The goal of this clinical trial is to evaluate an AI-integrated Emotional Granularity Growth intervention (AI-EGG) designed to enhance resilience and improve quality of life in young and middle-aged colorectal cancer (CRC) survivors. Emotional granularity refers to the ability to clearly identify and differentiate subtle emotional experiences, which may help individuals regulate emotions more effectively and build resilience after cancer treatment.

The main questions it aims to answer are:

• Does the AI-EGG intervention improve resilience in CRC survivors compared with routine psychological care?
• Does the intervention improve emotional granularity, emotion regulation ability, and quality of life?
• Is the AI-EGG intervention feasible and acceptable for young and middle-aged CRC survivors?

Researchers will compare the AI-EGG intervention group to a control group receiving routine psychological care and standard educational materials to see whether the intervention leads to better psychological outcomes.

Participants will:

• Complete baseline assessments measuring emotional granularity, emotion regulation, resilience, and quality of life
• Be randomly assigned to either the intervention group or the control group
• In the intervention group, engage in a 4-week AI chatbot-based program focusing on emotional identification, differentiation, regulation, and reflective practice (at least two sessions per week)
• In the control group, receive routine psychological care and standard educational materials
• Complete post-intervention assessments immediately after the 4-week program and again at a 1-month follow-up
• Some participants in the intervention group will be invited to complete interviews about their experience of the program

This study is a randomized controlled trial designed to evaluate the feasibility, acceptability, and preliminary effectiveness of an AI-integrated Emotional Granularity Growth intervention (AI-EGG) in improving resilience, quality of life emotional granularity and emotion regulation among young and middle-aged colorectal cancer (CRC) survivors.

CRC survivors in early and middle adulthood often experience persistent psychological adaptation difficulties after completion of primary treatment, even when clinical disease is stable. These difficulties include reduced resilience, impaired emotion regulation, and decreased quality of life. Emotional processes are considered central to post-treatment psychological adaptation, and resilience is conceptualized as a key psychosocial outcome reflecting individuals' ability to adapt to cancer-related adversity.

Emotion regulation is an important determinant of resilience; however, existing interventions typically focus on general emotion regulation strategies without directly targeting the individual's ability to differentiate and label emotional experiences. Emotional granularity, defined as the ability to distinguish and accurately label discrete emotional states, is proposed as a cognitive-affective mechanism that may enhance emotion regulation effectiveness and thereby support resilience.

The AI-EGG intervention is developed based on this theoretical framework and is delivered via an online chatbot platform. The intervention is structured as a 4-week program that provides guided, interactive training in emotional granularity. The chatbot system delivers standardized yet interactive modules focusing on emotional identification, emotional differentiation, emotion regulation, and reflective emotional processing. Participants are encouraged to engage with the system at least twice per week, with additional voluntary engagement supported by the platform.

The study adopts a parallel-group randomized controlled design. Participants are randomly assigned to either the AI-EGG intervention group or a control group receiving routine psychological care and standard educational materials. The intervention is delivered remotely through a mobile-based chatbot system, enabling flexible and repeated engagement in real-life contexts.

Outcome assessments are conducted at baseline, immediately after the 4-week intervention period, and at 1-month follow-up. These assessments evaluate changes in emotional granularity, emotion regulation, resilience, and quality of life over time.

In addition to quantitative evaluation, a qualitative component is conducted among a purposively selected subset of participants in the intervention group. Semi-structured interviews are used to explore participants' experiences of using the AI-EGG system, perceived changes in emotional granularity and resilience, engagement patterns, and perceived acceptability of the intervention.

Overall, this study aims to provide preliminary evidence on whether an AI-based emotional granularity training program can improve emotional granularity and support resilience in CRC survivors, and to inform the development of scalable digital psychosocial interventions in oncology care.

Inclusion Criteria:

• (1) young and middle-aged adult patients (age in the range of 18-60 years);
• (2) patients diagnosed with CRC;
• (3) patients have completed primary curative-intent treatment for colorectal cancer, and are currently in a stable post-treatment or maintenance phase of care without evidence of active disease progression or acute treatment-related instability;
• (4) patients able to use a smartphone and agree to participate in the study.

Exclusion Criteria:

• (1) patients who have been informed of their cancer diagnosis due to family decision to withhold information;
• (2) patients suffering from severe comorbidities or conditions that may affect participation or assessment, such as significant cognitive impairment (e.g., dementia, severe memory loss), psychiatric disorders, or other serious medical complications;
• (3) those who are participating in other psychological intervention studies.

A final decision regarding sharing individual participant data (IPD) has not yet been finalized. This is due to the need for further internal discussions and consultations with all collaborating institutions, particularly regarding compliance with cross-border data transfer regulations under China's Personal Information Protection Law (PIPL) and ethical guidelines governing participant privacy. Additionally, the feasibility of robust de-identification processes and secure data-sharing mechanisms is being evaluated to ensure participant confidentiality. The final decision will be determined before study completion and will align with journal requirements, participant consent agreements, and institutional policies.

Outcome assessors are blinded to participants' group allocation to reduce assessment bias. Participants are assigned unique study identification codes, and outcome data are collected and analyzed using de-identified datasets to maintain blinding at the analysis stage. The research coordinator responsible for randomization and allocation is not involved in outcome assessment. Although participants are aware of their group assignment due to the nature of the behavioral intervention, access to the AI-EGG intervention is restricted through individual login credentials to minimize cross-group contamination. Backend usage logs are monitored for technical monitoring purposes only and are not accessible to outcome assessors.