Approximately half of patients with synchronous metastatic colorectal cancer are asymptomatic from the primary tumor.1 Primary tumor resection is indicated for symptoms including bleeding, obstruction, and perforation, but the role for asymptomatic primary tumor resection has been debated. Current National Comprehensive Cancer Network (NCCN) Guidelines advise against asymptomatic primary tumor resection but acknowledge randomized clinical trial data are desired.2 Theoretic benefits of primary tumor resection include prevention of obstruction, bleeding, and other complications during subsequent chemotherapy. Prior retrospective studies demonstrated an association between primary tumor resection and increased survival, although were plagued by inherent selection bias.3, 4, 5

A large retrospective propensity-matched study of stage IV CRC undergoing palliative primary tumor resection (with or without subsequent systemic therapy) versus systemic therapy alone in the Netherlands concluded that primary tumor resection was associated with improved overall survival.3 The study included 10,371 patients treated from 2008-2011. Median overall survival after primary tumor resection (17.2 months) compared to systemic therapy (11.5 months) was significantly higher (HR 0.44, P < 0.001). Thirty-day mortality following primary tumor resection was 9% and was higher in geriatric patients. The same 30-day mortality (9%) was reported after initiation of systemic therapy, which was comparable across all age groups. The authors concluded that primary tumor resection should be considered for younger patients without significant comorbidities.

Multiple systematic reviews and meta-analyses favored primary tumor resection over systemic therapy alone.4, 5 Clancy et al. identified 21 studies including 44,226 patients reporting outcomes after primary tumor resection versus systemic therapy alone in stage IV CRC with unresectable metastases. primary tumor resection was associated with a lower risk of mortality (OR 0.28, 95% CI 0.165-0.474, P < 0.001) and a mean survival of 6.4 months in favor of resection (95% CI 5.025-7.858, P < 0.001). The authors noted significant cross-study heterogeneity. In addition, given that the source studies were almost universally retrospective, selection bias may have driven the results.

More recently, Simillis et al. performed a systematic review and meta-analysis of primary tumor resection followed by systemic therapy versus systemic alone. The authors identified 77 studies with almost 160,000 patients. The study included all patients undergoing primary tumor resection (both symptomatic and asymptomatic). A statistically significant survival difference of 3.86 months (HR 0.69) was reported in a subgroup analysis of asymptomatic patients who underwent resection. Primary tumor resection resulted in 4.5% perioperative mortality and 22.4% morbidity (10.2% major). Patients receiving systemic treatment alone experienced 21.7% morbidity (obstruction 14.4%, anemia 11.0%, hemorrhage 1.5%, perforation 0.6%, other adverse event requiring surgery 15.8%). Surgery that did not involve resection of the primary tumor resulted in 10.6% perioperative mortality and 21.7% morbidity (7.9% major). The authors concluded that primary tumor resection in patients with incurable CRC resulted in limited improved survival without a significant increase in morbidity and should be considered on an individual basis. The prematurely terminated RCT NCT01978249 investigated the role of primary tumor resection in metastatic incurable CRC and reported significantly improved 2-year cancer-specific survival after primary tumor resection (72.3% vs 47.1%, P = 0.049).6 However, the study accrued only 48 patients from 2013 to 2016. Further, 2-year overall survival was not significantly different between the two groups.

Critics of primary tumor resection in the setting of unresectable stage IV CRC argue the perioperative morbidity and mortality following surgery outweigh any potential benefit. Further, resection of the primary tumor may result in a stoma. In an asymptomatic patient, the morbidity of a stoma or post-operative complication may decrease quality of life. Two large randomized clinical trials were published in 2021 evaluating primary tumor resection followed by systematic therapy versus palliative systemic therapy alone. Both studies fail to support a role for primary tumor resection in nonresectable stage IV disease.7, 8

Kanemitsu et al. evaluated whether primary tumor resection before chemotherapy improved overall survival in CRC patients with an asymptomatic primary tumor and synchronous unresectable metastases. This multi-center study in Japan included 165 patients who were randomized to chemotherapy alone (mFOLFOX6 plus bevacizumab or CapeOX plus bevacizumab) or primary tumor resection followed by chemotherapy. The study was terminated early after interim analysis on the basis of futility, as future expected accrual would not demonstrate a benefit to primary tumor resection. Overall survival in the primary tumor resection plus chemotherapy arm was 25.9 months (95% CI 19.9 to 31.5) and in the chemotherapy arm was 26.7 months (95% CI 21.9 to 32.5) with a median follow-up of 22 months. The hazard ratio of primary tumor resection plus chemotherapy with regard to overall survival was 1.10, 95% CI 0.76 to 1.59, one-sided P = 0.69. Three postoperative deaths occurred in the primary tumor resection arm, and primary tumor resection was associated with more severe chemotherapy-related adverse events. The authors concluded primary tumor resection followed by chemotherapy showed no survival benefit over chemotherapy alone and should not be considered for patients with unresectable stage IV CRC with an asymptomatic primary tumor.

The CAIRO4 study evaluated the difference in 60-day mortality in CRC patients with synchronous, unresectable metastatic disease and an asymptomatic primary tumor between primary tumor resection followed by chemotherapy versus chemotherapy alone.8 A secondary aim was the identification of risk factors for 60-day mortality. This multi-center study conducted by the Danish and Dutch collaborative recruited 196 patients who were randomized to chemotherapy alone (fluoropyrimidine-based plus bevacizumab) versus primary tumor resection followed by chemotherapy. Sixty-day mortality in the primary tumor resection arm was significantly higher at 11% (95% CI 6-19%) compared to the chemotherapy arm at 3% (95% CI 1-9%), P = .03. In a per-protocol analysis, 60-day mortality remained significantly higher in the primary tumor resection arm (10% versus 2%, P = .048). Factors significantly associated with 60-day mortality in the primary tumor resection arm included elevated serum lactate dehydrogenase, aspartate aminotransferase, alanine aminotransferase, and/or neutrophils. In CAIRO4, 14% of the 97 patients who underwent surgery received a stoma, the majority of which were colostomies.

CAIRO4 was criticized for changing both the primary outcome of interest and for failing to predefine the secondary outcome prior to initiation of the trial. The initial protocol published in 2014 proposed a primary outcome of overall survival, not 60-day survival.9 Authors of the CAIRO4 publication noted that overall survival results will be forthcoming once additional data are accrued.8 CAIRO4 was also criticized for inadequate recruitment. Initial power calculations called for a study population of 360 patients, whereas the published study recruited only 196.

There are common limitations to the study of whether or not the primary tumor should be resected in asymptomatic stage IV CRC. Study accrual in RCTs is challenging, as some patients who are fit to undergo surgery desire this intervention. Even in the face of data suggesting no survival benefit to removal of the primary tumor, patients often prefer a proactive treatment approach and do not understand the logic of leaving a malignancy in situ. In almost all non-randomized studies, selection bias is unavoidable, as surgery is generally not favored in patients who are high risk surgical candidates or are perceived to have aggressive tumor biology. Patients of advanced age or with significant co-morbidities are rarely candidates for surgery at all and are therefore treated systemically. These high-risk patients have shorter expected survival at baseline. Selection bias is therefore a real limitation in any observational study. While a RCT study design should eliminate this bias, the issue of patient preference may impact recruitment. The heterogeneity of stage IV disease is an additional challenge as a patient with small, bilobar liver metastases and a well-differentiated tumor carries a very different prognosis than a patient with advanced carcinomatosis and a poorly differentiated signet ring cell primary.