Palliative resection of the primary tumour improves survival in incurable metastatic colorectal cancer

Studies show conflicting results on whether primary tumour resection (PTR) in metastatic colorectal cancer (mCRC) prolongs survival. The aim of this study was to analyse prognostic factors and the effects of PTR on overall survival (OS) in mCRC patients.


Introduction
3][4][5] The location of the primary tumour site is also of importance.Patients with a right-sided colorectal tumour have a shorter OS compared with patients with a left-sided colorectal cancer. 6,7][10][11] Whether palliative surgery of the primary tumour resection (PTR) in mCRC prolongs survival is a topic of debate, and studies show conflicting results.][14][15] A few randomized controlled trials (RCT) have terminated due to difficulties to recruit patients, 16,17 but one recently published RCT showed no effects of PTR on OS in patients with mCRC. 18However, several RCTs are ongoing and awaiting final results. 19,20opulation-based cohort studies provide the opportunity to assess the real-world applicability of current clinical practices.The Region Halland in western Sweden has a stable population of 350 000 people and has three hospitals sharing the same digital medical records (Table 1).
This study aims to show prognostic factors and the effect of PTR on OS in mCRC patients.

Cohort and classifications
In this population-based cohort study, we identified patients treated with standard chemotherapy and targeted therapy against mCRC between 2008 and 2019.Patients offered the best supportive care were not included.Patients were retrieved from the digital medical records in Region Halland, western Sweden, and were followed from detection of metastatic disease until death or 31 December 2021.Follow-up was complete due to low migration rate from the region and the use of national personal identity numbers.All patients were discussed at a multidisciplinary conference.The therapeutic approach was either palliative or neoadjuvant.Date of mCRC diagnosis was set when biopsy was taken for the colorectal tumour.When biopsy was not taken, date of mCRC diagnosis was instead set when CT scans detected metastases.
Clinical characteristics of the patients and pathological data of the tumours were retrieved from the medical records and registered in a database.Tumour location was registered as right-sided (C18.0 and C18.2-C18.4),left-sided (C18.5-C18.7 and C19.9), and rectal tumour (C20.9)according to WHO ICD10-codes.The database registered the site of distant metastases, comorbidities and performance status according to the Eastern Cooperative Oncology Group (ECOG) at the time of diagnosis.

Statistical analysis
The chi-square test to compare categorical variables and the Mann-Whitney U-test were used for continuous variables.Kaplan-Meier OS data estimates were performed and compared with a two-sided log-rank test (Mantel-Cox).We used Cox regression univariate and multivariate analyses to identify factors predicting OS and crude hazard ratios (CHR) and adjusted hazard ratios (AHR) were calculated with 95% confidence intervals (CI).Statistical significance was set at P-values less than 0.05.Values are presented as mean AE standard deviation (SD) if nothing else is indicated.We used the IBM SPSS Statistics version 26 (IBM Corp., Armonk, NY, USA) and GraphPad Prism program 9.1.0(GraphPad Software, Inc., San Diego, USA) to analyse data.

Ethics
This study was performed in accordance with ethics standards of the 1964 Helsinki Declaration and approved by the Regional Ethics Review Board in Gothenburg (reference number 2019-01961).All Data are given as mean AE SD for continuous variables and proportions (%) for categorical variables.The chi-squared test was used to compare categorical variables and the Mann-Whitney U-test was used to compare continues variables between the two treatment arms.PTR, primary tumour resection.
participants were still alive at the time of the study signed an informed written consent form.

Results
In this population-based cohort study, 347 mCRC patients treated with chemotherapy and/or targeted therapy were initially identified.Neoadjuvant (n = 108) and recurrent (n = 41) patients were excluded (Fig. 1).The study population comprised 188 palliative mCRC patients of whom 119 patients (63%) had palliative chemotherapy and/or targeted therapy.The remaining 69 patients (37%) also underwent PTR.The median time from diagnosis of mCRC to resection of the primary tumour in the PTR group was 9 days.Acute PTR was performed in 25 patients (36%) due to complications of the primary tumour.Indications for PTR in the acute setting were obstruction, perforation and acute bleeding from the tumour.PTR was performed electively in 44 patients (64%) due to anaemic bleeding, obstructive symptoms, or as a prophylactic palliative resection.OS was longer in patients undergoing PTR compared with patients not undergoing surgery.Median OS was 22.9 AE 1.9 months for the PTR group and 14.5 AE 1.5 months for the non-operated group (P < 0.01; Fig. 2).PTR was also an independent variable correlating with better (P < 0.05; Table 2).Patients undergoing PTR were exposed to significantly higher cumulative doses of fluoropyrimidines compared with patients not undergoing surgery (P < 0.001), but not to other chemotherapeutic drugs or targeted therapies (Table 3).
Patients with PTR had better performance status (P < 0.001), fewer metastatic sites (P < 0.05) and lower CEA levels (P < 0.001) at baseline compared with patients not undergoing palliative resection (Table 1).Furthermore, there were more left-sided colon tumours and fewer rectal tumours in the resection group (P < 0.001; Table 1).In the study population, 55 patients had rectal cancer of whom 33 patients (60%) were treated with radiotherapy as local tumour control.We found neither differences in gender, age, comorbidities, nor differences in proportion of patients with lung and liver metastases and proportion of patients with RAS mutations in the palliative primary resection group compared with the non-surgery group.
Left-sided colorectal cancer (P < 0.001) and rectal cancer (P < 0.001) were associated with better OS than right-sided colorectal cancer.Poor performance status at baseline and liver metastases (P < 0.01) were associated with worse OS.Metastasectomy had no effects on OS (P = 0.13).Due to missing values, univariate and multivariate analyses for carcinoembryonic antigen (CEA) were calculated separately based on N = 148.Pathological CEA (CEA≥5) was significantly associated with worse OS in the univariate analysis, that is, HR = 1.85 (1.20-2.87;P < 0.01) but not in the multivariate analysis, that is, HR = 1.08 (0.66-1.78;P = 0.77).

Discussion
This population-based cohort study showed prolonged OS for palliative mCRC patients undergoing PTR.Poor performance status, right-sided CRC and presence of liver metastases were factors associated with shorter OS.
The strength of this study is that we were able to include and follow up with almost all mCRC patients treated with palliative chemotherapy for more than 10 years thanks to low migration rate and the Swedish personal identity number system.The two assessed colorectal clinics in Halland treat all patients living in Halland but is not a referral centre and therefore our study population is representative of the general population.Since PTR was performed only within 9 days following diagnosis the clinical outcome of palliative chemotherapy did not seem to contribute to the decision to recommend PTR.Generally, $45% of rectal tumours respond to chemoradiotherapy by downstaging, followed by lesser obstruction symptoms. 21Although one-third of the patients in the non-surgical group had palliative radiotherapy because of rectal cancer, we still observed a prolonged OS in the operated group after PTR, which further strengthens our results.Multiple studies have shown that Crude hazard ratio (CHR) and adjusted hazard ratio (AHR) were estimated with univariate and multivariate Cox regression analyses, respectively, and the 95% confidence intervals (CI) are indicated.Synchronous tumours = two or more tumours in two or more locations in the colon.3][24][25] Radiotherapy for rectal cancer may have positive effects on OS and therefore could constitute an option for surgery, but long-term effects on survival are at present unknown. 21,23,26s we anticipated, the PTR patients had better performance status and fewer metastatic sites at baseline and the risk for selection bias can still be a weakness and source for misconstruction.Even after correction for covariates, patients undergoing PTR may be healthier than patients not undergoing PTR.In contrast to PTR, no effects of metastasectomy on OS could be shown in the multivariate analysis.To note, our cohort of 69 PTR patients was considered palliative at baseline.Thus, patients eligible for neoadjuvant chemotherapy and metastasectomy were excluded in this cohort.Consequently, only a few patients underwent metastasectomy in our studied cohort and therefore our study could have been underpowered to assess the effects of metastasectomy.Furthermore, only patients treated with chemotherapy or targeted therapy were included and not patients offered best supportive care at baseline.
8][29] Our findings are in line with previous retrospective studies 15,27,28 showing prolonged OS following PTR.In contrast, a RCT in Japan with 165 patients showed no effects of PTR on OS in patients with asymptomatic, synchronous unresectable mCRC. 18Another RCT in Korea showed a higher 2-year cancerspecific survival rate for patients treated with PTR followed by chemotherapy in patients with asymptomatic mCRC compared with chemotherapy alone.The study was though small with only 48 patients included and terminated in 2018 because of difficulties in participant enrollment. 30A large meta-analysis from 2019 including 159 991 patients from 77 studies showed an improved OS for PTR of 7 months, which is close to the 9.2 months in OS gain with PTR in our study. 31urthermore, a systematic review by Yefei Shu et al recently published based on 2805 mCRC patients showed prolonged OS for PTR in asymptomatic mCRC patients. 32An individual-participant-data (IPD) meta-analysis of 3423 patients from 8 RCTs, showed better OS for patients treated with PTR regardless of which type of chemotherapy and/or targeted therapies that had been used. 154][35] Both the French CLIMAT RCT study 33 and the Chinese RCT 34 are inactive since 2017 with unknown recruitment status, and, two other RCTs have been terminated due to difficulties to recruit patients. 16,17The Dutch CAIR04 RCT has not yet published any final results for PTR on OS, but showed in an initial analysis a higher 60-day mortality in the PTR group compared with the non-operated group.This aim was though not predefined in the study protocol. 35his population-based cohort study showed prolonged OS in mCRC treated with PTR.Previous studies have mainly evaluated patient-related prognosticators for PTR.We suggest assessment of tumour-related prognosticators for PTR with focus on genetic subsets of mCRC including microsatellite instability-high (MSI-H)/ mismatch repair deficient (dMMR).

Conclusions
This population-based cohort study shows prolonged OS in mCRC treated with PTR.PTR was offered to mCRC patients with prognostically favourable disease.

Fig. 2 .
Fig. 2. Primary tumour resection (PTR) in palliative mCRC.Kaplan Meier survival data in palliative chemotherapy patients treated with or without PTR.

Table 2
Factors associated with overall survival in the palliative metastatic colorectal cancer cohort

Table 3
Treatment with chemotherapy and targeted therapy in the palliative mCRC cohort Data are given as mean AE SD for continuous variables and proportions (%) for categorical variables.The chi-squared test was used to compare categorical variables and the Mann-Whitney U-test was used to compare continues variables between the two treatment arms.PTR, primary tumour resection.