Transanal total mesorectal excision (TaTME) versus laparoscopic TME for MRI-defined low rectal cancer: a propensity score-matched analysis of oncological outcomes

This propensity score-matching analysis is the first study that compares TaTME with lapTME for low rectal cancer based on a strict anatomical definition. By including only MRI-defined low rectal cancers, we have focused on a highly selected patient group, known to pose greater technical difficulty and increased risk of poorer oncological and clinical outcomes. Moreover, by propensity matching for nine potential confounding factors, homogeneous groups for comparison were created.

This study did not show a significant decrease in R1 resections after TaTME.

However, we were able to demonstrate a significant decrease in conversion rate from 22.0% in the laparoscopic group, to no converted cases in the TaTME cohort (P < 0.001). So far, six other comparative matched studies on this topic have been published [19‐24], and only the study by Chen and Persiani [20, 24] included more TaTME cases (n = 50 and n = 46, respectively). However, none of these studies focused specifically on low rectal cancer. Similarly to our findings, Persiani et al. also reported a significant decrease in conversion rates in TaTME versus lapTME, from 19.6 to 0%, respectively. Low conversion rates for TaTME were also reported in the recent update from the International TaTME registry (4.3%) and the largest systematic review to date by Deijen et al. (3%) [25, 26]. The study by Oostendorp et al. [27] confirmed the promising conversion rates in TaTME, as the authors report on a conversion rate of 2% compared to 13.7% in a comparative cohort of four recent large RCT’s on lapTME [7‐9, 28]. The conversion rate in our comparative laparoscopic cohort was 22%, which is comparable to other studies describing high conversion rates in laparoscopic TME, ranging from 12 to 50% [29‐31]. The low conversion rates observed in TaTME have the potential to improve oncological outcomes, as several studies found increased risk of loco-regional recurrences and decreased disease-free survival rates in patients after converted laparoscopic procedures [11, 12]. Moreover, decreasing the conversion rate in TaTME also influences the long-term morbidity with a lower rate of incisional hernias and adhesional obstructions [32].

The hypothesis that TaTME improves positive margin rates was not confirmed in this study, although there appeared to be a trend towards lower R1 rates (from 12.2 to 4.9%). The studies by Chen, De’Angelis, Marks and Velthuis et al. [21‐24] showed a similar trend in decreased rate of positive margins after TaTME, although also not significant. The only comparative study reporting on a significant decrease in involved margins was the study by Chang and coworkers [19]; 0 of 23 cases in TaTME versus 4 of 23 cases after lapTME (P = 0.037). The reported positive resection margin rates in the first two reports from the international TaTME registry based on 634 and 1540 TME resections for cancer, were 2.7% and 4.9%, respectively [25, 33]. The systematic review by Deijen et al. [26] analyzing 794 cases from 33 studies, reported a positive margin rate of 4.7%. These results suggest that TaTME has the potential to decrease R1 rate, an important predictor of LR [34], when compared to lapTME, as three recent randomized controlled trials found R1 rates in laparoscopic TME surgery varying from 7 to 12% [7‐9].

In this study, we matched for nine factors that are known to increase case difficulty and subsequently influence the risk of R1 resection. The Mercury II study group recently published a stratification model of pre-operative features on MRI that influence development of local recurrence after resection for rectal cancer [35]. Those risk factors on MRI are extra-mural vascular invasion (EMVI) status, an anteriorly located tumor, predicted CRM involvement and tumor height < 4 cm from anal verge. Correspondingly, we matched for tumor height < 1 cm from the anorectal junction. We were not able to match for the other factors; however, our baseline characteristics show that these variables were comparable between both groups, except for EMVI status, which was not routinely reported for all patients early on in the control group.

Over the past decades, surgical management for rectal cancer has changed. Nevertheless, whether by open or laparoscopic surgery, TME for low tumors encounters many challenges, which can be potentially overcome with TaTME. When approaching from below, patient characteristics seem to have less impact on the procedure, as a multivariate analysis in the study by Penna et al. [33], failed to show any patient characteristics to be of influence on the risk of poor specimen outcome in TaTME. Also, a clearer visualization may result in less traction on the specimen and more accurate TME dissection, possibly providing better quality TME and negative margins, a more accurate distal transection, and lower conversion rates compared to laparoscopic surgery. Furthermore, the technique allows for a double pursestring anastomosis, avoiding difficulties with stapling the rectal stump, and therefore may have the added potential to lower the anastomotic leak rate [36, 37].

This hypothesis on AL was not seen in our results. The AL rate of 18% in our TaTME cohort might seem relatively high. However, when interpreting these rates, tumor height must be taken into account. A large study by Borstlap et al. [38] found an early AL rate of 13.4% in patients who underwent low anterior resection, but this increased to 20% beyond 30 days. An independent risk factor for AL was a distal tumor (≤ 3 cm from ARJ), with an odds ratio of 1.88. The study by Bertelsen and coworkers [39] published on a five to six time increase in AL risk for low rectal tumors. The recent registry study on anastomotic leaks in TaTME by Penna et al. [25] published on an early AL rate of 7.8%, which is within an acceptable range compared to previously reported incidences in colorectal surgery [40].

The learning curve effect regarding the new TaTME approach could be the reason why more significant differences were not seen favoring TaTME, in particular for histopathological outcomes and anastomotic leakage, but will hopefully improve with increasing surgeon experience. TaTME is not an easy surgical technique, and surgeons have to adapt to the unfamiliar anatomical view from below. Our institution regularly receives visitors and is a training center for surgeons aiming to learn TaTME. This might be the reason that operative time was not decreased in the TaTME group, while other comparative studies [22, 41‐43] have reported on the ability of TaTME to effectively decrease operative time when performed with a two-team approach.

We acknowledge that this study has limitations. First, the retrospective view on prospectively collected data. By propensity score-matching for nine factors known to make a procedure more difficult, we attempted to create two comparable groups; nevertheless, we could not match for unknown confounders. Further, a large confounder in this study is time, as the inclusion period for the conventional cohort starts in 2006. This may have negatively influenced the outcomes of this group, as surgeons were still on their learning curve for laparoscopic TME, particularly in the early part of this period. However, we compared results to TaTME from the start of implementing this technique in our institution, and so included the early stages of the surgeons learning curve for both procedures. We could not account for possible other improvements in management over time, which might have skewed results. Third, the sample size is relatively small, resulting in a lack of statistical power. Finally, this study only reports on short-term outcomes. Currently, the literature on TaTME still lacks long-term oncological and functional data. Longer follow-up, on large patient cohorts, is needed to provide such data, hence the importance of the international TaTME registry [14]. Recently, an international randomized clinical trial (COLOR III [44]) comparing transanal to laparoscopic TME for rectal cancers started patient recruitment and will be completed in the next 5 to 8 years. While awaiting these results, non-randomized comparative studies are the best level of evidence available.

This study aimed to compare results of TaTME with lapTME in a highly selected patient group by only including MRI-defined low rectal tumors. No significant difference was found in R1 rate. This novel technique warrants further experience and prospective randomized studies to ensure oncologic safety and assess long-term functional results.