Intensive surveillance instead of completion surgery is being more widely adopted after local resection of high-risk T1 colorectal cancer (T1CRC) for early detection of recurrences. However, due to limited evidence, the optimal surveillance intensity remains unknown. This study examined the preferences and attitudes of high-risk T1CRC patients and their physicians toward surveillance. 

We conducted a multicenter cross-sectional survey in five Dutch hospitals. Eligible patients had an histologically confirmed high-risk T1CRC treated with local organ-preserving resection after January 2018. Physicians actively involved in T1CRC management were also invited. Participants completed an Adaptive Conjoint Analysis (ACA) to evaluate the relative importance of surveillance attributes (mortality, endoscopy, imaging, laboratory testing and incidental findings) in context of T1CRC follow-up. Participants compared pairs of hypothetical surveillance scenarios varying across these attributes and selected their preferred option. Choices revealed the relative importance of each attribute and the trade-offs they are willing to make. Furthermore, participants completed the Medical Maximizer–Minimizer Scale (MMMS), a validated tool assessing individuals’ general tendency to favor more (maximizers) or less (minimizers) medical intervention across all healthcare decisions. Patients additionally completed the Cancer Worry Scale (CWS), and the EQ-5D-5L to assess cancer-related worry and quality of life.

Based on ACA results, we developed a preference simulator, a tool integrating utility scores from individual participants to predict the proportion of patients or physicians likely to favor specific surveillance strategies under varying clinical scenarios. 

In total, 104 patients and 40 physicians participated. Patients had a mean age of 66.1 years (Standard Deviation (SD) 9.0), with balanced sex distribution (48.9% male). Most patients had one histological high-risk feature (66%). Adjuvant therapy was given in 57.9%, mostly surgery (70.5%). Median follow-up was 46 months, with 6% recurrence (3% local, 3% distant).

Among patients, 55.3% were medical maximizers (MMMS score ≥40) and 44.7% minimizers (MMMS score <40). All physicians were medical minimizers. Overall, 39.4% of patients reported high levels of cancer worry (CWS score ≥20) and the mean EQ-5D-5L score was 0.90 (SD 0.13).

Mortality was the most important ACA attribute for patients and physicians (relative importance of 31.4 and 32.3, respectively), followed by endoscopy frequency (21.3 and 20.5). Imaging frequency (17.2 and 16.6), risk of incidental findings (15.3 and 19.3), and lab testing (14.8 and 11.4) had lower but non-negligible relative importance.

Our simulator enabled modelling and comparison of multiple surveillance scenarios. Two scenarios were modeled in our simulator: (1) standard surveillance according to Dutch guidelines, consisting of five endoscopies, five imaging sessions and ten laboratory tests over five years, with 3% mortality and 20% risk of incidental findings; and (2) reduced surveillance, consisting of three endoscopies, three imaging sessions and five laboratory tests over five years, with 6% mortality and 15% risk of incidental findings. Under these assumptions, 91.7% (Standard Error (SE) = 0.7%) of patients and 83.0% (SE = 2.0%) of physicians preferred the reduced surveillance strategy. A further comparison was made between (1) guideline surveillance and (2) a lower-intensity scenario, consisting of two endoscopies, two imaging sessions and five laboratory tests over five years, with 10% mortality and 15% risk of incidental findings. In this comparison, 84.5% of patients (SE = 1.5%) and 62.4% of physicians (SE = 4.4%) favored the reduced surveillance strategy.

Patients and physicians were willing to accept higher mortality in exchange for reduced surveillance, despite ranking mortality as the most important attribute. Using our simulator, a modest reduction in surveillance with higher mortality risk was preferred by most patients (91.7%) and physicians (83.0%), with patients showing stronger preference. This highlights the importance of considering patient perspectives in T1CRC surveillance. Our simulator can help predict and incorporate these perspectives in surveillance guidelines.