BRAF mutant appendiceal adenocarcinoma differs from colorectal cancer but responds to BRAF-targeted therapy

Cohort generation

The MD Anderson adapted version of the Palantir-Foundry software system was used to perform an automated query of the MD Anderson GI Medical Oncology database to identify patients with AA treated between 2004 and Dec 2023. Eligible patients had a pathologic diagnosis of AA, mucinous adenocarcinoma, signet ring adenocarcinoma, or goblet cell adenocarcinoma along with a history of BRAF mutational profiling in the MD Anderson Molecular Diagnostic Laboratory (MDL), a CLIA-certified diagnostic laboratory. A cohort of patients with CRC was similarly created (Supplementary Fig. 1). This study was approved by the University of Texas MD Anderson Cancer Center Institutional Review Board (protocol 09-0373). A waiver of informed consent was granted in accordance with US federal regulation 45 CFR § 46.116 (Common Rule) and the study’s minimal risk to patients. This study was conducted in accordance with the ethical guidelines stated in the Declaration of Helsinki. We followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cohort studies (Supplementary Table 4).

Data collection

Mutation profiles and other easily accessible information (age, sex, date of diagnosis, date of last follow-up, date of death, labs for tumor marker data, and microsatellite instability) were collected using the Palantir-Foundry software system. Patients with germline TP53 mutations (5 patients in the BRAF^WT CRC) were grouped with non-germline TP53 mutations for presentation. A manual chart review was performed on BRAF-mutant AA patients to confirm the specific BRAF mutation. BRAF variant class assignments were determined by following prior literature^11,36. Histological features and tumor grade were collected in a semi-automated fashion using a regular expression on pathology notes and confirmed by manual review. The MD Anderson clinical panels for microsatellite stability were used to define microsatellite status. These panels were either PCR-based or immunohistochemistry-based. We prioritized PCR results over IHC results, meaning that if a patient was tested with both IHC and PCR, the PCR results were used to define microsatellite status.

Overall survival was determined by calculating the difference between the date of diagnosis and either the date of last follow-up or the date of death. Median overall survival (OS) for each group was determined using the Kaplan-Meier method; p-value by the Cox Proportional Hazard model.

Treatment information was collected via manual chart review. In several cases, the notes only included the month of treatment initiation and cessation. In these cases, the 1st was assigned as the day of initiation. For analysis, 5-FU and its prodrug capecitabine were grouped.

Treatment response determination

Radiographic best response for BRAF^V600E treatment were manually determined using radiologist interpretations of abdominal or thoracic CT scans with or without contrast, abdominal MRI scans with or without contrast, and FDG PET/CT scans. When data were available, the RECISTv1.1 criteria³⁷ were used to determine the best radiographic response to treatment. Due to inadequate information in patient notes, the modified peritoneal RECIST criteria (mpRECIST) could not be used (3).

Biochemical response for treatment was determined using tumor marker information. Normal Tumor marker (TM) levels and biological response were defined as previously described^31,38.

Statistics

Baseline characteristics of the study population were summarized by frequency, and differences were compared using the chi-square test. Time-to-event endpoints were visualized using Kaplan-Meier curves, and differences in OS were compared using Cox’s proportional hazards model.

All statistics were performed using R (v4.3.2) in RStudio (2023.09.01).