Systemic and local treatment strategies for initially unresectable colorectal cancer liver-only metastases

Abstract

For patients with colorectal cancer liver-only metastases (CRLM), the only potentially curative treatment is local treatment by surgical resection and/or thermal ablation. However, only around 25% of patients have CRLM that are directly amenable to local treatment. Patients with initially unresectable CRLM may qualify for local treatment after reducing the tumour size by induction systemic treatment. Part I of this thesis focuses on selecting the optimal systemic treatment for patients with initially unresectable CRLM. Chapters 1 and 2 describe the randomised phase 3 CAIRO5 trial, which included 530 such patients. A unique aspect of this study was the liver expert panel that evaluated resectability of CRLM. In 294 patients with right-sided or RAS or BRAFV600E mutated primary tumours, FOLFOXIRI plus bevacizumab led to a longer progression-free survival and higher complete local treatment rate, at the cost of more toxicity, compared with FOLFOX or FOLFIRI plus bevacizumab. No difference in overall survival was observed. In 236 patients with left-sided and RAS and BRAFV600E wild-type primary tumours, we observed no difference in progression-free survival, local treatment rates, and overall survival between adding panitumumab versus bevacizumab to FOLFOX/FOLFIRI, whereas panitumumab was associated with more toxicity. After complete local treatment, 40% of patients had an early recurrence within six months and about a third of patients had early recurrence without salvage local treatment options. In these patients, for whom only palliative treatment remains, overall survival was worse compared with patients without early recurrence or those who received local treatment after early recurrence. Adjuvant chemotherapy, administered in about 40% of patients who underwent complete local treatment, was associated with longer overall survival. Current clinically available parameters are not sufficient to predict which patients benefit from FOLFOXIRI versus FOLFOX/FOLFIRI, both plus bevacizumab (chapter 3). In conclusion, FOLFOX or FOLFIRI plus bevacizumab are valid treatment options with less toxicity compared to regimens with FOLFOXIRI or anti-EGFR agents for patients with initially unresectable CRLM. Part II addresses the decision-making on local treatment strategies for patients with initially unresectable CRLM. Results from chapter 4 show considerable variability between liver expert panel surgeons in resectability assessments and technical local treatment planning. The unwarranted variability as found in this study should be reduced to ensure that all patients have the same probability of receiving local treatment regardless of which hospital they are treated in. In addition to the need for consensus guidelines on resectability criteria and technical approaches, the introduction of liver expert panels in daily practice may help overcome unwarranted variability which is supported by results from chapter 5. Chapter 6 presents a propensity score-matched analysis demonstrating that one-stage surgery is associated with longer overall survival compared to two-stage surgery, with comparable postoperative morbidity and mortality. Part III of this thesis is dedicated to the identification of patients with initially unresectable CRLM who benefit from local treatment. Currently available clinical factors (chapter 7), the consensus molecular subtypes classification (chapter 9), and RECIST and morphologic response (chapter 10) were not associated with early recurrence after local treatment for initially unresectable CRLM. Pathologic response was associated with early recurrence, but this is only available after local liver treatment. In chapter 8, we show that preoperative detectable ctDNA was associated with early recurrence outcomes and overall survival. Preoperative ctDNA is the first factor that has shown such a strong prognostic value. Nevertheless, early recurrence (without the option of salvage local treatment) does not occur in a third of patients with detectable ctDNA, so ctDNA is not sufficient on its own to select patients for local liver treatment. Hence, novel preoperative parameters are warranted to predict early recurrence and prevent potentially futile liver surgery.