This multi centre, prospective, longitudinal observational cohort study is designed to validate a novel chemotherapy toxicity prediction tool—the CARG-GI score—specifically for older adults (more than or equal to 60 years) with gastrointestinal (GI) malignancies. The new score is a simplified modification of the original Cancer and Aging Research Group (CARG) score, addressing limitations observed in GI cancer patients, where the original model tends to overestimate risk.
Rationale The original CARG score incorporates 11 variables but is time-intensive and less practical in high-volume cancer centers. Importantly, in GI malignancies, the score inherently categorizes patients into intermediate or high risk, limiting its discriminatory ability. Retrospective analysis of 701 GI cancer patients at Tata Memorial Hospital demonstrated that a reduced set of parameters: Mobility-Tiredness (Mob-T) score, hearing impairment, stage (I to III vs IV), site of primary tumor, and chemotherapy dosing: predicted grade 3 to 5 treatment related adverse events (TRAEs) more accurately than the original CARG score (AUC 0.75 vs 0.59, p less than 0.0001).
Study Objectives
Primary Objective: Assess the incidence of grade 3 to 5 TRAEs and its association with CARG GI stratification (low, intermediate, high risk).
Secondary Objectives:
Compare predictive accuracy of the CARG GI score versus the standard CARG score using AUC analysis.
Assess the association of risk groups with treatment modifications and delays.
Study Design:
Type: Multicenter, prospective, observational validation study.
Sites: Tata Memorial Hospital, Parel, and ACTREC (with potential for expansion).
Duration: 18 months.
Sample Size: 408 patients, calculated using Riley et al.’s method for external validation of prediction models, anticipating 53 percent event rate and accounting for 15 percent attrition.
Population: Patients more than or equal to 60 years with histologically confirmed stage I to IV GI cancers (esophageal, gastric, colorectal, hepatobiliary, pancreatic, gallbladder, and rarer GI cancers) planned for systemic chemotherapy (with or without targeted/immunotherapy).
Eligibility Criteria
Inclusion: ECOG PS 0–2, adequate hematologic/renal/hepatic function, ability to provide informed consent.
Exclusion: Contraindications or hypersensitivity to chemotherapy drugs, or severe comorbidities precluding systemic therapy.
Methodology: All patients will undergo baseline geriatric assessment, including CARG and CARG GI parameters. Data captured will include demographic variables, comorbidities, Mob-T score, hearing status, tumor site, stage, and chemotherapy dosing (standard vs reduced). Patients will then be followed during the first 3 months of systemic therapy.
Toxicity Assessment:
TRAEs defined as grade 3–5 hematologic and non-hematologic toxicities (per NCI CTCAE v5.0).
Relevant endpoints: anemia, neutropenia, febrile neutropenia, diarrhea, nausea/vomiting, mucositis, neuropathy, fatigue, hepatotoxicity, hand-foot syndrome, emergent hospitalizations, and sudden cardiac death.
Events adjudicated by treating oncologists/PIs for attribution to chemotherapy.
Statistical Analysis:
Baseline characteristics: Descriptive statistics; chi-square tests for categorical variables, t-tests for continuous variables.
Risk stratification: Low (0 to 13 points), intermediate (14 to 21), high (more than 22).
Predictive performance: ROC curves, AUC comparison between CARG and CARG-GI via DeLong method.
P-value less than 0.05 considered statistically significant.
Analyses conducted in R v4.2.1.
Ethical and Regulatory Aspects:
Conducted per ICMR guidelines and Declaration of Helsinki.
Informed consent mandatory; clear explanation of study objectives, risks, and voluntary participation documented.
No additional interventions beyond standard care; non-interventional observational design.
No reimbursement/compensation provided. |