Background:
The prevalence of early-invasive colorectal cancer (T1 CRC) has seen a steady increase, attributed in part to the enhanced population colorectal cancer screening. Concurrently, there has been a significant rise in the adoption of endoluminal treatment (ET) as a non-invasive curative treatment strategy, propelled by technological advancements in endoscopic technologies. Despite a trend towards ET, clarity on the recurrence rates of T1 CRC remains elusive, and comprehensive large-scale studies on long-term outcomes are limited.
Methods:
In this retrospective cohort study, we utilized data from the Taiwan Cancer Registry from 2004 to 2017 to identify all T1 CRC cases after the implementation of the nationwide CRC screening program. These T1 CRCs were categorized based on the treatment approach—either ET only or surgical resection (SR). Those treated surgically were further stratified into T1N0, T1N1, and T1N2, according to the N stage. The recurrence rate of each category of T1 CRCs and the proportion of recurrence patterns were calculated. Multivariable analysis was used to compare the risk of recurrence across the different T1 CRC categories. The follow-up on the survival status of patients with recurrences was extended until 2020, allowing for the assessment of CRC-related mortality rates.
Results:
Out of the total 18,053 T1 CRC cases, 4,075 patients (21.7%) received only ET (ET group), and 13,978 T1 CRCs underwent SR and followed by systemic therapy if N was positive [with T1N0 constituting the majority (70.41%), followed by T1N1 (7.10%), and T1N2 (0.78%)]. (Figure 1A) The overall recurrence rate was 3.8%, with the ET group at 5.7%, T1N0 group at 2.9%, T1N1 group at 15.9%, and T1N2 at 15.6%. The distribution of recurrence patterns is shown in Figure 1B. Most recurrences in the T1N2 group occurred within three years post-treatment, whereas recurrences in other categories predominantly occurred within five years (Figure 2A). Compared to the T1N1 group, after adjusting for age, gender, and year of treatment, the adjusted hazard ratio (aHR) for recurrence was 2.18 (95% CI=1.84-2.58) for the ET group, 2.30 (95% CI=1.42-2.39) for the T1N1 group, and 5.89 (95% CI=3.86-8.98) for the T1N2 group. The lowest CRC-related mortality rate for recurrent cases was in the T1N0 group at 43.8%, followed by ET group at 50.0%, T1N1 group at 59.7%, and T1N2 group at 60.9%. The survival curves for recurrent cases are shown in Figure 2B.
Conclusions:
T1 CRC patients treated only with ET had a 2.18-fold higher risk of recurrence compared to T1N0 patients who underwent SR, indicating potential undertreatment of deeply invasive lesions that might require surgical intervention. These findings underscore the need for precise colonoscopic diagnostic assessment of invasion depth and accurate pathological analysis to identify patients at risk of lymph node metastasis.

Figure 1. The attributable proportion of different categories of T1 CRCs and their recurrence patterns
A: Attributable proportion of different categories of T1 CRCs
B: Recurrence pattern of different categories of T1 CRCs
Figure 2. Long-term outcomes of T1 CRCs
A: Cumulative recurrence across different categories of T1 CRCs
B: Survival after recurrence across different categories of T1 CRCs