Endoscopic tattoo localization for colorectal lesions has become a crucial component of surgical planning to ensure accurate identification of lesions prior to surgical resection. With the increasing trend toward laparoscopic procedures and the inherent inability to manually palpate lesions intraoperatively, precise preoperative lesion localization is paramount [1,2,3,4]. Despite increased use of endoscopic tattooing for surgical planning, rates of tattoo placement, practice patterns, and documentation has been reported to vary widely in the literature [5, 6]. Although colonoscopy accurately localizes lesions preoperatively in 88.7% of cases, the addition of endoscopic tattooing not only allows identification of small tumours that may not be identified on abdominal imaging, it also enables intraoperative localization of lesions in 97.9% of cases [1].

However, localization errors rates still vary and can have a profound impact on patient care. Prior studies have reported that errors in localization have led to tumour misidentification at time of surgery, on-table intraoperative endoscopy, conversion from laparoscopic to open, significant changes to the operative plan, and even resection of the wrong segment of bowel [7,8,9]. Although preoperative tattoo placement has been found to decrease localization error at time of surgery, preoperative tattooing is currently not standard practice and error rates in localization still exists despite tattoo placement (3.8–34%) [2, 7, 10, 11]. A systematic review that included 2578 patients found that when preoperative endoscopic tattooing was performed, there was a significant reduction in localization error compared to conventional colonoscopy without tattooing on pooled analysis [9.5% (95% CI 5.7–13.3, I2 = 73.1%) versus 15.4% (95% CI 12.0–18.7, I2 = 84.9%), respectively] [2]. Despite the use of endoscopic tattooing for nearly seven decades, standardized guidelines for endoscopic tattooing and reporting does not exist [12,13,14].

The aim of the present study is to determine the current rate of endoscopic tattoo placement, tattoo location in relation to the lesion, and factors associated with preoperative tattoo placement at a high volume Canadian tertiary care centre. We hypothesize that the rates of tattoo placement and location continue to vary and support the need for evidence-based standardized guidelines for endoscopic tattooing for colorectal lesions requiring surgical resection.

Materials and methods

Study population

A retrospective chart review was performed on all patients who underwent elective colorectal surgery for a colorectal neoplasm (cancer or polyps) at St. Boniface Hospital in Winnipeg, Manitoba, Canada between January 1, 2007 and December 31, 2017. St. Boniface Hospital is a tertiary centre that performs the highest volume of colorectal surgery in the province. The study protocol was approved by the Health Research Ethics Board at the University of Manitoba and St. Boniface Hospital.

All patients who underwent elective colorectal resection for a benign or malignant lesion were eligible for the study.

Data collection

Data specific to the diagnosis and management of all patients with colorectal lesions requiring surgical resection were collected. Data were collected and managed using REDCap (Research Electronic Data Capture) secure online platform. Information was grouped into five categories (1) patient demographics and history, (2) presentation, (3) endoscopy-specific factors, (4) surgical intervention, and (5) perioperative factors. All dictated and hand-written clinic and referral notes, endoscopy reports, operative reports, pathology reports, and diagnostic imaging reports located in the paper and electronic charts were reviewed. A thorough review of relevant endoscopy-specific factors were collected and included the specialty of the initial endoscopist (Gastroenterologist or Surgeon) and the site at which the endoscopy took place (urban or rural). Urban locations included tertiary hospitals and community facilities. A community facility included a hospital or clinic located within 0–49 km in distance from a tertiary site. Rural facilities were defined as a hospital or clinic located greater than 50 km away from tertiary care centres.

Particular focus was placed on the preoperative location of the lesion and if endoscopic tattooing was performed at initial (index) endoscopy. Data specific to the location of the tattoo (proximal, distal, or both), if the location of the lesion was described, and what descriptors were used (anatomical landmarks, tattoo, distance from anal verge, image capture, scope guide) were also collected. The anatomical location of the lesion was defined by the bowel segment (right colon, cecum, ascending colon, hepatic flexure, transverse colon, left colon, splenic flexure, descending colon, sigmoid colon, rectosigmoid, rectum). Tattoo localization and error rate was defined as the difference in greater than one anatomical segment from where the tattoo was placed at index endoscopy compared to findings at time of surgery.

Statistical analysis

Standard descriptive statistics were used to describe the overall cohort of patients. Univariate analysis was performed to identify variables associated with tattoo placement at index endoscopy. A full multivariate logistic regression analysis model was performed to determine predictors of tattoo localization at index endoscopy. A subgroup analysis was undertaken to stratify predictor variables for colonic versus rectal lesions and determine if there were any dichotomous findings when separated by location. The Bhapkar test of marginal homogeneity was used to establish the lesion localization error rate. Data analyses were conducted with R version 3.5.1. Statistical significance was set at a p-value less than 0.05 and a 95% confidence interval.

Results

Patient demographics

A total of 1385 patients who underwent elective colorectal resection for a polyp or cancer during our study period were identified. Patients were excluded if they underwent emergency surgery (n = 7), palliative surgery (n = 13), surgery for anal cancer (n = 8), resection of a small bowel tumour (n = 25), or resection of an appendiceal tumour (n = 12). Patients who did not have a lower endoscopy performed prior to surgery (n = 14), had missing documentation or were misclassified (n = 244) were also excluded. After excluding 323 patients, a total cohort of 1062 patients were included for analysis. Baseline patient characteristics are shown in Table 1. The majority of patients were male (56%, n = 596) with a mean age of 68 years (SD 12.3) and a mean body mass index (BMI) of 28.8 (SD 6.3). The majority of patients lived in an urban area (76%, n = 807) and had undergone previous abdominal surgery (51%, n = 543). Inflammatory bowel disease (2.6%, n = 28) and hereditary familial disorders (3.6%, n = 28) were uncommon.

Table 1 Patient demographics
Full size table

Endoscopy-specific factors

The majority of index endoscopies were performed at an urban location (90%, n = 901), which included both tertiary and community facilities (Table 2). The index endoscopist was most commonly a Surgeon (53%, n = 553) compared to a Gastroenterologist (47%, n = 484). The majority of patients had lesions located in the right colon (49%, n = 517), followed by the rectum (23%, n = 238), and the left colon (20%, n = 216). Descriptors at index endoscopy included what the endoscopist chose to document in the report.

Table 2 Endoscopy-specific factors
Full size table

The majority of lesions were reported as being tattooed at index endoscopy (57%, n = 542) (Table 2). There was no statistically significant difference in rates of tattoo placement by a surgeon at index endoscopy compared to a Gastroenterologist (OR: 0.89, CI 0.69–1.16, p = 0.43). Although tattoo location in relation to the lesion was not reported in the majority of cases, when documented, placement in relation to the lesion often varied (Fig. 1).

Fig. 1
figure 1

Tattoo placement at index endoscopy for patients undergoing elective colorectal resections for benign and malignant neoplasms (2007–2017)

Full size image

Surgery-specific factors

Intraoperatively, the conversion from a laparoscopic to an open surgery occurred at a rate of 12% (n = 74) (Table 3). Reasons for this were documented as being due to issues encountered during surgery, adhesions, or bleeding. Intraoperative endoscopy occurred in 5.2% of cases (n = 56). The most common surgical procedure performed was a right hemicolectomy (52%, n = 554), followed by an abdominoperineal resection (14%, n = 150). For those who had clear documentation of the tattooed location in the operative report, anatomical segment differed from that reported on index endoscopy compared to intraoperative location by a rate of 36% (n = 123). In patients who had a tattoo placed preoperatively, the tattoo was documented in the operative report as not visible in 8% of cases (n = 247).

Table 3 Surgery-specific factors
Full size table

Predictors of tattoo placement at index endoscopy

Multivariate logistic regression analysis was utilized to determine the independent predictors of tattoo placement at index endoscopy (Table 4). Lesions located in the transverse colon (OR: 1.93, 95% CI 1.04–3.59, p = 0.04) were predictive of tattoo placement at index endoscopy. In addition, patients who underwent surgery after 2010 (2011–2014: OR: 1.88, 95% CI 1.31–2.68, p = 0.001; 2015–2017: OR: 2.87, 95% CI 1.93–4.26, p < 0.001), underwent a planned laparoscopic procedure (OR: 1.69, 95% CI 1.22–2.33, p = 0.001), and had their index endoscopy performed in an urban setting (OR: 5.92, 95% CI 3.23–10.87, p < 0.001) were at higher odds of having a tattoo placed at index endoscopy. On the other hand, patients with rectal lesions had decreased odds of having a tattoo placed at index endoscopy (OR: 0.29, 95% CI 0.19–0.44, p < 0.001).

Table 4 Multivariate analysis: predictors of tattoo placement at index endoscopy (full model)
Full size table

When performing a subgroup analysis to stratify lesions by location, we found slightly different predictors for tattoo placement at index endoscopy for colonic as compared to rectal lesions. We found that surgeries performed after 2015 (OR: 4.10, 95% CI 1.61–10.45, p = 0.003) and urban location (OR: 2.75, 95% CI 1.01–7.45, p = 0.04) were significant predictors of tattoo placement at index endoscopy for lesions located in the rectum (Table 5). There was no change in significant predictors for tattoo placement at index endoscopy when a subgroup analysis was performed for colonic lesions as compared to the full model.

Table 5 Multivariate analysis: predictors of tattoo placement at index endoscopy (rectal lesions)
Full size table

Discussion

This study presents the largest series to date investigating endoscopic tattoo localization prior to elective colorectal resection. Prior studies have included sample sizes between 10 and 310 patients [2]. We have identified that lesion location and planned laparoscopic surgery are factors that contribute to preoperative endoscopic tattoo placement. Importantly, endoscopic tattooing techniques varied widely and colonoscopy reports contained suboptimal documentation. In addition, endoscopists who are located in urban facilities are much more likely to tattoo lesions preoperatively. These factors provide important insights into the inconsistencies of current endoscopic practices and lack of standardized guidelines related to endoscopic reporting.

Endoscopic tattooing has been shown to decrease localization error rates, thus leading to fewer changes to the operative plan [2, 6, 9]. However, the literature is fraught with concerns regarding variations in tattooing practices and lack of clear guidelines. Issues related to endoscopic tattooing include lack of tattoo placement, variability in tattoo technique and location, and lack of consistent reporting practices. Such issues have led to increased rates of repeat preoperative endoscopy due to wide variations in localization error, need for intraoperative endoscopy, significant changes to the operative plan, and ultimately the concern of poor oncologic resection or not being able to identify the tumour altogether [2, 5, 6, 9, 15]. In a retrospective cohort study of 203 patients conducted by Fernandez et al., 64.5% of patients underwent preoperative endoscopic tattooing prior to colonic resection to facilitate intraoperative lesion localization [9]. In contrast, a multicentred retrospective cohort study of 244 patients performed by Al Abbasi et al. demonstrated a preoperative tattoo rate of only 21.4% [15]. Although we found that more than half of lesions at index endoscopy underwent endoscopic tattoo placement (57%), it is unclear why over 40% of lesions were not tattooed. Lesion location, type of surgery, individual practice patterns, and study time period may have influenced our findings.

Our study demonstrated that lesions located in the transverse colon were associated with an increased rate of tattoo placement at index endoscopy, whereas lesions located in the rectum were associated with reduced rates of tattoo localization on multivariate analysis. This is an important finding as the location of the lesion within the transverse colon will greatly influence the planned resection (for example extended right hemicolectomy for proximal transverse lesions, extended left hemicolectomy for distal transverse lesions, or a transverse colectomy for mid transverse lesions). This is consistent with previous studies that identified transverse and left-sided lesions as independent predictors for alterations in surgical management and described routine consideration for preoperative tattooing of distal lesions to reduce modifications to the planned procedure [9]. In our study, rectal lesions made up 23% of our patient population with 14% of patients undergoing abdominoperineal resection. This likely reflects referral patterns and centralization of low rectal tumours to higher volume centres [16]. Our high proportion of rectal lesions may have influenced the overall rate of tattoo placement at index endoscopy by lowering our rates. Clinically, some endoscopists feel that tattoo localization is not necessary for right-sided lesions which may be close to anatomical landmarks such as the appendiceal orifice or ileocecal valve. Others believe that rectal lesions palpable on digital rectal exam do not require tattooing. For rectal tumours, this might be due to the theoretical risk that tattoo placement may increase submucosal fibrosis and potentially risk the plane of dissection during surgery or endoscopic mucosal resection [17, 18]. However, others consider tattooing at initial diagnosis an important step in preoperative planning as a rectal lesion may disappear with neoadjuvant therapy [19].

Still, rates of endoscopic tattooing vary, with previous studies identifying endoscopic tattoo rates that range between 21.4% and 65.1% [4, 9, 15]. In addition, there are ongoing inconsistencies in tattooing technique and placement in relation to the lesion which may contribute to localization errors. Although endoscopic tattooing has been in existence for nearly 70 years as a safe and reliable method to mark colorectal lesions prior to surgery, standardized guidelines still do not exist [12, 13]. Institutional recommendations include tattooing between 1and 5 cm distal to the lesion in multiple quadrants circumferentially to ensure identification of the lesion intraoperatively [14, 20]. However, Gastroenterologists and even Surgeons tattoo proximal, distal, and sometimes even both, making localization of lesions confusing at time of surgery [5, 6]. We found that tattooing practices varied with respect to where a tattoo was placed in relation to the lesion at index endoscopy. Tattoos were placed distal (27%), both proximal and distal (4%), and exclusively proximal (2%). Importantly, the tattooed location was not reported in the majority of cases (67.3%). These data are consistent with a retrospective chart review conducted by Spaete et al. that found endoscopic reports often lacked the location of the tattoo in relation to the lesion (78%), and for those who did have a location described, 47% were placed distal, 39% were placed both proximal and distal, and 9% were proximal [5]. In addition, we found that for patients who had a tattoo placed at index endoscopy and had this location documented at time of surgery, the location varied by greater than one anatomical segment in 36% of patients (n = 123). Although our data include specific anatomical segment (such as cecum, ascending colon, hepatic flexure as described earlier) a difference in greater than one anatomical segment may not greatly impact the initial planned resection (for example a right hemicolectomy for right-sided lesions). As described in the literature, it is still important to consider that variations in tattoo location can lead to tumour misidentification, inability to visualize the lesion, and major changes to the operative plan, and thus may profoundly impact patient care [7, 9].

We have identified that laparoscopic surgery is associated with increased rates of tattoo placement at index endoscopy (OR: 1.69, 95% CI 1.22–2.33, p = 0.001). Likewise, more than half of the planned procedures were laparoscopic (59%). This is consistent with data supporting the increasing trend towards the use of laparoscopic surgery for colorectal neoplasms, as well as, recognition of the importance of tattoo localization prior to minimally invasive surgery [1, 4, 6]. Accurate identification of the lesion prior to surgery is instrumental in precise operative planning and tumour identification at time of surgery. This is a fundamental component of laparoscopic procedures given the inherent inability to manually palpate tumours and the lack of haptic feedback. Consistent tattooing prior to laparoscopic surgery is an important area that warrants future initiatives to establish standardized guidelines that outlines consistent and accurate endoscopic tattoo placement and reporting.

Finally, we found that preoperative endoscopic tattooing was higher in urban compared to rural facilities, however, it is unclear if this is due to a physician’s level of experience, specialty of training, or if patient characteristics or location of the lesion impacted these findings. What it is evident is that this highlights an important area of focus for future educational initiatives. Physicians who perform endoscopy in rural centres may include General Practitioners, community Surgeons, or community Gastroenterologists. Educational initiatives may enhance knowledge translation and bridge gaps in communication with an effort to streamline care, increase rates of preoperative tattoo localization, and reduce unnecessary repeat procedures for patients undergoing elective colorectal resection.

Limitations

The results of our study must be interpreted in the context of several limitations. The retrospective nature of our study design is subject to missing variables, misclassification bias, and unknown potential confounding variables. However, the majority of data collection was performed by a single individual (OH) to mitigate these possibilities. Additionally, as our sample population was obtained from a single academic Canadian institution, our results may not be generalizable to other centres who perform elective colorectal resection for benign and malignant lesions. Notably, Manitoba is unique in its high proportion of surgical endoscopists performing the initial endoscopy, which differs from other provinces where the majority of endoscopies might be performed by Gastroenterologists. This creates a unique opportunity to explore tattoo localization for lesions requiring surgical resection at our institution.

Conclusion

Preoperative endoscopic tattooing for elective colorectal resection is becoming increasingly performed. Lesion location, laparoscopic procedures, and site of endoscopy were found to be predictive of tattoo placement. However, tattooing method, placement in relation to the lesion, and documentation lacked consistency. Endoscopic tattooing remains a crucial component of preoperative planning and is vital to surgical localization and resection. This study highlights the need for standardized tattooing practices and reporting amongst endoscopists.