Yara Backes,1 Tom C. J. Seerden,2 Rosanne S. F. E. van Gestel,1 Onno Kranenburg,3 Inge Ubink,3 Raymond M. Schiffelers,4 Demian van Straten,4 Malu S. van der Capellen,1 Simone van de Weerd,1 Wendy W. J. de Leng,5 Peter D. Siersema,1,6 G. Johan A. Offerhaus,5 Folkert H. Morsink,5 Winesh Ramphal,2 Jochiim Terhaar Sive Droste,7 Anja U. G. van Lent,8 Joost M. J. Geesing,9 Frank P. Vleggaar,1 Sjoerd G. Elias,10 Miangela M. Lacle,5 and Leon M. G. Moons1
Background and Aims
In patients who have undergone surgery for colorectal cancer (CRC), 3% have recurrence of (metachronous) CRC. We investigated whether tumor seeding during colonoscopy (iatrogenic implantation of tumor cells in damaged mucosa) increases risk for metachronous CRC.
In a proof of principle study, we collected data from the Dutch National Pathology Registry for patients with a diagnosis of CRC from 2013 through 2015, with a second diagnosis of CRC within 6 months to 3.5 years after surgery. We reviewed pathology reports to identify likely metachronous CRC (histologically proven adenocarcinoma located elsewhere in the colon or rectum from the surgical anastomosis). For 22 patients fulfilling the inclusion criteria, we ascribed the most likely etiology to tumor seeding when endoscopic manipulations, such as biopsies or polypectomy, occurred at the location where the metachronous tumor was subsequently detected, after endoscopic manipulation of the primary tumor. We collected clinical data from patients and compared molecular profiles of the primary and metachronous colorectal tumors using next-generation sequencing. We then examined the source of seeded tumor. We tested whether tumor cells stay behind in the working channel of the endoscope after biopsies of colorectal tumors, and whether these cells maintain viability in organoid cultures.
In total, tumor seeding was suspected as the most likely etiology of metachronous CRC in 5 patients. Tumor tissues were available from 3 patients. An identical molecular signature was observed in the primary and metachronous colorectal tumors from all 3 patients. In 5 control cases with a different etiology of metachronous CRC, the molecular signature of the primary and metachronous tumor were completely different. Based on review of 2147 patient records, we estimated the risk of tumor seeding during colonoscopy to be 0.3%–0.6%. We demonstrated that the working channel of the colonoscope becomes contaminated with viable tumor cells during biopsy collection. Subsequent instruments introduced through this working channel also became contaminated. These cells were shown to maintain their proliferative potential.
In an analysis of primary and secondary tumors from patients with metachronous CRC, we found that primary tumor cells might be seeded in a new location after biopsy of the primary tumor. Although our study does not eliminate other possibilities of transmission, our findings and experiments support the hypothesis that tumor seeding can occur during colonoscopy via the working channel of the endoscope. The possibility of iatrogenic seeding seems low. However, our findings compel awareness on this potentially preventable cause of metachronous CRC.