; 95% Hazard Ratio CI: 1.2-8.4) (Figure 7). We were unable to demonstrate a correlation between the biomarkers and number of metastases using the Spearman correlation coefficient.
Dependence of the disease free interval after metastasectomy on the number of metastases.
Discussion
Generally accepted requirements for the indication of PM are the radical removal of the primary tumour, absence of extrapulmonary metastases, except for liver metastases suitable for resection and the acceptable risk of surgery (4, 5). Another important requirement is the ability of the patient to undergo resection in terms of their pulmonary function and tolerance to general anaesthesia. Synchronous or metachronous occurrence of CRC liver metastases is no longer considered a contraindication to PM because the sequential surgery of the liver and pulmonary lesions provides relatively successful results (6-8). Those are more apparent in metachronous occurrence rather than synchronous, in other words, if the liver resection precedes lung surgery. These patients reached the median survival of 83 months in our clinic (5). Conversely, PM should not be indicated in patients with disease progression despite their preceding cancer therapy or with a preoperatively found lesion of the hilar and mediastinal lymph nodes, as their prognosis is poor (9, 10). In relation thereto, provision of neoadjuvant therapy is subject to discussion. In the event of positive response, it should be followed by metastasectomy in the second period of surgery. Current outcomes are however not clear-cut, which is why some authors recommend SMLA during PM for management of the LN lesions (11, 12). Metastasectomy of synchronous, bulky (>3 cm), multiple or bilateral metastases with short DFI from surgery of the primary tumour and preoperative elevation of biomarkers are regarded as problematic due to a poorer prognosis (10, 13, 14). For our decision making, the key criteria were: absence of tumour lesions in the intrathoracic lymph nodes, positive response to previous cancer therapy, possibility of attaining R0 resection and functional tolerance of a patient to surgery.
Dependence of the disease free interval after metastasectomy on the preoperative serum level of the polypeptide specific antigen (TPS).
According to numerous studies, the OS and DFI results in PM performed using a traditional thoracotomy approach, compared to minimally invasive methods, are similar. However, minimally-invasive procedures have a higher risk of omission and may leave the preoperative unidentified metastatic lesions in situ due to the absence of pulmonary palpation. This may happen in every fifth patient according to some studies (4, 15-17). In this respect, our approach remains conservative. We prefer PM to thoracotomy, which is also reflected by a very low number of minimally-invasive procedures in the study population (3%). In our opinion, and as agreed with Yano and Pfannschmidt, the minimally-invasive approach may be accepted only in some individuals presenting solitary metachronous peripheral metastases with DFI of years after the primary surgery, which had been diagnosed using modern high-sensitivity CT scanners so that the probability of additional hidden lesions is minimal (10, 18).
Overall survival after metastasectomy.
Dependence of the overall survival after metastasectomy on the number of metastases.
The principal requirement of PM is always to achieve R0 resection. Incomplete resections have a significantly poorer prognosis and should be avoided (19). The standard procedures in pulmonary metastasis surgery are limited non-anatomical resections (wedge or stapler) or precise laser excisions with a safety resection edge of at least 3 mm in small lesions and 8-10 mm in larger lesions (20). The procedures are sparing the healthy lung parenchyma and therefore pose no limitation during disease recurrence or repeated surgeries. They are also recommended as they significantly improve the survival rates (2, 21, 22). Anatomical pulmonary resections (segmentectomy, lobectomy and pneumectomy) do not provide better long-term survival. In terms of morbidity, the outcomes are even worse. In addition, they cause limitations to potential repeated surgeries for recurrent metastases due to insufficient residual functional lung capacity. Despite this, they have their place in the surgical therapy of pulmonary metastases, but are only applied for the central or multiple lesions when a minor surgery is not technically feasible or does not guarantee the necessary anti-cancer radicality (23). It was the central location of metastases and their higher number that was the cause of the relatively high number of anatomical pulmonary resection in the study population (30%). If a radical resection of the metastatic lesion cannot be performed for various reasons, it can be treated with RFA during which we prefer a transparietal access under CT guidance (4).
Dependence of the overall survival after metastasectomy on the preoperative serum level of the polypeptide specific antigen (TPS).
The frequency of lesions to the hilar and mediastinal LN due to dissemination of cancer is described within the range of 8 to 47% in CRC pulmonary metastases with a demonstrable worse survival, which does not reach five years (11, 19, 24). In such a high percentage, the SMLA should be performed as a standard procedure in order to ensure precise staging or choice of adjuvant cancer therapy, especially if the sensitivity of preoperative diagnostic CT or PET/CT examination is only 35% (1, 11, 18, 19, 24, 25). However, the impact of SMLA on long-term survival remains unclear, even in case of LN involvement due to cancer dissemination (11, 19). In our study population, this impact was on the lower end of the specified range. Two out of 3 patients with disseminated metastatic lesions to the LN in the pulmonary hilum (LN sampling always performed) died and out of 3 operated patients with the involvement of ipsilateral mediastinal LNs (2 SMLA, 1 sampling) 1 patient who underwent SMLA, died. Due to the low number of patients and ambiguous results in our study population, we cannot draw any general conclusions.
Dependence of the overall survival after metastasectomy on the preoperative serum level of the carbohydrate antigen 19-9 (CA 19-9).
According to the various sources in literature, the five-year survival rate after radical removal of CRC pulmonary metastases ranges between 40-68% (2, 3, 6, 7, 16, 21, 22, 25, 26). On the other hand, the five-year survival rate of untreated metastatic lesions does not reach 5% (1). Significant prognostic factors, which influence the long-term survival rate, primarily include the following: histology and location of the primary tumour, length of occurrence of the metastatic process in relation to the primary tumour (synchronous vs. metachronous), duration of DFI after surgery of the primary tumour, CEA levels before and 1 month after the metastasectomy, history of the liver metastases, R0 resection of metastasis, number and size of metastases and lastly, condition of intra-thoracic LNs (2-4, 21, 26, 27). The best results of long-term survival may be expected in solitary metastases, metachronous and small size metastases, with DFI >1 year since the primary tumour resection, in R0 metastases resection, in the absence of lesions into the intrathoracic LN, in well differentiated tumours or in colon cancer, when the biomarkers level is normal prior to metastasectomy, and in the absence of liver metastases. Adjuvant chemotherapy after PM prolongs survival (28, 29). The risk of relapse after radical metastasectomy is negatively influenced by DFI within 1 year after the primary tumour intervention, by CEA ≥5 μg/l prior to metastasectomy, size of the largest metastasis >3 cm, bilateral metastases and a history of extrapulmonary metastases (3, 14, 22). Our results are similar to the data in the literature in terms of the risk factor related to the number of metastases. Its significance was demonstrated both for the overall survival rate and for the DFI values. Unlike the literature data, however, we did not verify the correlation between OS, DFI and elevated CEA levels, but we demonstrated a correlation between the levels of the cytokeratin biomarker TPS and the carbohydrate antigen CA 19-9. The findings correlate with our previous studies on colorectal and lung cancer (30). It becomes apparent that TPS is primarily a marker of tumour's proliferation activity and hence a significant marker of disease progression (31, 32).
Conclusion
Radical PM, as a potentially curative method, clearly prolongs the survival of patients with CRC pulmonary metastases. Optimal candidates are those who have a small number of metachronous metastases for a long-term period without any symptoms after primary tumour treatment and normal preoperative levels of biomarkers. Patients, with tumour dissemination to intrathoracic LN and disease progression despite the provided cancer therapy, should not be indicated for surgery. A complete R0 resection of the pathological lesion with the maximum effort in order to save healthy pulmonary tissue is essential. It enables re-operation in case of a relapse and is highly recommended. SMLA should be integrated into every pulmonary metastasectomy. The number of metastases and the preoperative TPS values are decisive prognostic factors influencing OS and DFI.
Acknowledgements
Supported by Institutional Grant No. 96-44 from the University Hospital Pilsen.
- Received April 15, 2014.
- Revision received June 3, 2014.
- Accepted June 5, 2014.
- Copyright© 2014 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved
References
In this issue
Jump to section
Related Articles
Cited By...
- The Role of TPS and TPA in the Diagnostics of Distant Metastases
- Factors Affecting Survival in Patients with Lung Metastases from Colorectal Cancer. A Short Meta-analysis
- Short-term Survival of Patients with Lung Metastases from Colorectal and Non-colorectal Cancer Who Underwent Pulmonary Metastasectomy