Introduction

Colorectal carcinoma (CRC) is the third most common cancer as well as the third leading cause of cancer-related death in Europe and the USA [1]. Due to continuous treatment improvements including multimodal treatment options and improved surgery concepts, CRC is nowadays associated with a good prognosis.

However, about 50% of the patients with CRC had at diagnosis or develop during follow-up a metastatic disease, which significantly impairs the prognosis [2]. Next to the most common location in the liver, colorectal organ metastases also often affect the lung. Colorectal lung metastases (CRLM) occur in approximately 10% of patients with CRC [3, 4]. The 5-year survival of patients with CRLM is currently estimated about 40–60% [5, 6].

While standardized multimodal therapy concepts have been established for liver metastases, the best management for lung metastases remains controversial [7]. This is also reflected in existing guidelines, which usually contain hardly any recommendations for the therapy of lung metastases [8]. There is especially an ongoing discussion, if and when pulmonary metastasectomy is indicated. In an era of increasingly improved chemotherapy and radiotherapy as alternative treatment options, the usefulness of lung metastasis resection is being questioned. This is fueled by the recently published randomized controlled PulMiCC trial, which failed to show a benefit from pulmonary metastasectomy compared to an observational approach [9]. Likewise, a current review and meta-analysis including ten studies could reveal no significant survival difference between the surgical and non-surgical management of CRLM for recently published data and recommended further studies to improve evidence [7].

The aim of this study was to evaluate whether pulmonary metastasectomy of CRLM is associated with a survival benefit for affected patients and to identify patient subgroups that might benefit most from surgical therapy.

Material and methods

The prospectively maintained Cancer Registry of the Department of Surgery was used to identify patients for this retrospective analysis. All adult patients with colorectal lung metastasis, who were treated for any step of therapy at the department of general and visceral surgery of University Hospital between 01.01.2000 and 31.12.2018, were included to this study. Patients with missing data regarding the therapy of lung metastasis were excluded (n = 2). In summary, data from 418 patients with CRLM were analyzed.

Clinical and epidemiological data, treatment, histopathological findings and follow-up data were retrieved partially from the Registry for Colorectal Carcinomas (ERCRC) and partially from the clinical information system. The detailed documentation of the histopathological examinations allowed the classification of all carcinomas in accordance with the 8th edition of the UICC TNM classification [10]. The median follow-up time for the patients was 19 months [range 0–183 months].

This retrospective study was approved by the local ethics committee (23–219-Br). The work has been registrated at the German Clinical Trial Registry (DRKS00032938) and reported in line with the STROCSS criteria [11].

Treatment of CRLM and follow-up

Management of CRLM was decided for each patient individually by a multidisciplinary tumor board according to the current evidence at the time of treatment.

The surgical approach was chosen according to the location and number of pulmonary nodules. In most patients, pulmonary metastasectomy was performed during one surgery (63 patients, 55%); 33 (29%) and 19 patients (17%) needed two or more than two surgeries. Most patients received open surgery (75%) mostly as segmentectomy (56%) with thoracic lymphadenectomy (65%) (Supp tab. S1). Additional perioperative chemotherapy, radiotherapy or chemoradiotherapy was performed in 40% (46 patients), 7% (eight patients) and 10% (11 patients).

Non-surgical management included the performance of a chemotherapy in 209 patients (69%), of a chemoradiotherapy in 16 (5%), of a radiotherapy in six patients (2%) and an observational non-surgical management in 71 patients (24%). Chemotherapy regimens were selected according to the evidence-based German guideline for colorectal cancer that was valid at the time of treatment and taking into account previous chemotherapies.

All patients were offered a close follow up, beginning quarterly, then semi-annually and finally at least annually. Follow-ups contained physical examination, determination of carcinoembryonic antigen (CEA) levels and preferable a chest and abdomen CT scan or an abdominal ultrasonography and chest X-ray.

Statistical analysis

SPSS® Version 28 (IBM, Armonk, NY) was used to analyze the data. Comparisons of metric and ordinal data were calculated with the Student t test or Mann–Whitney U test. The Chi-square test was used for categorical data. Overall survival (OS) was calculated for the period between the date of diagnosis of CRLM and the date of death or last follow-up. Possible factors related to the overall (OS) of patients were tested using univariate and multivariate analysis. Variables with a p ≤ 0.05 in univariate analysis were used for multivariate analysis by Cox regression model. Survival curves were plotted using the Kaplan–Meier method and compared with the log-rank test. A p value ≤ 0.05 was considered statistically significant.

Results

Dataset of patients

A total of 418 patients met the inclusion criteria and were analyzed. One hundred fifteen patients underwent a surgical therapy of their CRLM, and 303 received a non-surgical management of their CRLM.

All patients with surgical therapy received R0 resection. Of the patients with additional thoracic lymphadenectomy during pulmonary metastectomy (75 patients), 21% had histologically proven thoracic lymph node metastasis (Supp. tab. S1).

Patient characteristics

Baseline characteristics of the included patients at the time of first diagnosis and about the primary tumor are presented in Table 1. Significant differences in baseline characteristics contained a lower age (61 vs. 62 years, p = 0.036), a higher proportion of rectal carcinoma (71 vs. 54%, p = 0.003) and less synchronous metastasis (40 vs 54%, p = 0.012) in the surgical group (Table 1).

Table 1 Baseline characteristics of patients with colorectal lung metastases
Full size table

Lung metastases specific characteristics

Lung metastases occurred in median 16 months after first colorectal carcinoma diagnosis and were mostly metachronous (71%), bilateral (65%) and multiple (70%). Patients with a surgical therapy of CRLM were significantly more likely to have a unilateral (65 vs. 23%, p < 0.001) and singular (44 vs. 16%, p < 0.001) pattern of CRLM, had a significantly lower CEA values (2.7 vs. 10.2 ng/ml, p < 0.001) and showed significantly fewer other existing metastases (27 vs. 73%, p < 0.001). Chemotherapy was performed more often in the non-surgical patient group (72 vs 50%, p < 0.001) (Table 2).

Table 2 Lung metastases specific characteristics of patients with colorectal lung metastases
Full size table

Prognostic factors for overall survival

Two- and 5-year overall survival (OS) from the time of diagnosis of lung metastasis was 78.2% and 54.6%, respectively, in our cohort. Univariate analysis indicated that the location of CRLM (uni- vs. bilateral, p = 0.005), the number of CRLM (p < 0.001), the tumor situation at diagnosis of CRLM (no previous and existing metastasis vs. previous metastasis, now R0 vs. existing other metastasis, p < 0.001; Fig. 1) and the therapy approach (surgical vs. non-surgical, p < 0.001; Fig. 2) significantly influence the overall survival after diagnosis of CRLM (Table 3).

Fig. 1
figure 1

Overall survival of patients with colorectal lung metastasis according to tumor situation at diagnosis of lung metastasis (no previous metastasis vs. previous metastasis, now R0 vs. existing other metastasis)

Full size image
Fig. 2
figure 2

Overall survival of patients with colorectal lung metastasis according to therapy performed (surgical vs. non-surgical therapy)

Full size image
Table 3 Prognostic factors of patients with colorectal lung metastases for overall survival (OS)
Full size table

Multivariate analysis revealed that a surgical resection of CRLM (HR 4.51 (95% CI = 2.33–8.75, p < 0.001) and the absence of other metastases (HR 1.79 (95% CI = 1.05–3.04), p = 0.032) were significant independent prognostic factors regarding the OS (Table 3).

Subgroup analysis

Subgroup analysis of all subgroups identified as prognostic factors in univariate analysis (location of CRLM, number of CRLM, tumor situation at diagnosis of CRLM) showed that patients with unilateral (p < 0.001) as well as bilateral CRLM (p < 0.001), with one (p = 0.003) as well as two to five CRLM (p < 0.001) and with no previous and existing other metastasis (p < 0.001) as well as with previous, but R0-resected metastasis (p = 0.027) as well as with existing other metastasis (p < 0.001) benefit from surgical therapy of CRLM. Only in the group with more than six CRLM surgical resection of CRLM reached no significance (p = 0.144). Again, surgical therapy of CRLM was associated with a significant better overall survival in still more selected subgroups (one CRLM and no other metastasis; ≥ 2 CRLM and existing other metastasis; only patients with combined liver and lung metastasis) (Supp. tab. S2). Surgical details as well as preoperative suspicion and histologically proven existence of thoracic lymph node metastasis did not affect the overall survival in the surgical subgroup (Supp. tab. 1).

Discussion

The optimal therapy of CRLM remains an ongoing controversial topic, especially against the background of increasingly improved chemotherapeutic and radiotherapeutic therapy options. The current data situation on this topic is divergent, which justifies the need for further evidence [7].

The 2- and 5-year overall survival (OS) from the time of diagnosis of lung metastasis was 78.2% and 54.6%, respectively, in our cohort, which is similar to reported survival rates in the literature [5, 6, 12].

Our data suggest that the surgical therapy of CRLM is associated with a significant survival advantage for patients with CRLM compared to a non-surgical treatment in a selected patient cohort (3). These results are in line with several other study results, which show an approximately 40–50% better 5-year survival following pulmonary metastectomy for CRLM [5, 13,14,15,16,17,18]. In contrast, there are some studies, which revealed comparable survival outcomes in surgical and non-surgical treated patients [9, 19, 20]. These results mainly concern especially the more recent ones and are explained by the advent of novel chemotherapeutic regimes and interventional therapy options. However, if we only include the more recent data (from 2010) in our collective, the significant survival benefit of the surgical therapy of CRLM remains.

An important aspect regarding these data, which allows the survival advantage in the surgical group to be interpreted with caution, may be a significant selection bias favoring the surgical group. Comparing the baseline and lung metastasis specific characteristics between the surgical and the non-surgical group, those selected for pulmonary metastasectomy more frequently had prognostically favorable characteristics (e.g. lower metastatic disease rate at first diagnosis, less bilateral pattern and lower number of CRLM, lower rate of existing other metastasis next to CRLM). However, in multivariate analysis including these mentioned divergent parameters between the groups pulmonary metastasectomy remain an independent highly significant prognostic parameter. Moreover, we performed a subgroup analysis of all identified significant prognostic factors (location of CRLM, number of CRLM, tumor situation at diagnosis of CRLM) to reduce the influence of potential selection bias and to identify subgroups that might benefit most from surgical therapy. With the exception of patients with more than six CRLM, surgical therapy was associated with a significant survival advantage in all subgroups formed. Combinations of two prognostic factors reflecting low and high tumor burden (one CRLM and no other metastasis; ≥ 2 CRLM and existing other metastasis) also led to a significant prognostic benefit of surgical therapy of CRLM. In all subgroup analysis the percentage of surgically treated patients was higher in subgroups with less tumor burden reflecting again a selection bias, which limits the validity of the data. However, our data suggest that with a rational selection like in our collective also patients with high tumor burden can benefit from pulmonary metastasectomy. One factor that may be underrepresented in our analysis, because it cannot be determined with sufficient validity from the available data, is the general condition of the patient when deciding whether or not to treat the lung metastases surgically.

Not investigated in our study, but another important aspect to consider in the decision between surgical and non-surgical therapy for colorectal lung metastases is the impact on quality of life. Patients who undergo surgery may experience periods of chemotherapy-free time, which may improve their quality of life. However, it is important to acknowledge that surgery carries the risk of postoperative morbidity, which can negatively affect both quality of life and even survival [21].

Next to surgical resection of CRLM we identified the presence of existing other metastasis at diagnosis of CRLM as independent prognostic parameter in multivariate analysis. Patients with previous, but now not anymore existing metastases, had a similar survival as patients without any previous or existing metastases underlining the importance and usefulness of effective metastasis treatment in colorectal carcinoma.

Additionally, in the univariate analysis, the location of colorectal lung metastases (unilateral vs. bilateral) as well as the number of metastases had an impact on overall survival. These findings are consistent with previous studies that have described the predictive influence of these factors on patient outcomes [12].

The present study has several limitations. First, it is a single-center study which has the advantage of a homogeneous therapy concept, but makes it difficult to generalize the results. Second, the retrospective analysis of prospective recorded data may have incurred some bias. Third, the number of patients is collected over a long period of 18 years. Therapy options improved over the years especially concerning adjuvant chemotherapy. However, there was no significant difference of the results between earlier und the later years of data period. Fourth, retrospective data about this topic are always influenced by a selection bias, which has already been discussed in detail above. Randomized controlled trials are needed to overcome this limitation.

Conclusion

The present study confirms the importance of surgical therapy of CRLM, as it is associated with a significant better overall survival in a selected patient cohort—even in patients with high tumor burden. Randomized controlled trials are needed to clarify this important question in a patient collective without selection bias.