Backgrounds: A multicenter randomized phase III trial (NCT02605265) proved that adding irinotecan guided by UGT1A1 to capecitabine-based neoadjuvant chemoradiotherapy significantly increases complete tumor response. The treatment toxicities were increased but tolerable.

Purposes: This study aims to identify the predictive biomarkers (from patients' tumor biopsy samples and peripheral blood samples before neoadjuvant therapy) for predicting the response and toxicities to neoadjuvant therapy to stratify patients and optimize treatment strategy.

OBJECTIVES:

Primary:

• Establish a predictive model for response based on tissue RNA and plasma exosome RNA
• Establish a predictive model for toxicities based on tissue RNA and plasma exosome RNA

Secondary:

• Internal validation of the established predictive models
• External validation of the established predictive models

OUTLINE:

-Treatment: Patients receive neoadjuvant therapy and surgery per the protocol. Samples collection Tumor tissue and peripheral blood will be collected prior to neoadjuvant therapy.

-Grouping: Response: Patients will be dichotomized into two groups based on the TRG. TRG of 0-1 is defined as good response. TRG of 2-3 is defined as poor response.

Toxicities: Patients will be dichotomized into two groups based on the grade of AEs. No grade 3-4 toxicities occurs during neoadjuvant therapy is defined as light toxicities. Grade 3-4 toxicities occur during neoadjuvant therapy is defined as heavy toxicities.

-Predictive Model Construction: Using RNA sequencing method to obtain the whole genome transcription profiles of the tumor tissue and plasm exosome RNA. Compare the gene expression differences between the two response groups and the two toxicity groups. Predictive models of response and toxicities are constructed.

-Internal Validation: Patients treated at Fudan University Shanghai Cancer Center (N=50) per the protocol will be enrolled as the internal validation cohort. Samples of tissue and plasm will be collected and analyzed. The performance of the model will be evaluated by the correlation of the predicted response/toxicities and the actual response/toxicities.

-External Validation: Patients treated at Liao'ning Cancer Hospital & Institute (N=50) and Harbin Medical University Cancer Hospital (N=50) per the protocol will be enrolled as two external validation cohorts. Samples of tissue and plasm will be collected and analyzed. The performance of the model will be evaluated by the correlation of the predicted response/toxicities and the actual response/toxicities.

Inclusion Criteria:

• Pathological confirmed adenocarcinoma
• Clinical stage T3-4 andor N+
• The distance from anal verge less than 12 cm
• No suspicious metastatic disease (M1)
• ECOG PS 0-1
• UGT1A1*28 6/6 or 6/7
• No previous anti-cancer therapy

Exclusion Criteria:

• Pregnancy or breast-feeding women
• Serious medical illness
• Baseline blood and biochemical indicators do not meet the following criteria: neutrophils≥1.5×10^9/L, Hb≥90g/L, PLT≥100×10^9/L, ALT/AST ≤2.5 ULN, Cr≤ 1 ULN
• DPD deficiency
• UGT1A1*28 7/7

Stage T3 or T4 and/or N+ rectal adenocarcinoma cancer eligible for neoadjuvant chemoradiation